Are there any resources specificly for WJEC PY3 Statistical Tests? I struggle so much with them. Help me break it down and revise better. (I've added detail below)

  • 0 votes

I'm studying Psychology at A2 level, WJEC board and really struggle with Statistical Tests for the PY3 exam: Wilcoxon, Man Whitney, Sign Test, Chi Squared and Spearman's rank. Are there any resources that specifically explain these, breaking them down preferabley? Please and Thank You :)

Posted Sat 14th April, 2012 @ 17:45 by Nutmeg

1 Answer

  • 1 vote

To answer the questions about the statistical tests you have to ask yourself a set of questions;

1. Are you testing a difference between two conditions (i.e. pre-test/post test) or a relationship between two variables?

  • If the answer is difference then ask yourself; 2. is the design independent groups (including matched pairs) or repeated measures?

- If the answer to that is repeated then 3. what type of data does it use?

- Nominal = Sign
- Ordinal/interval = Wilcoxon

- If the answer in independant, again 3. what data does it use?

- Nominal = Chi-Squared

- Ordinal/interval = Man Whitney

  • If the answer is relationship than 2. what type of data does it use?

  - Nominal = Chi-Squared

- Ordinal/Interval = Spearman's rank

You don't need to know how to do them or any more than be able to describe why you would use them. Don't forget that you just state the design of the study, what you are testing (difference or relationship) and then mention the type of data used and you'll get all the marks.

Remember the rules about p</=0.05. They like to put them in the paper because they are hard questions. Just remember;

  • If P</=0.05 then the difference between the two gaps is significant so the hypothesis can be accepted.
  • If P>/=0.05 then the difference between the two groups is not significant so the null hypothesis is correct and any difference that has occurred is due to chance. 

Another thing to remember is the R rule;

  • If there is an R in the name of the test, the observed value should be greater than the critical value for the results to be significant (Spearman's Chi-Squared)
  • If there is no R in the name of the test, the observed value should be smaller than the critical value (Wilcoxon, Mann Whitney, Sign)

Hope this is of some help :) 

Flo x

Answered Sat 14th April, 2012 @ 19:10 by Flo