COLLEGE OF EDUCATION AND BEHAVIOURAL SCIENCES

DEPARTMENT OF EDUCATIONAL PLANNING AND MANAGEMENT

Course Inst. Wakgari Tasisa (Asst. Prof and PhD Cand. Educational Policy and Leadership
studies, AAU

Statistics in Education

3rd Year, EdPM Students@ EdPM

Addis Ababa, 2023
 CHAPTER ONE: THE NATURE OF STATISTIC METHODS IN EDUCATIONAL
RESEARCH IN
Learning Objectives
After completing this chapter, you should be able to
Illustrate the concept, nature and characteristics of educational
statistics
Demonstrate knowledge of statistical terms.
Differentiate between the two branches of statistics.
Identify types of data.
Stipulate level of measurement for each variable.
Explain how statistics can be used and misused in educational
research.
1.1. Nature, Concepts, Definitions and Basic Terms of Statistical
May 29, 2025Methods in Education
 In ancient times, the governments used to collect information regarding
the population and property or wealth of the state.
 This information was necessary to achieve political purposes of rulers
to assess the manpower and to introduces new taxes or levies.
In modern times, collection of data and analyzing them are not limited
to some particular sphere of human activity.
 Quantitative information is not only necessary to run the
government or non-government organization but also it is the
backbone of the study of
agriculture and education  political sciences
Biology and Psychology  economics,
 electronics  business and
May 29, 2025 Medicine and sociology  commerce
 The objective of statistics is to
 make an inference about a population of interest based on information
obtained from a sample of measurements from that population.
By statistical data we mean
 numerical statement of facts
while statistical methods will refer to
 complete information of the principles and techniques used in collecting
and analyzing such data.
The word statistics has been defined by several statisticians in two
different but at the same time two allied senses.
1.Singular sense- statistics as statistical methods.
2.Plural sense-statistics as statistical data.

May 29, 2025

 1. Statistics as Statistical data( plural)
it is equivalent to referring to numerical facts, figures or
statistical data, but not all numerical data are statistics.
These can be described using counts or
measurements.
 Example
 Count
number of teachers in a college by department.
Number of schools in a region.
Number of students in departments.
Number of hotels in towns.
Measurement
Age, height, temperature, weight, etc.
Statistical data may include numbers such as averages, proportions,
  To be statistical data, they must
be set (or sets) of information
show significant relationship among themselves
have been collected, compared, analyzed and interpreted in a
systematic way.
2. Statistics as a statistical methods ( Singular)
Literally mean methods that are applied for data collection,
analysis and interpretation of numerical observation.
To sum up, statistics in education can be defined as the science that
deals with the methods of
collecting, classifying, presenting, analyzing, comparing and
interpreting numerical data collected through light of any sphere of
May 29, 2025
Other functions of Statistics
1. To Present Facts in Definite Form
• The facts are given in a definite form. If the results
are given in numbers, then they are more convincing
than if the results are expressed on the basis of
quality.
• E.g. [1]Unemployment is increasing at faster rate, [2]
Population in 2004 would be 15% more as compared
to 1990.
2. Precision to the Facts
• Help in condensing the data into important
figures.
• In other words statistics helps in simplifying
complex data to simple-to make them
understandable[graph, diagram, average &
coefficients etc.]
 Statistics function Contd…
3. Comparisons
• Simplifying the data, it can be correlated as
well as compared.
• Comparison is one of the main functions of
statistics as the absolute figures convey a
very less meaning.
4. Formulation and Testing of Hypothesis
• The statistical methods help us in formulating
and testing hypothesis or a new theory.
• E.g. to study whether introducing good role
model children is effective in controlling
children’s behavior or not
 Statistics function Contd…
5. Forecasting
• Statistics helps to predict the future course of
action of the phenomena.
• We can make future policies on the basis of
estimates made with the help of Statistics.
• We can predict the demand for goods in 2005
if we know the population in 2004 on the basis
of growth rate of population in past. The
statistics help in shaping future policies.
Statistics function Contd…
6. Policy Making
• Statistics helps to frame favorable policies
• How much food is required to be imported in 2007?
• It depends on the food-production in 2007 and the
demand for food in 2007.
• On the basis of forecast the government forms the
policies about food grains, housing etc.
7. It Enlarges Knowledge
• Enables one to enlarge his/her horizon of
knowledge
• Widens his/her thinking and reasoning power.
• It also helps him to reach to a rational conclusion.
Statistics function Contd…
8. To Measure Uncertainty
• Future is uncertain, but statistics help the various authorities to
make correct estimation
• So the uncertainty could be decreased.
• As we have to make a forecast we have also to create trend
behaviors of the past, for which we use techniques like
regression, interpolation and time series analysis.
 1.2. Major Steps in Statistical Study
 A statistical investigation is an investigation conducted according to
the statistical technique (A search for knowledge).
The main steps in statistical investigation are

Collection of Data

Organization (classification) of
Data

Presentation of Data

Analysis of Data

Interpretation of Data
May 29, 2025
 1.3. Basic Terms in Statistical Methods in Education
1. Population and Sample
i. Population (universe)
is the collection of all items (individuals, objects, measurements)
in a particular study.
A population is the set of all measurements of interest to the
sample collector.
Population may be finite or infinite.
ii. Sample
 A sample is any subset of measurements selected from the
population.
numerical characteristic of the sample
May 29, 2025
calculated using the sample
May 29, 2025
 2. Statistic and Parameter
i. Statistic (not to be confused with statistics)
is a measure used to describe some characteristics of a sample
(that is numerical values obtained from a sample).
Such values are denoted by Latin letters such as:
X  Arithmetic mean of a sample.
S2  Variance of a sample.
S Standard deviation of a sample
ii. Parameter
 is a numerical value (measure) used to describe from
some
May 29, 2025 characteristics of a population.
 numerical characteristic of the population
calculated using the whole population
Such value are denoted by Greek letters such as
  Arithmetic mean of a population
2 variance of a population
  Standard deviation of a population
3. Census and Sample survey
i. Census
is a complete enumeration of the entire population.
 It is the method of collecting information from the population.
ii. Sample survey
 is an investigation that consists of only a portion of the population
May 29, 2025
(sample).
 a) Population and sample refer to the objects that are sources of
information.
b) Census and sample survey are methods or approaches employed
to collect data.
c) Statistic and parameter are results.
Univariate data set
Consists of observations on a single attribute
Multivariate data
Consists of observations on several attributes
Special case: Bivariate data
 the relationship or association between two variables (‘variate’ is
another word for variable.)
May 29, 2025
 4. Sample Frame
 is a list of entire population from which items can be selected to
form a sample.
 listing of all the units in the population
5. Variables and Attributes
i. Variables
The term variable is derived from the word ‘vary’ that means to
differ or change.
Hence, variable means the characteristic that varies, differs, or
changes from person to person, time to time, place to place, etc.
A variable refers to a quantity or attribute whose value varies from
one investigation to another.
May 29, 2025
 Types of variables
(a). Discrete variable data
Variables that are capable of taking only an exact value and not any
fractional value are termed as discrete variables.
is a variable, which assumes only specified values (usually integral
values) on its individual observation.
That is those results from counts.
All qualitative (categorical) variables are discrete.
Example- Number of customers, family size, number of inhabitants,
etc.
(b) Continuous variable data
 is a variable which can theoretically assume any real number (as a
value) such as decimals.
May 29, 2025
A continuous variable can assume all values between any two given
 Examples of continuous variables are the heights of adult men, body
temperatures of rats, and cholesterol levels of adults.
b. Attributes
are those which cannot be described numerically
these are qualitative (or categorical) characteristics.
Example 1.6. Color of eye, color of hair, nationality, sex, marital status,
etc. are attributes (or attribute values).
1.4. Pilot Study (Test)
 A pilot study is a mini-version of a full-scale study or a trial run done in
preparation of the complete study.
called a ‘feasibility’ study.
It can also be a specific pre-testing of research instruments, including
May 29, 2025 questionnaires or interview schedules.
  Wolman and Kruger (1999:146) listed the following three values
of a pilot study:
1.To detect possible flaws in measurement procedures (including
instructions, time limits, etc) and in the operationalisation of
independent variables.
2. to identify unclear or ambiguous items in a questionnaire
3. The non-verbal behaviours of participants in the pilot study
may give important information about any embarrassment or
discomfort experienced concerning the content or wording of items
in a questionnaire.

May 29, 2025

  Other advantages of pilot studies include the following
 can give advance warning about where the main research
project can fail.
 indicates where research protocols might not be followed.
 can also identify practical problems of the research procedure.
 indicates whether proposed method or instruments are
inappropriate or too complicated.
 Pilot studies can be both Validity and Reliability of the research
instruments.
1. Validity ( Expert judgement)
 Validity is the degree to which an instrument measures what it

May 29, 2025 supposed to measure.

  It is the
 appropriateness,
 meaningfulness, and
 usefulness of the decisions researchers will make based on the
information they collect from their respondents using tests and
other instruments.
There are several ways to estimate the validity of a test including
content validity,
concurrent validity, and
predictive validity.

May 29, 2025

 Types of Validity
a. Content Validity
the most important type for most certification and licensure
programs is probably that of content validity.
Content validity is a logical process where connections between the
test items and the job-related tasks are established.
b. Concurrent Validity ( statistical- Latent variables)
Concurrent validity is a statistical method using correlation, rather
than a logical method.
Once the tests have been scored, the relationship is estimated
between the examinees’ known status as either masters or non-
masters and their classification as masters or non-masters (i.e., pass or
May 29, 2025
 The stronger the correlation is, the greater the concurrent validity of the
test is.
c. Predictive Validity
Another statistical approach to validity is predictive validity.
This approach is similar to concurrent validity, in that it measures the
relationship between examinees' performances on the test and their
actual status as masters or non-masters.
d. Face Validity
Like content validity, face validity is determined by a review of the items
and not through the use of statistical analyses.
Unlike content validity, face validity is not investigated through formal
procedures and is not determined by subject matter experts.
May 29, 2025
 To wind up, the validity of a test is critical because without sufficient
validity, instrument of data collection scores have no meaning.
2. Methods of Estimating Reliability
Reliability is the level of internal consistency or stability of the
measuring device overtime.
It is the degree of consistency between two measures of the same
thing.
Moreover, reliability is an index that estimates dependability
(consistency) of scores.
It is the prerequisite to validity because if you are not measuring
something accurately and consistently, you do not know if your
inferences are valid.
May 29, 2025
Classical theory of Reliability
 In the simplest case, we think of each observed
score as being made up of a “true score (T)” and
an “error score ( E)” such that the true score is
unobservable, and the term can be a bit
misleading .
The true score is the portion of the observed
score not affecting by random error.

By making the assumptions, we can use the

variance of a test or any measurement is
S2x= S2t + S2e………. 2

May 29, 2025

 S2t= variance of a group of individual’s true
scores
S2e = variance of a group of individual’s error
scores

Theoretically, reliability (rxx) is defined as the

ratio of the true score and observed score
variances.

rxx=s2t/S2o……… 3

The symbol rxx is used for reliability because of

May 29, 2025 so many reliability are computed by the Pearson

 xx

of the same trait.

Equation (1), (2) and (3) are basic formulas from
which most of the commonly written expressions
concerning reliability and the standard error of
measurement .
Standard error Measurement
is the standard deviation that would be obtained by a
series of measurements of individual scores.

S2x = S2t + S2o………… given

S2t = S2x - S2e…………….4
rxx= S2t/ S2x………………. 3

rxx= S2x - S2e/ S2x…………. Substitution

May 29, 2025 rxx= S2x/ S2x - S2e/ S2x
 To solve standard error of Measurement
S2t = S2x - S2e……………4
S2x= S2t + S2e
S2e= S2x - S2x rxx
S2e= S2x- S2e
S2e= S2x( 1- rxx)
rxx= 1- S2e/ S2x
Se= + √S2x(1-rxx)
rxx= 1- S2e/ S2x
Se= + Sx(√(1-rxx)
S2x rxx= S2x - S2e
This is an estimate of the measure of individual
variability.
Eg Supposed the test with the reliability of 0.90 has
the standard deviation of 15pont . Calculate the
May 29, 2025
standard error of the test.
 The reliability of a test is 0.75.
The standard deviation of observed scores is 10.0.
Compute the standard error of measurement

Estimates of Reliability
1. Measure of Stability
2. Measure of Equivalence
3. Measure of Internal Consistency
Split half
Kuder Richardson Estimates
Coefficient of Alpha

May 29, 2025

 1. Stability (Test-retest)
is a measure of reliability obtained by administering the same test
twice over a period of time to a group of individuals.
The scores from Time 1 and Time 2 can then be correlated in order
to evaluate the test for stability over time.
A coefficient of stability is obtained by correlating scores from the
same test of a group of individuals on two different occasions.
In general, as the time gap between measures increases, the
correlation between the scores becomes lower.
2. Equivalence (Parallel forms)
Parallel tests are defined as tests that have identical true scores
and identical person-specific error variances.
May 29, 2025
  Parallel tests must measure the same construct or underlying
trait.
 Parallel tests must measure the same construct or underlying
trait.
 The equivalent forms estimate of reliability is obtained by giving
two forms (with equal content, means, and variances) of a test to
the same group on the same day and correlating these results.
3. Internal Consistency
 It is the most common type of reliability since it can be estimated
from giving one form of a test once.
 There are three common types of internal consistency:
 Split-half method,
May 29, 2025
 Sampling items that are equivalent is very difficult.
Lack of cooperation, motivation, fatigue, and boredom by
examinees on a second administration of a test present additional
weaknesses in using these methods to estimate reliability.
a. The Split-Half Method
Split-half reliability is a statistical method used to measure the
consistency of the scores of a test.
As can be inferred from its name, the method involves splitting the
items into halves and correlating respondents' scores on the two
halves of the test.
Each test half should be of similar difficulty level.
May 29, 2025
 This technique should not be used with speeded tests.
This is because all students do not answer all items, a factor that
tends to increase the correlations between the items.
2 roe
rxx 
(1  roe )
Where roe = the Pearson correlation between the half-test scores
on the odd and the even items.
rxx= estimated reliability of the half test
The Spearman-Brown split half method assumes the two halves
have equal standard deviations.
Eg. If two halves of a test correlated 0.60, the estimated reliability
of the whole test will be the reliability coefficient of internal
May 29, 2025 consistency for a whole test whose split half tests are correlated by
 b. Coefficient Alpha
Cronbach's alpha is a measure of internal consistency, that is, how
closely related a set of items are as a group.
It is considered to be a measure of scale reliability.
A “high” value for alpha does not imply that the measure is
unidimensional.
Likert scale (e.g rate 1 to 5)
 Short answer
Partial credit
Alpha (α) = estimated reliability of the full-length test,
n = number of items
S2t = variance of the whole test (standard deviation squared)
Σs2i = sum the variance for all n items
May 29, 2025
 Table 2. George and Mallery’s (2003. p.221)scale
Cronbach’s Alpha Values Internal consistency
α > 0.9 Excellent ( Very high/ strong)
0.8 < α < 0.9 Very Good (High/ strong)
0.7 < α < 0.8 Moderate (Acceptable
0.6 < α < 0.7 Questionable
0.5 < α < 0.6 Poor
α < 0.5 Unacceptable

May 29, 2025

Two Major types of Statistics
Descriptive statistics: Inferential statistics:
• Devoted to summarization • using sample data to make
and description of data. an inference about a
• Consists of methods for population of data.
organizing and • drawing and measuring the
summarizing information. reliability of conclusions
• Includes methods like about population based on
Construction of graphs, information obtained from
charts, and tables, and the a sample
calculation of various • includes methods like point
descriptive measures such estimation, interval
as averages, measures of estimation and hypothesis
variation, and percentiles. testing which are all based
on probability theory.
Reasons Why Psychologists do basic statistics

1. Reasons-1:
• Statistics is deeply intertwined with research design.
• If you want to be good at designing psychological
studies, you need to at least understand the basics
of stats.
2. Reason-2:
• Understand reports of results presented in statistical
analyses in psychological literature/article.
• Understanding a certain amount of statistics crucial
for psychologists.
•

3. Reason-3:
• There is a big practical problem with being
dependent on other people to do all your statistics.
• Basic statistics knowledge and skills demanded in
almost any real life situation.
Basic terms in statistics
1)Data (plural) are measurement or observations. A
data set is a collection of measurement or
observations.
2)Datum (singular) is a single measurement or
observation and is commonly called as score or row
score.
3)Parameter is a value, usually a numerical value that
describes a population. A parameter is usually derived
from measurements of the individuals in the population.
4)Statistic is a value, usually a numerical value that
describes a sample. A statistic is usually derived from
measurements of the individuals in the sample.
5)Discrete variable consists of separate, indivisible
categories. No values can exist b/n two neighboring
categories.
6)Continuous variable, there are an infinite number of
possible values that fall b/n any two observed values. A
Scales of Measurement in Statistics
1.Nominal Scale
• Nominal means “having to do with names.”
• Classifying individuals into categories that have different
names but are not related to each other in any systematic
way.
• E.g. Classifying students according to his or her
major[sociology, Psychology, etc.]
• Allow us to determine whether two individuals are different,
but they do not identify either the direction or the size of
the difference.
• Classifying people by race, gender, or occupation is also
example of nominal scale measured variables.
• The room numbers are simply names and do not reflect any
quantitative information. Room 109 is not necessarily
bigger than Room 100 and certainly not 9 points bigger.
• The use of numerical values as a code for nominal
categories E.g. Male=0, Female=1 [the numerical values
are simply names and do not represent any quantitative
difference]
Scales of Measurement
Continued…..
2. Ordinal Scale
• Ordinal scale categories not only have different names but
also are organized in a fixed order corresponding to differences
of magnitude.
• Consists of a series of ranks (first, second, third, and so on)
• Categories are identified by verbal labels like small, medium,
and large drink sizes at a fast-food restaurant
• There is a directional relationship between categories.
• Helps to determine whether two individuals are different and
you can determine the direction of difference.
• Ordinal measurements do not allow you to determine the size of
the difference between two individuals.
• If Birhanu is placed in the low-reading group and Eyerus is
placed in the high-reading group, you know that Eyerus is a
better reader, but you do not know how much better.
Scales of Measurement
Continued…..
3. Interval Scale
• Used when the distance b/n observations is
measurable, equal, and ordinal, but a true zero point is
unnecessary.
• What is a true zero point? A true zero point occurs when
a scale cannot measure any observations
• Is quantitative, contains measurably equal distances
between observations, but lacks a true zero point.
• Basic mathematical operations[+, -, X, and division can
be performed on data collected from interval scales.
• Placement of zero on this scale does not mean that
information stops at that point; Temperature of 0°
Fahrenheit.
• Each degree entry on a thermometer is equally far
apart from every other degree,
• A temperature of 62° is objectively higher (and warmer)
than 60°, and the same 2° differences exists between
53° and 55°-temperatures that are in turn cooler than
the first pair.
Scales of Measurement---Interval
scale contd…
• Many personality scales, intelligence (IQ) measures,
educational tests, and rating scales use an interval scale.
• All tests, quizzes, and exams are based on an interval scale of
some type
• Obtaining a score of zero on a test of verbal skills, such
performance does not necessarily imply the complete absence
of relevant verbal ability?
• By interval data one cannot claim that one measure is twice or
three times the magnitude of another.
• 20° winter day in Ambo is not 3 times colder than a 60° day in
Adama,
• A score of 50 on a scale that measures depressive symptoms is
at greater risk for depression than a score of 25, but it would be
incorrect to say that one is twice as depressed as the other.
Scales of Measurement
Continued…..
4. Ratio Scale
• Incorporates all of the properties found in the previous
three scales, as well as an absolute zero point.
• It ranks observations, contains equal and meaningful
intervals, and has a true zero point.
• A zero point is meaningful because it indicates a true
absence of information.
• E.g. measurement on a ruler, miles reading on
speedometer/automobile no in motion
• Existence of a true zero point on a ratio scale enables
users to describe measurements in terms of numerical
ratios.
• Weight is a ratio scale, so that a 2-ton object is to a I-
ton object as a 4-ton object is to a 2-ton object (i.e.,
each is twice the weight of the other).
• Similarly, where height is concerned, a 6-ft person is
twice as tall as a 3-ft person.
Scales of Measurement---Summary

• Most scales used by behavioral scientists are

interval scales.

• Although nominal and ordinal scales are used

in research, they lend themselves to a
relatively small number of statistical tests.
 Scales of Measurement---Summary

•y