hi all, am a law student trying to come up with a table to show the revenues earned by companies over the particular period of time. i have a group of companies say 100, i need to choose a number of them. probably 30%. is that a good sample? what is the name of this method of sampling - someone told me it is random sampling. (i don't have any clue about stats but am learning the 5 months to submission - not good!) my question is:

1) how do i describe my methodology - is 'random sampling' an academic method?
2) if i use random sampling is my study qualitative/quantitative or what? i also intend to hand out a number of questionairres.
3) are there any books i can read on this?

this data forms only a small part of my thesis and its only serves to elaborate what am already saying by providing background information. i rely on 'argument'/ theory to make my thesis using comparative case-study. i also look at stuff from other disciplines. how do i describe my method? any views are helpful.

hailey (love you all. xx)

Hey Hailey! The type of sampling depends on how you are going to choose the 30 or so companies. Are you planning to use a particular method to choose which of the companies you're going to include? Whether your study is quantitative or qualitative doesn't really depend on the type of sampling you use- it's more to do with the type of data you have/are going to collect. For example, interview data is often analysed using a qualitative approach, whereas questionnaire data is often analysed statistically and that would usually be a quantitative approach. That's only a very simplistic answer- there are exceptions. Any decent stats book should have a discussion of these queries! Best, KB

thanks KB. I was going to using the randomising function on the calculator. i think its quantitative am using them because although i have interviews they are much less and informal than the questionairres. i will get a hold of a stats bk.

is 30% considered a decent sample?

Sample size depends on
(i) what you're measuring/what your hypothesis is
(ii) how much variation you expect to see between your study subjects (in this case companies)

When you get hold of a textbook suitable for your field look at 'sample size calculations' and that should help.

As a rough guide, if the answers to your survey question are numbers (e.g. revenue) or can be coded into categories (yes/no, high/low/medium) then you can do statistical analyses and it's a quantitative study. If your responses are free text and then you're picking out common phrases/themes etc it's qualitative. I suspect your work will be quantitative (or you can make sure you ask the questions in such a way that it is).

BTW - my perspective is as an epidemiologist and for me this stuff really matters. For you, maybe the sample size can be estimated - you could sweat and toil to make all of this perfect and then it doesn't matter to your examiners. As a general rule with any stats, if the stats matter and you're not an expert, it's a good idea to run it past a statistician in your department before you go too far.