Assessment techniques must meet four technical criteria before they can be considered scientifically acceptable measures of individual differences in people’s enduring qualities. These criteria are standardisation, norms, reliability, and validity. Let us deal with each of these and understand what these terms mean.
A key concept in the measurement of personality dimensions is that of standardisation. This concept refers to the uniform procedures that are followed in the administration and scoring of an assessment tool. For instance, in self-report scale, the examiner must make every effort to ensure that subjects read and understand the printed instructions, respond to the same questions, and stay within any stated time limits. It also involves information (in the manual) about the conditions under which the assessment test should or should not be given, who should or should not take the test (sample group), specific procedures for scoring the test, and the interpretative significance of the scores.
Norms: The standardisation of a personality assessment test includes information concerning whether a particular “raw score” ranks low, high, or average relative to other “raw scores” on the test. Such information, called test norms, provides standards with which the scores of various individuals who take the test later can be compared. Usually, the raw scores on a test are converted into percentile scores, which indicate the percentage of people who score at or below a particular score. Thus, test norms permit the comparison of individual scores to a representative group so as to quantify the individual’s relative rank standing to others.
Reliability: Any test whether personality or intelligence or aptitude, etc. should have reliability and this should be demonstrated. Reliability means that repeated administrations of the same test or another form of test should yield reasonably the same results or scores. Thus, reliability refers to the consistency or stability of an assessment technique when given to the same group of people on two different occasions.
This kind of reliability is termed as test-retest reliability (Anastasi, 1968) . To determine test retest reliability, the scores from the first administration are correlated with those of the second by a simple correlation procedure. The magnitude of the resulting correlation coefficient gives us an estimate of the test’s consistency over time. Although there are no fixed guidelines about acceptable levels of reliability, the reliability coefficients for most psychological tests are above +.70. The closer this statistic approaches +1.00, the more reliable the test is. In other words, when retested, people’s scores should match their first scores quite closely.
A second kind of reliability is determined by splitting the test into two sets (e.g., odd-numbered items versus even-numbered items), summing people’s scores for each set, and correlating the two sets of summed scores with each other. The correlation between these sets is termed split half reliability.
From MPC-003 Personality: Theories and Assessment – IGNOU
Assessment techniques must meet four technical criteria before they can be considered scientifically acceptable measures of individual differences in people’s enduring qualities. These criteria are standardisation, norms, reliability, and validity. Let us deal with each of these and understand what these terms mean.
A key concept in the measurement of personality dimensions is that of standardisation. This concept refers to the uniform procedures that are followed in the administration and scoring of an assessment tool. For instance, in self-report scale, the examiner must make every effort to ensure that subjects read and understand the printed instructions, respond to the same questions, and stay within any stated time limits. It also involves information (in the manual) about the conditions under which the assessment test should or should not be given, who should or should not take the test (sample group), specific procedures for scoring the test, and the interpretative significance of the scores.
Norms: The standardisation of a personality assessment test includes information concerning whether a particular “raw score” ranks low, high, or average relative to other “raw scores” on the test. Such information, called test norms, provides standards with which the scores of various individuals who take the test later can be compared. Usually, the raw scores on a test are converted into percentile scores, which indicate the percentage of people who score at or below a particular score. Thus, test norms permit the comparison of individual scores to a representative group so as to quantify the individual’s relative rank standing to others.
Reliability: Any test whether personality or intelligence or aptitude, etc. should have reliability and this should be demonstrated. Reliability means that repeated administrations of the same test or another form of test should yield reasonably the same results or scores. Thus, reliability refers to the consistency or stability of an assessment technique when given to the same group of people on two different occasions.
This kind of reliability is termed as test-retest reliability (Anastasi, 1968) . To determine test retest reliability, the scores from the first administration are correlated with those of the second by a simple correlation procedure. The magnitude of the resulting correlation coefficient gives us an estimate of the test’s consistency over time. Although there are no fixed guidelines about acceptable levels of reliability, the reliability coefficients for most psychological tests are above +.70. The closer this statistic approaches +1.00, the more reliable the test is. In other words, when retested, people’s scores should match their first scores quite closely.
A second kind of reliability is determined by splitting the test into two sets (e.g., odd-numbered items versus even-numbered items), summing people’s scores for each set, and correlating the two sets of summed scores with each other. The correlation between these sets is termed split half reliability.
From MPC-003 Personality: Theories and Assessment – IGNOU