1
$\begingroup$

I have a sample of 65 participants. 20 chose A. 18 chose B. 14 chose C and 13 chose D. What is the suitable test to determine if there is significant difference in the proportions?

$\endgroup$

1 Answer 1

1
$\begingroup$

It's not completely clear that I have understood correctly, but I think you're asking whether the proportions choosing A, B, C and D are consistent with the population proportions being $(\frac{1}{4},\frac{1}{4},\frac{1}{4},\frac{1}{4})$.

If that's the case, this is just a chi-squared goodness of fit test.

(Searching on chi-squared goodness of fit test or chi-square goodness of fit test will turn up many hits here, a few of which will relate to your kind of problem.)

Since this looks like the particular numbers may be from a self-study question, I won't do the problem itself, but I will explain the calculation of the test statistic.

Given the 65 participants you compute the expected count in each cell, $E_i = 65/4,\, i=1,2,3,4$. The observed counts, $O_i$ are given in your question.

You then compute the chi-square statistic, $X^2 = \sum_{i=1}^4 \frac{(O_i-E_i)^2}{E_i}$, which under the null hypothesis has a $\chi^2$ distribution with three degrees of freedom.

Most stats packages can do this work for you, but it's a few moments to do by hand if you have chi-square tables.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.