# Thread: Comparing 2 percentages - how?

1. ## Comparing 2 percentages - how?

Hi again guys,
I need to compare two percentages.

I have the percentage 12% and the percentage 38.5%

I need to state if they are significantly different.

I've been told I need to do a chi-square test but I'm not sure how exactly I'm meant to do that.

Laura

2. Originally Posted by laura marie
Hi again guys,
I need to compare two percentages.

I have the percentage 12% and the percentage 38.5%

I need to state if they are significantly different.

I've been told I need to do a chi-square test but I'm not sure how exactly I'm meant to do that.

Laura
You can use an exact Binomial test;

Online reference:

http://en.wikipedia.org/wiki/Binomial_test.

SPSS;

Other references:

Clopper, C. J. & Pearson, E. S. (1934). The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika, 26, 404–413.

William J. Conover (1971), Practical nonparametric statistics. New York: John Wiley & Sons. Pages 97–104.

Myles Hollander & Douglas A. Wolfe (1973), Nonparametric statistical inference. New York: John Wiley & Sons. Pages 15–22.

good luck,

I only have these four numbers.

The 38.5% is out of 52 participants.
The 12% is out of 114 participants.

I am using SPSS but I don't know how to arrange the numbers to run the test. It is very very frustrating!

4. Originally Posted by laura marie

I only have these four numbers.

The 38.5&#37; is out of 52 participants.
The 12% is out of 114 participants.

I am using SPSS but I don't know how to arrange the numbers to run the test. It is very very frustrating!
Lets see if this helps:

You're sample size is probably sufficient (this is subjective) to allow the use of a Chi-squared and/or G-test.

These test might be less difficult to compute and are explained in great detail on the web.

Heres a resource:
http://udel.edu/~mcdonald/statgtestind.html

5. Originally Posted by laura marie

I only have these four numbers.

The 38.5% is out of 52 participants.
The 12% is out of 114 participants.

I am using SPSS but I don't know how to arrange the numbers to run the test. It is very very frustrating!
You can do this in Excel (sorry I don't know SPSS). A very simple-minded Excel worksheet is attached, pardon its look.

You know that 38.5% of 52 participants = 20 are "successes" in row 1
You know that 12% of 114 participants = 14 are "successes" in row 2

(See the ACTUAL table, rows 1-3)

Thus you know that 34 of 166 participants (20.5%) are successes overall. So using this overall % of successes, you can determine the EXPECTED table (see rows 5-7)
eg 20.5% of 52 participants = expected 10.6506 "successes"
and 20.5% of 114 participants = expected 23.3494 "successes"

Chi-square test compares the actual successes vs expected successes in row 1, and in row 2 (alternatively, you can say it compares the actual non-successes vs expected non-successes, since successes + non-successes = participants).

Excel has a CHITEST function (see cell A12).

 Tweet

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts