Help
Options
Didn't know it?
click below
click below
Knew it?
click below
click below
Don't Know
Remaining cards (0)
Know
retry
shuffle
restart
0:00
Embed Code - If you would like this activity on your web page, copy the script below and paste it into your web page.
Normal Size Small Size show me how
Normal Size Small Size show me how
PDAM #7
| Question | Answer |
|---|---|
| What are allowed variable types for a regression (2) | 1, outcome/ DV must be continous 2, predictors/IV can be either continous or categorical |
| What are 7 Regression assumptions | 1, non-zero variance 2,additivity and linearity 3, independence of observation 4,no milticollinearity 5,independence of errors 6,homoscedasticity 7,normality of errors |
| What is non-zero variance | the variable must not be constant (not only males in a sample when gender is a predictor) |
| What is multicollinearity | when there is a strong correlation between two or more predictors in a regression model (like male 0,1 and female 0,1) |
| What are 3 methods to detect multicollinearity and its "rules of thumb" | 1, correlation matrix inspection >> r>.7 indication and .8 problematic 2, variance inflation factor >> over 5 indication over 10 problematic 3, tolerance of VIF >>> over .2 indicator over .1 problematic |
| How does Variance Inflation Factor and Tolerance work | VIF = 1 / 1-R^2(of other factors) Tolerance = 1 / VIF |
| How to test autocorrelation (error correlation with each other) (2) | 1, with Durbin-Watson test 2, result of Durbin-Watson test should be between 1-3 |
| What is positive and negative autocorrelation | positive autocorrelation = one value high , next is also high or one is low and another is low negative autocorrelation = positive error then negative error and again repeativily |
| What does autocorrelation indicates | there is something in the model we are missing |
| What is homoscedasticity violation (2) | 1, heteroscedasticity > higher values have higher variation 2, there is no standart test on that only visualy in residuals we see a triangle like shape |
| What is homogeneity of variance | when we are comparing MULTIPLE SAMPLES their VARIANCES SHOULD BE SIMILAR |
| What is difference between Factorial ANOVA and One-Way ANOVA | Factorial ANOVA = multiple IV (diet + training > weight loss) One-Way ANOVA = one IV ( diet > weight loss) |
| What are interaction effects | change in effect (direction or strengt) of a predictor if another predictor changes |
| What is mean centering when observing interaction in regression (2) | 1, some variable "cannot" be interpreted when being =0 so we set its average value as default point (average=0) 2, +10 mean average +10 and -10 mean average -10 |
| How to check if adding interaction is meaninfull for the model (2) | 1, checking whenever R^2 increases 2, is the new interaction term statistically significant |
| What is 1, spotlight analysis 2,floodlight analysis | 1, spotlight analysis = splitting the graph into multiple lines of X effect acording to the different M values 2,floodlight analysis = showing the function as whole with different condition (like X*M when M <50 and X*M when M > 50 |
| What are 3 possible predictors in interaction analysis | 1, continous * continous 2, continous * dichotomous 3,dichotomous * dichotomous |
| What are covariates in the regression analysis (2) | 1, also called control variables 2, covariates remove noise from other variables |
| What is standartized beta | β(standartized) = β(unstandartized) * (SDx/SDy) where β is coeffitient of x |
Created by:
semester1.IBWL