topic Comparing Percentages between Treatment Groups in Statistical Procedures

Comparing Percentages between Treatment Groups

Sharam — Thu, 17 Dec 2020 04:05:53 GMT

Hi there.

I had an experiment with 6 treatment groups (T1-T6). There were 96 animals in each treatment group (N=96). I want to compare the Mortality% between the 6 treatments statistically. The data are shown below. Could you please advise? Thank you!

Treatment Mortality%

T1 2.1%

T2 4.2%

T3 0%

T4 2.1%

T5 7.3%

T6 4.2%

Re: Comparing Percentages between Treatment Groups

PaigeMiller — Thu, 17 Dec 2020 12:40:07 GMT

You could do a Chi-Squared test, and/or a Fisher's EXACT test, as shown. I am lazy and only typed the first three treatment groups.

data have;
    input treatment $ died $ n;
    cards;
T1 Y 2
T1 N 94
T2 Y 4
T2 N 92
T3 N 96
;

proc freq data=have;
    tables treatment*died/chisq exact;
    weight n;
run;

The Chi-squared test Prob is greater than 0.05, and in the Exact test the Pr<=P is also greater than 0.05, indicating no difference. If you run this on all 6 groups, you may get a different answer.

Re: Comparing Percentages between Treatment Groups

Ksharp — Thu, 17 Dec 2020 13:09:52 GMT

ANOVA or Kruskal-Wallis test ?

Re: Comparing Percentages between Treatment Groups

Ksharp — Thu, 17 Dec 2020 13:15:50 GMT

http://support.sas.com/kb/24/188.html

Re: Comparing Percentages between Treatment Groups

Sharam — Thu, 17 Dec 2020 22:35:58 GMT

Thanks!