Tests are useful selection devises in that they uncover qualifications and talents that can’t be detected otherwise. They can be used to predict how well one would perform if one is hired, why one behaves the way one does, what situational factors influence employee productivity etc. Tests also provide unbiased information that can be put to scientific and statistical analysis.
However, tests suffer from sizeable errors of estimate. Most psychological tests also have one common weakness, that is we can’t use scales which have a known zero point and equal intervals. An intelligence test, for example starts at an arbitrary point, where a person may not be able to answer questions properly. This does not mean that the person is totally lacking in intelligence. Likewise, a person who is able to answer all the 10 questions correctly cannot be called twice as intelligent as one is able to answer only 5 questions correctly. If the test has commenced at some other point, where there were easier questions their scores might have been different. Tests also fail to elicit truthful responses from testees. To compound the problems further test results are interpreted in a subjective way by testers and unless these tester do their homework well, the results may not be reliable.
Standards for Selection Tests:
To be useful as predictive and diagnostic selection tools, tests satisfy certain basic requirements:
Reliability: test scores should not vary widely under repeated conditions. If a test is administered to the same individual repeatedly he should get approximately identical scores. Reliability is the confidence that an indicator will measure the same thing every time. A yardstick that measures 60 inches tall every time it is used is reliable. Generally speaking HR managers should choose instruments that have the following types of reliability:
Test retest reliability: where the techniques gives the same results when administered repeatedly to the same person. For example if a person takes the same hearing test twice during the week and receives the same result the test retest reliability is high.
Inner rater reliability: where the instruments give the same results when used by two or more different raters. For examples, if two supervisors interview the same workers and evaluate the worker similarly, the interview has high inner rater reliability.
Intra Inter reliability: where the technique gives the same results repeatedly used by the same rater to rate the same behaviors or attitudes at different times.
Validity is the extent to which an instrument measures what it intends to measure . In a typing test, validity measures a typist’s speed and accuracy . To determine whether it really measures the speed and accuracy of a typist is to demonstrate its validity. The question of determining the validity of a selection test thus, has a lot to do with later performance on the job (known as criteria): a selection tool such as a test, or an interview is referred to as a predictor and what to measure –job performance is a criterion). If a person has done well both in selection and subsequently on the job, the test selection would be accepted as a valid technique for selection .
Qualified people: tests require a high level of professional skill in their administration and interpretation. Professionals, technicians are needed for skilled judgmental interpretations of test scores.
Preparation: A test should well prepared. It should be easy to understand and simple to administer.
Suitability: A test must fit the nature of the group on which it is applied. A written test comprising difficult words would be fruitless when it is administrated on less educated workers.
Standardization: Norms for finalizing test scores should be established. There must be prescribed methods and procedures for administering a test and for scoring or interpreting it.
Source: HRM VSP