Reliability of Psychological Tests

Introduction

            A definition of psychological tests includes two important terms i.e.  Reliability and Validity. These are the psychometric properties or characteristics of a test. The credibility and generalizability of a psychological test depends upon these two properties. They help in the objective evaluation of a test and ensure that it makes sense. In other words a test is considered standardized only when its reliability and validity are established.

Meaning

            In psychometrics ‘Reliability’ means ‘Consistency’ (agreement between two independently derived sets of scores).

Definitions

            “The ability of a psychometric test to yield consistent results from one set of measures to another” (Freeman, 1965).

            “Reliability refers to the consistency of scores obtained by the same persons when re-examined with the same test on different  occasions, or with different sets of equivalent items, or under other variable examining conditions” (Anastasi, 1976).

Purpose in Technical Terms

            To estimate what proportion of the total variance of test scores is error variance (Anastasi, 1976).

Purpose in General Terms

            To indicate the extent to which individual differences in test scores are attributable to “true” differences in the characteristics under consideration and the extent to which they are attributable to chance errors (Anastasi, 1976).

Needs of Reliability

            To address the possibility of

(i)        Chance errors,

(ii)       Internal defects of the instrument,

(iii)      Measurement and interaction among individual differences, and

(iv)      To underlie the computation of the error of measurement of a single score.           

Conditions Affecting Reliability

(i)        Actual or true differences among individuals.

(ii)       Differences in specific abilities, 

(iii)      Test taking skills of subjects,

(iv)      Effect of practice of coaching,

(v)       Natural/normal fluctuations in performance,

(vi)      Personal characteristics of subjects, 

(vii)     Physical conditions such as light, heat etc.

(viii)    Unpredictable or chance factors,

(ix)      Fortunate guessing of answers, and

(x)       Competence of examiners and their scoring style (Freeman, 1965).

Methods of Estimating Reliability

1.         Test-Retest – The same form of the test is administered within a week or two twice to the same group of subjects. Time is important if administered less than a week the practice effect due to recall can compromise the reliability co-efficient and in more than two weeks’ time the growth and development can affect the results. Also known as ‘temporal reliability’.

2.         Retest using two equivalent forms – Two separate but equivalent in all respects such as no. of items, difficulty level, item homogeneity, standard deviation etc. of the test to same subjects. This lessens the practice effect due to recall because of different items. But it is little bit of difficult to design and make two similar forms.

3.         Split Half – The test items of a single test are subdivided into two equivalent and separate forms. The two scores obtained are correlated and reliability coefficient is calculated.  It can be achieved only when the test items are arranged in increasing difficulty level. The most practiced method of splitting a test is Odd-even method i.e. odd number items in one test and even number items on the second test. Also known as “Odd-even’ reliability.

Factors affecting Reliability 

(i)        Range of ages,

(ii)       Range of scores,

(iii)      Time interval between testing,

(iv)      The effect of practice and learning,

(v)       Reliability of subtests, and

(vi)      Consistency of scores.

References:

1.         Anastasi, A (1976). Psychological Testing. New York: Macmillan Publishing Co.

2.         Freeman, F. S. (1965). Theory and Practice of Psychological Testing. New Delhi: Mohan Primlani for Oxford & IBH.