8 Testing For Associations And Regressions

Question 1

Strength of association

Answer 1

R=0 no correlation
R=0.1-0.4 weak correlation
R=0.41-0.6 moderate
R=0.61-1 strong

Question 2

Type of an association

Answer 2

Linear- as one variable increase/decrease the other decrease/increase

Curvilinear- u curve, increases to a certain point (staff cheerfulness- good to a certain point)

Spurious relationship- ice cream and rate of drowning, there is an association but buying ice cream doesn’t actually influence drowning

Question 3

Scatter gram

Answer 3

Interval and ratio

Question 4

Pearson correlation

Answer 4

Interval/ratio, p value, correlation coefficient

Question 5

Spearman correlation

Answer 5

Ordinal
Correlation coefficient
P Val

Question 6

Correlation

Answer 6

Describes the strength of a relationship between variables. No distinction between independent and dependent variables.

Question 7

Regression

Answer 7

Describes the strength of a relationship between variables and predicts the relationship. Introduces concept of dependent and independent variables.

Linear regression can be conceptualised as finding a line of best fit between observed scores

Categorical IV need to be converted into dummy variables (0,1)

Question 8

Bivariate regression

Answer 8

Only one variable as a predictor (IV)

Question 9

Adjusted r squared (coefficient of determination)

Answer 9

Estimate the percentage of explained variance in a dependent variable

High r2 = smaller error

R2=0.147 = model explains 15% of variance

If VIF scores are above 5 there is multicolinearity (not good)

Question 10

Multiple regression types

Answer 10

Enter- all IV estimated at once

Stepwise - IVs are removed by software to derive to the model that explains most variance in DV

hierarchical- IVs are entered in pre determined blocks