W4

Question 1

how to analyse predictors of the degree of treatment success

Answer 1

by linear regression analyisis

Question 2

how to analyse predictors of treatmnet success (yes no)

Answer 2

logistic regression

Question 3

what is the main difference between the linear regression and logistic regrssion

Answer 3

wether the DV is nominal or continous

Question 4

General interpretation b’s?(0,1,2)

Answer 4

• b0 = expected score on dep. var (y) if both indep. var x1 and x2 equal 0
• b1 = difference in expected score on dep. var (y) if indep. var x1 increases by 1 while indep. var x2 remains constant
• b2 = difference in expected score on dep. var (y) if indep. var x2 increases by 1 while indep. var x1 remains constant

Question 5

2 steps of a simple linear regression

Answer 5

Step 1: Whole model better than mean model? (ANOVA)
* Compares the whole model (b0, b1) with the mean model (b0)
Step 3: How to interpret the individual predictors? (Coefficients)
* Compares b1 (and b0) with ‘0’ (is b>0 or <0?).

Question 6

what are the 3 steps of MULTIPLE LINEAR REGRESSION IN SPSS (ENTER)

Answer 6

Step 1: Whole model better than mean model? (ANOVA)
* Compares model 1 (b0, b1) with mean model (b0)
* Compares model 2 (b0, b1, b2) with mean model (b0)
Step 2: Extended model better than previous model? (Model
Summary)
* Compares model 1 (b0, b1) with mean model (b0)
* Compares model 2 (b0, b1, b2) with model 1 (b0, b1)
Step 3: How to interpret the individual predictors? (Coefficients)
* For model 1, compares b1 (and b0) each with ‘0’ (is b>0 or <0?)
* For model 2, compares b1, b2 (and b0) each with ‘0’ (is b>0 or <0?)

Question 7

how can we tell how good is a model

Answer 7

with R square, it shows the amount of explained variemce in dep variable by predictor/s
Rsqr in simple regression
- 0.01 - small
- 0.09 - mid
- 0.25- large

Rsqr in multiple regression
- 0.02 - small
- 0,13 - mid
- 0.26 - large

Question 8

how to see how good is a predictor

Answer 8

standardized beta - effect of predictor on dep variable when both predictor and dep are standardized
- u can compare within a model and between diff models

Question 9

Logistic regression:

Answer 9

P(yi) = f{b0 + b1*x(i)} + error(i)
probability (P) that the outcome (y=1) occurs at a certain value of the predictor (x) for subject i
* P can vary between 0 (= very unlikely the outcome (y=1) occurs) and 1 (= very likely the outcome (y=1) occurs)

Question 10

T or F

SPSS in the log regression calculate how good the model fits

Answer 10

false,
* SPSS calculates the inaccuracy of the model fit
* Comparison model fit with predictor with baseline model fit
* If model with predictor is less inaccurate (= better fit) than baseline model, then include predictor in the model!

Question 11

what are the 2 steps of a aimple logistic regression in SPSS

Answer 11

Step 1: Fit extended model less inaccurate (= better) than previous model?
* Calculates inaccuracy of fit baseline model
* Calculates inaccuracy of fit model (with 1 predictor)
* Compares fit model with fit previous model (= Step)
* Compares fit model with fit baseline model (= Model)
Step 2: How to interpret the individual predictor(s)?
* Provide direction of effect of individual predictor(s)

Question 12

MULTIPLE LOGISTIC REGRESSION IN SPSS
(ENTER) 2 steps

Answer 12

P(y) = f{b0 + b1x1(i) + b2x2} + error(i)
Step 1: Fit extended model less inaccurate = better) than
previous model?
* Calculates inaccuracy of fit baseline model
* Calculates inaccuracy of fit model 1 (with 1 predictor)
* Calculates inaccuracy of fit model 2 (with 2 predictors)
* Compares fit model 2 with fit previous model (= Step)
* Compares fit model 2 with fit baseline model (= Model)
Step 2: How to interpret individual predictor(s)?
* Provide direction of effect of individual predictors
5

Question 13

what does it mean that a variable is centered

Answer 13

the average is substracted from each score in the variable