Testing, like any other area, attracts its share of misconceptions and objections. Of course, some of these contain a kernel of truth.

In this section we address some of the most common comments BTPA members hear in the course of their daily activities and try to explain our point of view.

You may disagree with some of our answers or have other views which you’d like to express. In that case, do contact Ian Florance, BTPA secretary on ian.florance@btinternet.org. We hope this area will develop into as genuine dialogue between test users, developers and publishers and that the discussion below will stimulate you to get in touch.

Testing is expensive
Some users comment: “ I already use interviews, CVs and other techniques; why would I want to spend MORE money on testing which, anyway, looks pretty expensive to me.”

Like any other sort of product or service, testing comes in many different types with many different pricing structures. You can buy a simple test for high volume selection exercises and will get change out of a £5.00 note. You can also use a very complex test with rich output that can serve as the basis for a six month senior executive programme and spend over £100. You can buy tests in single copies, on a license basis and with discounts for quantity. So, to say that testing is “ highly priced “ is like saying that books are highly-priced: some are, some aren’t.

A more important question is whether tests are valuable. It’s true that there is an up-front investment in training, experience gathering and in the materials themselves . However tests make selection and promotion decisions more accurate ( see other answers below for this point ). They also serve as the basis for more effective development of individuals and groups and for putting together more efficient teams. So, although there is an upfront investment in testing, btpa argues that this is paid back many times in a number of areas ( and Return on Investment analyses support this ):

• reducing wrong recruitment decisions
• increasing good recruitment decisions
• increasing the speed with which successful candidates get up to speed, because they’ve bee n recruited to do that.
• reducing time spent on recruitment processes by cutting down the number and increasing the effectiveness of recruitment
• increasing effective internal promotion which is cheaper than external recruitment
• increasing efficient working
• saving the costs of tribunals and employment litigation (see below)

The precise savings will vary according to your particular business but it’s estimated that any wrong recruitment costs at least first year package to rectify and the indirect costs in missed opportunities or underperformance are much greater. US companies report around 30% savings in recruitment using on-line testing. Considering the use of tests is a good time to look at the hard economics of your recruitment: how much does t cost; how often do you get it right; are you in line with employment law; what might your business look like if your key recruitment exercises had been done better?

Testing is a “ flavour of the month”
It’s true that there’s been a huge upsurge in interest in psychometric testing in the past 20 years but attempts to describe the differences between individuals go back to 400 BC and psychometric tests were offered to the public in the UK at the end of the 19th century. The armed forces started using tests in recruitment after the First World War.

The fact that some of the tests still widely used in business were created over 50 years ago and the use of such instruments was a major element in the adoption of “scientific management “ confirms that testing is a subject with a long academic history and has been used “ in anger” for over a century.

People don’t like being tested
Some people think of school exams as soon as the word test is mentioned. This is no bad thing as a bit of adrenaline helps performance ( see http://www.btpa.org/Test_Takers_Area/How-to-do-best-at-tests on this site ). But btpa members and other publishers report that once people have taken their first test they are often fascinated by the results and look forward to finding more out about themselves. Our case studies (http://www.btpa.org/case-studies) show how professional feedback from testing can enthuse a culture. If testing is done well, even unsuccessful candidates feel they’ve been treated well and learnt something from the process,. This is good for your brand.

Testing is biased against certain groups
All decisions about other human beings have some degree of bias in them. The trick is to minimise this as much as possible and, above all, to know the likelihood that a decision is being biased by unrelated factors. Tests, far from being particularly biased, are designed to minimise bias and provide hard statistical data which can alert employers if they are missing out on really good candidates because of flawed recruitment or internal development processes.

To take an example, most companies use interviews. Few interviewers know how subjective unstructured interviews are and that interviewers will make up their mind about candidates almost as soon as they’ve sat down, their decision being based on issues unrelated to the purpose of the interview. At its worst, decisions can be affected by outright gender, racial, religious and other sorts of discrimination. These failures affect company performance by missing opportunities to recruit the best person for the job. They also open up recruiters to legal action. It shouldn’t need to be said that all prejudice of this kind is unfair to candidates.

You can revise interview strategies to overcome this effect to some extent: highly structured interviews, for instance, are fairer by providing a “ level playing field” for all candidates.

Good tests, on the other hand, are designed to be fair to all candidates: to be objective and, if chosen carefully, focussed on issues directly relevant to the job in hand. Administered in a standardised way psychometric tests approximate to scientific method in making decisions about other people. More than that they provide a hard audit trail which can be used to create a virtuous circle of improvement for HR procedures, defend particular decisions, examine the fairness and success of recruitment and development.

But they key issue here is that a good instrument (one that is ft-for-purpose, a prerequisite of unbiased testing) is only as good as its use, and the care with which it is chosen. Even a very robust test can “mis-fire” if used casually or unprofessionally.

In some situations group differences do occur in performance on certain tests. For example in personality terms women are frequently found to score higher than men (on average) on some of the people-oriented traits (e.g. empathy), and lower on more ‘tough minded’ traits such as competitiveness. To take another example, some ethnic minority groups tend to score higher, on average, in Numerical Reasoning tests than the majority group do, whilst other ethnic minority groups tend to score lower.

Differences of this type can result in what is known as ‘adverse impact’ where a fixed selection criterion (e.g. numerical test pass mark of 20) is applied, and a significantly lower proportion of one group reach the pass mark as compared with the majority.

Adverse impact should not, however, be confused with bias or discrimination. It is simply the case that some traits and job-related characteristics tend to differ among different groups of candidates, for a variety of reasons including culture, background and education.

Where adverse impact is apparent, for example in volume selection situations, it is important for the test user and test publisher to work together to ensure that the particular usage and application of the test in question is fair, valid and defensible for the job role in question.

Where tests are used in combination with other measures, adverse impact may be less of a concern given that any particular test is only one of multiple sources of information contributing to the overall decision about the individual.

Testing is unfair since it tries to sum up a complex person with the answers to a few questions.

Testing doesn’t sum up a complex person with a few questions: it looks in-depth at a specific aspect of a person – his or her work-related personality, numerical ability or characteristic motivation – and gives very precise measurements of that area – either how important a particular characteristic is to that person, how well he or she scores in comparison with a relevant group of people or how likely it is that he or she will act in a certain way, for instance.

It’s worth making a particular point about personality tests. These don’t provide an exhaustive description of personality: they predict how a person is characteristically prone to act. Someone may tend to be calm, a bit of a loner and very good at detail in most situations; this doesn’t mean that under certain situations he or she will not lose his or her temper, gate crash a party or talk in generalities. Personality tests should also ideally be used in conjunction with a feedback discussion where the meaning of the particular scores can be interpreted in the context of the individual’s experience.

Testing can replace all other forms of judging people
As can be seen from the above answers this is simply not true. The btpa recommends that tests always be used with other forms of information and assessment, among which are:

• CVs
• structured interviews
• application letters
• references
• job simulations
• in-tray exercises

To get a rounded view of a person. Different techniques have different strengths and weaknesses and are better or worse at getting at different areas of a person’s relevant skills, experience, personal-style and values.

Testing is only used in recruitment
In fact, the huge growth in testing has been fuelled by its use in development: for instance, assessing people before a training course to see what they need to learn and how best they will learn it. Post training testing clarifies the success of an intervention. Testing is used in a wide variety of situations:

• in recruitment to set up a short list or as part of the final round of interviewing and assessment
• in development, whether that’s training, counselling, mentoring or on-the-job learning
• to put together new teams, solve problems with existing ones or to improve an already excellent team. Increasingly tests are used to maximise the effect of virtual teams or teams which cross organisational boundaries
• as the basis for organisational development, either to audit the human potential in the existing organisation or to look at attitudes throughout a workforce which might be affecting performance.

It costs a fortune to be trained in testing
As with tests, test training comes in different packages with different price structures. btpa’s members provide a full range of testing:

• 5 day courses which introduce you to a range of tests and testing techniques and provide access to the nationally recognised Level A and B qualifications based on the British Psychological Society’s standards
• Test-specific courses that allow you to use a certain companies instruments or approaches
• Skills courses – in feedback or test administration

Which path you chose depends on your and your company’s needs and goals. It’s worth noting:

• Training is crucial to using tests. The more you know about testing and the particular tests you use, the more valuable they’ll be to your organisation
• Test expertise is a highly marketable skill in modern organisations.

Testing measures intelligence
Intelligence ( or IQ ) tests were very popular but are now rarely used in businesses due to their controversial nature ( though its worth pointing out that IQ measures are good predictors of work success and you can expect you see more debates in the future about the role of intelligence in organisations ).

In essence, if you can define an aspect of a human being you can create an assessment to measure it; though how easy that will be, how accurate the measurement will be and how predictive that human characteristic will be of job success will depend on the clarity of the concept and the excellence of the test design. Tests are generally classified into two main groups. As their name suggests, Tests of maximum performance assess you working at your best; examples include tests of numerical, verbal and diagrammatic reasoning. Maximum performance tests tend to have right or wrong answers, and to be strictly timed.

Tests of typical performance are normally called questionnaires or indicators and do not have right/wrong answers. They are designed to assess your typical style, behaviour and preferences and require you to assess yourself. Examples include measures of personality, motivation and interests. Typical performance questionnaires are usually untimed.

Published tests exist to measure among other things:

• Personality (in the workplace, at home and to assess clinical states)
• Abilities (with numbers, diagrams, abstract concepts, mechanical systems, words)
• Specific office skills (such as checking, filing and using particular computer programmes)
• Attainment (knowledge of a particular area)
• Leadership and entrepreneurship
• Creativity
• Preferred roles in a team
• Integrity and ethics
• Values, beliefs
• Motivation

You have to be a psychologist or consultant to use tests.
No, you can train to use tests and some of the most effective and expert users of tests in the UK are HR staff, owner-managers of smes, training staff and line managers.

Tests are easy to cheat
Recent research in the USA shows the lengths to which graduates, in particular, are prepared to go to get a job, even though that job might be the wrong one. Cheating and lying affects all the sources of information a recruiter must deal with: CVs, references, application letters , tests and interview answers.

Yes, there are instances of people “cheating” at tests and the use of the internet as a test delivery medium raised fears that this would increase. But test publishers have been investing heavily in providing you the existing armoury of techniques to prevent cheating and manipulation of the results.

• Tests of knowledge and ability with right and wrong answers are obviously most prone to cheating. The development of item banks testing (a way of delivering tests with exactly the same characteristics but different questions and answers to different people) has grown hugely. Use of the same or similar tests at shortlist and interview stage show where the person claming to be answering a test was in fact someone entirely different. Increasingly sophisticated recording of peoples’ performance on internet tests will may also show if they’re looking up answers.
• Tests without right or wrong answers such as personality tests have long had sophisticated statistical and design techniques to highlight where a pattern of answers suggests manipulation.
• This is an area of considerable concern whatever information you use to judge an applicant or worker and the btpa is committed to spreading good practice in this area.