Test - Psychometric Test - Inventory - Psychological Test - Survey - Questionnaire - Instrument - etc. etc. All of these terms are used to describe a series of questions, designed to help measure knowledge, skills, aptitudes, behaviours, attitudes and personality traits. Here are two other definitions given by the British Psychological Society. 'A psychological test is an instrument designed to produce a quantitative assessment of some psychological attribute or attributes.' 'A psychological test is any procedure on the basis of which inferences are made concerning a person's capacity, propensity or liability to act, react, experience, or to structure or order thought or behaviour in particular ways.' The word ‘test' is the term most widely used by people other than psychologists. That is why we chose to use that term in this Guide. However, we do like the words ‘questionnaire' and ‘inventory'. So when we think that it is appropriate we will use these descriptions. You will find that we tend not to use the term 'psychometric' too much. Hope all this does not confuse you!

You judge the quality of a test in much the same way as you would judge a ‘good interviewer'. A good test, like a ‘good interviewer', will pose many questions rather than just a few Common Sense Rule Asking more questions = more answers = more information

Like a good interviewer, a good test needs to be realistic. People completing questionnaires as part of a selection or assessment programme often try to ‘beat the system'. Many people frequently and naturally seek to impress. Some even, on rare occasions, attempt to make a ‘bad impression'. Like a good interviewer, a well-constructed test anticipates this behaviour. Common Sense Rule Be ‘street-wise' rather than ‘ivory tower'

Good tests use interrogation techniques. For example, they use a technique called ‘masking ‘. This involves embedding significant and critical questions amongst non-relevant questions. Questions are also posed in an unexpected rather than an anticipated sequence. Incidentally, research involving testing the same people more than once, established that people are less likely to lie or evade the truth when answering questionnaires. This is probably because a test question demands a definite answer to a limited range of choices whereas questions posed at interviews are more often open ended and debatable. Do not be a textbook interviewer! Common Sense Rule Change your pace, change your style, and try not to let the interviewee know what is coming next. Remember it's a battle of wits. Study your 'enemy'. Read all the books giving hints about - 'How to Get Through a Tough Interview.' Good candidates will!

We vigorously stress that no selection, development or training decisions should be made solely on the basis of a test result. Like a good interviewer, a good test reserves judgement. It does not take any single answer, or small group of answers, at face value. A good test looks for consistency – a pattern of responses to questions that is consistent rather than contradictory.

A good test only promises to provide ‘clues'. Clues to help you structure your interview/s and assessment investigation. A clue does not become a fact until you have very carefully searched for a lot of evidence to confirm – or contradict - the test result.

For example, if a test result has given you a clue that a person is assertive, you should be observing assertive behaviour and getting information that confirms assertiveness. Find out what the person has actually done in the past rather than asking them what they would do, or may do, in the future. A careful study of their past record, references, job and life progress should provide evidence to confirm assertiveness. If it does not, you must question the accuracy of the test result. WYGIWYS - Never trust a test result (clue) until you have at least six ‘bits' of supporting evidence. As previously stated, we vigorously stress that no selection, development or training decisions should be made solely on the basis of a test result. We recommend that you consider using the Rudyard Kipling Technique - the RKT: ‘I keep six honest serving–men (They taught me all I knew); Their names are What and Why and When and How and Where and Who. I send them over land and sea, I send them east and west; But after they have worked for me, I give them all a rest.' Rudyard Kipling

A good test should compare the person being tested against a large number of other people very similar to him or her. For example. You are very rich - You are very poor You are very tall - You are very short You are very young - You are very old. Any of these statements could be true about you. It just depends against whom you are being compared. WYGIWYS and Rudyard Kipling Technique (RKT) will help you find out which is right. (See above section.) Common Sense Rule Put your money on the comparison that is being made, rather than the glossiness and length of the test report. You would not expect a 10-year-old's answers to a maths paper to be compared against the answers given by someone with a degree in maths - would you? Always ensure that your people are being compared against a large number of people as much like themselves as possible. Remember – ‘like with like' increases accuracy. 'Like compared with unlike' distorts the score and reduces accuracy.

The authors of this Guide strongly recommend and encourage disclosure of test results. We believe that people have the right to this information. We also believe that they should be given an opportunity to express their opinions on the accuracy of test results. Always disclose and discuss a person's test scores. Let them know who you are comparing them against. This should either be your own company norm or an appropriate occupational norm group. Tell them why you consider the areas you are measuring to be important. Obviously, you refrain from doing this until you have carried out most of your other investigations. You don't want to affect their behaviour or answers to interview questions. Common Sense Rule People are more likely to trust you if you do not keep scores a secret and if you tell them why you are measuring a particular aspect of personality or behaviour. If the test you are using is 'good' - and the group against which test takers are being compared is appropriate - there is absolutely no reason why you should not give the test taker his or her results. Indeed, in some countries, the law does require you to disclose test results. Hopefully this law will soon be universal. We believe that everyone has the right to have access to the results of any test they complete.

They need a good test because they are only human. As a good interviewer, you will know that during an interview it's virtually impossible to ask identical questions, in an identical manner, in the same sequence and elicit a definite response. Your voice intonations, facial and body signals will almost certainly influence, and possibly distort, the answers. You may also, dare we say it, have some biases and stereotypes embedded in your brain. If you don't think this is possible try the Australian/Japanese Experiment when you next run a meeting relating to selecting or developing people. Common Sense Rule You are only human. You may be a good judge of character but, unless you are Mr. Spock, we doubt that you can ever be as clinical as a good test.

Given the right set of circumstances and budget, anyone can design a good test. There should not be any mystique surrounding the creation of a good test. It is easier and less costly to develop tests to measure knowledge, skills, aptitudes and attitudes than it is to measure personality traits. Tests relating to personality can take a long time to develop and validate. This can be very costly. Consequently, you have to consider the cost-effectiveness of creating such tests very carefully. In a recent article* Dr. Robert McHenry, Chairman of OPP - one of Britain's largest test publishers - said "developing a test could cost up to £100,000". He also "... urged HR practitioners to use tests in conjunction with other selection methods such as role play, in-tray exercises and structured interviews." We totally agree with that advice. Indeed, we vigorously stress that no selection, development or training decisions should ever be made solely on the basis of a test result. *'People Management' published by the Chartered Institute of Personnel Development

A recent article in the Daily Telegraph reported on various selection methods and their accuracy in predicting job success. Selection Method against Degree of accuracy at predicting job success: Normal interviews 17% Structured interviews 33% Personality test 42% Ability/skill tests 54% Assessment Centres 65% Note: Assessment Centres almost always include Ability and Personality tests.

If tests are being used, make sure that they definitely relate to competencies being measured and have not just been included out of unquestioned habit. If tests are not being used, you might ask the organiser of the assessment centre ... Why not?