reciprocal transformation, 340–341
resampling, 342–343
square-root transformation, 340
trimmed samples, 341
when to transform and how to choose a transformation, 342
treatment effects (lambda), 323, 642–643, 652–654
expected cell frequencies, 653–654
interpreting the model, 654
ordinal variables, 654
treatment of ties, 677
trend, 402
trend analysis, 402–408
unequal intervals, 408
trimmed means, 35–36, 383
trimmed samples, 341
true-score analysis of covariance, 611
Tscores, 79
Tukey HSD (honestly significant difference) test, 391
Tukey’s test, 391
two-tailed test (nondirectional test), 100
two variable relationships, 520–521
two-way classification variables, contingency table
analysis, 145–148
2 3 2 tables are special cases, 147
calculation of chi-square, 146–147
correcting for continuity, 147
degrees of freedom, 147
evaluation of x^2 , 147
expected frequencies for contingency tables, 145–146
Fisher’s exact test, 147–148
Fisher’s exact test versus Pearson’s chi square, 148
two-way contingency tables, 631–637
conditional equiprobability model, 633–635
equiprobability model, 633
mutual dependence model, 635–636
two-way factorial design, 414
type I errors, 96–99
type II errors, 96–99
type I SS, 595
unbalanced design, 441
unbiased estimator, 46
unbiasedness, 46
unconditional probability, 117
underlying model, 319–321
assumptions, 320
homogeneity of variance, 320
independence of observations, 321
normality, 321
the null hypothesis, 321
unequal intervals, 408
unequal sample sizes, 234, 332–334, 444–446, 580, 592
effective therapies for anorexia, 332
and heterogeneity of variance, 392
uniform (rectangular) distribution, 181
unimodal distribution, 27
unweighted means, 444, 598validity, 294
variable, 4
variance, 40
variance estimation, 323
variance sum law, 204
Venn diagrams, 537
VIF (variance inflation factor), 527
violations of assumptions, 334–336weighted average, 206
Welch procedure, 335–336
Welch-Satterthwaite solution, 214
when to transform and how to choose a transformation, 342
whiskers, 50
Wilcoxon’s matched-pairs signed-ranks test, 679–682
normal approximation, 681–682
ties, 680–681
Wilcoxon’s rank-sum test, 673–678
Mann-Whitney Ustatistic, 678
the normal approximation, 675–677
null hypothesis, 677
treatment of ties, 677
Wilcoxon’s test and resampling procedures, 677–678
Wilcoxon’s test and resampling procedures, 677–678
Winsorized samples, 39, 341
worked examples
SPSS analysis, 211–212
writing up the results, 167, 213–214, 241, 330
WSD (wholly significant difference) test, 391Yates’ correction for continuity, 147zscore, 72768 Index