Investigating relationships

Question 1

correlation coefficient, r

Answer 1

measures strength of a relationship between 2 continuous variables

Question 2

r=0.9

Answer 2

strong positive linear relationship

Question 3

r=0.01

Answer 3

no linear relationship

Question 4

r=-0.9

Answer 4

strong negative linear relationship

Question 5

regression

Answer 5

association between 2 variables: estimating the line of best fit through the data minimising the sum of squared residuals

Question 6

when is regression useful

Answer 6

1. look for significant relationships btw 2 variables

2. predict a value of one variable for a given value of the other

Question 7

residuals

Answer 7

differences btw observed and predicted

Question 8

how to check the relationship btw independent & dependent variable is linear?

Answer 8

scatter plot

Question 9

how to check the variance of the residuals around the predicted responses are the same?

Answer 9

scatter plot of standardised predicted values & residuals

Question 10

how to check the residuals are independently normally distributed?

Answer 10

plot residuals in histogram

Question 11

what does funnelling shape show

Answer 11

problems when checking residual values against predicted

Question 12

independent variable

Answer 12

explanatory/predictor variable

Question 13

dependent variable

Answer 13

outcome variable

Question 14

parametric tests

Answer 14

assume data follows normal distribution

Question 15

non-parametric tests

Answer 15

based on ranks

Question 16

when are non-parametric tests used

Answer 16

data is ordinal 
data doesn't follow shape/distribution, e.g. normal
skewed plot of data 
influential outliers
small sample size

Question 17

what to do when data is not normally distributed?

Answer 17

1. use non-parametric tests

2. transform dependent variable

Question 18

how to deal w positively skewed data

Answer 18

take log of dependent variable to produce normally distributed values