Faculty of Health Sciences

Department of Psychology

Academic performance and student dropout

Results from two studies in upper secondary and higher education in
Northern Norway

—
Rannveig Grøm Sæle
A dissertation for the degree of Philosophiae Doctor – April 2016
Academic performance and student dropout

Results from two studies in upper secondary and higher education in

Northern Norway

Rannveig Grøm Sæle

Department of Psychology

Faculty of Health Sciences

UiT The Arctic University of Norway

Dissertation for the degree of Philosophiae Doctor

April 2016

© Rannveig Grøm Sæle

 Contents

Acknowledgements ....................................................................................................................iii

Abstract .......................................................................................................................................v

Sammendrag (abstract in Norwegian)....................................................................................... vii

List of papers .............................................................................................................................. ix

Abbreviations...............................................................................................................................x

Introduction ............................................................................................................................... 1

Effects of education ................................................................................................................ 1

Employment. ...................................................................................................................... 1

Social consequences. .......................................................................................................... 1

Health. ................................................................................................................................ 1

The Norwegian educational system ....................................................................................... 3

Upper secondary education. .............................................................................................. 3

Higher education. ............................................................................................................... 4

Academic performance .......................................................................................................... 5

Dropout .................................................................................................................................. 5

Predictors for academic performance and dropout ............................................................... 8

Demographics. .................................................................................................................... 8

Academic factors. ............................................................................................................... 8

Psychological factors. ....................................................................................................... 11

Aims and hypotheses ............................................................................................................... 14

Methods ................................................................................................................................... 15

Young Will ............................................................................................................................. 15

Learning in Higher Education ............................................................................................... 16

Missing data.......................................................................................................................... 17

i
 Statistical analyses ................................................................................................................ 19

Item Response Theory and the LP scale. .......................................................................... 19

Measures .............................................................................................................................. 20

Outcomes. ........................................................................................................................ 20

Predictor variables. ........................................................................................................... 21

Procedure. ........................................................................................................................ 21

Ethical considerations of the projects .................................................................................. 23

Summary of papers .................................................................................................................. 24

Paper I ................................................................................................................................... 24

Analyses and results. ........................................................................................................ 24

Discussion. ........................................................................................................................ 25

Paper II .................................................................................................................................. 27

Analyses and results. ........................................................................................................ 27

Discussion. ........................................................................................................................ 28

Paper III ................................................................................................................................. 29

Analyses and results. ........................................................................................................ 29

Discussion. ........................................................................................................................ 29

General discussion ................................................................................................................... 31

Implications .......................................................................................................................... 35

Strengths and limitations ..................................................................................................... 38

Future research .................................................................................................................... 40

Conclusions .............................................................................................................................. 40

References ................................................................................................................................ 42

ii
 Acknowledgements

I have been very lucky to have three outstanding supervisors. Oddgeir – you have been

flexible, positive, creative and supportive. You have understood and approved what I wanted to find

out and helped me do it in the best possible way. I have learned very much from you about

methodology, statistics and all the other research stuff, and for that, I am forever grateful. Tore – you

have welcomed me warmly into the Ung vilje project and provided new thoughts and ideas to my

own project in a splendid way. Tove – wonderful Tove! You always help me remember what I really

want with my research and why it is important. Thank you all for guiding me towards thinking, acting

and writing like a researcher.

Charlotte and Karl Ottar, my fellow PhD students in the Ung vilje project: Thank you for doing

solid groundwork and data collections, and thank you for letting me take part in your project.

Charlotte, I very much enjoy the collaboration we have, and I appreciate all our long and valuable

conversations during these years.

Colleagues and fellow PhD students at the fantastic Department of psychology: Thank you for

long-lasting one-to-one conversations, interesting and stimulating discussions and for just being kind,

clever and great. It’s not too bad being in a work environment filled with psychologists. Every lunch

break teaches me something new. Torstein: Thank you so much for listening to premature ideas,

frustrations and the occasional happy outburst during coffee breaks and runs. Talking things over

with you always leads me one step further.

Thanks to all the students participating in the studies and to the people at the institutions

providing me with data (UiT The Arctic University of Norway, Troms County, DBH), making this

research possible. And thanks to the best English teacher in the world, Anne Jedi Dahl, for

proofreading the introduction.

iii
 Finally, I want to thank my family. My parents, who taught me the importance of curiosity,

commitment, knowledge and education. My dear husband and friend, Kurt Henrik, who is always

there for me, supporting every thing I do. Thanks for being you. And, at last, my wonderful kids, Signe

and Sondre: Thank you for bringing me unbelievable joy and just enough worries to keep my mind on

something else, every day.

Tromsø, 14.04.2016

Rannveig Grøm Sæle

iv
 Abstract

Education is important for the individual and for society. Higher education is associated with

a range of positive factors for work life and social and personal well-being, as well as for the economy

of a society. Lower levels of education, and especially dropout from education, come with both

personal and public costs, and are associated with unemployment, lower salaries, more crime and

more use of welfare benefits. One of the most important and stable predictors of educational

dropout is academic performance, typically measured by grades.

This thesis aims to investigate factors related to academic performance and student dropout.

It comprises two projects. The first project, Young Will (Ung vilje), is based on a sample of

adolescents entering upper secondary school in the autumn of 2010. The second project, Learning in

Higher Education, studies a sample of students entering university in the autumn of 2013.

Academic performance: Several predictors were examined in relation to academic

performance, measured by grade point average. In the upper secondary sample, the most important

predictors for lower grade point averages were male gender, enrolment in the vocational track,

lower self-reported task solving abilities, lower educational ambitions, less promotion focus and

more literacy problems. In the university sample, the most important predictors for lower grade

point averages were less reported use of deep and strategic learning approaches and lower incoming

grades. Literacy problems and procrastination were not significant predictors in the university

sample.

Dropout: Predictors for student dropout, measured two years after enrolment in a study

programme at university, were higher age, less use of surface learning approaches, fewer hours

spent on studying, lower grades achieved in the first year and enrolment in study programmes with

shorter durations. Deep and strategic learning approaches were not significant predictors. We also

included literacy problems, procrastination and burnout, but these constructs did not predict

dropout.

v
 The findings indicate that educational institutions should focus not only on what students

learn, but also on how they learn. At upper secondary level this includes supporting students with

literacy problems and paying attention to low academic performance. At the higher education level,

it seems important to foster productivity and commitment from the beginning of the study

programme.

vi
 Sammendrag (abstract in Norwegian)

Utdanning er viktig for den enkelte og for samfunnet. Høyere utdanning er assosiert med en

rekke positive faktorer for arbeidsliv og for den enkeltes livskvalitet, samt for økonomien i

samfunnet. Lavere nivåer av utdanning, og særlig frafall fra utdanning, har både personlige og

offentlige omkostninger, og er forbundet med arbeidsledighet, lavere lønn, mer kriminalitet og mer

bruk av velferdsgoder. En av de viktigste og mest stabile prediktorer for frafall fra utdanning er

akademiske prestasjoner, vanligvis målt ved karakterer.

Denne avhandlingen tar sikte på å undersøke forhold knyttet til faglige prestasjoner og

studentfrafall. Den består av to prosjekter. Det første prosjektet, Ung vilje, er basert på et utvalg av

ungdom som startet i videregående opplæring høsten 2010. Det andre prosjektet, Læring i høyere

utdanning, studerer et utvalg av studentene som startet på universitet høsten 2013.

Akademiske prestasjoner: Flere prediktorer ble undersøkt i forhold til akademiske

prestasjoner, målt ved karaktersnitt. I videregåendeskole-utvalget var de viktigste prediktorene for

lavere karaktergjennomsnitt å være mann, å gå på yrkesfag, lavere selvrapporterte evner til

selvstendig å løse oppgaver, lavere utdanningsambisjoner, mindre grad av promotion fokusert

motivasjon og større grad av lese- og skrivevansker. I universitetsutvalget var de viktigste

prediktorene for lavere karaktergjennomsnitt mindre rapportert bruk av dyp og strategisk

læringstilnærming og lavere karakterer fra videregående. Lese- og skrivevansker og utsettelsesatferd

var ikke signifikante prediktorer i universitetsutvalget.

Frafall: Prediktorer for studentfrafall, målt to år etter at studentene var oppmeldt ved

universitetet, var høyere alder, mindre bruk av overflatiske læringstilnærminger, færre timer brukt

på studiene, lavere karakterer i det første året og oppmelding til studieprogrammer med kortere

varighet. Bruk av dyp og strategisk læringstilnærming var ikke signifikante prediktorer. Vi undersøkte

også lese- og skrivevansker, utsettelsesatferd og utbrenthet, men disse konstruktene predikerte ikke

frafall.

vii
 Funnene tyder på at utdanningsinstitusjonene bør fokusere ikke bare på hva elevene lærer,

men også på hvordan de lærer. På videregåendenivå innebærer dette å støtte elever med lese- og

skriveproblemer og vie oppmerksomhet til lave skoleprestasjoner. I høyere utdanning er det viktig å

fremme produktivitet og skoleengasjement fra starten av studiet.

viii
 List of papers

I. Sæle, R. G., Sørlie, T., Nergård-Nilssen, T., Ottosen, K.-O., Goll, C. B., & Friborg, O. (2015).

Demographic and psychological predictors of grade point average (GPA) in North-

Norway: A particular analysis of cognitive/school-related and literacy problems.

Educational Psychology, 1-22. doi:10.1080/01443410.2014.998630

II. Sæle, R. G., Dahl, T. I., Sørlie, T., & Friborg, O. (2016). Relationships between learning

approach, procrastination and academic achievement among first year university

students. Higher Education, resubmitted manuscript

III. Sæle, R. G., Dahl, T. I., Sørlie, T., & Friborg, O. (2016). Predictors of student dropout two

years after enrolment at university. Submitted manuscript

ix
 Abbreviations

ASSIST Approaches and Study Skill Inventory for Students

ASQ Adolescent Stress Questionnaire

BSCS Brief Self Control Scale

CFA Confirmatory factor analysis

ECTS European Credit Transfer and Accumulation System

EFA Exploratory factor analysis

GPA Grade point average

GRFM General Regulatory Focus Measure

HSCL Hopkins Symptom Checklist

IPS Irrational Procrastination Scale

IRT Item response theory

LP Literacy problems

MAR Missing at random

MBI Maslach Burnout Inventory

MCAR Missing completely at random

MNAR Missing not at random

PAQ Personal Attributes Questionnaire

RSA Resilience Scale for Adults

x
 Introduction

The impact of education for individuals and for society is massive. First and foremost,

education builds competence. It enables individuals to pursue a professional career and provides

society with a proficient workforce. But the implications span even more broadly, as the level of

education may also impact employment, social engagement and health.

Effects of education

Employment.

Employment status and income are associated with level of education. People who drop out

of upper secondary education work less and have lower incomes than people with higher education

(Rouse, 2007). For every year of schooling a person completes, earnings rise with approximately 10

percent. More highly educated individuals also receive fewer social insurance benefits than adults

with lower education (Belfield & Levin, 2007; K. A. A. De Ridder, Pape, et al., 2012; OECD, 2013;

Putnam, 2015; Steingrímsdóttir et al., 2012).

Social consequences.

Holding a higher education degree is associated with more engagement in society and in the

local community. People with a higher education also participate more in politics and are more likely

to vote than people with only upper secondary education or less, making education important also

for the function of democracy (Putnam, 2015; Rumberger, 2011; U.S. Census Bureau, 2015). School

dropouts are more likely to be engaged in anti-social and criminal behaviour (Moretti, 2007;

Rumberger, 2011). Moreover, these patterns tend to reproduce themselves in the family, making it

harder for children of parents with lower educational levels to move upwards in socio-economic

status (Putnam, 2015).

Health.

1
 Adults with higher education report higher life satisfaction, have better health and live longer

than adults with lower education (Belfield & Levin, 2007; K. A. A. De Ridder, Pape, et al., 2012; OECD,

2013; Putnam, 2015; Steingrímsdóttir et al., 2012). It seems that many of the health benefits of

education – leading, for instance, to a longer life expectancy – are actually due to the education itself

(Muenning, 2007). The influence of educational attainment on health works through both a cognitive

and an occupational pathway. Higher cognitive abilities are associated with healthier behaviours. This

may be related to increased acknowledgment of health warnings and recommendations, for example

regarding exercise and diet. More highly educated people also tend to have a better comprehension

of doctors’ instructions, superior understanding of health bureaucracies and a greater capacity for

developing good coping strategies against stress. In addition, social networks influence health

behaviour (Christakis & Fowler, 2008). Holding a higher education degree may create or strengthen

ties to social networks with other people who have higher education and lead more healthy

lifestyles, thus additionally increasing the probability of healthier behaviours in the individual.

The occupational pathway includes the impact of type of occupation, type and amount of job

benefits and level of income (Muenning, 2007). These may moderate the level of stress experienced

and whether or not stressors become vulnerabilities. School dropouts are more likely to perform

physically burdensome or hazardous labour than people holding a higher education degree. The

effects of higher levels of stress among school dropouts is a higher risk of cardiovascular disease,

cancer, infectious disease and diabetes.

The consequences of dropout are prominent for the individual, but they also affect society,

as upper secondary school dropouts contribute only 40% of the tax revenues contributed by

graduates (Rouse, 2007). In addition, inadequate education increases costs of health care, criminal

justice and other public services (Belfield & Levin, 2007).

Given the importance of education for the individual, society and the nation, understanding

the barriers to achieving an education is paramount. One important barrier is poor academic results,

2
which may hinder further educational advancement and opportunities for skilled or professional

jobs. However, a more important concern is if the student drops out of education completely. The

aim of the present thesis is hence to examine aspects related to academic performance (exam

grades) and dropout from education.

The Norwegian educational system

Education in Norway is free and open for all. The first ten years of schooling are mandatory,

and after that, all adolescents have the right to continue into upper secondary level. Students

achieving the necessary academic requirements may then continue into higher education, without

costs.

While our society has good systems for letting people start an education, we do not yet have

sufficient ways of securing that they complete that education. Currently, about 90% of adolescents

aged 16-18 enter upper secondary school (Statistics Norway, 2015b), but only 71% of these have

completed when their legal study right expires five years later (Statistics Norway, 2015c). The

number of students entering higher education is increasing (OECD, 2015; Statistics Norway, 2016).

Here, the problem of not attaining an academic degree is much larger as about two out of five

students entering higher education have not completed any degree ten years later (Statistics

Norway, 2012).

Upper secondary education.

The year 1994 represents a change in the educational system in Norway, when all

adolescents between ages 16 and 19 received a right to secondary education through the Reform 94

policy. One of the aims of the reform was to improve throughput, but it also increased the number of

adolescents entering secondary education. The share of 16-year-olds in the first year of secondary

education increased from 79% in 1993 to 92% in 1994 (Statistics Norway, 1996; Støren, Skjersli, &

Aamodt, 1998). Among students entering upper secondary education in 2009, 71% had completed

3
schooling five years later (Statistics Norway, 2015c). This is an increase of 2.3 percent points since the

1994 cohort, and most of the increase is achieved among boys in vocational study tracks. Still, more

girls than boys complete upper secondary school even today (76 vs. 66%), and the completion rates

are higher in the general track than in the vocational track (83 vs. 58%). The three counties in

Northern Norway have the lowest rates, with Troms and Nordland at 65% and Finnmark at 54%.

However, the strongest discriminator seems to be grade levels. In Norway, grades from lower

secondary school are summed up, divided by the number of subjects completed, and multiplied by

ten, resulting in school points ranging from 0 to 60 (Statistics Norway, 2015c). The importance of

grades from lower secondary school for further education is clearly illustrated as follows: Among

students with 55 school points or more, almost all (99%) complete upper secondary education, while

among students with less than 25 school points, only 14% do.

Higher education.

Public higher education in Norway is open to all students who have completed upper

secondary school either in the general track, or in the vocational track with special supplemental

studies. There are no registration fees, except for a common fee to the student union of about €60

per semester. Students may apply for a student loan, which makes economic support from parents

less important. A student will receive about €10500 per year, and if the student does not live with

her parents, 40% of this sum is transformed to an educational grant every year – if she passes all the

exams. The payments typically do not cover all everyday expenses and it is common to have a part-

time job while studying. Still, the threshold for entering university is low, and in 2014, one in three

persons between 19 and 24 years of age was enrolled in higher education (Statistics Norway, 2015a).

Sixty percent of all students enrolled were women. However, ten years after entering a higher

educational programme, 41% of the students had not achieved an academic degree (Statistics

Norway, 2012).

4
 Academic performance has been established as one of the most central predictors also for

dropout at university (Araque, Roldán, & Salguero, 2009; Jia & Maloney, 2015; Paura & Arhipova,

2014). Moreover, poor academic performance has been found to be a mediator of other important

predictors, such as socio-economic status and demographics, deviant behaviour, socialization in

school and in the family (Battin-Pearson et al., 2000) and mental disorders (Esch et al., 2014).

Academic performance

Academic performance is normally measured by the teachers’ assessment of student

learning. Teacher judgment may be biased due to its subjective aspects. Unclear expectations and

poor understanding of the learning outcomes among teachers and examiners may lead to unreliable

evaluations (Dahl, 2006). A review by Hoge and Coladarci (1989) found that correlations between

teacher judgments and objective, standardised tests ranged from .28 to .92. However, the median

correlation was .66, indicating adequate validity of teacher assessment as a measure of performance.

Assessment may take different forms, such as written or oral exams, or observation of

behaviour and acquired skills (Schunk, 2014). The evaluation may be expressed as a number (e.g. 78

out of 100) or a percentage (78%). However, most grades are given as either pass/fail or a letter. In

Norway, the letters A-B-C-D-E-F are used, where A reflects an outstanding performance (Norwegian

Ministry of Education and Research, 2008). An E reflects passing with only the minimum

requirements fulfilled and an F is a fail.

Dropout

The definitions of dropout from higher education are diverse. Some studies investigate

dropout after the end of a certain study programme, while others examine dropout rates during the

study period (Grau-Valldosera & Minguillón, 2014). Studies investigating dropout during the study

period may include students who have withdrawn from a study programme; students who were

registered, but who did not turn up; students who did not complete course work and students who

5
did not renew their enrolment for the next year. Some studies also set a cut-off point for consecutive

semesters without registration, because students may take breaks from studying and return later.

Grau-Valldosera and Minguillón (2014) found that in online courses, with a five percent error

threshold, students not registered for 3-5 consecutive semesters are likely to be real dropouts.

Forty years ago, Tinto (1975) proposed a theory of school dropout, emphasising the role of

integration, and separating academic integration from social integration. Background factors in the

model include family background, individual attributes such as age and gender and previous

schooling. Previous schooling includes past educational experiences, in terms of achievement, i.e.,

grades, but also in terms of other experiences related to school. Family background is important

because parental education, work status and attitudes towards education influence the students’

choices and opportunities. People with higher education have higher incomes, a larger formal and

informal network from which their children may benefit and more time and resources to support

their children in education and career planning (Putnam, 2015). This involvement may also

strengthen the integration of their children at school and in society, as adolescents growing up with

more highly educated parents tend to be better informed in their meetings with school systems and

bureaucracy.

Tinto’s model then proceeds into two systems, an academic and a social one. Both systems

concern commitment and how this influences the dropout decision. The decision to quit school is

often based on a long process, and educational commitment and participation are key factors in this

process (Frostad, Pijl, & Mjaavatn, 2014; Reegård & Rogstad, 2016; Rumberger & Lim, 2008). It is not

uncommon among university students to doubt the education chosen and to consider quitting

(Xuereb, 2014). Reasons given by the students themselves for having doubts include too high

workload or pressure, failed exams, unmet expectations, managing competing priorities, family and

work, and financial stress. Reasons for staying despite the doubt are commitment and a wish for goal

fulfilment. Several researchers have proposed categories of dropout students based on different

6
types of risk factors. Heublein (2014) identified three types of students which mirror the most

important reasons found by Xuereb (2014), with different elements characterising the process for

each type. The first type drops out due to performance-related issues. The process starts with poor

grades from upper secondary school and poor knowledge about studying, and continues with

inadequate study support, tough demands and exam failures. The second type drops out because of

motivational causes. Their process starts with unmet expectations or with not studying their

favourite subject, and continues with a lack of motivation and perceived relevance of the education,

which leads to decreased interest and willingness to perform. The third type drops out because of

financial causes. In this thesis, primarily reasons that characterize the first student group are

examined.

Academic performance is one of the most important and stable predictors for dropout

(Bowers, 2010; Casillas, Robbins, Allen, & Kuo, 2012). An array of factors also explain variance in both

academic performance and dropout, as described in the next section. Finally, the two constructs tend

to predict similar individual outcomes later in life, such as employment status, job performance and

salary (Belfield & Levin, 2007; French, Homer, Popovici, & Robins, 2014; Kuncel, Hezlett, & Ones,

2004). Hence, academic performance and dropout are related phenomena, and both should be

addressed in the pursuit of identifying underlying tentative causes (i.e., predictor variables).

In this thesis, grade point averages (GPA) were used as the outcome measure in Papers I and

II. In Paper III, student dropout two years after enrolment was used as the outcome variable. We

included predictor variables covering two main domains in order to examine their importance for

academic performance and dropout. Academic factors were GPA and study behaviour (representing

academic commitment in Tinto’s model), literacy problems, learning approaches and other factors

thought to influence the learning process. Psychological factors were motivation and educational

aspirations, stress and burnout, personality-related factors and mental health. In addition, some

demographic factors were included.

7
Predictors for academic performance and dropout

Demographics.

Gender is a stable predictor of dropout, as more males than females quit their studies

(Alexander, Entwisle, & Horsey, 1997; Battin-Pearson et al., 2000; Casillas et al., 2012). Female

students also achieve higher grades than males (Freudenthaler, Spinath, & Neubauer, 2008;

Pomerantz, Altermatt, & Saxon, 2002). In addition, socio-economic status, parents’ level of education

and their engagement with and attitudes towards their children’s education add to the picture,

regarding both performance and dropout (Alexander et al., 1997; Casillas et al., 2012; De Witte &

Rogge, 2013; Lundetræ, 2011; Sirin, 2005; Strom & Boster, 2007). Among Norwegian students whose

parents hold a higher education degree equivalent to a Master’s Degree, 87% complete upper

secondary school, compared to 46% of students whose parents have only completed compulsory

schooling (Statistics Norway, 2015c).

Academic factors.

Academic factors represent concepts likely to be related to the learning process. These are

learning approaches, procrastination and difficulties with reading and writing.

Learning approaches.

Learning results in change, most notably in terms of changes in knowledge (Mayer, 2011),

but also in terms of learning new skills or applying new behaviours (Schunk, 2014). Both types of

changes are the results of practice or experience and last over time. How do these changes occur?

Behaviourism offers several explanations of change in behaviour, but the change in knowledge is

more difficult to examine and explain.

In the 1970s, a number of researchers in Sweden (Marton & Säljö, 1976), Australia (Biggs,

1979) and the UK (Entwistle, Hanley, & Hounsell, 1979) interviewed, tested and surveyed students in

order to examine the process of learning. They discovered that students have different approaches

8
to learning tasks. The initially identified – and still most well-known – approaches were called deep

and surface (Marton & Säljö, 1976). As the terms indicate, students using the first approach go deep

into the material to be learned, while students using the second only scratch the surface. The deep

learner searches for meaning and understanding. For the surface learner, on the other hand, it is not

that important to understand and achieve a true change in knowledge. The emphasis is on rote

learning and on remembering enough to pass exams and avoid failing. The deep learner tends to be

intrinsically motivated – the surface learner extrinsically.

A few additional approaches and labels have been proposed, for instance the strategic

learning approach (Entwistle et al., 1979). The strategic learner may use both the deep and the

surface approach depending on the learning task. The goal of the strategic learner is to achieve

success. She is organised and manages time and resources effectively to reach this goal.

Deep and strategic learning approaches seem to be associated with higher performance

(Duff, Boyle, Dunleavy, & Ferguson, 2004; W. A. Reid, Duvall, & Evans, 2007), but one study found

that the association is weaker among male students (M. Byrne, Flood, & Willis, 2002). A surface

approach is negatively associated with performance (Diseth, 2007; Duff et al., 2004; W. A. Reid et al.,

2007), but this relationship may be curvilinear (Diseth, 2002). Other studies have failed to find any

association (Campbell & Cabrera, 2014; Gijbels, Van de Watering, Dochy, & Van den Bossche, 2005).

One explanation may be poor validity of the deep learning measure used in the studies. Another

explanation may be that the performance assessment did not rely on deep conceptual knowledge.

The mixed results in the learning approaches literature make the connections between learning

approaches and learning outcome worth closer examination.

9
Procrastination.

To procrastinate is to “voluntarily delay an intended course of action despite expecting to be

worse off for the delay” (Steel, 2007, p. 66). Procrastination is also called a self-regulatory failure

(Ferrari, 2001). The behaviour is irrational, because the procrastinating person knows that it is

unwise to procrastinate and that it will probably lead to an unwanted or at least poorer result. Still,

procrastination is prevalent (Harriott & Ferrari, 1996), especially in the student population (Goda et

al., 2015). It has been associated with a range of behaviours and emotions. The most relevant

correlations for the topic of this thesis are lower self-reported scores on conscientiousness,

organisation, achievement motivation and self-control (Steel, 2007). In addition, the positive

correlations with self-handicapping, boredom-proneness, impulsivity and distractibility are

important, because these aspects are likely to interfere with the learning process. The associations

between procrastination and academic performance are small, but significant. However, the high

prevalence of procrastination and the strong associations with other negative behaviour make it

relevant and important to further investigate the consequences of procrastination among students.

Literacy problems.

Educational attainment is lower among adolescents with learning difficulties than among

their peers (Wagner, Newman, Cameto, Garza, & Levine, 2005). Learning difficulties are associated

with poorer educational outcomes (Gerber, 2012) and dropout (Dunn, Chambers, & Rabren, 2004;

Goldstein & DeVries, 2011; Wagner et al., 2005). Yet, students with learning difficulties are entering

upper secondary and higher education at an increasing rate (Goldstein & DeVries, 2011). The

demands for having an education to get a job are increasing in the entire society (Wagner et al.,

2005). Higher education institutions must thus be prepared to meet the needs of students with

learning difficulties (Gregg, 2007). In this thesis, the focus is on reading and writing difficulties,

labelled literacy problems (LP). Such problems are characterised by difficulties with visual, temporal

and phonological processing and lower processing speed, in addition to poorer metacognitive

10
awareness and memory (G. Reid, 2009). These factors are increasingly important in higher levels of

education, with more material to read and understand in less time.

Psychological factors.

Motivation.

Regulatory focus theory is based on self-regulated learning and motivation. Two self-

regulatory processes proposed by Higgins (1997) are promotion and prevention, based on the

hedonic principle to seek pleasure and avoid pain. A desirable end-state is wanted, but the

motivation for reaching it differs. A promotion focus stimulates sensitivity to presence or absence of

positive outcomes and accomplishments, while the prevention focus is characterised by a sensitivity

towards negative outcomes and failure. A student with a promotion focus wishes to reach a goal and

is motivated by feelings of fulfilment, while a student with a prevention focus is trying to avoid failure

and is preoccupied with safety precautions. The two foci also have dissimilar associations with a

range of other measures. For example, job satisfaction correlates positively with promotion and

negatively with prevention and only promotion correlates (positively) with job engagement (Lanaj,

Chang, & Johnson, 2012). A fit between type of motivation and task characteristics enhances

performance (Chalabaev, Major, Sarrazin, & Cury, 2012; Shah, Higgins, & Friedman, 1998).

Stress.

Another element likely to influence study and learning processes are subjective perceptions

of stress, in particular school-related stress. Stress is the reaction to a provoking factor, a stressor,

and the reaction occurs when demands are exceeding the capacity of the individual (Lazarus &

Folkman, 1984). School-related or academic stressors may be fear of failure, i.e., exam anxiety.

Reactions to stress likely to influence the study situation, such as lack of concentration, inability to

begin work and constant fatigue, have been reported to be experienced weekly by more than 50% of

students in high school (Feld & Shusterman, 2015) and are also highly prevalent among university

11
students (Abouserie, 1994). Stress and stressful life events are linked to poor grades (Pluut, Curşeu, &

Ilies, 2015; Struthers, Perry, & Menec, 2000), a decline in grades (Liu & Lu, 2011) and dropout (Hess

& Copeland, 2001; Schmeelk-Cone & Zimmerman, 2003). Often, the workload itself is not the

problem, but the perceived lack of time (Nonis, Hudson, Logan, & Ford, 1998).

Burnout.

Work stress over time may cause burnout. Burnout comprises three components:

Exhaustion, cynicism and inefficacy (Maslach, 1982; Maslach & Leiter, 2007). Exhaustion is a basic

dimension, and implies a depletion of resources, both emotionally and physically. Cynicism is an

interpersonal dimension, which often includes a loss of idealism, and refers to a negative response to

work. Inefficacy is a self-evaluating dimension, which implies a feeling of loss of competency in the

job. Burnout among students is called educational or learning burnout (Schaufeli, Martinez, Pinto,

Salanova, & Bakker, 2002). Examining burnout with regard to dropout is thus relevant.

Personal attributes.

The gender gap in academic performance is widely recognised (Battin-Pearson et al., 2000;

Casillas et al., 2012; Pomerantz et al., 2002). Female students perform better academically than male

students do, and they more often complete their studies. Differences in cognitive abilities do not

explain this difference, and neither do the Big Five personality traits (Spinath, Eckert, & Steinmayr,

2014). Certain attributes associated with either masculinity or femininity may explain individual

differences in achievement motivation (Spence & Helmreich, 1978). Being more independent and

competitive corresponds to the stereotypical perceptions of masculinity, while interpersonal skills

and cooperation abilities are more typically associated with femininity. However, other typically

masculine traits, such as self-efficacy (Choi, 2004), predict achievement (Richardson, Abraham, &

Bond, 2012). Examining how masculine and feminine traits relate to performance may increase

understanding of the gender differences.

12
Resilience.

The resilience construct is used to denote a favourable outcome (or positive adaptation)

despite exposure to difficult life circumstances (Luthar, 2006). However, the construct also refers to a

range of protective factors that may facilitate resilience, such as individual, social, family and

community factors (Friborg, Barlaug, Martinussen, Rosenvinge, & Hjemdal, 2005). Adolescents who

adapt well despite adversity are adept in actively making use of the resources they have available

(Werner, 1993), which has been shown to positively predict academic performance (Werner & Smith,

2001). Assessing the role of the protective factors that an individual may use to cope with

psychosocial stressors is relevant.

Self-control.

Self-control is the ability to override a dominant response and to regulate thoughts and

behaviours (Bandura & Vandenbos, 1989; D. T. D. de Ridder, Lensvelt-Mulders, Finkenauer, Stok, &

Baumeister, 2012). The importance of self-control for a range of domains has been established for a

long time (Baumeister, Heatherton, & Tice, 1994; D. T. D. de Ridder, Lensvelt-Mulders, et al., 2012;

Mischel, Shoda, & Peake, 1988). This includes a positive association with cognitive competence and

academic performance (Mischel et al., 1988; Tangney, Baumeister, & Boone, 2004). Self-control is

important for education by facilitating the ability to finish tedious tasks or assignments and to persist

in pursuing an academic goal.

Mental health.

Depression and anxiety are more frequently reported among university students than in the

general population, and average prevalence reports are about 30% (Ibrahim, Kelly, Adams, &

Glazebrook, 2013). Students tend to report that their mental problems affect their performance in

school (Eisenberg, Gollust, Golberstein, & Hefner, 2007; Keyes et al., 2012). However, the predictive

value of depression on academic performance varies in the literature. Richardson et al. (2012) found

in their meta-analysis that depression was negatively associated with GPA, but this association was

13
not significant. The role of mental health is also given increased attention in dropout research

(Jackson, 2009). A recent systematic review (Esch et al., 2014) confirms that mood and anxiety

disorders seem to predict dropout, but this relationship may be mediated by behaviour problems

and academic performance. The significance of mental health on academic outcomes is still

uncertain (Esch et al., 2014; Richardson et al., 2012) and needs further examination.

Aims and hypotheses

The aim of the thesis is to examine the role of academic and psychological factors for

academic performance and dropout in upper secondary school and in higher education.

Paper I. Aims: To examine factors predicting GPA in an upper secondary school sample.

Additionally, to examine the psychometric properties of a self-report measurement for LP and to

validate it against GPA. Hypotheses: Having more literacy problems, less educational ambition, more

mental health problems, a prevention rather than a promotion regulative motivational focus and

fewer masculine personality traits predict lower GPA in a sample of Norwegian adolescents.

Paper II. Aims: To examine predictors of GPA in a first-year university student sample and to

investigate the relationships between learning approaches, procrastination and GPA. Hypotheses:

Deep and strategic approaches will positively predict GPA, while surface learning and procrastination

will negatively predict GPA. Procrastination will negatively predict deep and strategic learning

approaches.

Paper III. Aims: To examine factors predicting student status after the first two years of

university studies. Hypotheses: Higher age, male gender and low parental education will predict

dropout, as will procrastination, literacy problems, stress and burnout. Deep and strategic learning

approaches will predict staying, along with hours spent studying, ECTs passed and GPA.

14
 Methods

The thesis comprises data from two projects. The upper secondary sample was recruited

through the project Young Will (Ung vilje) and the university sample was recruited through the

project Learning in Higher Education. Both studies follow a particular group of students

longitudinally, thus representing a prospective cohort design. The information was mainly collected

using surveys, but registry data were also included.

Young Will

The Young Will project aims to identify causes and consequences of dropout in upper

secondary education in Northern Norway, and to provide suggestions for action which may prevent

dropout. The project is funded by the RDA: Regional differensiert arbeidsgiveravgift (regional

differentiated employers' national insurance contributions) of Troms County, and the Sparebank 1

Northern Norway foundation. The project is anchored at the Department of Clinical Medicine at the

Faculty of Health Sciences at UiT The Arctic University of Norway and is led by prof. Tore Sørlie.

The sample consists of 1676 students (69% of the total number of students) entering upper

secondary school in Troms County in the fall of 2010, and the students will be followed until 2020. As

the registry data on dropout rates after five years were not available when Paper I was prepared, the

analyses in this paper are entirely cross-sectional. A description of the participants, the procedures

and flow of participants are reported in Paper I. The Young Will data material contains a broad range

of variables and the project group members use different methodologies (e.g., quantitative and

qualitative) and different professional perspectives (e.g., psychological and educational) to address

the research questions. The paper in the present thesis includes factors related to the learning

processes, while other aspects captured in the qualitative analyses are reported elsewhere (Ottosen,

Goll, & Sørlie, 2016).

15
Learning in Higher Education

The aims of the second project correspond to the aims of Young Will, but the sample is

comprised of 428 university students (which equals approximately ten percent of the new students),

drawn from the group of students entering UiT The Arctic University of Norway in the fall of 2013.

Descriptions of participants and procedures are given in Papers II and III. Flow of participant

recruitment is presented in Figure 1. These students will also be followed for ten years, until 2023,

through the use of questionnaires and registry data1. The project is not externally funded.

4709 invited to the study

4154 did not answer 555 consented

430 completed the

questionnaire

2 withdrew their consent 428 included in the study

Figure 1. Flow of participants in the Learning in Higher Education study

1
Some of the data used are retrieved from the Database for Statistics on Higher Education (DBH) at
the Norwegian Social Science Data Services (NSD). Neither of the institutions delivering data to DBH nor the
NSD are responsible for use and analysis of the data nor the interpretations based on the data.

16
Missing data

Missing data are non-observed – empty – entries in a data matrix, due to for example

nonresponse or attrition (Little & Rubin, 2014). As scientists, we try to understand and explain the

world through observations and these empty entries may impede our ability to do so (McKnight,

McKnight, Sidani, & Figueredo, 2007). Possible consequences are biased estimators, leading to

inaccurate and wrong conclusions. Depending on why data are missing, validity and generalizability

of the results may be threatened.

Thus, it is important to identify the reasons for why data are missing (D. B. Rubin, 1976).

Usually three categories are identified (Allison, 2012; D. B. Rubin, 1976; L. H. Rubin, Witkiewitz,

Andre, & Reilly, 2007; Schafer & Graham, 2002). Data missing completely at random (MCAR) are data

missing without any pattern or without being dependent on any variable in the dataset, neither the

variable where data are missing itself nor others. Missing at random (MAR) implies that data are

missing because of a relation the variable has to the other observed variables. If these variables are

controlled for in the analysis, the missing data will be MCAR (Manly & Wells, 2015). The third

category is missing not at random (MNAR), or non-ignorable missing. Missing data in this category

will be dependent on the variable itself, or on unobserved variables. If we label the variable with the

missing data Y and the other observed variables X, then the data will be MCAR if the probability of

the missingness is related neither to Y nor X, MAR if it is related to X, and MNAR if it is related to Y.

Missing data may be handled in different ways, either by simply excluding cases with missing

data, replacing missing data with the mean, estimating missing values by regression or using multiple

imputation (MI) or expectation maximization (EM) methods (Tabachnick & Fidell, 2013). If the

number of missing data points is low, deleting cases is acceptable. However, it may cause a

considerable drop in sample size and power if the number is high (L. H. Rubin et al., 2007). Missing

data on the predictor variables in this thesis were imputed using the EM method, conducted in Prelis

8.8 (Jöreskog & Sörbom, 2005). This procedure has two steps, expectation (E) and maximization (M).

17
In the first step, missing values are predicted through an expectation based on observed data and

correlations, and in the second, a maximum likelihood estimation is performed until convergence is

reached (Tabachnick & Fidell, 2013).

The participants in the Young Will project completed the initial questionnaire on paper. This

led to missing data in some cases and variables because participants forgot or did not wish to

answer. Little’s MCAR test for all variables was non-significant: χ2(17111) = 14708.019, p = 1.000,

indicating that the data were MCAR, as the missingness was not related to the variables themselves.

The total proportion of missing data on predictor variables was 7.5%. We imputed values on cases

containing less than 40% missing data on the included latent variables and factors. The distribution of

missing data and the imputation procedure are further described in Paper I. The outcome variable in

Paper I was GPA from lower secondary school, collected from the county’s upper secondary school

registry. These had missing data from 19.2% of the cases due to enrolment in a Steiner school (where

students are not graded), grade exemptions because of different dysfunctions or language difficulties

or admittance to upper secondary school on grounds other than lower secondary GPA. These missing

data points were partly related to high scores on literacy problems, as Little’s MCAR test with GPA

and the literacy problem items was significant: χ2(244) = 483.831, p < .000. We did not impute

missing data on this variable.

Participants in the Learning in Higher Education project provided all questionnaire data

through an online service where all questions were made mandatory to complete in order to

proceed. Hence, there were no missing data on these variables. Some participants responded that

they did not know their parents’ education level. These responses were treated as missing. Student

status per semester after enrolment was retrieved from the national database (DBH). Missing data

on this variable imply that the student was not registered in any study programme in higher

education in Norway; hence, had quit or was on a break, and the student was thus treated as a

dropout student. Grades from upper secondary school and on completed exams at university were

18
initially collected from the university registry. Normally, upper secondary grades are used for

admittance to university; however, some students are admitted on other grounds, causing missing

data for 30% of the cases. Some students quit before their first exam or received only pass/fail

grades, causing missing data on university GPA for 11% of the cases. We did not have enough

information to impute missing data on these variables; therefore, we included GPA lastly in the

regression models to avoid it biasing the other coefficients until the last step.

Statistical analyses

Correlation and regression analyses were performed using SPSS v.22, as was the exploratory

factor analysis (EFA). Confirmatory factor analyses (CFA) and item response theory (IRT) analyses

were performed in Mplus (Muthén & Muthén, 1998-2010). Mediation and moderation analyses were

performed in SPSS using the Process macro (Hayes, 2013). A more thorough description of the

statistical procedures is available in the appropriate subsections of the respective papers.

Item Response Theory and the LP scale.

In Paper I, we used item response theory - or latent trait theory - when we examined the

eight items that constitute the LP measure. The method is not as well known as classical test theory;

hence, an elaborated justification and explanation of the IRT follows here. The items and the results

of the analysis are reported in Table 1, Paper I.

Classical test theory (CTT) does not reveal how well a particular item discriminates between

students of different levels of ability (Crocker & Algina, 2006; Embretson & Reise, 2000). The factor

loading that CTT provides is used independently of ability levels, whereas the discrimination

coefficient in IRT may vary depending on the latent trait variable, i.e., ability. Moreover, IRT

estimates these trait levels based on both individual responses and item properties, thus improving

the generalizability of the results. The LP items were scored dichotomously; thus a unidimensional,

logistic IRT model was appropriate. Up to three item parameters may be estimated: difficulty,

19
discrimination and guessing. Item difficulty or location (b) refers to how difficult it can be expected to

be for students to answer yes to the question. If the location parameter is located low on the x-axis,

then the question is easy to answer. An interpretation of this is that students may agree to the item

even if LP levels are low. Conversely, if the item location is high, the required unobserved trait (i.e.,

LP problems) needs to be high for students to answer affirmatively. Item discrimination (a) is the

slope of the regression line and provides information about how well the item discriminates between

students experiencing low and high levels of literacy problems. Steeper slopes indicate better

discrimination. The guessing parameter (c) was not estimated since socially desirable responding was

considered low and an objectively correct answer does not exist.

Measures

Outcomes.

In the first two papers, academic performance was our primary concern. Grades, being a

more stable and robust measure, seem to be more important than scores from single tests or exams

(Rumberger & Lim, 2008); hence a composite measure like GPA is appropriate. The GPA is calculated

from all the grades a student receives during a given time period. When merging all grades into one,

we assume that they all share the properties of a latent construct. In a paper concerning the use of

GPA in research studies, Bacon and Bean (2006) found that internal consistency reliability for all

courses ranged from .84 at the end of the first year of a study programme to .94 at the end of the

fourth year. Earlier GPA is the best predictor for later academic performance, confirming GPA as a

valid measure (Salvatori, 2001).

In the first paper, discussing the upper secondary sample, we used GPA from lower

secondary school. The GPA variable was comprised of all grades from lower secondary school, which

were summed up and divided by the number of school subjects. These grades were decided a few

months before the collection of questionnaire data. Hence, the outcome measure was collected prior

20
to the predictor variables. However, we did not have access to grades achieved during upper

secondary school. Nevertheless, grades are found to be relatively stable over time (Casillas et al.,

2012) and the retrospective time frame was short; hence we consider the outcome variable valid.

In the second paper, which is concerned with the university sample, grades from the first

year were used as the outcome variable. The predictor variables were collected during the early

autumn of 2013, while the grades were collected after the final exams in the summer of 2014.

For university level, the duration and form of the education varies more than in upper

secondary school. In addition, students may change study orientation or fail to pass subjects, making

it essential not only to investigate the end-point, but also to examine what happens during the

educational process (Arias Ortiz & Dehon, 2013). In the third paper, student status after 2 years at

university was used as outcome.

Predictor variables.

Table 1 presents an overview of the instruments used in the three papers. Further

descriptions, validity reports and reliability measures are found in the papers in which they are used.

Procedure.

The predictor variables listed in Table 1 are based on self-reports. Self-reported data are

statements people make about themselves (Schunk, 2014). Instruments used in questionnaires

relying on self-reports need to be valid and reliable; hence, validity and reliability for the instruments

are discussed in the papers where they are used.

In the Young Will study, participants completed the questionnaire on paper. They were asked

to do this during class, and they spent approximately an hour completing it.

In the Learning in Higher Education study, the self-report data were collected via a web-

based questionnaire. Some participants responded twice to the same questionnaire. However, they

21
 were identifiable through their ID number, and only the first or the most complete response was

saved for analysis.

Table 1.
Overview of instruments used in Papers I, II and III
Construct Instrument (reference) I II III

Learning Approaches and Study Skill Inventory for Students

X X
approaches (ASSIST; Entwistle, 1997; Tait & Entwistle, 1996)

Procrastination Irrational Procrastination Scale

X X
Academic (IPS; Steel, 2010)

factors Literacy Literacy problem scale,

problem developed by Karl Ottar Ottosen and Trude Nergaard for
X X X
the first data collection in the Young Will project (Sæle et
al., 2015)
General Regulatory Focus Measure (GRFM)
X
(Lockwood, Jordan, & Kunda, 2002)
Motivation
Q: What is the highest level of education you are
X
planning to complete?
Stress Adolescent Stress Questionnaire
X
(ASQ; D. G. Byrne, Davenport, & Mazanov, 2007)
Burnout Maslach Burnout Inventory for students
X
(MBI; Schaufeli, Salanova, Gonzalez-Roma, & Bakker, 2002)
Personality Personal Attributes Questionnaire
Psychological X
traits (PAQ; Spence & Helmreich, 1978)
factors
Resilience Resilience Scale for Adults
X
(RSA; Friborg, Hjemdal, Martinussen, & Rosenvinge, 2009)
Self-control Brief Self-Control Scale
X
(BSCS; Tangney et al., 2004)
General mental Hopkins Symptom Checklist
X
health (HSCL; Derogatis, Lipman, Rickels, Uhlenhuth, & Covi, 1974)
Q1: How do you currently evaluate your mental health?
Q2: How do you currently evaluate your physical X
health?

22
Ethical considerations of the projects

The Young Will project has been approved by the Regional committee for medical and health

research ethics – North (REK; 2010/1503-7) and by the county administration of Troms. All

participants signed a consent form. Some students were under 16 at the point of the first data

collection, and in these cases, their parents also signed the form. The participants are de-identified

by replacing names and social security numbers with an ID number. The identifier is kept in a locked

safe, separate from any data files and questionnaires. Completed paper questionnaires are kept in a

locked cabinet.

For the Learning in Higher Education project, approval from the REK North was not required,

but it was approved by the Data Protection Official for Research at the Norwegian Social Science Data

Services (NSD; 34867). Participants first completed an electronic consent form. They were then

assigned an ID number and received an electronic questionnaire (Software: Questback). The

identifier is kept in a password-encrypted file, only accessible to the PhD candidate.

Participation was voluntary in both studies, and participants were informed that they were

free to withdraw from the studies at any point, without giving any reason. They also received an

information letter about the background and aims of the studies and about protection of privacy.

23
 Summary of papers

Paper I

Demographic and psychological predictors of grade point average (GPA) in North-Norway:

a particular analysis of cognitive/school-related and literacy problems

Analyses and results.

The psychometric validation of the LP scale was conducted in two randomly split samples.

The first and second samples were used for EFA and CFA purposes, respectively. The EFA extracted

one factor, and the full scale had a Cronbach’s α of .84. The IRT analysis showed that six out of the

eight items discriminated well between the subjects (see Table 1 in Paper I). Although the two final

items discriminated less well, factor loadings were satisfactory and the items were retained.

Figure 2 (see Appendix) shows the item characteristics curves of the eight items. The

difficulty of the item (the b parameter) is decided from the point where the ability (or latent trait) of

a person to respond correctly (or affirmatively) to the item passes a probability of 0.5. The

discrimination property of the item (the a parameter) indicates how quickly a person crosses this

probability cut-off as ability (or the latent) increases. In Figure 3 (see Appendix), the item information

curves show where the items contribute most information in the trait distribution of LP.

We conducted a multiple linear hierarchical regression analysis to identify significant

predictors of GPA. Variables were included in the model in four blocks. The first block included

demographic variables such as gender, age, self-reported financial status in the family and current

educational track (vocational/general). The second block included cognitive/school variables: Task

solving ability, conflict with classroom rules and educational aspirations. The third block included

psychological/affective variables: Regulatory focus, resilience, personal attributes and a short

measure of mental health. Literacy problems were included in the fourth and last block together with

an interaction term (literacy × task solving ability). The most important predictors for higher GPA

24
were being enrolled in a general track (std β = .36), being female (β = -.16), having good task solving

abilities (β = .20), less literacy problems (β = -.16), lower social competence (β = -.12), a promotion

focused motivation (β = .10) and more social resources (β = .09). The interaction effect between LP

and task solving abilities was significant. Total variance explained was 55%.

A particular feature of the regression analysis was the examination of the predictive power of

the four blocks of predictors (demographic, school-related, psychological and LP) if entered as sole

predictors of GPA. Since the r-squared statistics were considerably more pronounced for the

demographic (R2=.39) and the cognitive/school-related (R2=.35) than for the psychological/affective

(R2=.20) and LP (R2=.18) predictors, the former two seem to be most important for GPA. However,

these can be controlled by the learner to a lesser extent, so the latter remain important when

considering what to try to change.

Discussion.

The gender gap in academic achievement is well documented (Pomerantz et al., 2002), and

our study confirmed that female students have higher GPA than male students. Possible reasons for

the gender differences may be home and school environment, peer culture, differences between the

genders in adaptation to the school environment, attitude towards school work and higher self-

discipline among girls (Duckworth & Seligman, 2006; Spinath et al., 2014; Warrington, Younger, &

Williams, 2000).

A higher degree of self-reported LP was associated with a lower GPA, which was expected

since low academic achievement is one of the characteristics of learning difficulties (American

Psychiatric Association, 2013). Higher self-perceived ability to solve tasks – an indication of self-

efficacy – was associated with higher GPA, and this relationship was moderated by LP. Experiencing

more LPs weakened the association. The relationship may also be interpreted the other way around,

i.e., to imply that higher self-perceived ability to solve tasks moderated the effect LP had on GPA.

25
Hence, enhancing self-efficacy and skills to cope with academic tasks may contribute to a reduction

in the negative effect LP has on GPA.

Having a promotion focus was associated with higher GPA, in line with existing research on

motivational strategies (Richardson et al., 2012). Aiming to achieve a good grade, instead of aiming

to avoid a poor grade, is a better strategy.

Lower social competence predicted higher GPA, which was perhaps a counterintuitive

finding. However, scoring high on social competence may indicate being more open to new

friendships and nursing relationships with others, that is, spending more time with friends and less

time on school work.

In summary, the reliability and validity of the LP scale were adequate, and at least six out of

the eight items discriminated well between the students. The last two items addressed problems

with reading comprehension and articulating thoughts on paper. More than one third of the students

responded affirmatively to these, whereas the prevalence of the other items ranged between 12.9%

and 25.7%. These two items therefore measure the more general aspects of literacy problems and

are potentially more related to comprehension than to decoding and spelling. Moreover, both

questions were double-barrelled, trying to capture two closely related aspects in a single item. The

first addressed problems with understanding and remembering sentences, while the other addressed

problems with articulating and verbalizing thoughts on paper. However, since the psychometric

properties of the total score were not compromised by retaining these two items in terms of test

score reliability, they were kept until other studies could prove them unnecessary.

26
Paper II

Relationships between learning approaches, procrastination and academic achievement

among first year university students

Analyses and results.

We performed a hierarchical regression analysis with GPA as the dependent variable.

Learning approaches were included as predictors in the first step, procrastination in the second and

self-control in the third. Relevant covariates likely to affect GPA – among others, literacy problems

and self-reported mental and physical health – were included in the fourth step. Finally, upper

secondary GPA was included. The analysis showed that upper secondary GPA (std β =.39) and deep

(β =.12) and strategic (β = .12) learning approaches predicted GPA, while procrastination and self-

control did not. Total variance explained was 23%.

Hierarchical regression analysis with deep approach as the dependent variable showed that

procrastination negatively predicted a deep learning approach, but it turned out non-significant after

adjusting for strategic and surface learning approaches. Both strategic (β = .27) and surface (β = -.13)

approaches predicted a deep approach. Among the covariates, age (β = .17), study techniques learnt

at university (β = .20) and planning to pursue a master (β = .14) or a PhD degree (β = .28) contributed

significantly positively to a deep approach. Total variance explained was 22%.

The next regression analysis was conducted with strategic approach as the dependent

variable. Procrastination (β = -.40) and self-control (β = .30) together explained 51% of the variance.

In addition, deep approach (β = .13), study techniques learnt at university (β = .11), perceived

usefulness of the study (β = .14) and educational aspirations of completing only single subjects (β =

.08) were significant positive predictors. Total variance explained was 58%.

A mediation analysis revealed that a strategic approach mediated the relationship between

deep approach and GPA, explaining 21% of the direct effect of deep approach on GPA.

27
 Discussion.

Having a deep learning approach, that is, seeking meaning in and understanding of the

material, is valuable for academic achievement as measured by GPA. In addition, a strategic

approach, that is, being structured and spending study time thoughtfully, is helpful. Moreover,

having a strategic approach mediated the effect of deep learning on GPA. We hypothesized that

procrastination would negatively predict GPA, but this was not supported. However, procrastination

was negatively associated with strategic approach and self-control, indicating a negative impact on

study habits and perhaps the learning experience, even if it does not influence the learning product.

Implications: Teaching methods promoting active learning and deeper understanding should

be investigated further. Students should be taught how to acquire study strategies that promote

deep and strategic learning approaches in order to optimize the learning process and the study

results. However, relatively little variance in GPA was explained.

28
Paper III

Predictors of student dropout two years after enrolment at university

Analyses and results.

Since student status as a dependent variable was dichotomously coded (stayer vs. dropout),

we conducted a logistic regression analysis. In the first block, demographics, age, gender and

parental education were included. The next block, study behaviour/experience, included learning

approaches, procrastination, the number of study hours during the first semester, literacy problems

and previous study experience. The third block included stress and burnout. The fourth block, study

context, included the level/duration of the course where the students were enrolled and the number

of European Credit Transfer and Accumulation System (ECTS) passed in the first year of studying. In

the last block, we included GPA achieved during the first year. Out of the 426 students included in

the analysis, 21.4% were not registered in any study programme in Norwegian higher education two

years after enrolment. Predictors for dropout included stress (OR = 1.55), higher age (OR = 1.06) and

being enrolled in short-term study programmes (OR = 2.74). Predictors for staying included higher

grades (OR = .59), spending more time studying (OR = .96) and higher scores on surface learning

approach (OR = .60). Procrastination, deep and strategic learning approaches, literacy problems and

burnout were not significant predictors.

Discussion.

The dropout rate after two years of study was 23%; hence action to prevent dropout should

already be taken during the first semesters. Spending more hours studying - study time - and getting

higher grades were particularly important for decreasing the odds of dropping out. Study stress and

being enrolled in a short-term study programme increased the odds of dropout. The predictive value

of stress was only evident if surface learning was simultaneously tested, thus indicating a suppression

effect.

29
 Strengthening the students’ feelings of academic integration to prevent dropout is

important. This may include teaching students learning skills, giving feedback to students and making

demands and expectations clear to them. One way of making demands and expectations clearer may

be to align objectives, teaching and assessment in courses. Welfare structures for reducing stressors

and preventive interventions for reducing negative stress reactions might additionally be beneficial.

30
 General discussion

In the upper secondary school sample, gender, promotion focus, students’ self-reported

ability to solve problems on their own and literacy problems were significant predictors for GPA in

the final year of lower secondary education. Previous grades and deep and strategic learning

approaches were significant predictors for higher GPA at university. Student status after two years at

university was predicted by age, surface learning, study time during the first year, type of study

programme enrolment and GPA. Deep and strategic learning approaches did not predict student

status two years later. Literacy problems and procrastination failed to predict both GPA and dropout

in the university sample.

Previous research was confirmed by the predictive value of gender (Alexander et al., 1997;

Battin-Pearson et al., 2000; Casillas et al., 2012; Freudenthaler et al., 2008; Legewie & DiPrete, 2012;

Pomerantz et al., 2002), current educational choice (Rojewski & Kim, 2003), motivation (Richardson

et al., 2012) and LP (Reiff, Hatzes, Bramel, & Gibbon, 2001) for GPA in the upper secondary sample.

Deep and strategic learning approaches have also previously been associated with academic

performance (Diseth, 2007; W. A. Reid et al., 2007), as we found in Paper II. The relationships found

in the learning approach literature are typically weak to moderate, corresponding to our findings.

The significance of GPA for dropout confirmed previous research (Araque et al., 2009; Jia &

Maloney, 2015; Paura & Arhipova, 2014; Xuereb, 2014). In Tinto´s (1975) theoretical model of

dropout, two integration paths are identified. In the academic path, the commitment is directed

towards goals, while in the social path it is directed towards the institution. Institutional commitment

is influenced by peer-group and faculty interactions, which lead to social integration. Goal

commitment is influenced by intellectual development and grade performance, which lead to

academic integration. Both endpoints influence the dropout decisions. The predictors examined in

this thesis support the importance of academic integration. Performance and time spent studying

decreased the odds of dropping out after two years.

31
 Other findings were unexpected. For example, we anticipated that parental educational level

would be a significant predictor for both GPA and dropout (Alexander et al., 1997; Casillas et al.,

2012; De Witte & Rogge, 2013; Lundetræ, 2011; Sirin, 2005; Strom & Boster, 2007). The non-existing

importance of parental education in our data suggests that the significance of social background on

academic outcomes may be weaker in Norway than in countries with greater income gaps and less

social equality (Lundetræ, 2011; Wilkinson & Pickett, 2010). This is not surprising given that Norway

has a long history of policies and regulations which favour equality at many levels, including

education, work and health systems.

Literacy problems have previously been associated with both GPA (Gerber, 2012) and

dropout (Dunn et al., 2004; Goldstein & DeVries, 2011; Wagner et al., 2005). We expected LP not

only to predict educational outcomes in the upper secondary sample, but also in the university

sample. We also anticipated that procrastination would be of significance for both GPA and dropout,

as it is associated with a range of aspects likely to influence learning and studying (Steel, 2007).

Furthermore, we hypothesised that learning approaches would predict dropout. There is a

positive association between deep and strategic learning approaches and higher GPA (Duff et al.,

2004; W. A. Reid et al., 2007), and a negative association between the surface learning approach and

GPA (Diseth, 2007; Duff et al., 2004; W. A. Reid et al., 2007). Moreover, GPA predicts dropout. Thus,

we anticipated that students scoring high on the deep and strategic approaches would have

decreased odds of dropping out, but our findings indicated no such relationship. Regarding the

surface learning approach, we expected it to increase the odds of dropping out, but found the

opposite. Finally, we expected that reported depression or anxiety symptoms or mental health

status would predict GPA (Esch et al., 2014; Jackson, 2009). However, mental health did not

contribute to GPA neither in the upper secondary sample nor in the university sample. The

association between mental health problems and academic performance has also previously been

shown to be low (Richardson et al., 2012).

32
 This thesis thus provides four new insights to the research on learning and educational

performance.

1) Reported literacy problems predict lower grades among upper secondary school students,

but they do not predict neither GPA nor dropout in the university sample. This may indicate that

students with literacy problems perform equally well as their peers at university, but not in lower

secondary school. This pattern is unexpected because of the increased demands for reading and

writing at a higher level – and often even in another language – at university. One explanation for the

lack of association in higher education may be that students with the most serious literacy problems

do not enter the university at all, so that the university sample includes fewer students with

considerable levels of literacy problems. In the upper secondary sample, we used a response format

with yes/no, while we changed this to a Likert scale (1-5) in the university sample, so this hypothesis

is unfortunately difficult to test properly. Another explanation may be that students are acquiring

more strategies to cope with their problems as they grow and continue their education, either by

themselves or with the assistance and support of teachers. A third explanation may be that students

in higher education choose their educational path more freely, making it possible to not choose

subjects or study programmes with the most challenging curricula and assignments.

2) The predictive value of problem solving abilities on GPA was moderated by LP, as more LP

weakened the link between problem solving and GPA. Another and perhaps more useful

interpretation is that problem-solving abilities weaken the link between LP and GPA. Problem solving

may be related to self-efficacy, which has previously been shown to be lower in adolescents with

learning difficulties (Hen & Goroshit, 2014). Strengthening self-efficacy may thus reduce the negative

effect LP has on academic performance. Teachers should continue to help students with LPs, but

combining this effort with promoting self-efficacy may prove to be advantageous.

3) Having a strategic learning approach mediated the well-established association between

deep learning and academic performance. Having a deep approach and a wish for understanding and

33
true learning perhaps motivates students to manage time and resources well in order to learn more.

Previously, Diseth (2002) has found interaction effects between deep and strategic approaches,

strengthening the hypothesis that both are important for performance. However, in our study, the

effect of deep approach on GPA was not dependent on the level of strategic approach – the strategic

approach simply transferred the benefits of the deep approach through a mediation effect

4) Deep and strategic learning approaches predicted higher GPA at first year at the university,

but not dropout in the second year. On the other hand, a surface learning approach predicted

dropout, but not GPA. This is perhaps a counterintuitive finding. Diseth (2002) found a curvilinear

relationship between the surface approach and performance which may contribute to a better

understanding of the association between surface approach and GPA. A low to moderate level of

surface approach may be beneficial for performance, although this does not explain why higher

levels of surface approaches predict staying. However, this curvilinear relationship was not found in

our data. It may be that the explanation is more closely connected to the course design and learning

objectives. Alignment between the elements in a study program and in the specific courses is

important. Biggs (1999) suggests that a lack of alignment may cause the students to adopt more of a

surface approach to learning. While we regard a deep approach to learning as the ideal approach for

understanding and comprehending a topic, courses where the goals and expectations are unclear

and the elements do not correspond may push the students towards the surface approach. They try

to cover the entire curriculum and do not know how to separate out the important things to learn.

Most study programmes consist of several smaller courses each semester, and to complete a course,

a minimum of one exam is needed. A surface approach may thus be beneficial in order to pass

exams, even if it has the cost of a weaker understanding of the topics and potentially lower exam

grades.

A lack of consistency between elements in a course may also contribute to the students’

performance stress, because it increases the perceived workload. The suppression effect that we

34
fund of surface approach on performance stress supports this interpretation. In addition, a surface

approach is characterised by a fear of failure, which is also likely to contribute to performance stress.

Total variance explained by GPA in the university sample was 23% and total accuracy for the

prediction of dropout was 74.5% (Nagelkerke R2 = .31). This is similar to other studies in the field.

Nevertheless, a lot of variance in both outcomes remains unexplained. We still need to investigate

other factors to understand university dropout better. Academic performance is the most important

factor for dropout, both in this study and in previous research. However, it may be relevant to

separate students who drop out because of low academic performance from students who quit for

other reasons, perhaps as a more voluntary action (Tinto, 1975). Recognising these two groups of

students may help identify predictors that are concealed when examining all dropout students as one

group.

In the upper secondary sample, 55% of the total variance in GPA was explained, while this

percentage was only 23% in the university sample. This difference between the two samples may

indicate that predicting academic performance is an even more complex issue in higher education

than in upper secondary school.

Implications

Predictors of educational performance at different levels are a mix of relatively fixed factors,

such as gender and literacy problems, and changeable factors, such as learning approaches,

regulatory focus, effort and performance. Several of these factors may function as warning flags

which schools and educational institutions should be aware of when considering whether special

actions to assist at-risk students should be taken. Knowing about such predictors is beneficial when

distributing resources, allowing teachers to focus more closely on students who may be at risk. A

student with poor grades, or a student with problems related to reading and writing, should be

monitored more closely than other students. However, to be able to act on such flags requires a

35
close follow-up of students, which is perhaps feasible in upper secondary school, but probably not in

higher education. Interventions targeting the whole student population may therefore be more

useful in higher education. Promoting active learning and understanding, teaching students study

techniques which work and helping them to identify the strategies which do not work very well and

to beat procrastination tendencies will be beneficial for most students, and certainly not harmful.

Such teaching also need not be time-consuming or resource tapping.

Barr and Tagg (1995) wrote about a paradigm shift twenty years ago. They described the old

paradigm as primarily concerned with instruction, whereas the new paradigm primarily focuses on

learning. The main difference is that the former considers instruction – or teaching – as both a

method and a goal. Within this paradigm, success is measured by the number of students enrolled

and the amount of teaching happening at institutions. The latter, on the other hand, separates

concerns about methods from goals. Within this paradigm, teaching is simply one of many methods

for realizing goals, and hence success must be measured by what students learn.

Focusing on student learning as the aim and the goal makes the method of teaching a more

central issue. How should we teach in order to most successfully help students learn? We have

showed that deep and strategic learning approaches predicted higher performance for first year

students in higher education. One concept promoting the deep learning approach is active learning

(Phillips & Graeff, 2014). Active learning represents educational methods where practice is an

important aspect. Active learning methods require students to not only listen, but to read, write,

discuss and be engaged in problem-solving activities (Bonwell & Eison, 1991). This includes analysing

and synthesising material in addition to doing meta-cognitive tasks and evaluating the learning

progress. These skills help students think about their own learning process and enable them to

change study habits which are not working optimally. A recent meta-analysis by Freeman et al.

(2014) examined 225 studies comparing active learning to traditional lecturing in STEM (science,

technology, engineering and mathematics) courses. They found that student performance increased

36
(average examination scores went up by six percent) and fail rates decreased in courses using active

learning methods. It seems that the optimal way for a student to learn is to be actively engaged in

her own learning process and to seek a deeper understanding of the material when needed.

A range of factors contributes to explaining dropout in the literature, illustrating the

complexity of reasons. In a follow-up study of the university sample in this thesis in the autumn of

2014, one year after enrolment, we asked the participants for their reasons for taking a break or

quitting. Despite the low response rate in this follow-up, which is an important reason for not

analysing these data specifically as part of the current thesis, the qualitative answers illustrate

considerable diversity in reasons for quitting. These answers are shown in Table 2 in the Appendix. A

future goal of studies examining a variety of dropout factors should therefore be to integrate such

retrospective reports to improve the understanding of academic performance among university

students.

As teachers, our mission is to help students learn (Barr & Tagg, 1995). In this mission,

integrating knowledge of how students learn best should be a natural and logical part of teaching.

Teaching students adequate study skills is important to help them master the transition between

upper secondary and higher education and to improve the learning process during education. In

addition, many occupations require that employees keep professionally updated through continuous

self-education. This may be hard to do when they no longer have university teachers and curricula to

tell them what to read and learn. Students need to learn how to learn to be better prepared for their

formal education, but this is also important for their ability to function well in future jobs which are

complex and require constant acquisition of new knowledge and skills.

The findings of the present thesis point teachers towards focusing on some specific issues.

However, these issues are not necessarily the same at every educational level. In upper secondary

school, students with literacy problems and/or low self-efficacy and students who have been in

conflict with classroom rules should be monitored more closely. At university, literacy problems

37
seem to be a smaller problem. Here, teachers should focus on promoting deep and strategic learning

for the students and encouraging them to put time and effort into studying from the beginning.

Strengths and limitations

A key strength of this thesis is that it includes registry and survey data on students both in

upper secondary school and at university. Strengths and limitations of the two studies are discussed

within the three papers. The most important issues are the following:

A major strength of the Young Will study is the response rate and the large sample size

relative to the population of Tromsø. Among the students entering upper secondary school in the

autumn of 2010 in Troms County, 2/3 responded to our questionnaire, which is good given the low

response rates that current population studies generally face. The broad selection of variables

measured is also a strength compared to previous research within this area. The main limitation is

the lack of access to grades from upper secondary school. Yet, we did have access to the registered

students’ GPA from lower secondary level (although data were missing for about one fifth of the

students), which they received two to three months prior to data collection. This implies that we are

making a backwards prediction. However, the time frame was short, and GPA is a fairly stable

measure which also predicts later GPA (Casillas et al., 2012).

In the Learning in Higher Education study, we also had registry data on GPA. In this study, we

had access to a national database (DBH), which allowed us to collect grades also from students

migrating to other institutions in Norway. The response rate was however poor as only about 10% of

the invited students responded. On the other hand, a strength is that we had access to the gender,

age and incoming GPA of all invited students, making it possible to compare the participants to non-

participants. As these differences were minor, the low response rate does not seem to have

considerably biased the results with regard to central demographic variables and previous academic

achievements. In addition, even though a low response rate may introduce bias when estimating

38
means or prevalence estimates, the problem is less pronounced in terms of covariance structures like

correlations and regression coefficients (Stormark, Heiervang, Heimann, Lundervold, & Gillberg,

2008). Furthermore, students from all faculties of the university were included in the study,

increasing generalizability.

Using self-report measures may represent a limitation because of response bias among

participants (Cheung & Rensvold, 2000). Participants may answer in ways that are socially desirable,

and we do not know whether the answers correspond to actual behaviour. Another possible bias is

acquiescence, i.e. systematically higher or lower responses. However, the influence of acquiescence

is normally small (Friborg, Martinussen, & Rosenvinge, 2006) and several of the measures included

items that were reversed, minimizing this effect.

The participants in the upper secondary sample were informed that they could withdraw at

any stage; however, the classroom setting may have made such a decision difficult to make. The

relatively long questionnaire and the potential social cost of withdrawing could have influenced the

quality of the answers given.

We used a web-based questionnaire in the university sample. These seem to yield the same

results as paper-based questionnaires (Gnambs & Kaspar, 2016; Gosling, Vazire, Srivastava, & John,

2004), and have several advantages. Missing data are easy to sidestep by making questions

mandatory to answer in order to proceed. Participants may use their device of choice and complete

the questionnaire whenever they have Internet access. When the questionnaire has been completed,

the participant need only push the send-button – there is no paper work to post or deliver. Web-

based questionnaires are also easier and cheaper to distribute to potential participants and answers

are automatically stored and transferred to a database, avoiding scanning and punching mistakes.

One possible limitation to web-based questionnaires is a lower rate of completed questionnaires

(Heiervang & Goodman, 2011). Other limitations mentioned in the literature include biased samples

39
and multiple responses (Heiervang & Goodman, 2011; Wright, 2005), but these claims have largely

been repudiated (Ekman, Dickman, Klint, Weiderpass, & Litton, 2006; Gosling et al., 2004).

Future research

The LP scale we tested had good psychometric properties. Further validation of the scale

would be constructive for future use, especially to test the two items with less favourable properties.

The lack of predictive value of LP on learning outcomes at university may be better explained

through prevalence studies at the different levels of education and through closer examination of

those students scoring high on LP.

Since a large portion of individual differences in academic performance in university students

was unexplained in the research conducted in this thesis, other factors which may influence the

learning process should be investigated. Data on executive functions, alcohol- and drug use and sleep

would perhaps shed more light on the relationships between learning approaches and learning

outcomes.

Reasons for quitting vary, and while some students drop out because of low grades or lack of

academic integration, others quit voluntarily. In future research, trying to separate these groups will

be beneficial for identifying predictors which better fit each group. More information about factors

related to teaching methods, study practices and types of exam in specific study programmes may

also contribute to a better explanation.

Conclusions

The benefits of a population with adequate education include a better national economy

because of higher tax income as well as fewer expenses for welfare, prison, hospitals and so on. The

benefits for the individual include a more stable work situation, better income, better health and

longer life-span. These benefits have been acknowledged for a long time. Norwegian adolescents

40
have the right to attend upper secondary school because of these advantages. We also have a low

threshold for entering higher education and a principle that education should be free of charge.

Equal opportunities for all to start an education are thus good. Still, the dropout rates are high, both

in upper secondary and in higher education, even after several years with a broad focus on the

problem from academia, the government, teachers and school administrations. Extensive research

on different aspects of educational attainment and performance is central in order to identify why

education matters so much and how to secure adequate levels of education for the population.

There are some fixed factors which are important for academic performance and student

status, and these may serve as flags or signals for teacher response. Literacy problems constitute

such a flag at lower levels of education, but not necessarily at the university level.

Interventions for the changeable factors do not need to be offered only to individuals. They

can be offered to large groups of students, as well. Helping all students learn which strategy works

best in any given situation is beneficial for helping all student study effectively. This thesis highlights

several strategies and abilities that are beneficial for academic performance. These are a focus on

achievement rather than on avoiding failure, work motivated by a wish for a deeper understanding

of the material, a feeling of positive self-efficacy related to ability being able to solve learning tasks

and the ability to plan, organize and structure the study environment effectively. Educational

institutions should help promote these strategies to students, paying more attention to how students

learn, not only what they learn. Students who learn how to learn may thrive better in future jobs that

require the constant acquisition of more complex knowledge and skills. The future educational

system should therefore pay more attention to helping students attain good study habits which both

facilitate the quantity (study time) and the quality of the learning process. In these studies we have

identified some suggestions on where to start.

41
 References

Abouserie, R. (1994). Sources and levels of stress in relation to Locus of Control and Self Esteem in
university students. Educational Psychology, 14(3), 323-330.
doi:10.1080/0144341940140306
Alexander, K. L., Entwisle, D. R., & Horsey, C. S. (1997). From first grade forward: Early foundations of
high school dropout. Sociology of Education, 70(2), 87-107. doi:10.2307/2673158
Allison, P. D. (2012). Handling Missing Data by Maximum Likelihood. Paper presented at the SAS
Global Forum, Orlando, Florida. Paper retrieved from
http://www.statisticalhorizons.com/wp-content/uploads/MissingDataByML.pdf
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders: DSM-
5. Washington, D.C.: American Psychiatric Association.
Araque, F., Roldán, C., & Salguero, A. (2009). Factors influencing university drop out rates. Computers
& Education, 53(3), 563-574. doi:http://dx.doi.org/10.1016/j.compedu.2009.03.013
Arias Ortiz, E., & Dehon, C. (2013). Roads to success in the Belgian French community’s higher
education system: predictors of dropout and degree completion at the Université Libre de
Bruxelles. Research in Higher Education, 54(6), 693-723. doi:10.1007/s11162-013-9290-y
Bacon, D. R., & Bean, B. (2006). GPA in research studies: An invaluable but neglected opportunity.
Journal of Marketing Education, 28(1), 35-42. doi:10.1177/0273475305284638
Bandura, A., & Vandenbos, G. R. (1989). Human agency in social cognitive theory. American
Psychologist, 44(9), 1175-1184. doi:10.1037/0003-066X.44.9.1175
Barr, R. B., & Tagg, J. (1995). From teaching to learning — A new paradigm for undergraduate
education. Change: The Magazine of Higher Learning, 27(6), 12-26.
doi:10.1080/00091383.1995.10544672
Battin-Pearson, S., Newcomb, M. D., Abbott, R. D., Hill, K. G., Catalano, R. F., & Hawkins, J. D. (2000).
Predictors of early high school dropout: A test of five theories. Journal of Educational
Psychology, 92(3), 568-582. doi:10.1037/0022-0663.92.3.568
Baumeister, R. F., Heatherton, T. F., & Tice, D. M. (1994). Losing control: How and why people fail at
self-regulation. San Diego, CA, US: Academic Press.
Belfield, C. R., & Levin, H. M. (Eds.). (2007). The price we pay: economic and social consequences of
inadequate education. Washington D.C.: Brookings Institution Press.
Biggs, J. (1979). Individual differences in study processes and the quality of learning outcomes.
Higher Education, 8(4), 381-394. doi:10.2307/3446151
Biggs, J. (1999). What the student does: Teaching for enhanced learning. Higher Education Research
and Development, 18(1), 57-75.

42
Bonwell, C. C., & Eison, J. A. (1991). Active Learning: Creating Excitement in the Classroom.
Washington, DC: George Washington University.
Bowers, A. J. (2010). Grades and graduation: A longitudinal risk perspective to identify student
dropouts. Journal of Educational Research, 103(3), 191-207.
doi:10.1080/00220670903382970
Byrne, D. G., Davenport, S. C., & Mazanov, J. (2007). Profiles of adolescent stress: The development
of the adolescent stress questionnaire (ASQ). Journal of Adolescence, 30(3), 393-416.
doi:10.1016/j.adolescence.2006.04.004
Byrne, M., Flood, B., & Willis, P. (2002). The relationship between learning approaches and learning
outcomes: A study of Irish accounting students. Accounting Education, 11(1), 27-42.
doi:10.1080/09639280210153254
Campbell, C., & Cabrera, A. (2014). Making the Mark: Are Grades and Deep Learning Related?
Research in Higher Education, 55(5), 494-507. doi:10.1007/s11162-013-9323-6
Casillas, A., Robbins, S., Allen, J., & Kuo, Y.-L. (2012). Predicting early academic failure in high school
from prior academic achievement, psychosocial characteristics, and behavior. Journal of
Educational Psychology, 104(2), 407-420. doi:10.1037/a0027180
Chalabaev, A., Major, B., Sarrazin, P., & Cury, F. (2012). When avoiding failure improves performance:
Stereotype threat and the impact of performance goals. Motivation and Emotion, 36(2), 130-
142. doi:http://dx.doi.org/10.1007/s11031-011-9241-x
Cheung, G. W., & Rensvold, R. B. (2000). Assessing extreme and acquiescence response sets in cross-
cultural research using structural equations modeling. Journal of Cross-Cultural Psychology,
31(2), 187-212. doi:10.1177/0022022100031002003
Choi, N. (2004). Sex Role Group Differences in Specific, Academic, and General Self-Efficacy. The
Journal of Psychology, 138(2), 149-159.
Christakis, N. A., & Fowler, J. H. (2008). The collective dynamics of smoking in a large social network.
New England Journal of Medicine, 358(21), 2249-2258. doi:10.1056/NEJMsa0706154
Crocker, L., & Algina, J. (2006). Introduction to Classical and Modern Test Theory. Mason, OH:
Cengage Learning.
Dahl, T. I. (2006). When precedence sets a bad example for reform: conceptions and reliability of a
questionable high stakes assessment practice in Norwegian universities. Assessment in
Education: Principles, Policy & Practice, 13(1), 5-27. doi:10.1080/09695940600563579
de Ridder, D. T. D., Lensvelt-Mulders, G., Finkenauer, C., Stok, F. M., & Baumeister, R. F. (2012).
Taking stock of self-control: A meta-analysis of how trait self-control relates to a wide range
of behaviors. Personality and Social Psychology Review, 16(1), 76-99.
doi:10.1177/1088868311418749

43
De Ridder, K. A. A., Pape, K., Johnsen, R., Westin, S., Holmen, T. L., & Bjørngaard, J. H. (2012). School
dropout: a major public health challenge: a 10-year prospective study on medical and non-
medical social insurance benefits in young adulthood, the Young-HUNT 1 Study (Norway).
Journal of Epidemiology and Community Health. doi:10.1136/jech-2011-200047
De Witte, K., & Rogge, N. (2013). Dropout from secondary education: All's well that begins well.
European Journal of Education, 48(1), 131-149. doi:10.1111/ejed.12001
Derogatis, L. R., Lipman, R. S., Rickels, K., Uhlenhuth, E. H., & Covi, L. (1974). The Hopkins Symptom
Checklist (HSCL): A self-report symptom inventory. Behavioral Science, 19(1), 1-15.
doi:10.1002/bs.3830190102
Diseth, Å. (2002). The relationship between intelligence, approaches to learning and academic
achievement. Scandinavian Journal of Educational Research, 46(2), 219-230.
doi:10.1080/00313830220142218
Diseth, Å. (2007). Approaches to learning, course experience and examination grade among
undergraduate psychology students: testing of mediator effects and construct validity.
Studies in Higher Education, 32(3), 373-388. doi:10.1080/03075070701346949
Duckworth, A. L., & Seligman, M. E. P. (2006). Self-discipline gives girls the edge: Gender in self-
discipline, grades, and achievement test scores. Journal of Educational Psychology, 98(1),
198-208. doi:10.1037/0022-0663.98.1.198
Duff, A., Boyle, E., Dunleavy, K., & Ferguson, J. (2004). The relationship between personality,
approach to learning and academic performance. Personality and Individual Differences,
36(8), 1907-1920. doi:http://dx.doi.org/10.1016/j.paid.2003.08.020
Dunn, C., Chambers, D., & Rabren, K. (2004). Variables affecting students' decisions to drop out of
school. Remedial and Special Education, 25(5), 314-323.
doi:10.1177/07419325040250050501
Eisenberg, D., Gollust, S. E., Golberstein, E., & Hefner, J. L. (2007). Prevalence and correlates of
depression, anxiety, and suicidality among university students. American Journal of
Orthopsychiatry, 77(4), 534-542. doi:10.1037/0002-9432.77.4.534
Ekman, A., Dickman, P. W., Klint, Å., Weiderpass, E., & Litton, J.-E. (2006). Feasibility of using web-
based questionnaires in large population-based epidemiological studies. European Journal of
Epidemiology, 21(2), 103-111. doi:10.1007/s10654-005-6030-4
Embretson, S. E., & Reise, S. (2000). Item response theory for psychologists. Mahwah, N.J.: Lawrence
Erlbaum Associates, Inc.
Entwistle, N. (1997). Approaches and study skills inventory for students (ASSIST). Edinburgh: Centre
for Research on Learning and Instruction, University of Edinburgh.

44
Entwistle, N., Hanley, M., & Hounsell, D. (1979). Identifying distinctive approaches to studying.
Higher Education, 8(4), 365-380. doi:10.1007/BF01680525
Esch, P., Bocquet, V., Pull, C., Couffignal, S., Lehnert, T., Graas, M., . . . Ansseau, M. (2014). The
downward spiral of mental disorders and educational attainment: A systematic review on
early school leaving. BMC Psychiatry, 14, 237. doi:10.1186/s12888-014-0237-4
Feld, L. D., & Shusterman, A. (2015). Into the pressure cooker: Student stress in college preparatory
high schools. Journal of Adolescence, 41, 31-42.
doi:http://dx.doi.org/10.1016/j.adolescence.2015.02.003
Ferrari, J. R. (2001). Procrastination as self-regulation failure of performance: Effects of cognitive
load, self-awareness, and time limits on ‘working best under pressure’. European Journal of
Personality, 15(5), 391-406. doi:10.1002/per.413
Freeman, S., Eddy, S. L., McDonough, M., Smith, M. K., Okoroafor, N., Jordt, H., & Wenderoth, M. P.
(2014). Active learning increases student performance in science, engineering, and
mathematics. PNAS Proceedings of the National Academy of Sciences of the United States of
America, 111(23), 8410-8415.
French, M., T. , Homer, J., F. , Popovici, I., & Robins, P., K. . (2014). What you do in high school
matters: High school GPA, educational attainment, and labor market earnings as a young
adult. Eastern Economic Journal. doi:10.1057/eej.2014.22
Freudenthaler, H. H., Spinath, B., & Neubauer, A. C. (2008). Predicting school achievement in boys
and girls. European Journal of Personality, 22(3), 231-245. doi:10.1002/per.678
Friborg, O., Barlaug, D., Martinussen, M., Rosenvinge, J. H., & Hjemdal, O. (2005). Resilience in
relation to personality and intelligence. International Journal of Methods in Psychiatric
Research, 14(1), 29-42. doi:10.1002/mpr.15
Friborg, O., Hjemdal, O., Martinussen, M., & Rosenvinge, J. H. (2009). Empirical support for resilience
as more than the counterpart and absence of vulnerability and symptoms of mental disorder.
Journal of Individual Differences, 30(3), 138-151. doi:10.1027/1614-0001.30.3.138
Friborg, O., Martinussen, M., & Rosenvinge, J. H. (2006). Likert-based vs. semantic differential-based
scorings of positive psychological constructs: A psychometric comparison of two versions of a
scale measuring resilience. Personality and Individual Differences, 40(5), 873-884.
doi:http://dx.doi.org/10.1016/j.paid.2005.08.015
Frostad, P., Pijl, S. J., & Mjaavatn, P. E. (2014). Losing all interest in school: Social participation as a
predictor of the intention to leave upper secondary school early. Scandinavian Journal of
Educational Research, 59(1), 110-122. doi:10.1080/00313831.2014.904420

45
Gerber, P. J. (2012). The impact of learning disabilities on adulthood: A review of the evidenced-
based literature for research and practice in adult education. Journal of Learning Disabilities,
45(1), 31-46. doi:10.1177/0022219411426858
Gijbels, D., Van de Watering, G., Dochy, F., & Van den Bossche, P. (2005). The relationship between
students’ approaches to learning and the assessment of learning outcomes. European
Journal of Psychology of Education, 20(4), 327-341. doi:10.1007/BF03173560
Gnambs, T., & Kaspar, K. (2016). Socially desirable responding in web-based questionnaires: A meta-
analytic review of the candor hypothesis. Assessment, 1-17. doi:10.1177/1073191115624547
Goda, Y., Yamada, M., Kato, H., Matsuda, T., Saito, Y., & Miyagawa, H. (2015). Procrastination and
other learning behavioral types in e-learning and their relationship with learning outcomes.
Learning and Individual Differences, 37, 72-80.
doi:http://dx.doi.org/10.1016/j.lindif.2014.11.001
Goldstein, S., & DeVries, M. (2011). The changing face of LD and ADHD in late adolescence and
adulthood. In S. Goldstein, J. A. Naglieri, & M. DeVries (Eds.), Learning and Attention
Disorders in Adolescence and Adulthood: Assessment and Treatment (2 ed., Vol. 636).
Hoboken, NJ, USA: John Wiley & Sons, Inc.
Gosling, S. D., Vazire, S., Srivastava, S., & John, O. P. (2004). Should we trust web-based studies? A
comparative analysis of six preconceptions about internet questionnaires. American
Psychologist, 59(2), 93-104. doi:10.1037/0003-066x.59.2.93
Grau-Valldosera, J., & Minguillón, J. (2014). Rethinking dropout in online higher education: the case
of the Universitat Oberta de Catalunya. International Review of Research in Open & Distance
Learning, 15(1), 290-308.
Gregg, N. (2007). Underserved and unprepared: Postsecondary learning disabilities. Learning
Disabilities Research & Practice, 22(4), 219-228. doi:10.1111/j.1540-5826.2007.00250.x
Harriott, J., & Ferrari, J. R. (1996). Prevalence of procrastination among samples of adults.
Psychological Reports, 78(2), 611-616. doi:10.2466/pr0.1996.78.2.611
Hayes, A. F. (Writer). (2013). SPSS Process [Computer software].
Heiervang, E., & Goodman, R. (2011). Advantages and limitations of web-based surveys: Evidence
from a child mental health survey. Social Psychiatry and Psychiatric Epidemiology, 46(1), 69-
76. doi:10.1007/s00127-009-0171-9
Hen, M., & Goroshit, M. (2014). Academic procrastination, emotional intelligence, academic self-
efficacy, and GPA: A comparison between students with and without learning disabilities.
Journal of Learning Disabilities, 47(2), 116-124. doi:10.1177/0022219412439325
Hess, R. S., & Copeland, E. P. (2001). Students' stress, coping strategies, and school completion: A
longitudinal perspective. School Psychology Quarterly, 16(4), 389-405.

46
Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 52(12), 1280-1300.
doi:10.1037/0003-066x.52.12.1280
Hoge, R. D., & Coladarci, T. (1989). Teacher-based judgments of academic achievement: A review of
literature. Review of Educational Research, 59(3), 297-313. doi:10.3102/00346543059003297
Ibrahim, A. K., Kelly, S. J., Adams, C. E., & Glazebrook, C. (2013). A systematic review of studies of
depression prevalence in university students. Journal of Psychiatric Research, 47(3), 391-400.
doi:10.1016/j.jpsychires.2012.11.015
Jackson, M. (2009). Understanding links between adolescent health and educational attainment.
Demography, 46(4), 671-694. doi:10.1353/dem.0.0078
Jia, P., & Maloney, T. (2015). Using predictive modelling to identify students at risk of poor university
outcomes. Higher Education, 70(1), 127-149. doi:10.1007/s10734-014-9829-7
Jöreskog, K. G., & Sörbom, D. (2005). LISREL 8.7 for Windows [Computer Software]. Lincolnwood, IL:
Scientific Softwqare International, Inc.
Keyes, C. L. M., Eisenberg, D., Perry, G. S., Dube, S. R., Kroenke, K., & Dhingra, S. S. (2012). The
relationship of level of positive mental health with current mental disorders in predicting
suicidal behavior and academic impairment in college students. Journal of American College
Health, 60(2), 126-133 128p. doi:10.1080/07448481.2011.608393
Kuncel, N. R., Hezlett, S. A., & Ones, D. S. (2004). Academic performance, career potential, creativity,
and job performance: Can one construct predict them all? Journal of Personality and Social
Psychology, 86(1), 148-161. doi:10.1037/0022-3514.86.1.148
Lanaj, K., Chang, C.-H. D., & Johnson, R. E. (2012). Regulatory focus and work-related outcomes: A
review and meta-analysis. Psychological Bulletin, 138(5), 998-1034. doi:10.1037/a0027723
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer.
Legewie, J., & DiPrete, T. A. (2012). School context and the gender gap in educational achievement.
American Sociological Review, 77(3), 463-485. doi:10.1177/0003122412440802
Little, R. J. A., & Rubin, D. B. (2014). Statistical Analysis with Missing Data. Somerset, US: Wiley-
Interscience.
Liu, Y., & Lu, Z. (2011). Longitudinal analysis of Chinese high school student’s stress in school and
academic achievement. Educational Psychology, 31(6), 723-729.
doi:10.1080/01443410.2011.600245
Lockwood, P., Jordan, C. H., & Kunda, Z. (2002). Motivation by positive or negative role models:
Regulatory focus determines who will best inspire us. Journal of Personality and Social
Psychology, 83(4), 854-864. doi:10.1037//0022-3514.83.4.854
Lundetræ, K. (2011). Does parental educational level predict drop-out from upper secondary school
for 16- to 24-year-olds when basic skills are accounted for? A cross country comparison.

47
 Scandinavian Journal of Educational Research, 55(6), 625-637.
doi:10.1080/00313831.2011.555925
Luthar, S. S. (2006). Resilience in development: A synthesis of research across five decades. In D. C. D.
J. Cohen (Ed.), Developmental psychopathology, Vol 3: Risk, disorder, and adaptation (2nd
ed.) (pp. 739-795). Hoboken, NJ, US: John Wiley & Sons Inc.
Manly, C., & Wells, R. (2015). Reporting the use of multiple imputation for missing data in higher
education research. Research in Higher Education, 56(4), 397-409. doi:10.1007/s11162-014-
9344-9
Marton, F., & Säljö, R. (1976). On qualitative differences in learning: I-outcome and process. British
Journal of Educational Psychology, 46(1), 4-11. doi:10.1111/j.2044-8279.1976.tb02980.x
Maslach, C. (1982). Burnout: the cost of caring. New York: Prentice Hall.
Maslach, C., & Leiter, M. P. (2007). Burnout. In G. Fink (Ed.), Encyclopedia of Stress (Second Edition)
(pp. 368-371). New York: Academic Press.
Mayer, R. E. (2011). Applying the science of learning. Boston: Pearson.
McKnight, P. E., McKnight, K. M., Sidani, S., & Figueredo, A. J. (2007). Missing data: A gentle
introduction. New York: Guilford Press.
Mischel, W., Shoda, Y., & Peake, P. K. (1988). The nature of adolescent competencies predicted by
preschool delay of gratification. Journal of Personality and Social Psychology, 54(4), 687-696.
doi:10.1037/0022-3514.54.4.687
Moretti, E. (2007). Crime and the costs of criminal justice. In C. R. Belfield & H. M. Levin (Eds.), The
Price We Pay: Economic and Social Consequences of Inadequate Education. Washington D.C.:
Brookings Institution Press.
Muenning, P. (2007). Consequences in health status and costs. In C. R. Belfield & H. M. Levin (Eds.),
The Price We Pay: Economic and Social Consequences of Inadequate Education. Washington
D.C.: Brookings Institution Press.
Muthén, L. K., & Muthén, B. O. (1998-2010). Mplus User's Guide. Sixth Edition. Los Angeles, CA:
Muthén & Muthén.
Nonis, S. A., Hudson, G. I., Logan, L. B., & Ford, C. W. (1998). Influence of perceived control over time
on college students' stress and stress-related outcomes. Research in Higher Education, 39(5),
587-605. doi:10.1023/a:1018753706925
Norwegian Ministry of Education and Research. (2008). Degree structure and grading system.
Retrieved from https://www.regjeringen.no/en/topics/education/higher-education/Degree-
structure-and-grading-system/id491287/
OECD. (2013). Education at a Glance 2013: OECD Indicators Retrieved from
http://dx.doi.org/10.1787/eag-2013-en

48
OECD. (2015). Education at a Glance 2015: OECD Indicators Retrieved from
http://www.keepeek.com/Digital-Asset-Management/oecd/education/education-at-a-
glance-2015_eag-2015-en
Ottosen, K. O., Goll, C. B., & Sørlie, T. (2016). The multifaceted challenges in teacher-student
relationships: A qualitative study of teachers’ and principals’ experiences and views regarding
the dropout rate in Norwegian upper-secondary education. Scandinavian Journal of
Educational Research, 1-15. doi:10.1080/00313831.2016.1147069
Paura, L., & Arhipova, I. (2014). Cause analysis of students’ dropout rate in higher education study
program. Procedia - Social and Behavioral Sciences, 109, 1282-1286.
doi:http://dx.doi.org/10.1016/j.sbspro.2013.12.625
Phillips, M. E., & Graeff, T. R. (2014). Using an in-class simulation in the first accounting class: Moving
from surface to deep learning. Journal of Education for Business, 89(5), 241-247.
doi:10.1080/08832323.2013.863751
Pluut, H., Curşeu, P. L., & Ilies, R. (2015). Social and study related stressors and resources among
university entrants: effects on well-being and academic performance. Learning and Individual
Differences, 37, 262-268. doi:http://dx.doi.org/10.1016/j.lindif.2014.11.018
Pomerantz, E. M., Altermatt, E. R., & Saxon, J. L. (2002). Making the grade but feeling distressed:
Gender differences in academic performance and internal distress. Journal of Educational
Psychology, 94(2), 396-404. doi:10.1037/0022-0663.94.2.396
Putnam, R. D. (2015). Our kids: the American Dream in crisis. New York: Simon & Schuster.
Reegård, K., & Rogstad, J. (Eds.). (2016). De frafalne: om frafall i videregående opplæring - hvem er
de, hva vil de og hva kan gjøres? Oslo: Gyldendal akademisk.
Reid, G. (2009). Dyslexia: A Practitioner's Handbook, Fourth Edition: John Wiley & Sons, Ltd.
Reid, W. A., Duvall, E., & Evans, P. (2007). Relationship between assessment results and approaches
to learning and studying in year two medical students. Medical Education, 41(8), 754-762.
doi:10.1111/j.1365-2923.2007.02801.x
Reiff, H. B., Hatzes, N. M., Bramel, M. H., & Gibbon, T. (2001). The relation of LD and gender with
emotional intelligence in college students. Journal of Learning Disabilities, 34(1), 66-78.
doi:10.1177/002221940103400106
Richardson, M., Abraham, C., & Bond, R. (2012). Psychological correlates of university students'
academic performance: A systematic review and meta-analysis. Psychological Bulletin,
138(2), 353-387. doi:10.1037/a0026838
Rojewski, J. W., & Kim, H. (2003). Career choice patterns and behavior of work-bound youth during
early adolescence. Journal of Career Development, 30(2), 89-108.

49
Rouse, C. E. (2007). Consequences for the labor market. In C. R. Belfield & H. M. Levin (Eds.), The
Price We Pay: Economic and Social Consequences of Inadequate Education. Washington D.C.:
Brookings Institution Press.
Rubin, D. B. (1976). Inference and missing data. Biometrika, 63(3), 581-592.
Rubin, L. H., Witkiewitz, K., Andre, J. S., & Reilly, S. (2007). Methods for handling missing data in the
behavioral neurosciences: Don't throw the baby rat out with the bath water. Journal of
Undergraduate Neuroscience Education, 5(2), A71-A77.
Rumberger, R. W. (2011). Dropping out: why students drop out of high school and what can be done
about it. Cambridge, Mass: Harvard University Press.
Rumberger, R. W., & Lim, S. A. (2008). Why Students Drop Out of School: A Review of 25 Years of
Research. Santa Barbara, CA: California Dropout Research Project.
Salvatori, P. (2001). Reliability and validity of admissions tools used to select students for the health
professions. Advances in Health Sciences Education, 6(2), 159-175.
doi:10.1023/a:1011489618208
Schafer, J. L., & Graham, J. W. (2002). Missing data: our view of the state of the art. Psychological
Methods, 7(2), 147-177. doi:10.1037//1082-989X.7.2.147
Schaufeli, W. B., Martinez, I. M., Pinto, A. M., Salanova, M., & Bakker, A. B. (2002). Burnout and
engagement in university students. A cross-national study. Journal of Cross-Cultural
Psychology, 33(5), 464-481. doi:10.1177/0022022102033005003
Schaufeli, W. B., Salanova, M., Gonzalez-Roma, V., & Bakker, A. B. (2002). The measurement of
engagement and burnout: A two sample confirmatory factor analytic approach. Journal of
Happiness Studies, 3(1), 71-92. doi:http://dx.doi.org/10.1023/A:1015630930326
Schmeelk-Cone, K., & Zimmerman, M. (2003). A longitudinal analysis of stress in African American
youth: predictors and outcomes of stress trajectories. Journal of Youth and Adolescence,
32(6), 419-430. doi:10.1023/A:1025934301468
Schunk, D. H. (2014). Learning theories: an educational perspective (6th ed.). Harlow: Pearson.
Shah, J., Higgins, E. T., & Friedman, R. S. (1998). Performance incentives and means: How regulatory
focus influences goal attainment. Journal of Personality and Social Psychology, 74(2), 285-
293.
Sirin, S. R. (2005). Socioeconomic status and academic achievement: A meta-analytic review of
research. Review of Educational Research, 75(3), 417-453. doi:10.3102/00346543075003417
Spence, J. T., & Helmreich, R. L. (1978). Masculinity & femininity. Austin: University of Texas Press.
Spinath, B., Eckert, C., & Steinmayr, R. (2014). Gender differences in school success: what are the
roles of students’ intelligence, personality and motivation? Educational Research, 56(2), 230-
243. doi:10.1080/00131881.2014.898917

50
Statistics Norway. (1996). Education Statistics - Upper Secondary Schools 1 October 1994. Retrieved
from https://www.ssb.no/a/histstat/nos/nos_c328.pdf
Statistics Norway. (2012). Education statistics. Throughput of students in tertiary education,
2010/2011. Retrieved from http://www.ssb.no/english/subjects/04/02/40/hugjen_en/
Statistics Norway. (2015a). Students at universities and colleges, 1 October 2014. Retrieved from
https://www.ssb.no/en/utdanning/statistikker/utuvh/aar/2015-05-04
Statistics Norway. (2015b). This is Norway 2015. What the figures say. Retrieved from
https://www.ssb.no/en/befolkning/artikler-og-
publikasjoner/_attachment/237252?_ts=14f4a08c970
Statistics Norway. (2015c). Throughput of pupils in upper secondary education, 2009-2014. Retrieved
from http://ssb.no/en/utdanning/statistikker/vgogjen/aar/2015-06-04
Statistics Norway. (2016). Facts about education in Norway 2016 – key figures 2014. Retrieved from
http://ssb.no/en/utdanning/artikler-og-
publikasjoner/_attachment/253129?_ts=15259e11408
Steel, P. (2007). The nature of procrastination: A meta-analytic and theoretical review of
quintessential self-regulatory failure. Psychological Bulletin, 133(1), 65-94. doi:10.1037/0033-
2909.133.1.65
Steel, P. (2010). Arousal, avoidant and decisional procrastinators: Do they exist? Personality and
Individual Differences, 48(8), 926-934. doi:http://dx.doi.org/10.1016/j.paid.2010.02.025
Steingrímsdóttir, Ó. A., Næss, Ø., Moe, J., Grøholt, E.-K., Thelle, D., Strand, B., & Bævre, K. (2012).
Trends in life expectancy by education in Norway 1961–2009. European Journal of
Epidemiology, 27(3), 163-171. doi:10.1007/s10654-012-9663-0
Stormark, K., Heiervang, E., Heimann, M., Lundervold, A., & Gillberg, C. (2008). Predicting
nonresponse bias from teacher ratings of mental health problems in primary school children.
Journal of Abnormal Child Psychology, 36(3), 411-419. doi:10.1007/s10802-007-9187-3
Strom, R. E., & Boster, F. J. (2007). Dropping out of high school: A meta-analysis assessing the effect
of messages in the home and in school. Communication Education, 56(4), 433-452.
doi:10.1080/03634520701413804
Struthers, C. W., Perry, R. P., & Menec, V. H. (2000). An examination of the relationship among
academic stress, coping, motivation, and performance in college. Research in Higher
Education, 41(5), 581-592. doi:10.1023/a:1007094931292
Støren, L. A., Skjersli, S., & Aamodt, P. O. (1998). I mål? Evaluering av Reform 94: Sluttrapport fra
NIFUs hovedprosjekt (NIFU Ed. Vol. Rapport 18/98). Oslo.
Sæle, R. G., Sørlie, T., Nergård-Nilssen, T., Ottosen, K.-O., Goll, C. B., & Friborg, O. (2015).
Demographic and psychological predictors of grade point average (GPA) in North-Norway: A

51
 particular analysis of cognitive/school-related and literacy problems. Educational Psychology,
1-22. doi:10.1080/01443410.2014.998630
Tabachnick, B. G., & Fidell, L. S. (2013). Using multivariate statistics. Boston: Pearson.
Tait, H., & Entwistle, N. (1996). Identifying students at risk through ineffective study strategies.
Higher Education, 31(1), 97-116. doi:10.2307/3447710
Tangney, J. P., Baumeister, R. F., & Boone, A. L. (2004). High self-control predicts good adjustment,
less pathology, better grades, and interpersonal success. Journal of Personality, 72(2), 271-
324. doi:10.1111/j.0022-3506.2004.00263.x
Tinto, V. (1975). Dropout from higher education: A theoretical synthesis of recent research. Review of
Educational Research, 45(1), 89-125. doi:10.2307/1170024
U.S. Census Bureau. (2015). Who Votes? Congressional Elections and the American Electorate: 1978-
2014. Retrieved from
http://www.census.gov/content/dam/Census/library/publications/2015/demo/p20-577.pdf
Wagner, M., Newman, L., Cameto, R., Garza, N., & Levine, P. (2005). After High School: A First Look at
the Postschool Experiences of Youth with Disabilities. : The National Longitudinal Transition
Study-2 (NLTS2).
Warrington, M., Younger, M., & Williams, J. (2000). Student attitudes, image and the gender gap.
British Educational Research Journal, 26(3), 393-407. doi:10.1080/01411920050030914
Werner, E. E. (1993). Risk, resilience, and recovery: Perspectives from the Kauai Longitudinal Study.
Development and Psychopathology, 5(04), 503-515. doi:doi:10.1017/S095457940000612X
Werner, E. E., & Smith, R. S. (2001). Journeys from childhood to midlife: risk, resilience and recovery.
Ithaca, N.Y.: Cornell University Press.
Wilkinson, R., & Pickett, K. (2010). The spirit level. London: Penguin Books.
Wright, K. B. (2005). Researching internet-based populations: Advantages and disadvantages of
online survey research, online questionnaire authoring software packages, and web survey
services. Journal of Computer-Mediated Communication, 10(3), 00-00. doi:10.1111/j.1083-
6101.2005.tb00259.x
Xuereb, S. (2014). Why students consider terminating their studies and what convinces them to stay.
Active Learning in Higher Education, 15(2), 145-156. doi:10.1177/1469787414527395

52


Appendix
 Item 1
4 Item 2
4 Item 3
8 Item 4
Item 5
4
Item 6
Item 7
3 Item 8
4
7
4 5
2 4
4 1
6
4
4

Figure 2. Item characteristics curve for the eight items included in the LP scale.

5
Item 1
4
Item 2
1 Item 3
4 Item 4
Item 5
Item 6
Item 7
Item 8
6
4
2
3
4
4

7
4
4 8
4 4
Figure 3. Item information curves for the items included in the LP scale
Table 2.

Reasons given for quitting university studies, N = 44

Reason N

I have completed the subjects that I wanted to complete 16

I wanted to work instead 11

Economic reasons 6

Health-related reasons 6

I wanted to study something else 4

I don’t think studying is for me 4

Family-related reasons (caretaker, illness in family) 4

The subjects were not interesting enough 2

Social reasons 1

Lack of housing near campus 1

Other 9

(for instance: not accepted at the study programme of wish, military service, travel)