question about frequencies in cases separated by grouping variables
Locked
question about frequencies in cases separated by grouping variables
 
|
Hello
list,
How do I use
grouping variables to get a frequency for cases within a larger data set?
I am an archaeologist studying stone tools from 15 sites and I coded the sites
from 1 to 15. Within each site are two categories of shape for tools
either diamond coded as 1, or lenticular (flat), coded as 2. So I have
coded tools from the first site that are diamond-shaped as 1.1 and tools that
are lenticular as 1.2. This goes on through site 15 being coded as 15.1
and 15.2. For metric variables such as width and thickness, I was able to
establish my codes as a grouping factor and get descriptive stats for each site
and each shape category. My question is, how do I get frequency counts for
qualitative variables like color and type of stone that aren't metric
values. Now frequency under descriptive statistics menu lists the counts
for the whole data set, but I would like counts for categories 1.1, 1.2, ....
15.1 etc. If I need to run syntax to do this task, please give suggestions
on how to run that since I've only used menus on SPSS . Thanks in advance
for any suggestions.
Cheryl
Fogle
Ph.D. candidate,
Department of Anthropology, University of New
Mexico
|
Re: question about frequencies in cases separated by grouping variables
|
|
Cheryl,
I think that you need to use either crosstabs so that you can see the crosstabulation of site-tool_shape by thickness category or color. Or, if you have thickness entered as the observed value you could use Means to see mean thickness by site-tool_shape. I'd like offer the suggestion that it would be (much) better to code site as one variable and shape as another. It will give you more flexibility because you can, for instance, look at the proportion of lenticular tools across sites. You can't do that now. I don't know anything about archeological data collection and analysis but I'd also suggest coding tool characteristics as separate variables and recording the observed values for those variables. You can come back later and, via recode operations, create new category variables for each 'raw score' variable. Using IF statements you can create composite categories combining, for instance, shape category and thickness category. Gene Maguin >>I am an archaeologist studying stone tools from 15 sites and I coded the sites from 1 to 15. Within each site are two categories of shape for tools either diamond coded as 1, or lenticular (flat), coded as 2. So I have coded tools from the first site that are diamond-shaped as 1.1 and tools that are lenticular as 1.2. This goes on through site 15 being coded as 15.1 and 15.2. For metric variables such as width and thickness, I was able to establish my codes as a grouping factor and get descriptive stats for each site and each shape category. My question is, how do I get frequency counts for qualitative variables like color and type of stone that aren't metric values. Now frequency under descriptive statistics menu lists the counts for the whole data set, but I would like counts for categories 1.1, 1.2, .... 15.1 etc. If I need to run syntax to do this task, please give suggestions on how to run that since I've only used menus on SPSS . Thanks in advance for any suggestions. Cheryl Fogle Ph.D. candidate, Department of Anthropology, University of New Mexico ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: question about frequencies in cases separated by grouping variables
|
In reply to this post by cheryl fogle
Cheryl
What other variables do you have and how have you
coded them? How many samples do you have? If you send me (a schema
of) your raw data file I'll have a better shot at it for you.
Assuming you have already generated a data set and
your 1.1 to 15.1 etc is called tools, try this
(untested):
Compute site = trunc ( tools
).
comment This yields site values 1 thru
15.
val lab site 1
'Stonehenge' 2 'Knossos'.
....etc etc whatever your sites are
called (don't forget full stop!)
compute shape = mod (tools
)*10.
comment This yields decimal point x
10 = values 1 and 2.
val lab shape 1 'Diamond' 2
'Lenticular'.
freq site
shape.
cro site by
shape.
You can be weaned away from menus towards syntax by
Jacqueline Collier's new book (Using SPSS Syntax: A Beginner's
Guide, Sage, 2010, which may not help with this particular example, but
is otherwise very useful for getting away from the menus and gives access to
facilities which simply aren't available via menus) or start from scratch
with the SPSS tutorials on my website (http://surveyresearch.weebly.com/survey-analysis-workshop-and-spss.html which
treat SPSS as a language from scratch, and are quite fun.)
HTH
John Hall (survey researcher, not archaelogist, but
did Arch and Anth at Cambridge 1962-63)
From: [hidden email]
|
Re: question about frequencies in cases separated by grouping variables
|
|
Too lazy to check manual: did it from memory of
many years ago, but glad it worked.
|
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |