Karteikarten
Tests of differences : Non-Parametric
Karten in diesem Stapel (17)
- 1Types of Non-parametric tests of differences1. Mann-Whitney U
2. Wilcoxon Signed Rank
3. Kruskal-Wallis
4. Friedman's ANOVA
5. Chi-Square - 2Why Mann-Whitney UExamine differences between two groups
-Comparisons of characteristics between groups at start of study
-Used for nonparametric data or data not normally distributed - 3Limitations of the Mann-Whitney U- Does not assume normal distribution
- Use with small sample size - 4Why Wilcoxon Signed Rankexamine differences within the same group
-Pretest/posttest designs in which subject act as own control
-Examines differences in the median
-Used for ordinal or nominal data - 5Limitations of Wilcoxon Signed RankNONE!
- 6Why Kruskal-WallisExamine difference between three or more groups
-IV has 3 or more levels
-Only 1 DV
-Used with parametric data that is not normally distributed or Ordinal and Nominal data - 7Limitations of Kruskal-Wallis- Does not assume normal distribution
- Use with small sample size
- Three or more groups will not tell you where difference is (need post hoc testing) - 8Why Friedman’s ANOVA- examine differences among repeated measure on subjects within the same group
- Repeated measures designs in which subjects act as own control
- Only one DV
- Designed for nonparametric data (ordinal and nominal) - 9Limitations of Friedman’s ANOVADoes not assume normal distribution
Use with small sample size
Three or more time points will not tell you where difference is (need post hoc testing) - 10Why Chi Squarerelationship between two variables
- Nominal data
- Frequencies - 11Limitations of Chi Square-Two sets of data are independent of each other
- 12Types of errors in statsType 1, 2
- 13Type 1 Error-conclude that a difference exists when in fact there is no difference (i.e., false positive)
α = probability of making a Type I error - 14Type 2 Errorconclude that no difference exists when in fact there is a difference (i.e., false negative)
β = probability of making a Type II error
inversely related to α - 15Power“the likelihood that one will detect a difference (relationship) when one exists”
- 16Methods for increasing power1. control extraneous variables & apply methods consistently
2. use homogeneous groups
3. increase sample size - 17Sample Size Determinants-desired level of power
-α level to be used in research
-estimate of effect size (clinically meaningful relationship or difference between groups)
Informationen
- Beschreibung
- Masters Research class Karteikarten am Tests of differences : Non-Parametric, erstellt von Alessandra poirier am 18/07/2017.
- Erstellt von
- Alessandra poirier
- Aufrufe
- 1
- Bewertung