Validity of Psychometric Tests

Measuring the Validity of Psychometric Assessment

The validity of psychometrics has been discussed for a long time. Arthur (2011) notes that the predictive ability of an assessment helps measure its success, and analytical ability helps in explaining the degree in which a score on a the psychometric assessment predicts how well a candidate will perform in the given position or in the given role. According to Arthur,(2011).

The validity of the assessment tests can be measured in quantitative terms by using 0 to 1 measurement scale, where the correlation of 0 will represent no prediction or no relation while the correlation of 1 stands for perfect prediction and the correlation above 0.3 is considered to be strong.  The validity of an unstructured interview has only measure to 0.2. On the other hand, a selection process that unites a general reasoning test, a personality questionnaire, a job simulation exercise, and a structured interview has predicted a validity of 0.7 to 0.8, which further on provides the tester with confidence in hiring the best candidate (Arthur, 2011). 

Fletcher,(2006) says that in order to make psychometrics valid, an organization is required to make changes in its assessment criteria every time in order to keep the balance of skills and personalities required at that time. For example, if an organization needs a person with good communication skills for a role, a tester will craft assessment that will help in assessing candidates on the basis of their communication skills or if they need people with more emotional stability than communication skills for another project, an assessment will be changed accordingly(Fletcher,2006).

Source: The Psychometric World,(2015).

Above video explain the concept of psychometric properties and learn about 3 main psychometric properties viz. reliability, validity and norms (The Psychometric World,2015).

However due to the complexity involved in designing such questionnaires and trials and testing of the questionnaire on a sample group of people within a company set up keeping the purpose behind the whole exercise confidential can be difficult companies can take the help of consultants for the purpose(Sayantany Paul,2013).

“At its essence, validity means that the information yielded by a test is appropriate, meaningful, and useful for decision making – the purpose of mental measurement” (Osterlind, 2010)

There are five types of validity:

Source: Edenborough,-(2005)The different types of validity.

Armstrong’s (2010) further elaborate above different types of validity as below.

• Predictive validity

the extent to which the test correctly predicts future behavior. To establish predictive validity, it is necessary to conduct extensive research over a period of time. It is also necessary to have accurate measures of performance so that the prediction can be compared with actual behavior.

• Concurrent validity

the extent to which a test score differentiates individuals in relation to a criterion or standard of performance external to the test. This means comparing the test scores of high and low performances as indicated by the criteria and establishing the degree to which the test indicates who should fit into the high or low-performance groups.

• Content validity

the extent to which the test is clearly related to the characteristics of the job or role for which it is being used as a measuring instrument.

• Face validity 

the extent to which it is felt that the test ‘looks’ right, ie is measuring what it is supposed to measure.

• Construct validity 

the extent to which the test measures a particular construct or characteristic. 

Aamodt (2015) argues that for the organization also knowing what tests are most commonly used but there is also need to know what is the best evidence what does a good test, Outside the simple logic there are technical criteria used by professionals when evaluating to estimate the strength of psychometric instruments. Validity refers to whether a test measures what it is supposed to measure (Edenborough,2005).

“There is plenty of evidence of the validity of tests and assessment centers.  Work in this area has been dominated by studies of US origin using US datasets and it would be valuable if more validation studies were available which used datasets from European and other non-American countries.  However, the evidence available does point quite strongly to the conclusion that psychometric tests are able to make valid predictions about job performance, across a broad range of different jobs” (Andrew Jenkins,2001).

Reference List

•  Andrew Jenkins (2001) Companies’ Use of Psychometric Testing and the Changing Demand for Skills: A Review of the Literature, Centre for the Economics of Education London School of Economics and Political Science. London WC2A 2A .

•          Arthur,(2011). The validity of psychometric testing: What is the most effective way of selecting the right candidate for the job.Accessed on May 20, 2020, from hfi.com: http://www.hfi.com/images/GrapevineArticles/psychometric%20testing%20grapevine%20mar%2011.pdf

•       Fletcher, S. (2006,). Feature: Does Psychometric Testing Work? . May 20, 2020, from trainingzone.co.uk.

•          Sayantany Paul,(2013), Use of Psychometric tests in recruitment. International Journal of Research in Management,  ISSN 2249-5908, Issue 3, Vol. 2 Available online on at<http://www.rspublication.com/ijrm/ijrm_index.htm>.

•          Armstrong’s (2010) A Handbook of Human Resource Management Practice 10th ed PP463-464.

•   ·    Aamodt, M., 2015. Industrial/organisational psychology: An applied approach. Nelson

  Education.

•  Robert Edenborough, (2005). Assessment methods in recruitment, selection, and performance. Kogan Page, London and Sterling, VA. 

•          The Psychometric World, (2015) YouTube. [Online]. Available          at:< https://www.youtube.com/watch?v=Y3tNKaNfb-w> May 25 2020.

Enable GingerCannot connect to Ginger Check your internet connection
or reload the browserDisable in this text fieldEditEdit in GingerEdit in Ginger×