Hypothesis Testing Roadmap For all tests:

p > 0.05 Fail to Reject Ho (null)

p < 0.05 Reject Ho
Hypothesis Testing
Chi-Squared
Ho: Two factors are independent
Ha: Two factors are dependent
Continuous Data Minitab:
Attribute Data Stat -Tables - Chi-square Test
Normality Test Which causes dependency? – P
Ho: Data is Normal Chart
Non Normal Ha: Data is NOT Normal Normal
Minitab:
Stat - Basic Stat-Graphical Summary -
Anderson-Darling P Value
2 or more samples 1 Sample 1 Sample 2 or more samples
Variance
Chi-Squared
Variance Ho: 1 =  target Test For Equal Variance
Ha:  1   target Ho:  1 =  2 =  3 = ...
Test for Equal Variance Mean to Variance Variance Delta
Minitab: Ha: at leastone is different
Ho: s1 = s2 = s3 = ... Target Stat – Basic Stats – Graphical between
Ha: at least one is different Minitab:
Summary 2
Minitab: Stat – Basic Stats – 2 Variances
Target falls within CI, then fail measures
Stat-Basic Stats-2 Variances H o: M 1 = M tar get Mean to OR Stat-Anova-Test for Equal
OR H a: M 1  M tar get to reject Ho. Variances
M initab:
Target
Stat-Anova-Test for Equal Bartlett's Test
Stat - Nonpar ametr ic - 1
Variances Sample-Sign (O R) Paired T
Levene’s Test Stat - Nonpar ametr ic - 1 Comparison Means
Sample-Wilcoxon
1 Sample T Test Ho: M1-M2 = 0 2 Samples
Ho: 1 =  target Ha: M1-M2  0
Medians Minitab:
Ha:  1   target
Stat - Basic
3 or more Samples
Minitab:
Ho: M1 = M2 = M3 = ... Stats – Paired t
Stat - Basic Stats - 1 Sample-T (Measurement Equal Unequal
Ha: at least one is different
Minitab: of the same Variances Variances
Stat - Nonparametric - Mann-Whitney (Similar to 2 thing 2 times)
sample T). P value appears as “Significant at”. One Way ANOVA
OR Ho:  1 =  2 =  3 = ...
Stat - Nonparametric - Kruskal-Wallis (OR) Ha: at least one is different
(More powerful when data is from many Minitab:
distributions) 2 Sample T Test – Equal 2 Sample T Test – Stat - Anova- One-way (or
OR Variances Unequal Variances one-way unstacked)
File
StatName Here
- Nonparametric - Mood’s Median (OR) Comparison  Tukey’s
Ho:  1 =  2 Ho:  1 =  2
(Compares Medians, not Means. Robust to
Ha:  1  2 Ha:  1   2
outliers)
OR Minitab: Minitab:
Stat - Basic Stats - 2-Sample T Stat - Basic Stats - 2- Must remove
Stat - Nonparametric - Friedmans variation. Can not
M1 = Median of sample 1, etc... (Compares Means using pooled Sample T
Std Dev) (Compares Means using do ANOVA with
 Assume equal variances each sample Std Dev) unequal
variances. 1