Week 4 Flashcards
What is a ANOVA test ?
An inferential parametric statistical test for comparing the means of 3 or more groups
* stands for Analysis Of Variance
What does ANOVA compare ?
ANOVA compares the amount of variance caused by the manipulation (IV) vs random error
What is the F-ratio for 1-way ANOVA ?
F = between groups Variance/Within groups variance
which equals
manipulation + error/error
What is One-way ANOVA ?
An inferential parametric statistical test for comparing the means of 3 or more groups using a between-subject design
* Only one independent variable
What is between-groups variance vs within groups variance ?
- Between group: An estimate of the effect of the independent variable and error variance
- Within-groups: An estimate of the population error variance
What occurs when the manipulation does or does not have an effect ?
- Does: F-ratio score is significantly above 1
- Does not: F-ratio score equals 1
What in involved in a One-way ANOVA source table
- Between: variability due to IV
- Within: inherrent variability (error)
- Total: use this an error check for out SS and df calculation
- SS: sums of sqaures
- df: degrees of freedom
- MS: mean squares
- Fcalc: the calculated value of F
- Fcrit: the critical (table value) of F
When is there a significant effect in a one-way ANOVA test ?
When Fcalc is greater than Fcrit
* at least one group that is different from the rest
What is used to calculate the effect size in a one-way ANOVA ?
eta-squared: n2
* SSB/SST
What are the assumptions of 1-way ANOVA ?
- The data (DV) are on an interval or ratio scale
- The underlying distribution is normally distributed
- The variance among the groups being compared are homogenous (homogeneity of variance)
When do you use multiple comparisons or post hoc tests ?
Use them to compare all possible pairs of groups to determine which ones differ significantly from each other
When do you use Turkey honestly significant difference (HSD) test ?
Used to make pairwise comparisons when the different levels have equal n
How do you use the HSD value?
- Find the mean of each group
- For each pair of comparison, if the value of difference is greater than the HSD, then there is a significant difference b/w this pair of means (ignore the signs)
