Correlations

Correlations

What is a correlation?

• a statistical technique which allows researchers to compare two sets of scores to see if there is a relationship b/w two co-variables

• there is no IV or DV

Correlations

What’s the difference b/w a correlation and an experiment?

• Experiments require manipulation of an IV and measuring the resulting change in the DV

  WHEREAS

• No variables are manipulated in correlations, the co-variables are measured + compared to look for a relationship

Correlations

What are co-variables?

The two factors/variables that are measured/collected by the researcher, then compared to each other

Correlations

What are the two ways of displaying correlations?

• Scattergrams

• Correlation co-efficients

Correlations

How are correlations shown on graphs?

• Scattergrams

• For each individual you obtain two scores which you use to plot one point

• The scatter indicates the degree of relationship

Correlations

What do you need when drawing a scattergram?

• A title

• One variable on each axis (either way around)

• Labelled axis’s

• Suitable scale, starting from 0

• A ruler

• A line of best fit

Correlations

What can you find out from a correlation?

Can only predict that there is or isn't a relationship (as another extraneous variable may be involved that is not known about) + how strong/weak it is

Correlations

What is positive correlation?

• As one variable increases (or decreases) so does the other

• They both do the same thing

Correlations

What is negative correlation?

• As one variable increases, the other decreases

• Both do different things

Correlations

What is no correlation?

Two variables have no relationship

Correlations

What type of correlation does this show?

Strong, positive

Correlations

What type of correlation does this show?

Strong, negative

Correlations

What type of correlation does this show?

no correlation

Correlations

what is the correlation coefficient?

• A numerical way to describe/display a the strength + direction of correlation/relationship b/w the co-variables

• Between -1 and +1

• Found by using statistical tests (eg: Spearman’s Rho)

Correlations

What does +1 mean for a correlation coefficient?

perfect positive correlation

Correlations
What does -1 mean for a correlation coefficient?

perfect negative correlation

Correlations

What does 0 mean for a correlation coefficient?

no correlation

Correlations

What is a strong/moderate/weak correlation coefficient for both positive and negative?

• Strong: ± 0.8

• Moderate: ± 0.5

• Weak: ± 0.2

Correlations

Evaluation: causation

Correlation doesn’t show causation → a strong correlation may suggests a relationship exists BUT doesn’t show which variable is CAUSING it, OR if another, unknown variable is involved

Correlations

Evaluation: highlighting potential causation

Cab highlight potential causation which researchers can use to make predictions + test experimentally to establish cause + effect

Correlations

Evaluation: ethics

Often limited ethical issues bcs the covariables often already exist + so the data is more easily accessible VS manipulating in experiments may lead to ethical issues SO can be an ethical alternative

Correlations

Evaluation: correlation coefficient = useful

useful for describing the direction/strength of relationships + can make it more easy to compare correlations precisely vs using just scattergrams