What analysis should I use?!
Brionny
Pathological/Reactive 0
Reactive 7
Reactive/Calculative 10
Calculative 10
Calculative/ Proactive 6
Proactive 9
Proactive/Generative 0
Generative 0
this is what I have. Pathological is no presence of
variable....Generative is lots of the variable. the numbers are how
many people thought that that much of the variable is present. I would
like to know if there is a significant difference between the number
of people who rated each variable. I guess... unless you see some more
valuable information that could be gleaned...
I am so lost!
Any help would be much appreciated
Bry
Art Kendall
What is a case? What are the variables? Are there subsets of variables?
What do you mean by "variable" in your post?
Is it something like this?
Each case is a person who gave yes or no answers to 5 variables
(Pathological Reactive Calculative Proactive Generative) for
several stimuli. The stimuli were microscope slides and the 5 questions
were asked about different bacteria (stimuli).
In response to a stimulus a person "checked or not" all that apply.
for example with 4 stimuli and the 5 variables for (Pathological
Reactive Calculative Proactive Generative) there would be 20
variables in the data.
Art Kendall
Social Research Consultants
Rich Ulrich
This is not at all clear, and my guess is different from Art's.
It was helpful to me that Art suggested *some* context -
even if it is the wrong one.
I think that there is an "ordinal" scale represented by
Pathological, Reactive, Calculative, Proactive and Generative;
and there are 4 more possible "scores" available which consist
of half-way between each of the 5 specific descriptors.
That is what you mean by saying that the numbers show
"how much of the variable is present" -- Generative is "a lot"
and Pathological is "none".
In all, there are 42 scores or ratings, and not *variables*.
- I must be totally misconstruing everything, if (as Art seems
to be guessing) the OP actually does mean "variables" in the
sense that these are 9 different categories, each of which
might be independently marked as Yes/No
Now, the only hypothesis that I can figure out seems to
be the rather trivial one, "Is there a uniform distribution
across these levels?" Well, there are 0 scores in the 4
extreme categories, and 42 in the middle 5, so the answer
is no. So what? Should someone bother with a test?
Doesn't the distribution reflect, in some fashion, the manner
of selection of the sample?
I doubt that anyone will be interested the counts being
unequal, unless there is some important information added
from somewhere else.
--
Rich Ulrich