Machine-translated by Google. Questions are authored in English.
Top

Psychological Testing Questions and Answers

Direction for [ Question No: 1 To 5 ] :

Read the following paragraph and answer the five questions which follow :

The notion of true score is at the heart of classical reliability theory. This model proposes that the obtained score on a variable is a sum of true score and the random error of measurement, the two components being independent. Using this model, various ways of computing reliability have been developed. One of the way of examining the reliability of a test is to test the homogeneity of test items or the internal consistency. The most widely used method of estimating the reliability is split-half technique, also called as odd-even reliability. In this context, Spearman Brown correction is important. Kuder-Richardson reliability and Chronbach Alpha are the two other methods of studying internal consistency.

Another important type of reliability is the interscorer reliability. This reliability is not useful for objective tests, but for certain types it is quite important. There are also other types of reliability such as parallel form reliability and test-retest reliability, the latter concerned with temporal stability of the test results.

1.

Reliability can be defined as

Exam Year(s)

UGC NET - 2013,

Answer: A

Reliability is the ratio of true score variance to total variance.

In classical test theory the obtained score is the true score plus random error. Reliability is the proportion of the observed variance that is due to the true scores:

\[ r_{xx} = \frac{\sigma^{2}_{T}}{\sigma^{2}_{X}} \]

The second option, error variance over total variance, is the unreliability, and the other two ratios do not have a meaning of this kind.

Please log in or register to report this question.

Discuss in Forum
2.

For a newly developed test, the scores on odd and even part correlated by 0.4. After applying Spearman-Brown formula

Exam Year(s)

UGC NET - 2013,

Answer: C

The corrected reliability would be greater than 0.4 and less than 0.8.

The Spearman-Brown formula corrects the half-test correlation to the reliability of the full test:

\[ r_{full} = \frac{2r}{1+r} = \frac{2(0.4)}{1.4} \approx 0.57 \]

That value lies between 0.4 and 0.8. It is not 0.8, which would arise by simply doubling 0.4, and the correction always raises the value, so it cannot be 0.4 or less.

Please log in or register to report this question.

Discuss in Forum
3.

Which of the following statement is true for tests with dichotomous items ?

Exam Year(s)

UGC NET - 2013,

Answer: C

For tests with dichotomous items Cronbach alpha and Kuder-Richardson reliability are equal.

The Kuder-Richardson formula 20 is a special case of Cronbach alpha for items scored one or zero, and when the items are dichotomous the two give the same value.

So alpha is neither greater nor smaller, and it does not vary from one to the other, which excludes the other three statements.

Please log in or register to report this question.

Discuss in Forum
4.

For which of the following types of tests interscorer reliability is relevant ?

1. Objective personality inventories
2. Projective tests
3. Achievement tests with descriptive items
4. Intelligence tests
Codes :

Exam Year(s)

UGC NET - 2013,

Answer: D

Interscorer reliability matters for projective tests and for achievement tests with descriptive items.

Where the scoring needs the judgment of the scorer, as with responses to inkblots or answers written in full sentences, different scorers may give different scores, so their agreement must be checked. That is statements 2 and 3.

Objective personality inventories and standard intelligence tests have fixed scoring keys, so the scorer makes no difference and interscorer reliability is not relevant to them.

Please log in or register to report this question.

Discuss in Forum
5.

If two groups are administered the same test with a time gap of eight weeks, we would be able to compute

Exam Year(s)

UGC NET - 2013,

Answer: D

None of the reliabilities named can be computed.

Test-retest reliability needs the same test given twice to the same group. Here two groups take the test with a gap of eight weeks, so there is no pair of scores for any individual and the test-retest correlation cannot be found.

Parallel form reliability needs two different forms, and interscorer reliability needs scores by two scorers. Neither is available either, so the last option is right.

Please log in or register to report this question.

Discuss in Forum
Direction for [ Question No: 6 To 9 ] :

Read the following passage and answer the five questions that follow :

The Classical Test Theory (CTT) assumes that every obtained score is a sum of the true score and the error of measurement. The earlier versions of the CTT, in their reliability conceptualization, ambitiously addressed the random errors of measurement. Accordingly, reliability can be conceptualized as a consistency between the true score and the obtained score. Reliability can be studied through various perspectives. In the test construction phase, internal consistency reliability is given heavy emphasis. Various reliability coefficients such as split-half reliability, Kuder-Richardson reliability, Cronbach alpha etc. have been developed for this purpose. One of the earlier development in CTT is the Spearman-Brown formula. The other important types of reliabilities include test-retest reliability and interscorer reliability. In addition to reliability, psychologists emphasize test validity. Content validity, concurrent validity, predictive validity, construct validity are some of the important types of validities in psychometric studies. Factor analysis has also been employed in evaluating the test validity. Once the reliability and validity are well established, the test constructor attempts to establish the representative and comprehensive norms.

6.

Reliability is conceptualized as

Exam Year(s)

UGC NET - 2013,

Answer: B

Reliability is conceptualised as the ratio of true variance to the sum of true variance and error variance, which is the same as true variance over the obtained (total) variance. In classical test theory the obtained score is the true score plus error, so the obtained variance is true variance plus error variance.

Obtained over true variance is the inverse of this and would exceed 1, true over error variance is a signal-to-noise ratio and not a reliability coefficient, and obtained over error variance is also not reliability. So the second option is correct.

Please log in or register to report this question.

Discuss in Forum
7.

Cronbach alpha is

1. larger than split half reliability coefficient
2. smaller than reliability coefficient
3. average of all possible split-half reliability coefficients
4. equal to Kuder-Richardson reliability for dichotomous items
Codes :

Exam Year(s)

UGC NET - 2013,

Answer: C

Cronbach alpha is the average of all possible split-half reliability coefficients, and it is equal to the Kuder-Richardson reliability for dichotomous items. So statements 3 and 4 are true.

Statement 1 is not true, since alpha is not necessarily larger than the split-half coefficient, as it is an average of them, and statement 2 is not true, as alpha is itself a reliability coefficient and is not defined as smaller than it. The third option is correct.

Please log in or register to report this question.

Discuss in Forum
8.

The interscorer reliability is especially important for

1. Projective tests
2. Achievement tests with short answers
3. Multiple choice tests
4. Judgements by multiple raters
Codes :

Exam Year(s)

UGC NET - 2013,

Answer: D

Interscorer reliability is especially important when scoring needs judgement. That is the case for projective tests (1), achievement tests with short answers (2), and judgements by multiple raters (4), so the answer is 1, 2 and 4 only.

Multiple-choice tests (3) are scored objectively with a key, so different scorers give the same score and the question of agreement hardly arises. The options that include statement 3 or omit statement 2 are therefore wrong.

Please log in or register to report this question.

Discuss in Forum
9.

Consider that a newly developed test for assessing abstract reasoning and a well established test of abstract reasoning are administered to a sample of 350 and a correlation of 0.25, significant at .01 level is obtained. It will be inferred that the new test has

Exam Year(s)

UGC NET - 2013,

Answer: A

The new test has unsatisfactory concurrent validity. Concurrent validity is shown by the correlation of the new test with a well-established test of the same ability administered at about the same time, and a satisfactory value is expected to be high.

A correlation of 0.25, though significant at .01 with a sample of 350, is low in size, so the new test shares very little with the established one. Predictive validity concerns a later criterion, which is not involved here. So the first option is correct.

Please log in or register to report this question.

Discuss in Forum
Direction for [ Question No: 10 To 10 ] :

Read the following passage and answer the four questions that follow :

A psychologist developed a new Anxiety Inventory of 25 items. The Cronbach alpha for the inventory was found to be 0.80. The inventory was administered to a group of 50 anxiety disorder patients before beginning their psychotherapy. The patients had a significantly higher mean than the normative group of normals. The same group of patients was retested after 12 weeks of psychotherapy. The post-test mean was significantly lesser than the pre-test mean. The new Anxiety Inventory was correlated with an established anxiety inventory and the correlation turned out to be 0.20. The norms of the inventory were developed in terms of normalized T scores.

10.

What would be the median T score for the new Anxiety Inventory ?

Exam Year(s)

UGC NET - 2012,

Answer: A

The median T score for the inventory would be 50.

T scores are normalised scores with a mean of 50 and a standard deviation of 10. Because the scores are normalised, their distribution is normal, and in a normal distribution the median equals the mean, which is 50. The values 100 and 120 do not fit a scale centred on 50, and the data given are enough to answer, so option D, saying they are inadequate, is wrong.

Please log in or register to report this question.

Discuss in Forum
Loading…

Topics

Other Subjects

Tags: Psychological Testing Questions and Answers

About Us

ExamsRepo

Copyright © 2026 ExamsRepo . All Rights Reserved . With