How to Choose the Right Statistical Tool in Prism

Type of Analysis

Compare Relationship Categorical Data Survival Analysis Sensitivity/Specificity

Means/Median Between X and Y

More Than Two Groups Correlation

Two Groups Simple Regression Multiple Regression Nonlinear Regression 2 x 2 Table 2 x Y Table Kaplan-Meier ROC Curve
(Drug A vs Drug B (Strength of
(One X) (More Than One X) (Dose Response)
vs Drug C) Relationship)

Same Subjects 100+ Nonlinear Model

Each Subject Normal Distribution Non-Normal Response is Response is Binary Response is Response is Binary Response is a Count All Expected Not All Expected Chi-square Test
Measured Twice Plus User Defined
Measured Once Distribution Continuous Continuous Cell Counts at Least 5 Cell Counts at Least 5
(Before/After) Models Available

Pairwise Differences Pairwise Differences Each Group Normal and Each Group Normal One or Both Groups Pearson Correlation Spearman Correlation Least Squares Logistic Regression Least Squares Logistic Regression Poisson Regression Response Continuous Response Continuous Response is a Count Chi-square Test Fisher’s Exact Test
Normal Non-Normal Variances Equal Non-Normal Regression Regression No Outliers With Outliers

Wilcoxon Matched Pairs Unpaired t test With

Paired t test Unpaired t test Mann-Whitney Test Least Squares Robust Regression Poisson Regression
Signed Frank Test Welch's Correction
(Parametric Test) (Parametric Test) (Nonparametric Test)* Regression
(Nonparametric Test) (Parametric Test)

One Factor (Drug) Two Factors Three Factors

(Drug, Diet) (Drug, Diet, Time)

Each Subject Measured Each Subject Each Subject Measured Each Subject Each Subject Measured Each Subject
More Than Once Measured Once More Than Once Measured Once More Than Once Measured Once

* The Mann-Whitney test is used to determine whether there are differences in the
medians of your two groups and assumes the two distributions have the same shape.
Normal Distribution Non-Normal Normal Distribution and Normal Distribution Non-Normal Normal Distribution Non-Normal Normal Distribution Non-Normal Normal Distribution Non-Normal Normal Distribution Non-Normal It can be applied to ordinal or continuous data.
Distribution Variances Equal Distribution Distribution Distribution Distribution Distribution
** The Kruskal-Wallis test is used to determine if there are differences in the medians
of your groups and assumes all the group distributions have the same shape. It can
be applied to ordinal or continuous data. The Kruskal-Wallis test cannot tell you which
specific groups are statistically different from each other. It only tells you whether
Brown-Forsythe or or not at least two groups are different.
Assume Sphericity Don't Assume Friedman Test ANOVA Kruskal-Wallis Test Assume Sphericity Don't Assume Consider ANOVA Consider Assume Sphericity Don't Assume Consider ANOVA Consider
Welch’s ANOVA Transformation
Sphericity (Nonparametric Test) (Parametric Test) (Nonparametric Test)** Sphericity Transformation (Parametric Test) Transformation Sphericity Transformation (Parametric Test)
(Parametric Test)

www.graphpad.com
ANOVA With G-G ANOVA With G-G
ANOVA or Mixed Model
ANOVA or Mixed Model ANOVA or Mixed Model Correction or Mixed ANOVA or Mixed Model Correction or Mixed
(Missing Data)
(Missing Data) (Missing Data) Model (Missing Data) (Missing Data) Model (Missing Data)
With G-G Correlation
With G-G Correction With G-G Correction
 How to Choose the Right Statistical Tool in Prism
Type of Analysis

Compare
Means/Median

More Than Two Groups Correlation

Two Groups Simple Regression
(Drug A vs Drug B (Strength of
(One X)
vs Drug C) Relationship)

Same Subjects
Each Subject Normal Distribution Non-Normal Response is Response is Binary
Measured Twice
Measured Once Distribution Continuous
(Before/After)

Pairwise Differences Pairwise Differences Each Group Normal and Each Group Normal One or Both Groups Pearson Correlation Spearman Correlation Least Squares Logistic Regression
Normal Non-Normal Variances Equal Non-Normal Regression

Wilcoxon Matched Pairs Unpaired t test With

Paired t test Unpaired t test Mann-Whitney Test
Signed Frank Test Welch's Correction
(Parametric Test) (Parametric Test) (Nonparametric Test)*
(Nonparametric Test) (Parametric Test)

One Factor (Drug) Two Factors

(Drug, Diet)

Each Subject Measured Each Subject Each Subject Measured Each Subject
More Than Once Measured Once More Than Once Measured Once

Normal Distribution Non-Normal Normal Distribution and Normal Distribution Non-Normal Normal Distribution Non-Normal Normal Distribution Non-Normal
Distribution Variances Equal Distribution Distribution Distribution

Brown-Forsythe or
Assume Sphericity Don't Assume Friedman Test ANOVA Kruskal-Wallis Test Assume Sphericity Don't Assume Consider ANOVA Consider
Welch’s ANOVA
Sphericity (Nonparametric Test) (Parametric Test) (Nonparametric Test)** Sphericity Transformation (Parametric Test) Transformation
(Parametric Test)

ANOVA With G-G

ANOVA or Mixed Model
ANOVA or Mixed Model ANOVA or Mixed Model Correction or Mixed
(Missing Data)
(Missing Data) (Missing Data) Model (Missing Data)
With G-G Correlation
With G-G Correction
 Relationship Categorical Data Survival Analysis Sensitivity/Specificity
Between X and Y

Multiple Regression Nonlinear Regression 2 x 2 Table 2 x Y Table Kaplan-Meier ROC Curve

(More Than One X) (Dose Response)

100+ Nonlinear Model

Response is Response is Binary Response is a Count All Expected Not All Expected Chi-square Test
Plus User Defined
Continuous Cell Counts at Least 5 Cell Counts at Least 5
Models Available

Least Squares Logistic Regression Poisson Regression Response Continuous Response Continuous Response is a Count Chi-square Test Fisher’s Exact Test
Regression No Outliers With Outliers

Least Squares Robust Regression Poisson Regression

Regression

Three Factors
(Drug, Diet, Time)

Each Subject Measured Each Subject

More Than Once Measured Once

* The Mann-Whitney test is used to determine whether there are differences in the
medians of your two groups and assumes the two distributions have the same shape.
Normal Distribution Non-Normal Normal Distribution Non-Normal It can be applied to ordinal or continuous data.
Distribution Distribution
** The Kruskal-Wallis test is used to determine if there are differences in the medians
of your groups and assumes all the group distributions have the same shape. It can
be applied to ordinal or continuous data. The Kruskal-Wallis test cannot tell you which
specific groups are statistically different from each other. It only tells you whether
or not at least two groups are different.
Assume Sphericity Don't Assume Consider ANOVA Consider
Sphericity Transformation (Parametric Test) Transformation

www.graphpad.com
ANOVA With G-G
ANOVA or Mixed Model Correction or Mixed
(Missing Data) Model (Missing Data)
With G-G Correction