Lecture 11 - Issues in design: Relationships

Question 1

What must be true for a correlation?

Answer 1

• Both variables must be continuous
• Nothing is manipulated
  -no IV or DV, just “variables”

Question 2

What hypotheses are there in correlational designs

Answer 2

Null

Two-tailed

One-tailed: Positive relationship

One-tailed: Negative relationship

Question 3

How to analyse correlations?

Answer 3

If parametric assumptions have been MET then Pearson’s correlation

If parametric assumptions have been VIOLATED then Spearman’s correlation

Question 4

When do we building predictive models

Answer 4

When analysing more than two continuous variables

Question 5

What is the terminology used for the variables for analysis

Answer 5

Outcome variable: variable BEING predicted

Predictor variable: variable USED to predict outcome

Question 6

What is the purpose of building a predictive model

Answer 6

Gives more in depth & detailed/accurate prediction of what the relationship between variables is

Doesn’t only say “there is a significant relationship” but provides more meaningful and tangible information

Question 7

What is the process of making a regression model

Answer 7

Starts off with a correlation
-eg “yes there is a significant positive correlation”

Uses one variable to predict the other

Line of best fit of the correlation forms the predictive model

Question 8

what is model variance & what is random variance

Answer 8

Model/explained variance: Variance explained by the line of best fit

Random/unexplained variance: Variance between line of best fit and raw data

if gaps between the lines and the dots are far bigger = more random variance

Question 9

What does model variance vs random variance mean

Answer 9

Model = More likely to be significant
vs
Random = Less likely to be significant

Question 10

What are confounds in relationships

Answer 10

Variable that can explain some variance
-can be model and/or random

Question 11

What variables can be confounds

Answer 11

Continuous

Binary categorical variables

Question 12

What are the four steps to writing critically about research

Answer 12

Clearly identify and explain issue
-include published research to support ideas

Explain how this may have influenced findings
-Be specific (over or underestimate?)

Consider alternative designs or improvements

Predict how this change may impact findings
-if new version of study was done

Question 13

What are the 3 steps when designing a correlational study

Answer 13

Hypothesis

Two variables

Think critically