Reliability

Question 1

Reliability

Answer 1

An estimate that is used to determine how consistent a measure is. Provides an estimate of error (high reliability = low error).

Question 2

Sources of Error in Tests

Answer 2

Construction (item sampling), administration (attention, motivation, tester characteristics) and scoring/interpretation (meaning of scores)

Question 3

Test-Retest Reliability

Answer 3

If a test is given to the same person at a later date, the scores should be similar (however, practice and learning may inflate later scores).

Question 4

Parallel Forms

Answer 4

Two versions of one test with equal means and SD’s (very difficult to achieve).

Question 5

Alternate Forms

Answer 5

Two versions of one test are designed to be parallel, but the means and SD’s are different.

Question 6

Split-half Reliability

Answer 6

Obtained by correlating two halves of one test to see if both halves measure the same thing (Spearman-Brown formula).

Question 7

Inter-item Consistency

Answer 7

Compare each item score with the overall test score to determine if the items are consistently tapping the same thing.

Question 8

Kuder-Richardson Reliability

Answer 8

Calculates internal consistency in dichotomous tests.

Question 9

Coefficient Alpha

Answer 9

Calculates internal consistency in multipoint tests.

Question 10

Inter-rater Reliabiity

Answer 10

Degree of agreement that exists between two or more raters.

Question 11

Standard Error of Measurement

Answer 11

SD of a theoretically normal distribution of test scores obtained by one person on equivalent tests. eg an IQ test has a SEM of 10, so a score can vary by + or - 10 and not be significantly different.

Question 12

Standard Error of Difference

Answer 12

Enables a determination of the size of the difference between two scores that is required for significance.