Quick reminder nonparametric tests
Download as PPTX, PDF6 likes1,646 views
Nonparametric tests use the median instead of the mean to calculate differences between groups. These tests have parametric analogues that use the mean. Some common nonparametric tests include the Wilcoxon test, Mann-Whitney U test, and Kruskal-Wallis test, which correspond to the t-test, independent samples t-test, and one-way ANOVA respectively. Z-tests are also nonparametric and used to compare proportions between samples and populations or between two samples.
1 of 2
Downloaded 54 times
More Related Content
What's hot (19)
Ad
Viewers also liked (20)
Ad
Similar to Quick reminder nonparametric tests (20)
More from Ken Plummer (20)
Quick reminder nonparametric tests
- 1. Nonparametric methods that test for differences between groups are the same as Parametric Tests except that they use the Median instead of the Mean in their calculations. Here are the non-parametric tests with their parametric analogues Non-Parametric Parametric One-Sample Wilcoxon Test Single-Sample t Mann Whitney U Test Independent Samples t-test Matched Paired Wilcoxon Test Paired Samples t-test Kruskal Wallis One-way ANOVA Friedman Test Repeated Measures ANOVA Comparing Non-Parametric Tests that Differentiate Between Groups Sample 1 Sample 2 Sample 1 – Time 1 compared with Sample 1 – Time 2 Sample 1 – Time 1 Sample 1 – Time 2 Sample 1 – Time 3 NEXT Sample 1 Sample 2compare Sample 3& Sample 1 compared with compared with POPULATION compare &
- 2. Z-tests are non-parametric tests used to investigate differences between proportions or percentages: Single sample z-test – compares a single sample of proportions with its supposed population or claim. Two sample z-test - compares two sample proportions. Sample - 85% say yes Population - 90% say yes Sample 1 – 84% say yes Sample 2- 86% say yes compared with Comparing Non-Parametric Tests that Differentiate Between Groups