Chapter 11: Measuring Research Variables SFM 651: Research Methods Dr. Johnson
Validity The degree to which a test or instrument measures what it is supposed to measure Logical – degree to which the measure obviously involves the performance being measured. Valid by definition Also called face validity Content – degree to which the test adequately samples what was being covered Usually educational Exam in class
Validity Criterion – degree to which scores on a test are related to some recognized standard MAP test Olympic Figure skating judges Predictive – can predict scores Construct - degree to which a test measures a hypothetical construct Looking at things that cant be observed Anxiety, creativity, attitude
Reliability The consistency or repeatability of a study If we do the same things again, will we get the same result? Ranges from 0.00 to 1.00 Closer to 0.00 means more error Closer to 1.00 means less error Intertester reliability Will the same tester get the same scores when observed.
ANOVA Analysis of Variance – called ANOVA Most common type of statistical analysis Looks at the Null hypothesis Tests the differences between two means General differences, not specific Often used in Reliability studies.
Stability Coefficient of reliability measured by the test-retest method on different days Will the same result occur again? More used on fitness tests Not so much on paper and pencil tests
Standard Measure of Error Scores fall within a range Standard Deviation * Square root (1.00-r) 22.4% ± 2.3%
Z Scores Basic standard score that converts raw scores to units of standard deviation in which the mean is 0 and SD is 1. Can compare 2 tests easily Vertical jump and push-ups
T-Scales Type of Standard score that sets the mean at 50 and standard Deviation at 10 to remove the decimal found in z scores and to make all scores positive. 1 SD above the mean = 50 + 10 = 60 1 SD below the mean = 50 - 10 = 40
Likert Scales Type of closed question that requires choosing from one of several scaled responses Usually 1-5 or 1-10 if number Usually strongly agree to strongly disagree is words
Errors Central tendencies error Raters give more scores in the middle. Observer Bias Error – expecting a certain response Halo effect – previous impressions influence ratings