What Test to Run – Help!
Six Sigma – iSixSigma › Forums › General Forums › Tools & Templates › What Test to Run – Help!
 This topic has 7 replies, 2 voices, and was last updated 9 years ago by MBBinWI.

AuthorPosts

June 4, 2012 at 3:23 pm #54088
Hi there,
I have a set of data with about 300 line items. With a Yes or No column and then a $$amount
The 2 variables are the following:
1. If the medical instrument has the updated accessories or not. So Yes or No
2. Average revenue over 3 yearsI am trying to see if there is a correlation between accessories to more revenue. Or if there is a statisical difference that exists?
I am not sure what test to run?? Is what I am looking for correct?
0June 4, 2012 at 7:53 pm #193517
MBBinWIParticipant@MBBinWI Include @MBBinWI in your post and this person will
be notified via email.@safarbags = you have a yes or no variable and then one which identifies a continuous variable on how much. Which test accomodates such a scenario?
0June 5, 2012 at 9:23 am #193521One of the most difficult things to do in any analysis if figure out what is my X and what is my Y. In your case, “Yes/No” is the X and you are trying to use this X to tell you something about the Average Revenue (the Y). Your X data is attribute and your Y is variable. Run a 1 factor ANOVA. (Minitab==>STAT==>ANOVA==>Oneway Use the MInitab help to understand the results. (if your p<=0.05 then X has a significant impact on Y (95% conf)
Hope that helps.0June 5, 2012 at 9:41 am #193522HI MBBin WI,
So is there no test that will accomodate this scenario? What about correlation?
0June 5, 2012 at 9:54 am #193523Hi Doug,
Thanks for the response. When I run the test I am not getting a p value. I am seeing only a rsquared
Can you email me at [email protected] so we can discuss? Thanks SO MUCH
0June 5, 2012 at 12:36 pm #193524
MBBinWIParticipant@MBBinWI Include @MBBinWI in your post and this person will
be notified via email.@safarbags – No, correlation must be between two contiuous variables.
While Doug gave you the more universal test, there is a more specific test that would be better in this situation. You have two criterion (Yes/No) and (I’ll assume) a continuous response variable (avg revenue for the past 3 years). For this scenario a 2 sample t test would be appropriate. If you had the avg revenue for the machines prior to and post upgrade, then you could do a paired t test which would be the best as you would then be comparing more specifically the before/after by machine.
0June 5, 2012 at 12:43 pm #193525Hi MBBinWI
Well there was no data for when it was upgraded or not. All I have is a file with Yes or No then the revenue. So I can use a 2 sample T test?
0June 6, 2012 at 7:14 am #193528
MBBinWIParticipant@MBBinWI Include @MBBinWI in your post and this person will
be notified via email.@safarbags – yes, and a graphical way to show the differences would be to use a box/whisker plot.
0 
AuthorPosts
You must be logged in to reply to this topic.