Intangible truths

Q. Examine the different methods of judging validity of the test items and the test as a whole.

8 thoughts on “Intangible truths”

  1. Validity is one of the critical stages of test construction and it measures ” whether and to what extent the test is able to measure that it intends to measure”. It is employed at overall test level and test-item level.

    Judging validity at test level:-
    (a) Construct validity: Construct validity is done at test level to establish the validity between test and the construct its measuring. Some of the methods that it follows:

    1) Convergent validity: In this method, we check to check the theoretical relationship established between test and construct is actually reflected in the actual scenario. It is tested, if there is a high correlation between test scores and corresponding construct manifestation.
    e.g. Subjects scoring above 140 in MMPI are termed “gifted” and often its reflected in their carrier and academic performance

    2) Discriminant validity: The constructs to which test is theoretically unrelated, should also not show any correlation between test scores and the respective construct manifestation.
    e.g. Tests which are measuring analytical intelligence should not have correlation with weight of individual

    3) Incremental validity: For the practical uses, test is said to have incremental validity, if its giving better results in measuring construct than the previous tests.

    (b) Criterion validity: It is done to establish to what extent the test is valid in measuring the construct it intends to measure.

    1) Concurrent validity: a.k.a. external validity is established by establishing strong correlation between the test scores and the scores on the proven and well established scales that are measuring the same construct
    e.g. The test scores of a person taking an Intelligence test should have strong correlation with proven tests like MMPI

    2) Predictive validity: it measures to what extent the psychologist is able to predict the construct accurately based on the test score
    e.g. Any score below 70 in MMPI is termed as ‘Mental retardation’ and it has legal backing too.

    Similarly, Judging validity at item level:-
    (a) Content validity: it seeks to establish the correlation at the item level and the construct, so that any superficial items can be eliminated.

    (b) Face validity: it seeks to establish to what extent the items are correlated with the construct. Helps in refining the quality of the content
    e.g. Items questioning the political interest should not be adjudged in analytical intelligence test even if they show some correlation

      1. Sir,
        I agree with simplification remark. but
        I did organize the structure as per the ask of question.

        Test level :-
        (a) construct validity – is there a real correlation between test and construct ?
        (b) criterion validity – if so, how correctly and to what extent ?
        Test item level:-
        (a) content validity – Are items are correlated with construct ?
        (b) face validity – Going beyond correlation, assessing quality of items

        Please, let me know if my understanding about these terms is incorrect.

        1. Does the criterion validity apply for the content or the test as a whole?… It actually applies for the both. Eg criterion key validation of items in MMPI and in Binet and Simon test. Moreover, Test as a whole also requires criterion validation. One criteria for this is predictive validation…
          Think more about this.

  2. Validity examines the degree to which the test measures what it intends to measure. It is established through :
    1. Content validity
    2. Criterion validity
    3. Construct validity
    First one measures item validity and the last two measures the validity of test as a whole.

    1. CONTENT VALIDITY: It includes:
    (i) Item validity- It measures the relevance of the content chosen to what the test intends to measure. Eg: does measuring mathematical skills(item) really measure intelligence in an IQ test
    (ii) Sampling validity- it measures the coverage area of what the test intends to measure eg: wescher scale can be said to have a better sampling validity than binet scale because it includes problem solving in addition to verbal and mathematical.

    2. CRITERION VALIDITY: This is established through
    (i) Predictive validity – this measures how much the test score is able to predict future outcomes eg: correlation between test score that measures mechanical apititude and performance of the employee in this area (criterion score) after 2years.
    (ii) Concurrent validity – it measures how much the test score correlate with another test measuring the same criteria. Eg: correlation between newly designed IQ test and wescher scale.

    3. CONSTRUCT VALIDITY : this establishes if what is being measured actually exists. Thus it is used to device test that intends to measure a criteria for the first time, that is theoretically defined and not operationally. This can be judged through:
    (i) Convergent validity
    (ii) Discriminant validity
    The test score should correlate with theoretically similar tests and should not correlate with irrelevant tests. Eg: numerical aptitude test should correlate with arithmetic reasoning test(convergent) and should not correlate with history test(divergent).

    1. Good analysis and answer.. The structure is nearly perfect. You may also include cross validation process for evaluating the validity of a test.

  3. Examine the different methods of judging validity of the test items and the test as a whole.
    Validity of a test means the test measure what it intends to measure.
    The different ways of measuring test and item validity are:

    1. Content Validity: a> Item validity: It measure whether the item chosen measures dimension of interest eg. Social behaviour can contribute as aspect of anxiety.
    b>Sampling Validity: Appropriateness of the sample with minimum error. It should represent various section of sample take. eg. Anxiety construct should be viewed from various dimensions, social, personal, physical aspects.

    2. Criteria Validity: The extent to which measure of a concept shows correlation with independent variable of same concept. Two type of criteria validity are:
    1. Predictive criteria: It is done after the score of the test is obtained. It ensure the predictive value. Eg. Good score in midterm test can predict good score in finals.
    2. Concurrent criteria: It measures how well score on one test relates to similar test on same concept. Eg IQ test by Binnet and Devaiation IQ.

    3.Construct validity: It validates that the new attribute measures what is being tested. It include:
    a. Convergent validity: The test or item must show positive correlation with similar test or item assessing same subject. eg IQ and Deviation IQ both show positive correlation and measures Intelligence.
    b. Divergent validity: Test shows correlation not only with similar test but dissimilar test.

    4. Cross validation: Positive result in initial test doesn’t guarantee same result next time. It is a way to check the method used one one sample also produces same result on other samples too. eg intelligence test should be valid for different sections of the society.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.