topic Re: P-Value calculation in SAS Programming

P-Value calculation

Dhana18 — Thu, 17 Oct 2019 17:07:58 GMT

HI I have the data like this and i would like to calculate p-value to to see if the populations differ statistically in terms of these variable (event).

	FEMALE
	Alert		Non-Alert
	#	%		#	%
TOTAL	11			202
GENDER
Male	0	0%		0	0%
Female	11	100%		202	100%
Male to Female	0	0%		0	0%
Unknown	0	0%		0	0%
RACE
African American	11	100%		176	87%
White	0	0%		8	4%
Other	0	0%		18	9%
ETHNICITY
Hispanic	1	9%		6	3%
Non-Hispanic	10	91%		191	95%
Unknown	0	0%		5	2%
GENDER SP
Male	9	82%		164	81%
Female	1	9%		26	13%
Both	1	9%		5	2%
Unknown	0	0%		7	3%
SYMPTOMS
Dysuria	0	0%		5	2%
Discharge	4	36%		49	24%
GC History
Yes	0	0%		21	10%
No	11	100%		179	89%
Unknown	0	0%		2	1%

Re: P-Value calculation

PaigeMiller — Thu, 17 Oct 2019 18:51:39 GMT

@Dhana18 wrote:

HI I have the data like this and i would like to calculate p-value to to see if the populations differ statistically in terms of these variable (event).

I don't see a variable named EVENT.

Please refer to the exact position(s) in the table that you want to compare via a statistical test.

I also don't understand the percents shown, for example, next to unknown, there are 2 "Non-Alert" and 1%? How do you get 1% out of 2 observations?

Re: P-Value calculation

ballardw — Thu, 17 Oct 2019 19:38:40 GMT

P-value implies a statistical test. Which specific test do you want to conduct? If you can provide exactly want kind of question you

I would guess that the easiest test might be a chi-square but you have so many empty cells or counts < 5 that the test may not be reliable that is one of the least sensitive to small sample size tests.

Re: P-Value calculation

Reeza — Thu, 17 Oct 2019 20:37:05 GMT

Same answer as last time, chi square tests will test if the counts are different between the groups. If you want to identify beyond that you can run a logistic regression on each one and see the outcomes. You may also be able to use the BINOMIAL option on PROC FREQ to get the comparisons. I think there's an example in the past month on here on how to compare the multiple levels via PROC FREQ.

Re: P-Value calculation

Dhana18 — Fri, 18 Oct 2019 12:56:11 GMT

I want to do a Fisher's exact test, Two -tailed. I want to see if the difference in these Alert and non-alert group is significant or not. Below is the example of my data. Please help.

Alert case

Non-Alert case

P value

ETHNICITY

Hispanic

1

3

Non-Hispanic

10

111

Unknown

0

2

GENDER SP

Male

9

93

Female

1

20

Both

1

2

Unknown

0

1

SYMPTOMS

Dysuria

0

3

Discharge

4

31

Re: P-Value calculation

PaigeMiller — Fri, 18 Oct 2019 13:57:17 GMT

This p-value can be computed from PROC FREQ, but I think we would need to see a portion of your actual data before we can provide code that you can use. Please follow these instructions to provide a portion of your data for us to see: https://communities.sas.com/t5/SAS-Communities-Library/How-to-create-a-data-step-version-of-your-data-AKA-generate/ta-p/258712