Research Methods Tests Flashcards
Which tests are parametric?
Independent t-test Paired sample t-test One way ANOVA Repeated measures ANOVA Pearson's Partial correlation Simple regression Multiple regression
Which tests are non-parametric?
Mann Whitney Test Wilcoxon's Matched Pairs Kruskal-Wallis Test Friedman's ANOVA Spearman's Rho Kendall's Tau
Which tests use independent samples and 2 conditions?
Independent t-test
Mann-Whitney
Which tests use paired samples and 2 conditions?
Paired sample t-test
Wilcoxon matched pairs
Which test use independent samples and 3 or more conditions?
One way ANOVA
Kruskal-Wallis Test
Which test use paired samples and 3 or more conditions?
Repeated measures ANOVA
Friedman’s ANOVA
When to use a paired samples t-test?
- One continuous outcome variable (DV)
- One categorical predictor variable (IV)
- Two levels in which the same ps are exposed to both levels
- Parametric
When to use an independent t-test?
- One continuous outcome variable (DV)
- One categorical predictor variable
- Two levels in which different ps are exposed to the separate levels
- Parametric
When to use Mann Whitney Test?
- One continuous dependent variable (DV)
- One categorical predictor (IV)
- Two levels
- Independent group design
- Non parametric
When to use Wilcoxon’s Matched Pairs?
- One continuous outcome variable (DV)
- One categorical predictor variable (IV)
- Two levels in which the same ps are exposed to both levels
- Non parametric data
What are ANOVAs?
- Analysis of variance
- Used to analyse data that includes 3 or more conditions
Solutions to alpha inflation
- Lower alpha but lose statistical power
- Compute an ANOVA
Types of ANOVA:
- One way
- Repeated measures
- Factorial
- Mixed measures
When to use One-way independent ANOVA?
- One continuous outcome variable (DV)
- One categorical predictor (IV)
- Three or more levels in which different ps are exposed to the separate levels
- Parametric
When to use Repeated Measures ANOVA?
- One continuous outcome variable (DV)
- One categorical predictor variable (IV)
- 3 or more levels in which the same ps are exposed to all of the levels
- Parametric
What are Factorial Designs?
Allow researchers to examine the effects of two or more IVs and their interactive influence on one DV at one time
When are factorial designs used?
- Two or more IVs
- Ps are exposed to each IV
What are the 3 types of factorial?
- Independent factorial ANOVA
- Repeated measures factorial ANOVA
- Mixed model factorial ANOVA
What are simple effects?
- For all factorial ANOVAs, an additional test is computed if interactions are significant (simple effects)
- Simple effects: effect of one factor or IV on the DV at the level of another factor (IV)
When to use Repeated Measures factorial ANOVA?
- One continuous outcome variable (DV)
- Two or more categorical predictor (IV)
- 2 or more levels in which the same ps are exposed to the separate levels
- Parametric
When to use Mixed Model Factorial ANOVA in SPSS?
- One continuous outcome variable (DV)
- Two or more categorical predictor variables (IV)
- Two or more levels in which, for one IV, the same ps are exposed to the separate levels and for at least one IV, different ps are exposed to the separate levels
- Parametric
What are the non-parametric alternatives to ANOVA?
- Kruskal Wallis test (independent samples)
- Friedman’s ANOVA (related samples)
- Cochran’s Q (related samples)
What is the Kruskall-Wallis Test?
Compares between the medians of 3 or more samples to determine if the samples have come from different populations
When to use Kruskal-Wallis Test?
- One continuous outcome variable (DV)
- One categorical predictor variable (IV)
- 3 or more levels in which different ps are exposed to separate levels
- Non parametric
When to use Friedman’s ANOVA?
- One continuous outcome variable (DV)
- One categorical predictor variable (IV)
- 3 or more levels in which the same ps are exposed to all of the levels
- Non parametric data
What is a correlation?
Examining the relationship between variables
What is a scatterplot?
A figure in which individual data points are represented by a point in a two-dimensional space
What does a positive correlation mean?
As one variable increases, the other also increases
What does a negative correlation mean?
As one variable decreases, the other also decreases
What is a non-linear relationship?
- The line of best fit is curved
- The relationship b/t variables is underestimated
What are the bivariate correlation subtypes?
- Pearson’s (parametric data)
- Spearman’s Rho (non parametric data)
- Kendall’s Tau (non parametric data)
When to use Pearson’s?
- One continuous outcome variable (DV)
- One continuous predictor variable (IV)
- Parametric
When to use Spearman’s Rho test?
- Two continuous variables
- Non parametric data
When to use Kendall’s Tau test?
- Two continuous variables
- Data is small with a large number of tied ranks
- Non parametric
What is a partial correlation?
- Relationship b/t two variables controlling for the effect of one or more additional variables
- Finds unique variance between two variables
- Parametric
When to use partial correlation test?
- One continuous outcome variable (DV)
- Two continuous predictor variables (IV)
- The one of interest and a “third variable”
- Parametric
How do you control for Type I error?
- Use the bonferroni correction
- .05 divided by number of additional tests
How do you know when a test is significant?
If the p-value is less than or equal to the significance level, the decision is to reject the null hypothesis
Compute One-way ANOVA (hand computations)
1) State hypotheses
2) Compute SStotal (s2)
3) Compute SStreat (nj= no. of ps in group)
4) Compute SSerror
5) Compute dftotal
6) Compute dftreat (k= no. of groups)
7) Compute MStreat
8) Compute MSerror
9) Compute Fobt
10) Compute effect size
11) Compute post hoc analysis
12) Look up crit value (make statistical decision)
Which tests use independent design?
Independent t-test
Mann-whitney U
One-way ANOVA
Kruskall-Wallis
Which tests use a paired samples design?
Paired samples t-test Wilcoxon matched pairs Repeated measures ANOVA Friedmans ANOVA Repeated measures factorial ANOVA
Difference between correlation and regression
In regression, researchers are interested in predicting Y from X
What does the line of best fit for regression tell researchers?
Regression tells researchers about the form of the relationship b/t two variables
When to use simple regression test?
- One continuous outcome variable (DV)
- One continuous predictor variable (IV)
- Parametric
What is multiple regression?
Uses multiple predictor variables (X1, X2, X3…) to predict the criterion variable (Y)
When to use multiple regression?
- One continuous outcome variable (DV)
- Two or more continuous predictor variables (IV)
- Parametric
When do you reject the null?
Obtained value > critical value, then you reject the null
For mann whitney and wilcoxon, reject null when obtained value < critical value
How to select Ws Mann Whitney
When sample sizes are equal, smallest value is taken
When unequal, Ws is taken from the smallest sample, if it is the bigger value, use Ws formula
When is bonferroni test significant?
0.05/ (no of additional tests)
p value has to be smaller or equal to^^ to reject null