Stats Flashcards
What type of test is mixed factorial ANOVA?
Parametric
What is the purpose of mixed factorial ANOVA?
To show the effects of each IV and also the main interactions
What two assumptions are made by mixed factorial ANOVA and how can they be tested?
Homogeneity of variance - Lavene’s test
Sphericity of covariance - Maulchys test
What types of formulas are used for a) BS b) WS and c) interactions?
a) BS
b) WS
c) WS
How are the F values for mixed factorial ANOVA reported?
F(between groups df, within/error df) = F-value, p= p-value
What is the purpose of a correlation?
To investigate relationships between variables (NO CAUSATION)
Is there a non-parametric alternative to mixed factorial ANOVA?
No
What are the parametric and non-parametric forms of a correlation test?
Non-parametric - Spearman’s
Parametric - Pearson’s
What are the conditions that must be present in order to carry out a Pearson’s correlation?
- Must have linear relationship
- Data interval/ratio and normally distributed (as it involves means and SDs)
- Must be free of outliers
What are the conditions that must be present in order to carry out a Spearman’s correlation?
- Monotonic relationship
- Data that is ordinal/interval/ration
- Outliers can be accepted
What is the unit of a correlation coefficient?
none
How can a graph be z-transformed?
z= (score-mean)/SD
What does having a z-transformed graph allow?
Direct comparison even if measured on different scales
What type of statistics is regression?
Inferential
What is the purpose of regression?
Test of association which allows us to make predictions/estimate how much to intervene, used when causal relationships are likely
What are the assumptions of regression?
- Linear, interval/ratio data which is normally distributed and free of outliers
- Homoscedasticity, the same degree of variation across all predictor variable scores
- That predictors are not highly correlated to one another
What are residuals?
The difference between the actual outcome score and the predicted score outcome
How is a simple regression reported?
R^2 = R(squared), F(df regression, df residual) = F-value p= p-value
What are the 3 ways in which the different variables in a multiple regression can be entered?
Simultaneous - all predictors entered at the same time
Hierarchical - predictors entered in a pre-defined order, used when informed by well defined theory
Stepwise - Entered in order of how well they correlate with the outcome, rarely used as is unstable
