Re: when is data considered "continuous" for parametric testing?

Hi Folks,
Your thoughful answers are helping a lot! (both in understanding what
to do and why). Thanks for taking the time to point me in the right
direction.

In my test, a person is given a series of 4 questions, each with a
binary response. I average these to get a result (either 0, .25, .5 or
..75, or 1). When I began to approach this analysis, just comparing one
group to another, I was simply doing a t-test. But that data is indeed
"ordinal" rather than "continuous" and I was concerned that I was not
fulfilling the assumptions (i.e. "continous"). I guess parametric
assumptions should really read "not categorical" rather than "must be
continuous. Or am I missing something? In a sense, there really is no
truly continous data, right? Just grades of ordinal: e.g. further
decimal points of ordinal data (e.g. 1-100 or 1.00 to 100.00). And then
there's binary data (0 or 1), which is both categorical (wrong or
right) and ordinal. But I reflexively do a chi square test there. Is it
fair to say that ordinal data can be treated as continous data, for
purposes of statistical testing, with exception of binary?

Thanks again!
Jeff

.

Loading