Chi-Square Test

Chi-Square Test

1. Analysis of nominal (category) data in freq.

2. Difference:- disparity (lack of similarity) between freq,. and expected freq.

   • bigger difference--> larger value Chi-square--> significant difference more likely
     • Chi-square value (observed - expected)squared / expected

3. Association:- looking at the relationship between the two variables. A test of the difference in proportions of frequencies-> disparity between the actual freq and expected freq

4. Calculate the expected freq. using our observed freq.

   • row total x column total / grand total (N)

5. There was [no] significant association between predictor variable and criterion variable, xsquared (df=__, N=__)=__,p=__ . Describe results.

   • Casually dressed witnesses were more likely to be judged as not credible and less likely to be viewed as credible, whilst the reverse was the case with the smartly dressed eyewitness.

6. Fisher Exact Probability Test

   • There was a significant association between predictor variable and crieterion variable, Fisher, p=__. Describe results
     • Children with ID were more likely to make false identifications and less likely to make correct rejections were presented with a target-absent line-up. The reverse was the case with the age-matched controls
   • Small sample test
   • 2x2 contingency table
   • Expected cell freq falls below 5 {note b}
   • Calculates an exact probability

7. McNemars Test

   • There was a significant increase [or decrease] in the predictor variable to criterion variable x[squared]=__, N=__, p=__.
   • 'Related' samples of freq- comparisons and not independent
   • Non-changing data is ignored -> each participants scores on 2 occassions
   • Within-subjects- participants have contributed obvserations to more than one cell in the contingency table

8. Contingency table

Images and files in the map