Skip to Main Content
Home Books Basic & Clinical Biostatistics, 5e

Appendix C: Flowcharts for Relating Research Questions to Statistical Methods

APPENDIX C: FLOWCHARTS FOR RELATING RESEARCH QUESTIONS TO STATISTICAL METHODS

| Download (.pdf) | Print
Flowchart to Use Research Questions
C–1 Is there a difference in means or medians for ordinal or numerical measures?
C–2 Is there a difference in means or medians for ordinal or numerical measures, three or more groups?
C–3 Is there a difference in proportions for nominal measures?
C–4 Is there an association?
C–5 Is there a difference in measures of association?
C–6 Are there two or more independent variables?

Figure C–1.

Is there a difference in means or medians for ordinal or numerical measures?

Number of groups: group 1 and group 2. group 1 branches into n greater or equal to 30, if yes leads to t (88 to 97) if no leads to normaly distributed, leads to t ()88 to 97 if yes and transform for t 106 to 107 or sign 107 to 109 if no. group 2: independent groups: yes leads to n greater or equal to 30 . it leads to t 118 to 123. it leads to transfor for 106 to 107 or sign 124 to 127 for no. independent groups if no leads to normalt distributed and leads to t paired 100 to 103 if yes and leads to transform for t 106 to 107 or wilcoxon signed ranks 109 to 110.
View Full Size||Download Slide (.ppt)
Number of groups: group 1 and group 2. group 1 branches into n greater or equal to 30, if yes leads to t (88 to 97) if no leads to normaly distributed, leads to t ()88 to 97 if yes and transform for t 106 to 107 or sign 107 to 109 if no. group 2: independent groups: yes leads to n greater or equal to 30 . it leads to t 118 to 123. it leads to transfor for 106 to 107 or sign 124 to 127 for no. independent groups if no leads to normalt distributed and leads to t paired 100 to 103 if yes and leads to transform for t 106 to 107 or wilcoxon signed ranks 109 to 110.
Figure C–2.

Is there a difference in means or medians for ordinal or numerical measures, three or more groups?

3 or more groups leads to independent groups. It leads to normally distributed, it leads to number of factors if yes and kruskal wallis for factor 1 160 to 161 if no. number of factors 1 factor leads to one way ANOVA 143 to 150, 2 or more factors lead to two way or other ANOVA 155 to 160. Independent if no leads to repeated measures ANOVA 159 and friedman 160.
View Full Size||Download Slide (.ppt)
3 or more groups leads to independent groups. It leads to normally distributed, it leads to number of factors if yes and kruskal wallis for factor 1 160 to 161 if no. number of factors 1 factor leads to one way ANOVA 143 to 150, 2 or more factors lead to two way or other ANOVA 155 to 160. Independent if no leads to repeated measures ANOVA 159 and friedman 160.
Figure C–3.

Is there a difference in proportions for nominal measures?

Number of groups: 1 group: n p and n (1 minus p) is greater than 5: z approximation 99 to 100. 2 groups: independent groups: if yes leads to small expected frequamncies: fishers exact test 133. if no leads to mcNemar of k. 3 or more: small expected frequencies: collapse categoties for x squared: 160 to 162, x squared 160 to 162, cochrans q not presented in this text.
View Full Size||Download Slide (.ppt)
Number of groups: 1 group: n p and n (1 minus p) is greater than 5: z approximation 99 to 100. 2 groups: independent groups: if yes leads to small expected frequamncies: fishers exact test 133. if no leads to mcNemar of k. 3 or more: small expected frequencies: collapse categoties for x squared: 160 to 162, x squared 160 to 162, cochrans q not presented in this text.
Figure C–4.

Is there an association?

Scale of Measurement for Both Variables leads to nominal: relative risk (44 to 46, 175 to 176), ordinal: spearmans rho, numerical: predicting one from another if yes: regression, if no pearson correlation.
View Full Size||Download Slide (.ppt)
Scale of Measurement for Both Variables leads to nominal: relative risk (44 to 46, 175 to 176), ordinal: spearmans rho, numerical: predicting one from another if yes: regression, if no pearson correlation.
Figure C–5.

Is there a difference in measures of association?

statistic: correlation coefficient: correlation 1 branches to t and Freshers z transformation. Correlation 2 leads to fishers x transformation and t. 1 regression line and 2 regression line from regression coefficient leads to t.
View Full Size||Download Slide (.ppt)
statistic: correlation coefficient: correlation 1 branches to t and Freshers z transformation. Correlation 2 leads to fishers x transformation and t. 1 regression line and 2 regression line from regression coefficient leads to t.
Figure C–6.

Are there two or more independent variables?

Scale of Measurement for Independent Variables: Scale of Measurement for Dependent Variable led to censored observations. It leads to Kaplan—Meier (195—197) or Actuarial (193—195), Mantel—Haenszel (199—201), Log Linear (221—222), Cox Proportional Hazard Model (222—224), and Analysis of Variance (143—160). Scale of Measurement for Dependent Variable: nominal leads to : logic Classification Methods (227—229), Confounding Independent Variable Cox Proportional Hazard Model (222—224), ANCOVA (215—218), and Multiple Regression (185—186, 209—215).
View Full Size||Download Slide (.ppt)
Scale of Measurement for Independent Variables: Scale of Measurement for Dependent Variable led to censored observations. It leads to Kaplan—Meier (195—197) or Actuarial (193—195), Mantel—Haenszel (199—201), Log Linear (221—222), Cox Proportional Hazard Model (222—224), and Analysis of Variance (143—160). Scale of Measurement for Dependent Variable: nominal leads to : logic Classification Methods (227—229), Confounding Independent Variable Cox Proportional Hazard Model (222—224), ANCOVA (215—218), and Multiple Regression (185—186, 209—215).

Pop-up div Successfully Displayed

This div only appears when the trigger link is hovered over. Otherwise it is hidden from view.

This site uses cookies to provide, maintain and improve your experience. MHE Privacy Center Close