When does difference in sample size become an issue?

[80past2]

I was comparing two different groups, and in one, my n was 1600, and the other n was around 700. I found pretty much all significant differences, but is that maybe due to sample size. I tried doing a random selection making the sample sizes equal (both around 700) and got more or less the same numbers and significance every time. Should I do anything else, or is this fine?
I also bootstrapped within one of these restricted sets and got about the same numbers.

 

[andrewr]

Do you need help checking distribution type by plotting it?

I was comparing two different groups, and in one, my n was 1600, and the other n was around 700. I found pretty much all significant differences, but is that maybe due to sample size. I tried doing a random selection making the sample sizes equal (both around 700) and got more or less the same numbers and significance every time. Should I do anything else, or is this fine?
I also bootstrapped within one of these restricted sets and got about the same numbers.

I am doing a plotting program to look at data-sets and check for normal-ness,
if you'd like, & your data is "near" normal -- you can attach a text file with the data (or a scaled version of it...to obscure what it is) that just lists the data values. eg:
7
3.5
11.0

etc;
And I could plot the data into 1% or 0.05% quantiles; like this:
converting binomial/normal distribution into quantiles and comparing against normal
and then I could post the graphs for you...
It will show skewness, and some information that could help identify what type of distribution it really is, but it's mostly to check Gaussian data...