Which test do I use with 2 variables and unequal testing pools?

kadify

New member
Joined
Sep 19, 2017
Messages
4
I'm trying to determine the statistical significance between 5 trials. The data is:


  • Control/Trial 1: 32 votes of yes out of 600 participants
  • Trial 2: 28 votes of yes out of 576 participants
  • Trial 3: 59 votes of yes out of 602 participants
  • Trial 4: 21 votes of yes out of 613 participants
  • Trial 5: 35 votes of yes out of 584 participants



So I'm trying to figure out which test applies here since the number of votes is an independent variable but the number of participants is also not constant. I want to compare all of the trials together to determine which trial was the most successful and therefore which trial (technically each trial is a company test on which method to use regarding a certain project) should be approved.

I thought it would be a one way ANOVA but now I'm wondering if it should be a two way ANOVA since the group size varies for each trial. Whichever method I do use though, I also can't figure out how to actually process the data (i.e. is there a calculator I can use to determine the significance of each trial (2-4) compared to trial 1 which is technically the control?


 

kadify

New member
Joined
Sep 19, 2017
Messages
4
No one has any insight? My main question, if it wasn't clear from the first post is which test is the most appropriate given my data?
 

tkhunny

Moderator
Staff member
Joined
Apr 12, 2005
Messages
10,030
Show your work, not just your musings. This normally leads to better and faster responses.

First impression: Your sample sizes are just barely different. I would simply ignore it.

What are you testing? Homogeneity? Come up with a Null Hypotheses and go from there.
 

kadify

New member
Joined
Sep 19, 2017
Messages
4
Show your work, not just your musings. This normally leads to better and faster responses.

First impression: Your sample sizes are just barely different. I would simply ignore it.

What are you testing? Homogeneity? Come up with a Null Hypotheses and go from there.

I don't really have any work yet because I'm stumped on which test I should use to do work with... This is a question I'm supposed to solve, but I don't know how to begin because I don't know which test I'm supposed to use. My goal is to figure out which trial is the most viable and results in the most votes of yes, percentage wise, compared to the sample size. Using common sense i can see that that is trial 3, but I don't know how to show that statistically trial 3 has the highest probability of getting the most yes votes.
 

kadify

New member
Joined
Sep 19, 2017
Messages
4
Wow what a fail

I have to say this was one of the least useful math help sites I've ever posted on. I got more help with this problem posting on yahoo answers and Facebook from random people than I did from anyone here =/ Anyway I was able to run my test, I had to analyze it as an A/B/n split test, for anyone that may ever have a similar question and happen upon this board post.
 

mmm4444bot

Super Moderator
Staff member
Joined
Oct 6, 2005
Messages
10,324
I have to say this was one of the least useful math help sites I've ever posted on. I got more help with this problem posting on yahoo answers and Facebook from random people than I did from anyone here =/
These boards are frequented by volunteers and other "random" people; as mentioned in the forum guidelines, there is no staff here and it could take awhile to receive helpful information.

We're glad that you were able to get what you needed, from other sites. :cool:
 

tkhunny

Moderator
Staff member
Joined
Apr 12, 2005
Messages
10,030
Everyone has rules. We appreciate it when ours are followed. This does not make us useless - excepting within the scope of your compliance. Sorry about that. We are who we are and you are who you are.
 

stapel

Super Moderator
Staff member
Joined
Feb 4, 2004
Messages
15,946
I'm guessing that the original poster (1) was given the complete solution when (2) s/he was not supposed to be getting any help at all. Because it's currently blocked as "under review" (or else the OP deleted it) on Yahoo. :wink:
 
Top