Programming the statistical procedures from SAS

Testing the significance for Email Marketing Control groups: Maybe with Proc Power?

Reply
Occasional Contributor
Posts: 7

Testing the significance for Email Marketing Control groups: Maybe with Proc Power?

[ Edited ]

I'm trying to figure out a way to test to see if my control groups for
email marketing campaigns will produce statistically significant
results. The group counts have already been set by marketing. The
control groups are refreshed monthly and not mailed email marketing
communication for the month and but the mailed groups are.

So, I'm comparing proportions from the two independent groups. I need
to make assumptions about the assumed proportion for each group.  I
know the sample sizes for each group, so I think I can use you can use
groupns= to specify the values, and use power=. to request the
computation for the power. Also, the assumed proportion for each group
is something  I'm having trouble deciding upon.

 

1. 0-12M Active customers (split by Email Only and DM/EM)

2. 13-24M Inactive customers

3. 25+M Lapsed customers

 

Controls

Segment  Description Segment Count
1 12M DM & EM Cohort5,000
2 13M -24mEM Cohort5,000
3 25M Cohort3,000

 

Mailed

Segment

  Description

 Segment
Count

1

12M DM & EM Cohort Balance

24,476

2

13-24M Cohort Balance

25,339

3

 25M Cohort Balance

26,190

 

 

I have some experience using Proc Power but not much. Below is an
example of some code which I have a feeling I should use but I'm not
if this is appropriate and where to plug in the values.

proc power;
twosamplefreq test=pchi
sides=
alpha=
groupns=
ntotal =
power = .;
run;

I've been looking at examples here:
http://support.sas.com/documentation/cdl/en/statug/68162/HTML/default/viewer.htm#statug_power_syntax...

Ask a Question
Discussion stats
  • 0 replies
  • 242 views
  • 0 likes
  • 1 in conversation