MARIAN POWER
A QUICK
GUIDE TO
HUMAN
RESOURCE
TESTING
Marian Power qualified as a psychologist in 1973, and has been registered since 1987. She has
worked as an organisational psychologist, equal opportunity manager and human resources
manager in federal, state and local government. Her roles encompassed recruitment and
selection, conflict resolution and management of grievances. Marian is currently employed as
consultant psychologist with the Australian Council for Educational Research providing advice
to psychologists and human resource professionals regarding the selection of the most
appropriate assessment instruments for their particular purpose. She also provides
accreditation training in the use of restricted tests.
Marian is an active member for the Australian Psychological Society, the College of
Organisational Psychologists, the Australian Association of Psychological Type, the Australian
Human Resources Institute and the Australian Association of Careers Counsellors.
A QUICK GUIDE TO
HUMAN RESOURCE TESTING
Can you afford to make the wrong recruitment selection decision?
The cost of a wrong selection decision can be up to one-and-a half times the salary of the job,
let alone the time taken in the re-hiring process.
Studies have shown that appropriate assessment tools enhance the chances of making a
good selection and recruitment decision. Testing is also important to the human resources
(HR) practitioner in a variety of other contexts, including team building, change management
and ongoing organisational needs.
A Quick Guide to Human Resource Testing is designed as an introduction, a refresher and a
quick reference guide for HR practitioners who use or plan to use assessment instruments in
any context. It includes explanations, tips, case studies and suggestions to help you get the
most out of your HR testing.
9

National Library of Australia Cataloguing-in-Publication data:
Power, Marian.
A quick guide to human resource testing.
ISBN 0 86431 458 2.
1. Employee selection - Australia - Handbooks, manuals, etc.
2. Employment tests - Handbooks, manuals, etc. 3.
Employees - Recruiting - Australia - Handbooks, manuals,
etc. 4. Employee selection - Law and legislation -
Australia. I. Title.
658.3110994
Visit our website: www.acerpress.com.au
Acknowledgement
In writing this guide, I would like to acknowledge the support I received from my husband, Adrian; and
Dominic, Stephen and Caithlin who freed computer time for me and provided encouragement.
Ralph Saubern, Test Publisher at ACER Press, offered frequent support and constructive advice. My
colleagues in Organisational Psychology and Human Resource Management have shared their experience
over twenty-five years, which has been an invaluable addition to my formal learning. Thank you all.
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page ii
Contents
Introduction 1
Section 1:Test Selection and Administration
When to Use Tests
2
Recruitment/Selection
Organisational Development
Career Planning
Types of Tests 4
Ability and Aptitude Tests
Achievement Tests
Personality Assessments

Normal Population
Mean
Range
Standard Deviation
Statistics 2: Test Scores 18
Changing a Raw Score to a Standardised Score
Percentile Ranks
Stens and Stanines
T-scores
Statistics 3: Reliability, Validity and
Measurement Errors
20
Reliability
Validity
Margins of Error
Section III: Reporting and Feedback
Reporting Results
22
Report Types
Using Reports for Feedback
Feedback to Candidates
Feedback to Managers
Section IV: Ethical and Legal Issues
Test Security and Candidates’ Rights
24
Equal Employment Opportunity
(EEO) Legislation
Discrimination in Testing
Direct and Indirect Discrimination
Appendix I: Frequently Asked Questions 26

read further about how testing can help you provide the best possible human resource services.
In the Appendices there is a section on frequently asked questions, a glossary of terms that will clarify
any technical jargon, suggestions for which test to use in particular selection contexts, and an annotated list
of tests together with icons that indicate appropriate usage areas.
Marian Power
1
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 1
In human resource management, many decisions are made based on information that is gathered and
presented – for example, information about strategy planning, leave arrangements, observance of
occupational health and safety regulations. When dealing with important decision making regarding
people and the workplace, information should be gathered from all possible reliable sources.
HR practitioners use testing to collect reliable, objective information in order to optimise the decision-
making in a range of situations. Table 1 details some of these situations, which are further discussed below.
tests
When to use
Section I: Test Selection and Administration
2
Table 1: Using Testing in Human Resource Management
Recruitment/Selection Organisational Development Career Planning
Screening
Recruitment
Order of merit for ongoing selection
Staff development
Organisation development
Promotion
Team building
Change management
Career choice
Career change
Redundancy support

The tests used in this phase are chosen because
they assess abilities, skills and attitudes that are
clearly related to the selection criteria for the job in
question. Verbal reasoning, numerical reasoning,
work style preference, manual dexterity, spatial
reasoning, and personality measures are examples
of tests that are commonly used. (See page 4 for
more detail.)
From these test results, organisations that have
almost continuous recruitment needs can use
candidates’ performance to compile an order of
merit for further reference. This allows for
candidates who achieve scores above a certain
predetermined level to be invited to participate
further in the selection process as their position on
this ‘ladder’ is reached.
Organisational Development
When discrepancies are noted between the skills
or attitudes of current employees and the
requirements of the organisation, a range of testing
protocols is available to assist the HR practitioner
implement change. These are largely chosen from
the same pool of assessments available for initial
recruitment purposes, and should relate to specific
needs as they arise. For example, if an employer is
concerned that staff are experiencing difficulty with
new financial reporting requirements, numerical
reasoning tests may be given to identify those
candidates who would benefit from training to
address any gaps in their skills. This process needs

Exploring career paths that accommodate these
preferences is a positive outcome of the process.
In the area of succession planning, tests that
assess abilities and skills required in jobs at a higher
level in the organisation are popular in assisting
managers plan for advancement of their staff.
Sensitivity is required in the management of this
process so that employees see it as constructive.
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 3
Section I: Test Selection and Administration
4
Ability and Aptitude Tests
Ability tests involve questions that require complex
sets of mental processes and are designed to test a
candidate’s natural ability in a particular area. Often
ability tests explore relationships between two or
more words, numbers or pictures and ask the
candidate to extend a pattern or make an assertion
based on an understanding of the relationship.
Aptitude tests are similar to ability tests, but are
designed to give an indication of a candidate’s likely
successful future performance on the attribute that
is being assessed.
In selecting personnel, ability and aptitude tests
are widely used as good general indicators of
someone’s potential to perform the duties of the job
to a satisfactory standard, and to demonstrate an
ability to apply knowledge gained in new situations.
Non-verbal tests are also often used when selecting
staff for positions that demand skills not directly

less desirable. An applicant for the police force
whose personality profile indicated an aggressive
component could be considered highly unsuitable;
positions involving ‘cold calling’ in sales often
attract extroverted personalities.
tests
Types of
Tests can be classified in a number of ways. One option would be to classify them according to what they
are assessing. For example, tests may be assessing optimum performance (as in ability or aptitude tests) or
practical knowledge (as in achievement tests). Alternatively, they may be assessing emotional responses to
gain a picture of typical response patterns or to identify a person’s preferences, likes and dislikes.
The following are some of the major categories of HR tests.
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 4
Table 2: Ability and Aptitude Test Types and Selection Criteria
Ability/Aptitude Type Example Selection Criteria
Numerical reasoning
Numerical estimation
Numerical checking
Verbal reasoning
Verbal checking
Abstract/non-verbal
reasoning
Spatial-visual
reasoning
Mechanical reasoning
ACER Select – numerical
ACER Test of Employment
Entry Mathematics
MOST Numerical estimation
MOST Numerical checking

Because of the sensitivity involved in interpreting
personality assessment results and in providing
professional feedback, these assessments are usually
available only to psychologists or people who have
successfully completed prescribed accreditation
training. Accreditation training is usually available to
HR practitioners and other professionals on an
instrument-by-instrument basis. If a psychologist is
not available to assist, a structured interview and
referee comments may provide helpful sources of
information on a candidate’s personal style.
Example 3 is from a personality inventory.
Interest Inventories
Candidates are sometimes asked to complete
vocational interest inventories to assist in placing
them in the most appropriate job. These inventories
are most effective when the person answers honestly
to give the most accurate picture of themselves.
These assessments are very helpful in career
management programs, working with people facing
redundancy or those voluntarily changing career
direction. Example 4 is from a career interest
inventory.
4
3
2
1
I’d rather go to a party than read a book.
Strongly Agree 1 2 3 4 5 Strongly Disagree
Find the two missing numbers in the following sequence.

than questions.
Content Type
Most HR tests present test questions either in
written format or in pictorial form. Written test
questions can either use words (verbal) or numbers
(numerical); pictorial test questions can use a
variety of pictures, diagrams, mazes, maps and
visual puzzles. Examples of the different options for
test content are listed below.
Numerical (or Quantitative)
Numerical test questions require the use of
numbers and numerical symbols and concepts.
They can be used in a variety of test types, including
tests of ability, achievement and aptitude. Some
numerical items require calculations, some require
pattern recognition, while others require the
candidate to check for errors.
Examples 4 and 5 on page 7 are numerical items.
Verbal (or Linguistic)
Verbal test questions are based on words and
textual information. They can be used in a variety of
test types, including tests of ability, achievement,
aptitude and personality. Some verbal items require
word knowledge or logical reasoning, while others
require reading comprehension ability. Others are
simply written statements to which the candidate is
asked to respond by indicating their level of
agreement or disagreement.
Examples 3, 6 and 8 on page 7 are verbal items.
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 6

1
2
3
4
5
6
7
8
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 7
Tests vary in the way they are presented, although the basic elements of questions, answers, scoring and
interpretation/reporting are always present.
The test manual is the primary source of information about the test and often contains administration
guides, score keys, and tables for interpreting data.
formats
Test
Section I: Test Selection and Administration
8
Questions
Questions are usually presented in a test, item or
question booklet. Sometimes these test booklets
also include space for candidates to record their
answers; sometimes the answers are recorded on a
separate answer sheet. When answers are recorded
separately, it means the test booklets can be reused
by the next candidate. The administrator should
ensure that such reusable test booklets have not
been marked or damaged in any way and that all
are collected at the end of each test session.
Often questions are presented in a multiple-
choice format. In this format there is usually a stem

Another type of answer sheet is the carbonised
sheet. Once the candidate has finished recording
their answers, the examiner removes the top layer
and the carbon copy of the answers is designed in
such a way as to facilitate simple and immediate
hand scoring.
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 8
A Quick Guide to Human Resource Testing
9
Score Keys
A score key provides the examiner with information
needed to score a candidate’s responses.
Sometimes the examiner needs to count the
number of correct and incorrect answers to obtain
a raw score, in which case a list of correct answers
will be provided. Sometimes each answer is given a
different value and the values need to be added to
obtain the raw score.
Score keys come in a variety of formats:
• in the test manual
• as a separate card or clear plastic overlay
• on the carbonised section of the answer sheet.
More frequently nowadays, score keys are part of a
software system into which the examiner transfers
the candidate’s responses. These systems
automatically score the responses and provide a
report on the candidate’s results.
Interpretation/Reporting
Once the test has been scored, the raw score needs
to be transformed into a standardised score (see

standard score comparison is also described so that
the examiner may select the most appropriate
comparison sector. See page 18 for more
information about standardised scores.
Personality Interpretive Report Jon Sample
(continued…) 29 February, 2004
Anxiety
According to his responses, Jon Sample is no more or less anxious than most people. He has a
tendency to trust people and therefore may not be as vigilant as others in examining people’s
motives.
Factor Sten 1 2 3 4 5 6 7 8 9 10
Anxiety – general
Stable
Trusting
Assured
Tense
Figure 1: Example of a Computer-generated Report
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 9
Selection Criteria
It is critical that any test can be clearly demonstrated
to relate to one or more of the selection criteria.
This will ensure that the HR practitioner receives
more relevant information from the test session
and that the candidate appreciates the relevance of
that test session to the final hiring decision. As well,
there are important ethical and equal employment
opportunity (EEO) considerations, which require
the tests to relate directly to the selection criteria of
a job. (See page 29 for a list of example selection
criteria and appropriate tests.)

performance in the job? Is a person who is very
change-oriented required? These and many other
characteristics may be assessed using a number of
tools including personality assessments and
interest inventories.
The table in Appendix III lists some common
selection criteria and examples of the types of tests
that might be used.
There are two main job-related issues that you need to consider in choosing the right test or tests for your
selection exercise. Firstly, you should consider the selection criteria for the job and secondly, you should
consider the general level of the job.
the job
Choosing the most appropriate test–
Section I: Test Selection and Administration
10
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 10
Job Level
Some basic and entry-level jobs require no formal
education or training, although they may require
on-the-job training. Factory assembly line or shop-
floor roles are in this category. Tests at a lower level
that assess general reading ability, numerical
checking, and speed and accuracy are most
relevant here. Scenario 1 below illustrates a typical
example.
Another class of job involves making decisions
based on the understanding of written and/or
verbal communication. For example, a customer
service officer is required to listen to concerns and
make decisions for future action based on his/her

Candidates are unlikely to have sat
for any tests since leaving school,
and less than 1 hour is available
for this part of the selection
process.
Test selection
An applied reading test is
recommended for the verbal
comprehension component of the
job.This test is used in technical
trade environments where there
is a need to read and understand
a limited range of materials,
such as union and OH&S notices
and company requirements. A
numerical checking test, used to
determine speed and accuracy
when reading numbers, is
recommended for the other major
requirement of the job.
These tests will take around
45 minutes in total to administer.
A team leader in a customer service
environment is being recruited.
Sound communication skills,
together with the ability to monitor
sales figures and report concerns to
management, are required. A desire
to help customers is essential.
Test selection

strategies.
Tests will take less than
2 hours to complete.
A Quick Guide to Human Resource Testing
11
Scenario 1
Scenario 2
Scenario 3
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 11
Time
Tests vary considerably in their administration
time. The test publisher or distributor typically
provides brief descriptions of their assessment
tools in their catalogues, which include information
on the purpose, cost and time of administering
each test. Obviously, if you have just one hour
available for testing, you cannot choose a test that
takes 50 minutes but addresses only one of the
three criteria you wish to assess. The numerical test
described in Scenario 2 on page 11 takes only
8 minutes to complete, whereas those for the
management role in Scenario 3 take 30 minutes
each.
Getting the best mix of tests that covers the
broadest range of assessable criteria within the
time available takes some thought and planning.
Budget
The cost of a poor staff selection decision will
usually far outweigh the extra money spent in
purchasing a more appropriate recruitment test.

scanned and scored directly by computer (as well as
by hand). These answer sheets are called OMRs (see
also page 8). If the testing program involves a large
number of candidates – for example, a recruitment
screening exercise – it may be appropriate to have
the answer sheets computer scored. This is a service
offered by the test publisher or distributor. Other
tests that do not have OMR answer sheets can also
be scored by the test publisher or distributor on a
fee-for-service basis.
If the testing program involves a small number
of candidates, hand scoring is usually faster and
cheaper.
Qualifications
Table 3 above describes typical levels of
qualification required for test administrators.
Professional development programs are available
that accredit new users to basic levels, and qualify
more experienced users to interpret specific tests.
In most cases, qualified psychologists may use any
published test instrument, although a few products
require specialised training even for psychologists.
A Quick Guide to Human Resource Testing
13
Table 3:Test Levels and Administrators
Test Level Typical Tests Test Administrators
High
Medium
Low
• Personality instruments

→ Basic ethics in testing.
Outcomes for participants will be:
→ A thorough grounding in the principles of test
administration and scoring
→ The opportunity to learn and practise the
administration and scoring of ability and
personality tests
→ The skills to assist qualified test users in
administering and scoring tests, so as to free
their time for interpretation and decision making
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 13
Environment
Let us consider the optimal environment
for conducting a testing session. The room should
be large enough to accommodate comfortably
the anticipated number of candidates. Give
consideration to the ventilation, lighting and
expected external noise levels. A candidate will not
give their best performance if the room is too hot or
cold, too crowded or noisy, or if it is too hard to hear
your instructions or see you or your assistant
administrators.
If you are conducting online testing, it is
essential that all computers are checked and
working well before candidates take their places.
Administrators
If you expect more than fifteen candidates, you
need an assistant to help with distributing and
collecting materials. Add another assistant for each
fifteen to twenty candidates beyond that number.

session should take, whether there will be any
breaks, where toilets are located, etc.
To gain the most useful information from the test session while being completely fair to all candidates, you
need to follow strictly the guidelines set down for the administration of the tests you are using. The Test
Manual or User’s Guide, which is an essential companion to any psychometric test, will contain a section
detailing instructions for the proper administration of the test you are using.
The following are important aspects to consider in setting up a test session.
best practice
Test administration–
Section I: Test Selection and Administration
14
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 14
It is wise to check if everyone can see and hear
you. Does anyone need reading glasses or need to
move to the front of the room because of a hearing
impairment? Is anyone feeling unwell? Again, the
aim is to tap into their best performance, so it may
occasionally be necessary to reschedule a
candidate.
Mobile phones are a distraction. Remind all
candidates that all phones need to be switched off
as a courtesy to everyone.
Using Aids
Many tests do not permit the use of calculators or
such aids during the test session. Again, candidates
should be reminded of this.
The Manual
Ensure that you deliver the formal instructions
exactly as they are printed.
When giving any test from the ability/aptitude

so that people may leave as soon as they have
completed all components of the testing session.
Collecting Materials
It is essential that all materials are collected and
accounted for before candidates leave the test
room. Copyright legislation prohibits the copying of
materials and, to maintain confidence that test
integrity is being preserved, materials need to be
counted in and counted out.
A Quick Guide to Human Resource Testing
15
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 15
Normal Population
The natural world includes many examples of the
so-called ‘normal population’, characteristics of
which are often described by the bell-shaped curve
commonly called the normal curve (see Figure 2
opposite). In a normal population, most of the
people are closer to the average measurement of a
given attribute than to the extremes. For example,
there are many more people who are about average
height than there are extremely tall or extremely
short people. This means that if you measure some
characteristic that is normally distributed, most of
the population will ‘bunch up’ around the middle.
This produces the distinctive bell-shape of the curve.
If you went into a busy suburban street and
measured the height of the first 100 adult males
who passed by (these comprise your ‘sample’
population) and plotted the frequency that each

This figure serves as a benchmark against which
other scores may be measured; for example, ‘his
score is well above the mean’, ‘she obtained the
same score as the mean’, etc. This is the most
frequently reported central tendency score cited in
manuals for selection tests.
To interpret test results properly, an understanding of some of the basic terms used in statistics is required.
It is important to be aware of their relevance to help you make sense of test results and decide whether the
data reported is sufficiently valid and reliable for your HR testing exercise.
In this section we will cover four basic statistical concepts:
• the normal population • the range
• the mean • the standard deviation.
the basics
Statistics 1
Section II: Test Interpretation
16
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 16
Range
It is also helpful to know how far candidates’ scores
are dispersed from the centre. Probably the most
commonly used term is ‘range’, which is the
distance between the lowest score obtained by a
candidate on a test to the highest score gained. For
example, on a test of numerical reasoning, the
mean score may be 68, but the range of scores may
be from 12 to 99. You will notice that the mean is not
the ‘middle’ of the range but the average of all
scores obtained. If the mean is much higher than
the middle of the range, it means that more people
are obtaining high scores on the test than people

standard deviation is very large, a score would have
to fall very far from the mean to be extreme. For
example, if in our height example, the standard
deviation was only 1cm, then a person 3cm taller
than the mean (that is, three standard deviations
from the mean) would be in the tallest 0.1% of the
population. If the standard deviation was 20cm, a
person would have to be 60cm taller than the mean
to be in the top 0.1% of the population.
We all have a good understanding of height and
how tall or short someone has to be before they
would be considered taller or shorter than usual.
As each test has its own way of measuring the
underlying attribute or quality that is being
assessed, we need statistics to fully understand the
meaning of any test score compared with the whole
population. Understanding the centre, the range
and the spread of scores is an important first step.
A Quick Guide to Human Resource Testing
17
Figure 2: Normal Curve
mean
height
Number of cases
2.1% 2.1%13.6% 13.6%34.1% 34.1%
standard deviations –3σ –2σ –1σ +1σ +2σ +3σ
height (cm) 145 155 165 175 185 195 205
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 17
Changing a Raw Score to a
Standardised Score

Percentile ranks are very useful for ranking
candidates in order of merit (especially with ability,
aptitude or achievement tests), and for simple
explanations of where a candidate’s score lies in
relation to the rest of the sample population.
Stens and Stanines
Sten (‘Standard Ten’) scores and stanine (‘Standard
Nine’) scores are other ways of comparing a
candidate’s performance with the whole
population. In both sten and stanine scores there
are numbered categories that cover the whole
population – stens have ten categories: 1 to 10;
stanines have nine categories: 1 to 9. A score of 9 or
10 indicates a very high level relative to the
reference group, while a score of 1 indicates a very
low relative level.
What does it mean if a person obtains a score of 11 correct responses on a test? If there are 11 items, that’s
pretty good! If there are 100 items, perhaps it is not so great.
In order to understand the meaning of a score of 11, the developers of test instruments provide users
with a standard against which they can measure the merits of a candidate’s performance. This puts them in
a much better position to make a fair comparison. The standard is based on a set benchmark that was
determined by assessing a comparison group, or sample, that is selected to represent the population. This
sample is often called the norm group.
test scores
Statistics 2
Section II: Test Interpretation
18
Quick guide to HR testing.qxd 1/4/05 10:09 AM Page 18
Raw Score Percentile Stanine T-Score Raw Score
Rank

use of HR tests is that standardised scores, not raw
scores, should always be used in reporting, so that
the candidates’ scores can be properly compared to
the reference population.
Remember – what does a score of 11 correct
answers really mean?
Figure 3 shows an example of a norm table with
raw scores and three different standardised scores.
The HR practitioner can use whichever of these
standard scores best suits the situation. Often norm
tables only provide one kind of standard score.
Figure 4 is a single graph showing the
relationship between the normal population,
mean, standard deviation and some commonly
used standardised scores.
A Quick Guide to Human Resource Testing
19
Figure 4: Normal Curve and Standard Scores
Number of cases
2.1% 2.1%13.6% 13.6%34.1% 34.1%
standard
deviations
mean
–3σ –2σ –1σ +1σ +2σ +3σ
50 55
20 25 30 35 40 45 60 65 70 75 80
0.1 1 2 7 16 31 50 69 84 93 98 99 99.9
t-score
percentile
rank