PIB and VPIB construct validity

HELENA KRIEL

 M.A.

INTRODUCTION

The concept of validation has been defined by various authors. Magnussen (1966: 123) defines the validity of an instrument as the accuracy with which meaningful and relevant measures can be made with it. The question is: does it measure the traits it intends to measure? Sundberg (1977: 43) defines validation as the property of a test that renders the obtained results useful. A procedure is not valid in general - it is valid in relation to something.

Cascio (1991: 149) argues that the traditional concept of validity, as “the extent to which a measurement procedure actually measures what is designed to measure”, is inadequate as it implies a procedure has only one validity, which is determined in a single study. In this regard Van Aswegen (1997: 19) feels that, particularly in selection research, many investigations are required to obtain thorough understanding of the interrelationships between scores from a particular procedure and other related variables. According to Binning and Barret (as quoted by Van Aswegen 1997: 19) three strategies can be used in this investigation process, namely content related, construct related and criterion related evidence of validity. The present concerns itself with construct validity.

CONSTRUCT VALIDITY

The essence of construct validity is to provide an answer to the question: “does the test measure the construct that it claims to measure?” It must therefore be shown that a measure correlates highly with other measures with which it could theoretically be expected to correlate. This process is described as convergent validation. A measure should also be shown not to correlate with measures with which it could be theoretically expected not to correlate. This is achieved by means of discriminant validation.

Correlations between PIB and VPIB Indices

A database consisting of 369 PIB and VPIB records was examined by the Department of Student Counseling at the Technikon Pretoria as a continuation of the investigation process into the situation-specific validity of PIB and VPIB. The PIB and VPIB indices utilized in assessing the potential of prospective students were inter-correlated. The indices that correlated significantly, are discussed.

• PIB 2 [Creativity] and VPIB 4 [Composition of Wholes]

A statistically significant correlation was found between the scores obtained on the Creativity index [PIB] and the Composition of Wholes index [VPIB].

r = 0.217* (*p = 0.041)

According to the developers (Erasmus & Minnaar, 1995) the capacity to compose is in essence creative. This view is confirmed by the above result.

• PIB 4 [Calculations] and VPIB 3 [Numerical Ability]

The scores obtained by respondents on the Calculations index [PIB 4] were correlated with their scores on Numerical Ability [VPIB 3].

r = 0.449** (**p = 0.000)

Both these indices comprise numerical manipulations of a basic nature. Scores reflected a high correlation in accordance with expectations.

• PIB 4 [Calculations] and VPIB 4 [Composition of Wholes]

The results of the Calculations index [PIB 4] were correlated with the results of the Composition of Wholes index.

r = 0.604** (**p = 0.000)

The Composition of Wholes index demands the ability to reason along concrete lines as regards figures and forms. The same trend towards concrete reasoning could be expected in the case of calculations.

• PIB 4 [Calculations] and VPIB 6 & 7 [Spatial Reasoning and Perception]

The following result was obtained when scores on Calculations were correlated with scores on Spatial Reasoning and Perception:

r = 0.425** [**p = 0.000]

This result probably indicates that Index 4 [PIB] also involves logical reasoning.

• PIB 5 [Mental Alertness] and VPIB 3 [Numerical Ability]

The results obtained on PIB Index 5 [Mental Alertness] were correlated with the results of VPIB Index 3 [Numerical Ability].

r = 0.391** (**p = 0.000)

According to Erasmus (1995: 2) Index 3 [VPIB] is mainly aimed at the respondents’ verbal potential to calculate. Mental Alertness assesses the abstract and general reasoning abilities. It can be expected that these scores will correlate.

• PIB 5 [Mental Alertness] and VPIB 4 [Composition of Wholes]

The applicants’ scores on Mental Alertness were correlated with their scores on Composition of Wholes.

r = 0.482** (**p = 0.000)

The capacity to compose involves both visual perception and logical reasoning. Both these could be expected to be based on the ability to reason, which forms part of Index 5 [Mental Alertness].

• PIB 5 [Mental Alertness] and VPIB 6 & 7 [Spatial Reasoning and Perception]

The Mental Alertness scores [PIB 5] were correlated with the result obtained with Spatial Reasoning and Perception [VPIB 6 & 7].

r = 0.358** (**p = 0.000)

Indices 6 & 7 [Spatial Reasoning and Perception] inter alia calls for the ability to select and to categorize. This ability also forms part of Mental Alertness.

• PIB 7 [Abstract Reasoning] and VPIB 3 [Numerical Ability]

The scores obtained with PIB 7 [Abstract Reasoning] were correlated with the VPIB 3 scores [Numerical Ability]

r = 0.304** (**p = 0.000)

Index 7 assesses abstract reasoning ability. It can be assumed that abstract reasoning will also be integral to calculations.

• PIB 7 [Abstract Reasoning] and VPIB 4 [Composition of Wholes]

Scores on Abstract Reasoning [PIB 7] were correlated with scores on Composition of Wholes [VPIB 4].

r = 0.218* (*p = 0.018)

Abstract Reasoning could be expected to be called for when parts are arranged in constituting wholes.

• PIB 7 [Abstract Reasoning] and VPIB 6 & 7 [Spatial Reasoning and Perception]

The results obtained on Abstract Reasoning were also correlated with scores on Spatial Reasoning and Perception.

r = 0.325** (**p = 0.000)

The ability to perceive, analyze and synthesize could be expected to call for the capacity to reason.

PIB-Intercorrelations

The correlation between scores obtained on various PIB indices were calculated. The indices that reflected significant correlations are now discussed.

• PIB 3 [Reading Comprehension] and PIB 4 [Calculations]

Scores obtained on Reading Comprehension and Calculations were correlated.

r = 0.499** (**p = 0.000)

Verbal ability could be expected to play a role, both in calculations and comprehension. This hypothesis seems to be confirmed by the obtained correlation.

• PIB 3 [Reading Comprehension] and PIB 5 [Mental Alertness]

Scores on PIB 3 and PIB 5 correlated significantly.

r = 0.522** (**p = 0.000)

Both dimensions necessitate linguistic reasoning ability. A high correlation could therefore be expected.

• PIB 3 [Reading Comprehension] and PIB 7 [Abstract Reasoning]

The scores obtained on the above-mentioned indices correlated significantly.

r = 0.362** (**p = 0.000)

The ability to assimilate information from the environment plays a role in both these indices. A significant correlation could be expected.

• PIB 3 [Reading Comprehension] and PIB 12 [Vocabulary]

Scores for PIB 3 [Reading Comprehension] and PIB 12 [Vocabulary] were correlated.

r = 0.429** (**p = 0.000)

It was expected that language ability will play a part in the performance on both these indices. This hypothesis was confirmed.

• PIB 4 [Calculations] and PIB 5 [Mental Alertness]

The scores obtained on the Calculations Index [PIB 4] were correlated with scores on Mental Alertness [PIB 5].

r = 0.671** (**p = 0.000)

Mental Alertness assesses abstract and general reasoning ability. Calculations demand reasoning along logical, mathematical lines and a significant correlation between the indices was expected.

• PIB 4 [Calculations] and PIB 7 [Abstract Reasoning]

A significant correlation was found between the scores obtained on Calculations and Abstract Reasoning.

r = 0.387** (**p = 0.000)

The ability to reason along logical, mathematical lines also calls for the ability to reason along abstract lines.

• PIB 5 [Mental Alertness] and PIB 7 [Abstract Reasoning]

The results of the Mental Alertness index were correlated with the results of the Abstract Reasoning index.

r = 0.511** (**p = 0.000)

It was expected that an ability to reason in terms of items and materials will play a part in the performance on both these indices. This hypothesis was confirmed.

• PIB 5 [Mental Alertness] and PIB 12 [Vocabulary]

The applicant’s scores on Mental Alertness [PIB 5] were correlated with their scores on Vocabulary [PIB 12].

r = 0.332** (**p = 0.000)

Both indices imply linguistic ability and a correlation between the scores was expected.

• PIB 9 [Interpersonal Relations] and PIB 21 [Assertiveness]

The following result was obtained when scores on Interpersonal Relations were correlated with scores on Assertiveness:

r = 0.407** (**p = 0.000)

A high score on Index 9 [Interpersonal Relations] usually indicates well-developed interpersonal and leadership skills. Such a person can be expected to be assertive.

VPIB-Intercorrelations

Relevant VPIB indices were inter-correlated. Statistically significant results are discussed below.

• VPIB 3 [Numerical Ability] and VPIB 4 [Composition of Wholes]

The applicants’ scores on Numerical Ability were correlated with their scores on Composition of Wholes.

r = 0.389** (**p = 0.000)

According to Erasmus (1995: 2) VPIB 3 inter alia indicates the ability to arrive at a solution that amounts to more than the mere sum of the parts. VPIB 4 demands the visualization of the end-result in order to reach the composed whole. Scores obtained on these indices could therefore be expected to correlate.

• VPIB 3 [Numerical Ability] and VPIB 6 & 7 [Spatial Reasoning and Perception]

The results obtained on VPIB 3 were correlated with the results on VPIB 6 & 7.

r = 0.353** (**p = 0.000)

The ability to reason along logical, mathematical lines, as well as to perceive, analyze and synthesize forms an integral part of the capacity to reason.

• VPIB 4 [Composition of Wholes] and VPIB 6 & 7 [Spatial Reasoning and Perception]

The applicants’ scores on VPIB 4 were correlated with their scores on VPIB 6 & 7.

r = 0.266** (**p = 0.000)

According to Erasmus (1995: 3&4), both VPIB 4 and VPIB 6 & 7 entail the ability to perceive, analyze and synthesize. These abilities seem furthermore to be fuelled by the general ability to reasoning effectively. The correlation could thus be expected.

CONCLUSION

From this study it appears apparent that the constructs measured by PIB and VPIB show a high degree of correlation in those cases where significant correlation, could be expected. There therefore seems to be significant evidence to claim construct related validity with the PIB.

LIST OF REFERENCES

Cascio, W.F. (1991). Applied Psychology in Personnel Management. 3rd Edition. Englewood Cliffs: Prentice Hall.

Erasmus, P.F. and Minnaar, G.G. (1995). Manual for Erasmus/Minnaar Potential Index (Intermediate Scale). Krugersdorp, PIA.

Magnussen, D. (1966). Test theory. Ontario: Don Mills.

Sundberg, N.D. (1977). Assessment of Persons. New Jersey: Prentice Hall Company.

Van Aswegen, M. (1997). The Standardisation of a Learning Potential Battery for the selection of Poorly Qualified Employees. Unpublished Thesis: University of Pretoria.