Chapter 1 Selection and Assessment

Chapter Overview
In this chapter we look at the selection system and at each of its component parts, seeing how they
interrelate and interreact. We look at the need to balance not only theoretical considerations, but practical
ones too. We also look at the fundamental issue behind selection the fairness and accuracy of our
methods. We then move on to examine the validity of the most well-established selection tools and
methods integral to the selection industry. These methods vary widely in their predictive validity that is,
their power to predict performance on the ob. !tudies on predictive validity need to be interpreted with
caution, since the performance criterion domain is still largely unclear and relatively understudied. "ost
research until very recently has concentrated on identifying and improving the predictive power of various
techni#ues, but this exercise is not theoretically framed. To this extent, we know very little about what
exactly each particular method is measuring. The more complex and multi-dimensional the method $for
example, interviews, biodata and assessment centres%, the less we know about what the tool is measuring
its construct validity. &ecause of the overwhelmingly 'technical( nature and preoccupation with selection
methods, this issue of construct-validity is relatively ill addressed. The atheoretical and incoherent nature of
selection is well illustrated in the biographical data domain. )esearchers are beginning to appreciate that
theoretical development is an essential step for progress to be made in the field of assessment and selection.
* relatively neglected area of research is on the more social psychological angle on selection. "ost of the
existing research, albeit sparse, is addressed to the micro-analytical influence of social processes on
selection bias $for example, non-verbal behaviour, candidate characteristics%. *n alternative, more macro-
analytic approach recogni+es the role of the selection system in managing the interface between employee
and employer as a social exchange process. This stance on selection has not been taken up in the literature
probably because it challenges the selection validation paradigm to address more dynamic considerations
beyond the issue of personob fit. The macro-analytical approach advocates the fundamental social nature
of the selection process and its role in negotiating the psychological contract of employment. ,n a similar
note, the role of the selection system in the formation of ustice perceptions is now being considered within
a new paradigm of research on applicant reactions. -ustice perceptions may be considered fundamental to
the content and character of the psychological contract formed between employee and employer.
The predictive model of assessment can be depicted as a logical and systematic process of eliciting
evidence on the basis of which a match to the ob and.or organi+ation can be udged $/igure 0%.
Figure 1.1 The selection system
Appendix 1: Job Analysis and Job Analysis Techniques
What is Job Analysis
-ob analysis can be described as 'any systematic procedure for obtaining detailed and obective information
about a ob, task, or role that will be performed or is currently being performed' $1arvey, 0220%. *lthough
Personnel or job specification (establish essential and
desirable traits)
Job analysis (establish job requirements) and identity
performance criteria (knowledge, skills, ability,
attributes)
Select a method for obtaining candidate evidence
atch evidence to criteria and select !best fitting"
candidate
#dentify biasing impact of
social processes and
minimi$e
%stablish utility of
procedures&
meta'analysis
cost'benefit
(andidate attraction
%stablish ultimate reliability and validity selection
procedures
M
i
n
i
m
i
z
e

b
i
a
s
:
)
s
s
e
s
s

r
e
l
i
a
b
i
l
i
t
y

a
n
d

v
a
l
i
d
i
t
y
*
m
a
+
i
m
i
$
e

b
o
t
h

d
u
r
i
n
g

i
m
p
l
e
m
e
n
t
a
t
i
o
n
essentially a ob analysis involves breaking down a ob into its component activities and re#uirements, it
can be conducted at several levels of analysis. Traditionally, ob analysis is conducted within a highly task-
oriented engineering tradition, describing ob obectives and technologies plus all associated physical,
mechanical and social demands.
The concept of job
Job analysis presupposes a job is a given rather than a concept or construct. Generally speaking, a job is a
set of tasks performed for a particular purpose that may or may not be bound up with a clearly defined role.
A role concerns how people interrelate with each other. A job is also different from an occupation, denoting a
set of similar jobs (for example, teaching, nursing). A job family, on the other hand, signals a group of similar
occupations (for example, public service professionals). Finally, a job may or may not be part of a career
involving a series of jobs (see Chapter 3).
The roles and tasks associated with a job can rapidly change. Also, managerial/professional jobs may
require the fulfilment of objectives as opposed to specific tasks, with interpretative licence in the way
objectives are translated into practice. In such instances, work can become more person-like (individually
defined and enacted) than job-like. Traditionally, job analysis does not account for variations in the overlap
between job and person factors. Many now argue that decomposing jobs into discrete tasks, skill and
attribute components isolates work from its broader context and that the appropriate analytic unit is the
workplace rather than the job.
3owadays, as well as deriving task information, ob analysis can identify the 'worker oriented( knowledge,
skills, abilities and personal attributes necessary to perform a ob successfully, that is, the processes
involved in doing the ob effectively. 1owever, in practice the difference between the two approaches is
not clear with considerable overlap in focus and levels of analysis. !ome also critici+e the worker-oriented
approach as having led to the development of too many ob analysis instruments. The issue is further
complicated by the use of specialist terminology.
!ome ob analysts argue that the two approaches should not be distinct and that systematic attempts should
be made to conceptuali+e and examine the interaction between ob, worker and the work environment.
The !ole o" Job Analysis
-ob analysis is multi-functional, playing a pivotal role in 1uman )esource "anagement $1)"%. !ome of
the maor uses of ob analysis information are as follows4
ob descriptions5
ob evaluation5
ob design and redesign5
specification of worker characteristics5
performance appraisal5
training design5
efficiency and safety5
workforce and human resource planning5 and
#uasi-legal re#uirements and obligations.
Choice o" #ethod o" Analysis
3owadays there are as many methods of ob analysis as there are uses, ranging from simple to
sophisticated, the latter being largely the province of specialists. It is generally agreed that no one ob
analysis techni#ue is likely to be suitable for all personnel purposes5 different methods are both
differentially effective and differentially practical for different human resource and personnel applications.
To ensure reliability and provide sufficient validity, researchers recommend a combination of methods
affording a 'customi+ed synthesis( of data. 6ach method will yield different types of information and a
different level of detail and depth.
7learly, practical considerations like costs and available time for both ob analyst and subect matter expert
$ob incumbent, supervisor% have an important bearing. 8oes one pursue a broad-based approach, gathering
information #uickly and cheaply using #uestionnaires, or pursue a more labour-intensive in-depth but
potentially costly approach9 There is also always the potential for overkill, 'a point of diminishing return at
which more fails to improve the : $accuracy and reliability of% : analysis( $8ipboye et al., 022;4 <=2%.
*lternatively, for legal reasons alone it could be argued, the more detail the better.
!ome methods are designed for a specific application. The -ob 6lements "ethod, for example, was
specifically designed to aid in the development of ob selection instruments. The -ob 7omponents
Inventory, on the other hand, was created primarily to assist in the development of vocational programmes
and career guidance. !ome methods have 'generic( human resource applicability, for example, the >osition
*nalysis ?uestionnaire, the *bility )e#uirement !cales, /unctional -ob *nalysis and the 7ritical Incident
Techni#ue. /urthermore, worker-oriented techni#ues can be used flexibly to make cross-ob comparisons
that are relatively insensitive to the task technologies involved in each ob.
Selection$relevant Job Analysis Techniques
Techni#ues range from 'simple( ob-information-gathering techni#ues yielding #ualitative information, and
'complex( more structured methods designed to elicit #uantitative information. The former include
observation, diary methods, critical incident techni#ues $7IT% and the repertory grid method, whilst the
latter include structured interviews, ob-specific #uestionnaires and task checklists and inventories. *nother
means of classification is by method of data elicitation4 observation, interviews and #uestionnaires. In task-
oriented observation, the ob incumbent is observed in the natural work setting, but the method is limited to
the analysis of overt worker behaviour including the procedures, e#uipment used, work aids and the
physical layout of the ob and work environment. It is often employed for studying obs involving repeated
manual operations.
The diary techni#ue, in contrast, re#uires ob incumbents to record their own work activities throughout the
working day, providing a rich source of ob information but affording limited control over what is recorded.
@eeping a diary moreover becomes yet another task demand and, if left unstructured, can be time-
consuming to analyse. It can be usefully applied nonetheless to the study of managerial and professional
obs.
Task analysis involves the compilation of a list or inventory of tasks associated with a ob$s% by analysing
relevant documentation $including ob descriptions%, observation and interviewing incumbents, supervisors
and subect matter experts $!"6s%. Task inventories can also generate worker-oriented information by
inviting the ob incumbent to rate each task item on a variety of scales $for example, fre#uency of
occurrence, relative time spent, difficulty level, importance% depending on the purpose of the analysis.
1owever, there is some debate about the reliability and validity of incumbent ratings. /or instance, ob
incumbents often exaggerate how many tasks they perform and how often and also conflate issues of task
difficulty with task dissatisfaction.
There are many pre-designed #uestionnaires available based on the task-inventory approach. These are
especially useful for comparing obs, since they employ standard sets of criteria regardless of the ob being
rated. ,ne of the most widely used worker-oriented commercial inventories applicable across all
occupations is the >osition *nalysis ?uestionnaire $>*?% $6xample 0% and the *bility )e#uirements
!cales $*)!% $6xample <%. The onus, however, is still on the selection specialist to select the most
appropriate tests.measures for the ob aptitude profile generated using worker-oriented techni#ues.
Positional Analysis Questionnaire example
The PAQ is a standardized measure containing 194 items built from a job components approach in which
job elements are linked to worker-oriented attributes. Items are organized into six main job dimensions
based on results from factor analysis across 2200 jobs, each of which is linked with an attribute profile
(information input, mental processes, work output, relationships with other persons, job context and the
other category). Each job dimension is broken down into several sub-divisions. For example, information
input unpacks into perceptual interpretation, input from representational sources, visual input from
devices/materials, evaluating/judging sensory input, environmental awareness, and use of various senses
(McCormick & Jeanneret, 1988).
As the level of analysis is fairly general, the PAQ has been criticized for failing to pick up on detail
(Aaamondt, 1996). The PAQ also presupposes a high reading level even in recent versions that have
attempted to rectify this. Also, many of the items in the PAQ are irrelevant for particular jobs, but a
disproportionately high number of do not apply responses can distort the analysis. The relative
inapplicability of the PAQ to white-collar jobs had prompted the development of an adapted inventory called
the Professional and Managerial Position Questionnaire (Page, 1988). Finally, due to the complexity of the
method, computer analysis is required to provide a rating for each of the six categories, an overall rating for
the job and a cognitive ability measure expressed as a General Aptitude Test Battery measure. The PAQ
does not claim to measure proficiency but is intended to provide a comprehensive evaluation of the content
of a job (.
The Ability Requirements Scales example
The ability requirements approach is a method for linking descriptions of job tasks with the generic abilities
required to perform them (Fleishman & Mumford, 1988). Specific tasks are said to require certain abilities,
and tasks requiring similar abilities can be so categorized.
The ARS were evolved from an identification of 50 abilities that consistently appear across jobs. Experts
developed seven-point behaviourally anchored rating scales on ability for each characteristic in the
taxonomy (Fleishman & Mumford, 1988: 922): For example, verbal comprehension denotes the ability to
understand English words and sentences (as opposed to verbal expression). A rating of 7 requires
understanding of complex, detailed information which contains unusual words and phrases, and involves
fine distinctions in meaning among words (for example, understanding a mortgage contract), whilst a rating
of 1 requires a basic knowledge of language necessary to understand simple communications (as in
understanding a comic book).
The ARS cover three domains: cognitive, psycho-motor, and physical abilities. Application involves
presenting the scales to a group of up to 20 so-called SMEs, coupled with job descriptions or a list of tasks
describing a job which are rated in terms of their ability requirements.
In contrast to these worker-oriented methods, /ine(s $02==% /unctional -ob *nalysis $/-*% focuses on task-
oriented behaviours specific to one job. The procedure begins with the generation of task statements that
are then rated in terms of how much of each of three functions $people, things and data% is perceived to be
involved and at what level of complexity. /or instance, against the 'people( function, 'taking instructions or
helping( is the lowest level of complexity indicated on the 'people( scale, whilst 'mentoring.counselling( is
the highest. ,ne advantage of the /-* is its simplicity, although it is also the simplicity of the classification
of ob functions that has been critici+ed $1arvey, 0220%.
Construct %alidity& Job !atin's and Job (erception
,ne of the problems with the practice of ob analysis is that little research has been conducted on the
reliability of ob incumbent information, on how ob ratings are made or on what influences their ob
perceptions $1arvey, 0220%. )esearch that does exist shows that ob analysis ratings are influenced by the
gender of the incumbent, tenure and performance level. "oreover, ob incumbent and supervisor often
depict the same ob completely differently, vary in their ratings of the same tasks and are also prone to say
that they are performing tasks they definitely do not.
,ne possible explanation for variations in the ratings made by !"6s is that the ob is truly heterogeneous
with respect to certain dimensions. Thus fire-fighting, for instance, may be very different in different areas
even within a single department, even though it is the same ob in principle. To account for this,
researchers are advised to identify commonalities across obs, whilst also being sensitive to true variations,
guarding against treating differences in perception as error. There is also some evidence that expert ob
analysts are not necessarily any better at providing accurate and reliable information than ob incumbents,
at least in the familiar ob domain. This may be due in part to the availability of common job
stereotypes. 7oovert $022A% proposed that ob knowledge is indeed organi+ed schematically, which
influences how ob information is assimilated, organi+ed and recalled.
-ob perceptions are central to understanding the products of ob analysis, yet few attempts have been made
to examine them. 1e proposes that we examine in more detail the way people perceive obs, in particular
how they encode, store and recall information about ob content and how they associate that content with
ob titles, attributes of obs such as worth, difficulty, complexity and interdependence, and attributes of
incumbents $knowledge, skills, abilities and other personal characteristics%. It should also be noted that obs
can be individuated over time such that the boundary between ob and person is no longer clear. !ome other
important issues for consideration in ob analysis research are presented below.
Research and practice in job analysis
An increase in the use of computer-based packages.
A greater emphasis on more generic job processes. For example, Cunningham and colleagues have
produced generic job profiles in both military and civilian contexts (Cunningham, Powell, Wimpe, &
Wilson, 1996). This could facilitate the flow of information and people across jobs, organizations and
occupations. Arguably there can be no pure description of a job or job family hived off from theories of
human performance. Harvey (1991) likewise notes that the future of job analysis rests in part on
building a taxonomy of job behaviours (for example, delegating, decision making) transcending
particular jobs or job families linking job behaviours to individual differences constructs (for example,
mental abilities, interests, personality traits).
Increases in team and group working, implying fuzzier distinctions between jobs and roles
Increases in the workplace use of technology, leading to changes in traditional job analysis dimensions
Increases in the diversity of the workforce (ethnic minorities, women, and so on) leading to an increase
in the importance and role of Equal Opportunities legislation.
Tas) Analysis
Task analysis has been described as dealing with 'issues associated with performance of human beings
interacting with tools:e#uipment and other human beings, and the world at large( $7hmiel, 022B%. Whilst
this definition may seem broad, it differentiates task analysis from other types of ob-based analysis by its
focus on performance5 thus, task analysis is concerned with what workers actually do and achieve, rather
than on the content of their obs. *pproaches to task analysis may take a variety of forms, including4
Taxonomic these are approaches that aim to describe various behaviours, or analyse
tasks.positions in a more rigorous way $for example, "c7ormick(s and -eannert(s$02BB% >*?, which
involves rating ob tasks in terms of psychological and contextual factors, for example, information
input, re#uired mental processes, work output, relationships and ob context%.
Co'nitive 1umancomputer interaction $17I% approaches focus largely on cognitive elements of
task performance, since these are the most crucial within such systems, for example, Task *nalysis for
@nowledge 8escription $T*@8%, and Coals, ,perators, "ethods and !election )ules $C,"!% $7ard,
"oran, D 3ewell, 02BA%.
*in) analysis 7hapanis $02=2% used link analysis to describe all possible links between tasks in a
target workplace.ob context together with the fre#uency with which they occur, weighting the links
for importance, and then combining weighting and fre#uency data to produce an indication of the
importance of each link within the target setting.
+ierarchical tas) analysis After *nnett and 8uncan $02EF%, this approach specifies system
goals $what the person is seeking to achieve%, operations $the activities re#uired to meet specified
goals% and plans $the conditions under which operations are carried out%. The techni#ue also draws on
the commonly used method of describing tasks in terms of hierarchical levels of significance, with
'packets( of subtasks making up each super-ordinate task within a given context. The approach is
described in detail by Wilson and 7orlett $022=4 0==0EE%.
6ach of these approaches to task analysis varied in its utility and validity depending on the nature of the
system and how it is intended to operate within the real world. * combination of a taxonomic approach $for
example, using the >*?% and an 17I perspective $for example, using T*@8 or C,"!% may provide
useful initial input to the turnkey design of a power station control centre, for example, where cognitive
functioning of operators is crucial, and computer systems must be as useable and interactive as possible to
provide the operators with the information they re#uire. * human.machine operated production line,
however, may benefit from the addition of link.linkage analysis in order that tasks may be weighted
according to importance within the overall production line system, with important task elements becoming
focus points for monitoring and.or safety features during design and construction of the system.
*s part of the task analysis process, the various skills or competencies re#uired of human operators to
successfully undertake each task will also need to be identified.
Appendix , -rom Job Analysis to (ersonnel Speci"ication
* ob description outlines the principal responsibilities and tasks associated with a particular ob, ordinarily
forming the basis for a contract of employment $although in practice, formal and up-to-date ob
descriptions are often difficult to find%, ob advertisement, recruitment, training needs analysis and
performance appraisal. -ob descriptions should be derived form ob analysis, although ironically ob
descriptions are often used as the starting point for ob analysis. Get research shows that ob descriptions
rarely reflect practical reality, particularly for professional obs that undergo continual evolution and
change $for example, 8iamond, &roadfinger, >ollach, D !ilver, 022=%.
!mith D )obertson $022A% identify six essential features of the ob description4
ob identification5
main purpose of ob5
responsibilities or tasks5
relationships with people5
physical working conditions5
pay and promotion.
-ob purpose and responsibilities are ideally written in the active voice in terms of outcomes or activities as
opposed to the personal characteristics re#uired for the ob $which is the >erson !pecification%. &oydell
$02FH% cautions against writing down every single thing that is done in the ob, recommending that ob
descriptions are relatively self-contained and with a definite end point. 1e also advises that tasks that have
a chronological order associated with them $that is, the order in which they are normally carried out% should
be outlined in the same order in the ob description. In some instances, it may also be necessary to divide
up activities into routine and non-routine, and also the working conditions $for example, 'a hot and noisy
environment(%. Tate $022;% critici+es the ob description as a poor way of introducing applicants to a ob
and advocates combining them with a realistic job preview $)->%, whereby applicants get to experience
the ob either vicariously $for example, on video% or first-hand $for example, by spending some time talking
to ob incumbents and shadowing them% before making a decision to take it on. ,n the other hand, even a
ob description is better than nothing at all, with some research demonstrating that vague, outdated or non-
existent ob descriptions can contribute to feelings of role ambiguity, ending up being a significant source
of stress to individuals.
The person or ob specification, on the other hand, specifies the knowledge, skills, abilities, experience,
and attributes or attitudes re#uired to perform the ob effectively. There is very little guidance to
practitioners on how to create a person specification, even though it is central to the production of selection
criteria. !o far there is no logical deductive way of transforming task characteristics into person
characteristics or behaviours. The issue is also not as straightforward as it may sound, given that 'task and
task components are not completely independent of the ob incumbent(. /or simple tasks, it is probably
fairly safe to assume that a defined set of abilities is necessary and sufficient for effective performance. /or
more complex tasks, however, the relationship between tasks and abilities may change, and one ability can
compensate for another. &oydell $02FH% advocates the drawing up of a grid of responsibilities and tasks
associated with a particular ob and then mapping the necessary knowledge, skills.abilities and social skills
$including attributes and attitudes% against them. The resulting matrix can then be content analysed to
identify the key knowledge, skills.ability and social skill areas.
,ne commonly used tool for generating person specifications is known as the 'F->oint >lan(, used to help
organi+e the information re#uired in person specifications. The seven points or criteria are as follows4
0. physical make-up5
<. attainments5
A. general intelligence5
;. special aptitudes5
=. interests5
E. dispositions5
F. home circumstances.
6ach set of information can then be classified as essential or desirable to the successful performance of the
ob.
>ersonnel specifications can often be critici+ed for depicting the ideal without any knowledge of how each
aspect of the proposed specification links with performance. There is a dearth of research on the production
of person specifications, their value in generating selection criteria or predictive validity. * taxonomy of
ob analysis methods produced by /leishman and ?uantance $02B;% could provide a useful basis for
identifying the kind of information re#uired to construct a ob description and person specification4 the
behavioural description $what exactly is done%, behavioural requirement $what exactly is re#uired%,
ability requirement $what abilities and attributes are re#uired%, and task characteristics $external to
the individual, like working conditions and tools%. Information on behavioural description and task
characteristics feed directly into the ob description, whilst behavioural and ability re#uirements feed
directly into the person specification. Ideally, then, for the purpose of constructing ob descriptions and
developing person specifications, ob analysis should yield information pertinent to all of the categories in
the taxonomy. *s an example, the case study below summari+es an exercise in producing a ob description
and person specification for the ob of 7opy )oom !upervisor from a class practical assignment in -ob
*nalysis.
Case Study 1.1: Job Analysis as a Basis for Producing a Job Description and a
Person Specification
Flynn (1996) conducted an analysis of the job of Copy Room Supervisor in the Admin Copy room of the
Reprographics Unit within a university Design & Print Centre. The units main functions are photocopying
and binding, production and print room support. The job incumbent and the Print & Reprographics Manager
provided input to the job analysis. A multi-level approach was used: task (specific tasks, procedures,
equipment, materials used), worker-oriented (work activities and behaviours) and personal characteristics
(knowledge, skills, abilities and personal attributes required for effective performance). The methods of
analysis selected were as follows:
Recorded observations and interview responses were pooled and then content analysed to produce a list of
tasks, activities and attributes. Fifty-one job statements were produced, organized under seven main
headings: planning and organizing, controlling and monitoring, decision making, administration, interaction
with others, equipment, security and safety. Job incumbent and manager independently rated each
statement on importance to effective operation of job and frequency of performance, using five-point scales.
It was found that the job incumbent and manager agreed overall in their ratings of importance and
frequency, although disagreements arose with respect to some administrative tasks and in how much
interaction with others was described.
1. ,bservation of job
incumbent
2. #nterview with job
incumbent
3. (onstruction of worker
activities questionnaire
and checklist
4. (ompletion of both
by incumbent and
manager
5. -ata analysis
6. Production of job
description and
person specification
The analysis was used as the foundation for the job description and person specification as follows:
JOB DESCRIPTION
Job Title: Copy Room Supervisor
Department: Design & Print Centre Unit: Reprographics
Line Manager: Print & Reprographics Manager
Staff: Two full-time and two part-time Copy Room Assistants
Hours of Work: Five days per week, 8.30am to 4.30pm
Job Purpose
Responsible for directing and controlling production of all photocopying and binding work within the Print &
Reprographics Departments copy rooms.
Main Duties and Responsibilities
Supervisory
Plans and organizes the three copy room schedules
Monitors production of all photocopying and binding work
Ensures work deadlines are maintained
Organizes delivery/collection of all completed work
Orders all photocopying materials and accessories required by copy rooms
Monitors copy rooms usage of materials and expenditure
Decides appropriate standards of copying and binding production
Maintains staff health and safety awareness in the use of machinery
Ensures security of copy room premises during hours
Overviews copy room staff performance
Updates Print & Reprographics Manager on copy room performance
Technical
Operates and tends Kodak computerized and standard bulk photocopiers
Operates fastback, comb and wire binding machines
Operates booklet maker, stitching and stapling machines
Operates guillotine and bulk shredder
Overviews training on newly-introduced equipment
Administrative
Maintains productivity reports for three copy rooms
Maintains records of supplies ordered
Maintains customer orders and payment records
Maintains staff details with respect to holidays, sickness, etc.
Customer-related
Calculates costs and provides quotes for all copying and binding equipment
Advises customer on their copying and binding requirements
Deals with customer complaints
PERSON SPECIFICATION
Knowledge
Familiarity with current photocopying and binding equipment
Appropriate comprehension and use of technical manuals
Knowledge of health and safety regulations relating to photocopying and binding equipment
Basic understanding of printing systems and services
Necessary Skills and Abilities
Effective communication skills: face-to-face and telephone
Basic clerical and administrative skills
Organize and manage time effectively and schedule work priorities
Operate well under pressure in warm and noisy working conditions
Commitment to quality
Value the importance of good customer service
Work well on own initiative as well as part of a team
Essential experience in photocopying and binding work
Essential managerial/supervisory experience
O level English and Mathematics or equivalent experience