Summary and study notes

Welke onderwerpen worden behandeld in het hoorcollege?

In dit hoorcollege worden de verschillende analyses in SPSS uitgelegd en voor gedaan die je moet kunnen voor het SPSS tentamen. 

Welke onderwerpen worden besproken die niet worden behandeld in de literatuur?

In dit college worden geen andere onderwerpen besproken dit niet worden behandeld in de literatuur.  

Welke recente ontwikkelingen in het vakgebied worden besproken? 

Dit is niet van toepassing.

Welke opmerkingen worden er tijdens het college gedaan door de docent met betrekking tot het tentamen?

Je hoeft geen simple effects analyse te kunnen maken, maar je moet wel de output kunnen lezen. 

Welke vragen worden behandeld die gesteld kunnen worden op het tentamen? 

De voorbeeldvragen in dit college zijn een goede representatie van de vragen die in het tentamen gesteld kunnen worden. 

Hoorcollege aantekeningen

Descriptive statistics 

What is the mean age of the participants? Analyze --> descriptive statistics --> descriptives --> mean. 

What is the maximum observed age? Analyze --> descriptive statistics --> descriptives --> maximum.

What is the number of participants with an age of 20? Analyze --> descriptive statistics --> frequencies. 

How many women have a low income? Analyze --> descriptive statistics --> crosstabs.

What is the Pearson correlation coefficient between age and depression? Analyze --> correlation --> bivariate.

Multiple regression

• Predicting the depressive symptoms of students, based on gender and age. Multiple regression. 
• Does income (use dummies, low is the reference category) add anything to the prediction of depression? You don’t need a dummy for the reference group. This is a hierarchical regression. 

Regression --> linear --> dependent and independent variables --> next --> add new independent variables --> statistics: R² change --> statistics: check for multicollinearity (collinearity diagnostics) --> save: standardized (2x), Mahalanobis and Cook’s. With mahalanobis and cook’s you get new variables to check the assumption with histograms. Cook distance below 1 is good. For homoscedasticity you have to check the residuals of the standardized values with a scatter plot. Homoscedasticity = variation must be the same around the zero-line for every score. 

Moderation

• Is IQ a moderator for the effect of age on reading ability? X = age, Y = reading abiity, M = IQ. 

Regression --> PROCESS --> Y, X and Moderation --> Moderation = model 1 --> Options: mean center, -1SD, Mean, + 1SD and p < .05.

When interaction is not significant, there is no moderation.

Mediation

• Is the effect of age on reading ability mediated by short term memory span? 

Regression --> PROCESS --> Y, X and Mediator --> Mediation = model 4 --> Options: total effect model. 

Look at the total, direct and indirect effect and if they are significant (for indirect effect look at CI and if the 0 is included). If there is a significant direct and indirect effect, there is partial mediation. If there is only indirect effect, there is complete mediation.  

ANOVA

• Is there a difference in insomnia when we look at the factors age (high, medium, low) and education (high, medium/low)? 
  • H0: no main effect of age on insomnia
  • H0: no main effect of education on insomnia
  • H0: no interaction effect of age and edcation on insomnia 

Y = insomnia, X = age and education. This is a factorial ANOVA. 

Analyze --> general linear model --> univariate (because 1 DV) --> plot: interaction of IV --> options: homogeneity tests, descriptive, effect sizes (and parameter). Levene’s test: look at the first line (mean) and if it’s not significant, then the assumption of homogeneity is met. Between-subjects effect: look at the main effects and interaction effect (significance). Plot: look at the visualizing of main and interaction effects. 

• If interaction is significant: run simple effects analyses. Only have to interpret the output. 

Simple effects analysis looks at the main effects. You have to type syntax for that. It tells you is the groups differ from each other. 

ANCOVA

• Is there a difference in aggression between patients receiving psychotherapy and CBT when we control for the aggression prior to treatment? First, test the assumption of homogeneity of regression coefficients! 

Analyze --> General linear models -->univariate --> X = treatment, Y = aggression after and covariate = aggression before --> Model: build factor (each factor you get main and interaction effects, but with covariate only main effects, therefore we need to build are own model) with two factors and the interaction. This is only to check if the interaction is not significant, so the assumption is not violated and we can run the ANCOVA. 

To run the ANCOVA --> EM Means: IV. If the covariate is significant, this is a influential factor on de DV. After controlling for the covariate, there is still significance, so there is still a difference. 

MANOVA

• Is there a difference between first, second and third year students in their knowledge of statistics, developmental psychology and social psychology? 

Analyze --> general linear model --> multivariate --> add DV’s and IV --> options: homogeneity, descriptive --> post hoc. Box M test: assumption of homogeneity, if it is not significant, the homogeneity is assumed. Multivariate test: only look at Lambda if the MANOVA is significant. Between factor: look which factor is significant.

• In addition: for the one significant DV (statistics) we will demonstrate a contrast test for comparing the mean in each year with the mean in the third year (i.e., a simple contrast). 

You want to compare each year with the third year = a simple contrast. Analyze --> general linear model --> univariate --> DV (the significant factor) and IV --> Contrasts: simple. 

Repeated measures

• Is there a difference in grades given by four different teachers to all pupils? 

Consists of only within-factors. Analyze --> general linear model --> repeated measures --> factor name and number of levels --> link the levels --> options: descriptives and effect size --> EM Means: factor, compare main effects and Bonferonni. Sphericity = the null hypothesis says there is sphericity, when its significant there is a violation of sphericity. Than you look at Epsilon how severe the violation is (1 is not severe). If its really low, look at MANOVA table. If its significant, there is a effect of teacher. Than you look in the pairwise comparisons which groups differ. 

Mixed design

• Is there a difference in the grades given by the four teachers to male and female students? 
  • H0: no main effect of teacher on grades (within subjects effect) 
  • H0: no main effect of gender on grades (between subjects effect) 
  • H0: no interaction of teacher and gender on grades 

Consists of within and between factors. Analyze --> general linear model --> repeated measures --> factor name and number of levels (number of conditions) --> link to the factors --> plots: graphs with within-factor on X-axis and separate lines for the between factor --> options: descriptive, effect size and homogeneity --> post hoc for between factor and EM means for within factor.