1What contributes to variance between IV levels?
- Manipulation of IV (treatment effects)
- Individual differences
- Experimental error
• Random error
• Constant error
2What contributes to variance within IV levels?
- Individual differences
- Experimental error
• Random error
3What does a t-distribution represent?
Distribution of sampled mean differences when the null hypothesis is true
4What represents normality under Levene's statistic?
- DV normally distributed
- Homogeneity of variance
- Sample size roughly equal
- Scores should be independent
5When should normality be assumed in a repeated measures paired t-test?
Where n>30
6Cohen's d
Measure of effect size, assesses the difference between two means in terms of standard deviation
7Family-wise error rate
The probability of making type I errors when performing multiple hypotheses tests
8Corrected significance
1-(1-level of confidence) to the power of the number of comparisons
9F ratio
variance between groups/variance within groups
10Model sum of squares
Sum of squared differences between IV level mean and grand mean
11Residual sum of squares
Sum of squared differences between individual values and corresponding IV level mean
12Degrees of freedom for model variance
K-1 (K=number of IV levels)
13Degrees of freedom for residual variance
N-K (K=number of IV levels)
14Partial eta squared
Amount of variance in the dependent variable uniquely explained by a single categorical independent variable. SSm/SSm+SSr
15Partial eta squared effect sizes
>.01 = small
>.06= medium
>.14= large
16What is the difference between contributions to variance in independent and repeated measures designs?
There are no individual differences as participants partake in all conditions
17Mauchly's test of sphericity
If Mauchly's is significant, use Greenhouse-Geisser
18Degrees of freedom for repeated measures one-way ANOVAs
Between - k-1
Within - dfM x (n-1)
19How can order effects be controlled?
Practice - pre-study practice
Fatigue - short experiments
Sensitisation - intervals between exposure
Carry-over - include a control group
20main vs. interaction effects
Interaction effects look at the interaction between the two IVs rather than them separately (main)
21F ratio of interaction
variance between IV1*variance between IV2 / variance due to error
22When should post-hoc tests be conducted on factorial ANOVAs?
When effect is significant AND the IVs have more than 2 levels
23Simple effects
The effect of one IV at one level of another IV
24Mixed ANOVA
at least one between subjects IV and one within subjects IV
25Between subjects IV in a mixed ANOVA
- Check Levene's
- No need for post hoc if only 2 levels
- Run individual t-tests for each level
26Within subjects IV in a mixed ANOVA
- No need to check Mauchly's if only 2 levels
- 2 separate paired t-tests for each level
27Bivariate linear correlation
- Examines relationship between 2 variables.
- 0.4-0.6 moderate
28Non-linear correlation
May hit a peak then decrease
29Parametric assumptions of bivariate linear correlations
- Both variables continuous
- Related pairs (x and y value)
- Absence of outliers
- Linearity
- Range restrictions (floor and ceiling effects)
30Covariance formula
Sum of individual scores - mean / n-1
31Pearsons r formula
cov(x,y)/Sx*Sy
32Shared variance
Proportion of separate variables that is shared
33Partial correlation
the correlation between two variables with the influence of a third variable statistically controlled for
34linear regression
Estimate how much Y (outcome/criterion variable) changes in terms of X (predictor/explanatory variable)
353 stages of linear regression
- Analyse relationship (strength and direction)
- Propose model to explain (line of best fit)
- Evaluating model
362 values in a regression model
a = intercept, value of y when x=0
b = slope, how much y changes when x increased by 1 unit
37What does a higher F score for a linear model show?
Regression model is good at predicting future values of y
38What is the difference between r and r squared?
r = strength of relationship between x and y
r squared = proportion of variance explained by the model
39multiple regression
Assess influence of several predictor variables and obtain measure of how much variance in outcome
40Zero-order squared
Variance explained by Xi, expressed as proportion of total variance in Y
41Part squared
unique variance in y explained by x
42Partial squared
Unique variance explained by Xi expressed as proportion of the variance in y that remains after variance explained by other predictors has been removed
43Shared order equation
Zero order squared - part squared
44Hierarchical regression
Predictor variables entered in order based on theoretical grounds. Examines the influence of predictor variable on outcome when controlling for other variables.
45Non-parametric equivalent of independent t-test
Mann-Whitney U test
46Non-parametric equivalent to paired t -test
Wilcoxon signed rank test
47Non-parametric equivalent to independent one-way ANOVA
Kruskal-Wallis test
48Non-parametric equivalent to repeated measures one-way ANOVA
Friedman's ANOVA
49When should chi-square tests be used?
When the data is nominal (measured on a categorical scale)
50What is the relationship between sample size and p value?
As the sample size increases the p value decreases
51Power
Ability to find a statistically significant effect where one exists.
= 1 - risk of type 2 error
52Factors that can decrease power
- Smaller effect size
- More conservative criterion significance level
- Smaller sample size
- Non-parametric tests
- Independent subjects design
- Two-tailed hypotheses
53What 3 components can be used to calculate power?
- Number of participants
- Effect size expected
- Significance level
54Degrees of freedom for independent t test
n-2
55Grand mean
the mean performance across all participants in a study
56SSm
difference between each IV level mean and grand mean calculated, squared and summed
58What does a smaller model variance mean?
The null hypothesis is less likely to be accepted
59Degrees of freedom for RM ANOVA
Total participants * number of conditions
60Degrees of freedom for all one-way ANOVAs
IV levels - 1
61Cell means
he mean of each cell comparing one level of a factor across a level of another factor. Examined by simple effects
62marginal means
Mean scores for each level of IV, averaged across all levels of other IV. Tested by main effects and post-hoc analysis.
63What does a significant Mauchly's result show?
Homogeneity
64What do larger residuals show?
Larger error of prediction