In other words the relationship between variables Correlations In other words the relationship between variables
Correlational Studies Offer us a way of establishing whether there is a relationship between two variables enable us to assess the strength of that relationship by calculating a correlation coefficient (R) The statistical test is known as the Spearman’s Rho Unlike experiments, correlational studies do not tell us about causal (cause and effect) relationships
Correlations are all about relationships! The difference between an experiment and a correlation is An experiment looks at cause and effect a correlation looks at the relationship between 2 variables- it doesn’t look at which variable has caused the other variable to change.
We assume the IV has an effect on the DV Examples If the 2 variables were stress and hours of sleep: An experiment would have 2 conditions:- IV - 1 group might be given a non-stressful task before sleeping. The 2nd group might be given a stressful task before sleeping. DV - Amount of hours of sleep would be recorded in both conditions. We assume the IV has an effect on the DV
Therefore there is no IV or DV in a correlation. However A correlation might investigate self-reported levels of stress and the number of hours sleep a participant had. The difference is that in an experiment something is manipulated, where as correlations are merely looking at 2 variables occurring at the same time. Therefore there is no IV or DV in a correlation.
Rules to correlations You have to look at the two variables to see if they are occurring at the same time. The 2 variables have to be measurable Question: On your white boards write down the term for making something measurable? answer operationalisation.
Examples of Correlation studies Number of subjects taken at 6th form and hours of homework Self-reported stress levels and personality type Out of school activities and grades Hours of TV watch and level of aggression
Can you come up with 3 ideas for a correlation? Activity Can you come up with 3 ideas for a correlation? Write them on your white board now
How do you know if the variables are correlated? There are 3 types of correlations positive and negative and zero 0 Positive-as one variable increases the other variable increases. Negative- as one variable increases the other variable decreases Zero – no relationship
Scattergrams The results from a correlation can be plotted on a scattergram. The scattergram will show whether there is a positive correlation, a negative correlation, or no correlation at all You can analyse the relationship by 1. Drawing a scattergram 2. Calculating a correlation coefficient (R)
Positive Correlation This is a positive correlation because as the number of hours watching TV increases so does the level of aggression. Line of best fit
Negative Correlation This is a negative correlation because as the temperature increases, the number of clothes worn decreases.
Uncorrelated relationships This scattergram shows there is no correlation between head circumference & IQ. This is also known as a zero correlation
The Correlation Coefficient Gives us a statistical method for assessing the strength of a correlation The sign (+ or -) tells you the direction of the correlation The number (between 0 and 1) tells you the strength of the correlation A perfect negative correlation would be -1
As one variable rises, so does the other What kind of correlation is this – write it on your whiteboards
The Type of Correlation is Positive Well done if you got this right What number would represent this perfect positive correlation? Answer ~ +1
What goes in the gap? Right it on your white board now As one variable rises, the other falls
The Type of Correlation is Negative Well done if you got this right
Cause and Effect? What can we conclude from this scattergraph? White board time again
What number would represent no correlation at all? A correlation between two variables does not mean that changes in one variable is causing changes in the other It simply means there is a relationship between the two variables that might be worth investigating further What number would represent no correlation at all?
Advantages and disadvantages Data may have already been collected Other factors which are ignored- not controlled for Ideal for showing a link between 2 measures No cause and effect Provides quantitative data Some problems of using a scale Can use observational data and self-report measures= high ecological validity. As some results are taken from a laboratory environment the results may have low ecological validity.
Examples Correlational Studies Research question: is there a relationship between stress and illness? Variable 1: scores on a stress questionnaire Variable 2: number of days absent from college since September Hypothesis: There will be a positive correlation between stress scores and number of absences Write the null hypothesis on your whiteboard now
Correlational Studies Research question: is there a relationship between absence from college and final AS marks? Variable 1: absences Variable 2: final AS mark (Max 300) Null Hypothesis: There will be no correlation between absences and AS marks Write the alternate hypothesis now
Why alternate and not experimental hypothesis?
Main task To write a testable hypothesis and a null hypothesis To design your own correlational study you will need To write a testable hypothesis and a null hypothesis To decide how you will measure your 2 variables. To know what your sample will be To know how you will carry out your research STOP
http://www.youtube.com/watch?v=A3b9gOt Qoq4