Research Methods ANOVA
Tukey Honestly Significant Difference Method - ✔️✔️- rather than using the type I error
probability rate of 0.05 for individual comparisons, it is reduced
- this way, the probability of a type I error in the set of multiple pairwise comparisons is
less than .05
- this is called the "familywise error rate"
Arcsine Transformation - ✔️✔️- ANOVA assumes normal distribution & equal variance
across groups
- this is not true of proportional data (eg. test scores), which tend to have less variance
at the low & high ends (eg. few ppl score below 20%)
- Arcsine transformation helps overcome this problem
Two-Way ANOVA - ✔️✔️Basic principles & interpretation follow one-way ANOVA, plus:
- interaction terms
- fixed & random factors
Degrees of freedom in 2-way ANOVA - ✔️✔️- number of groups for a given factor
minus 1, eg. 3 hair colour groups -> 2 degrees of freedom, plus 3 eye colour groups ->
another 2 degrees of freedom = 4 degrees of freedom for factor levels
- interaction: degrees of freedom for each group multiplied, in this case 2 x 2 -> 4
degrees of freedom for interaction
- within groups: total number of subjects minus number of cells in the design. in this
case 3 x 3 = 9 cells. If 10 participants per cell, 90 participants altogether. 90 - 9 = 81
degrees of freedom
Analysis of Variance - ✔️✔️A tool for analysing differences between groups of
measurements
- groups are based on a property of interest (eg. hair colour
Basic Principles of ANOVA - ✔️✔️ANOVA divides up the total variance in a set of
measurements into:
- a contribution from BETWEEN groups and
- a contribution from WITHIN groups
If between group variance is larger than within group variance, then the specified factor
has a significant effect on results
Tukey Honestly Significant Difference Method - ✔️✔️- rather than using the type I error
probability rate of 0.05 for individual comparisons, it is reduced
- this way, the probability of a type I error in the set of multiple pairwise comparisons is
less than .05
- this is called the "familywise error rate"
Arcsine Transformation - ✔️✔️- ANOVA assumes normal distribution & equal variance
across groups
- this is not true of proportional data (eg. test scores), which tend to have less variance
at the low & high ends (eg. few ppl score below 20%)
- Arcsine transformation helps overcome this problem
Two-Way ANOVA - ✔️✔️Basic principles & interpretation follow one-way ANOVA, plus:
- interaction terms
- fixed & random factors
Degrees of freedom in 2-way ANOVA - ✔️✔️- number of groups for a given factor
minus 1, eg. 3 hair colour groups -> 2 degrees of freedom, plus 3 eye colour groups ->
another 2 degrees of freedom = 4 degrees of freedom for factor levels
- interaction: degrees of freedom for each group multiplied, in this case 2 x 2 -> 4
degrees of freedom for interaction
- within groups: total number of subjects minus number of cells in the design. in this
case 3 x 3 = 9 cells. If 10 participants per cell, 90 participants altogether. 90 - 9 = 81
degrees of freedom
Analysis of Variance - ✔️✔️A tool for analysing differences between groups of
measurements
- groups are based on a property of interest (eg. hair colour
Basic Principles of ANOVA - ✔️✔️ANOVA divides up the total variance in a set of
measurements into:
- a contribution from BETWEEN groups and
- a contribution from WITHIN groups
If between group variance is larger than within group variance, then the specified factor
has a significant effect on results
