Question 749:

Asked on June 14, 2009

Tags: Chi-Square Goodnes of Fit Test

1

Answer:

No answer provided yet.

Files Available For Download

1. Q749_1_q749-chi-square.xls

For this question you need to perform a Chi-Square Goodness of Fit test. What you want to see is how well your sample data "fits" the population proportions. We expect some fluctuations from samples, but is it greater than chance alone to suggest our sample is not representative?

The test proportions are .70, .20 and .10 for the employee types and we're testing that against the observed proportions of .80, .12 and .08. We get a Chi-Square test statistic 2.514 and evaluate it against 2 degrees of freedom. This gets us a p-value of .2844 and is not significant (p > .05). 

The null hypothesis is that there is no difference in the actual and observed distributions. Since our p-value was above .05 we cannot reject the Null. We therefore conclude that our sample is representative. 

See the attached excel file for the calculations.