| A | B |
|---|
| NON-PARAMETRIC | measures which are not based upon the assumptions of the normal curve |
| KOLMOGOROV-SMIRNOV | test for comparing one sample design, ordinal scales |
| SIGN TEST | test for repeated measures, binary nominal scale; a version of the binomial distribution |
| MANN-WHITNEY | a rank sum test for two independent samples |
| FRIEDMAN | a ranks test for repeated measures |
| KRUSKAL-WALLIS | a rank sum test for more than two independent samples |
| SPEARMAN | a correlation coefficient based upon ranks |
| ROBUST | non-parametric tests are more cautious; they tend to avoid Type I error, therefore they are more |
| POWERFUL | parametric tests tend to avoid Type II error; they are more likely to reject the null and declare significance; therefore they are more |
| BINOMIAL | an exact non-parametric test appropriate for sample vs. norms, when all Bernoulli requirements are met |
| CHI SQUARE | an inferential statistic for nominal rows and columns data |
| RELIABILITY | consistency of measurement |
| VALIDITY | when the operational definition measures what it purports to measure |
| YATES | developed a correction formula for the two-by-two chi square |
| MCNEMAR | developed a formula for a repeated measures chi square |
| SENSITIVE | a measure with few false negatives |
| SPECIFIC | a measure with few false positives |
| RELIABILITY | consistency between measurements: the same subject gets a similar score on the same variable when measured again |
| VALIDITY | when a measurement succeeds in measuring what it intends to measure |
| FALSE POSITIVE | a case that is scored as having a characteristic, when actually it does not |
| FALSE NEGATIVE | a case that is scored as not having a characteristic, when actually it does |
