L13: KRUSKALL-WALLIS TEST Flashcards
(24 cards)
How do you analyse differences between three or more groups (e.g., music played at slow, medium, and fast tempos), when it meets the assumptions
Use a one-way ANOVA to compare the means across the three tempo groups (slow, medium, fast) to see if the length of music differs significantly between them
What do you do after finding a significant result in a one-way ANOVA?
Run post-hoc tests (like Tukey’s HSD) to determine which specific groups differ from each other, while controlling for Type I errors.
What is the Kruskal-Wallis Test and when is it used?
The Kruskal-Wallis Test is a non-parametric alternative to one-way ANOVA. It’s used when ANOVA assumptions are violated (e.g., non-normal data, unequal variances, unequal group sizes). It compares three or more groups without assuming a normal distribution.
What kind of design does the Kruskal-Wallis test apply to?
It applies to independent measures designs — where different people are in different groups.
How many groups can the Kruskal-Wallis test compare
It can compare three or more groups.
What is the Kruskal-Wallis test similar to?
It is similar to the Mann-Whitney U test, but while Mann-Whitney compares 2 groups, Kruskal-Wallis compares 3 or more
: What type of data does the Kruskal-Wallis test use?
It uses ranked (ordinal) data
What is the Kruskal-Wallis test an alternative to?
It’s an ordinal-data alternative to the one-way ANOVA
What is an example of when to use the Kruskal-Wallis Test?
When investigating whether tempo (slow, medium, fast) affects people’s duration estimates of a 45-second piece of music, using three separate groups (one per tempo)
What is the Shapiro-Wilk test used for?
To test if data distribution significantly differs from normal.
What does a significant Shapiro-Wilk test result indicate?
The data deviate from a normal distribution
Why should you use histograms alongside the Shapiro-Wilk test?
Because the test can be overly sensitive in large samples, so histograms help visually confirm normality
Can the Shapiro-Wilk test give misleading results?
Yes, in large samples, small deviations can produce significant results even if data are roughly normal
What does a non-significant Shapiro-Wilk test result mean?
Data are normally distributed; parametric tests are appropriate
What does a significant Shapiro-Wilk test result (p < .05) mean
Data are not normally distributed; use non-parametric tests.
If the Shapiro-Wilk test shows non-normality, which type of test should you use?
A non-parametric test, like the Kruskal-Wallis test.
Why continue with Kruskal-Wallis even if Shapiro-Wilk is non-significant?
For demonstration or consistency in a lecture on non-parametric tests and because it doesn’t assume normality so its safe and consistent espciecll when sample, size is small and data are ordinal o skewed
Helps to prevent wrong conclusions even if data does fit assumption perfectly.
Double checking that it is non-significant
What assumption does Levene’s test check?
Homogeneity of variance (equal variances across groups)
What does a non-significant Levene’s test result mean?
Variances are roughly equal; assumption is met
What does a significant Levene’s test result indicate?
Variances differ; assumption is violated
What is the most popular post-hoc test after Kruskal-Wallis?
Dunn’s test, especially when sample sizes are unequal
What are Bonferroni and Holm’s Bonferroni corrections used for?
To adjust for multiple comparisons, but they can be conservative and may miss true positives
After a significant Kruskal-Wallis test, what does Dunn’s test show?
Which specific groups differ significantly from each other.
How do you determine the direction of differences in Kruskal-Wallis post-hoc tests?
By comparing group median or mean scores
