Hi everyone,

I have a sample of prices of products across different competitors. I am trying to compare the pricing across each one with the goal of looking, which company is significantly cheaper?

To compare the prices I have already ran independent t tests on the average price. But now I want to compare count data of competitiveness. Basically my data is as follows

Pa -- Pb --- Pc ---- A cheapest --- B ---- Cheapest --- C ---cheapest

5 -- 8 -- 9 --- TRUE -- FALSE --- FALSE

I then count all the TRUE/FALSES and I'll end up with a distribution for example of (N = 100)

A : 20

B: 40

C: 40

Which represent the number of counts company A/B/C is cheaper than its competitors.

Because I am drawing from a sample of a larger population I am wondering whether I can look for significant differences between these count data. Especially to test whether it deviates from an assumed distribution of 33/33/33 (companies are equally as competitive).

I am familiar with models such as the Poisson distribution, but because I am not looking to use any other explanatory variables to explain causal relations I am looking for a statistical technique to compare this data (aka to something like a t test)