I thought I would post here for feedback on some early results that I have gotten for a research project I am working on. I got the idea for the topic from reading this forum. I am researching the viability of Graduate Degrees from colleges that have DETC accreditation for use as a credential for an Adjunct Faculty position at the Community College level (per several topics I’ve read here). I sent a two question survey to 1,100 institutions that identified themselves on the internet through their name as community colleges. I have received about 2% of the responses so far, would anyone be interested in what I am discovering?
Okay, I asked because with only 2% of the responses in, the results are only starting to approach statistically significant. I also want to point out that the survey methodology I’m following suggests a statistically significant sample size from a population of 1,100 should be around 280 responses. I don't have close to that many yet, but plan to try to reach that number. What I have preliminarily found is that within a 99% certainty, 95% (+-12%) of the answers from Community Colleges to the question: “Would you consider a qualified candidate for an Adjunct Faculty position, who's only academic credentials are from a legitimate college (or colleges) accredited by the DETC?” Will be: “Only if the degree granting institution is also accredited by a regional accrediting body.”
Sample sizes aren't statistically significant. Differences (between samples or between hypothesized population values and samples) are. You can find a significant difference with quite a small sample. The smaller the sample, the larger the observed difference must be to be considered significant. (Simply put; sample distribution also plays a role.) The larger the sample, the smaller the necessary difference before significance is established. It all has to do with the probability you'll find a difference where none (in the population(s)) truly exists. The numbers you present don't make sense to me statistically, at least as you present them. If you have tested this to the .01 level (as you assert, even if you don't use standard terms), you don't have a preliminary outcome. You have an outcome. In fact, if you've tested to the .01 level and found a significant difference exists, then the odds of this turning out to be incorrect (that no difference really exists) is 1 in 100. Stop spending money on sampling and get on with writing. Unless, of course, your sample really doesn't rise to that level. (Which I would suspect, given how you've couched the language.) Sorry, but if you run stats on this board, you'll probably hear from me. No offense intended.
No offense taken. I’ve couched the words because I’m still gathering input and survey results. My goal was 280 responses, I currently have 26. Every response except for one was: “Only if the degree granting institution is also accredited by a regional accrediting body.”
