For our next
assignment, we will be using the General
Social Survey 1972-2006 data file. This very large data set is
described in an article
from NY Times. To access the
data we can use the Survey Data and Analysis program at the University
of California, Berkeley. Just click on SDA.to
open the software. Click on "Analysis" and "Frequencies or
Crosstabulations" for this assignment. This is described on page 20 of
the Babbie book.
Your task is to produce a multiple bivariate table with one column
variable and five row variables one of which must be year and one
of which must be one of the PRES variables. .
The percentages should be row percentages. You should write a
paragraph summarizing the results in which you correctly describe some
of the percentages. The finished product should look like this Sample
Assignment. You should prepare the table in a word
processor, save it in *.doc or *.rtf format, and submit it to SAKAI or
turn it in on paper as instructed
You may use any variables you wish for your column variable. You
can
search for variables with the SEARCH facility. If you do not have
any idea, I
suggest
you use one of the ones selected by the New York Times for the column
variable. These include premarital sex,
trust in others, frequency of prayer, marijuana legalization, exciting
life, fear at night, the afterlife, spanking children, confidence in
institutions (just select one institution), happiness, abortion (pick
one item), gun permits, happiness of marriage, newspaper readership,
x-rated movies, homosexual teachers, social class memberhsip. You
should NOT use women & politics because we used that for the sample
assignment.
Your row variables may include variables such as age, sex,
religion, political party, year of interview, region of residence,
education, marital status, etc. HOWEVER, this data set has not
been
prepared for this purpose, so many variables need to be RECODED
first. To do this, click on "create variables" and
"recoding rules".
It shows how to recode AGE for this purpose. You need to
check each
of your row variables to see if it needs recoding. This
gives you the
flexibility to recode them as YOU wish rather than relying on someone
else's choices. However, once you create a recoded variable it is
available for others to use, so I may be able to create some and leave
them there for you. If I do, I'll add their names to these
instructions here:
Then you need to do a cross tabulation for EACH of your row
variables. You may
wish to
print these out for your convenience or you can work on the
screen.
You take the results for all the variables and type them into one big
table for your assignment.
When you choose "year" as a variable, you should get the same results
the
New York Times got. If your variable was asked in more than five
years, you may just select key years rather than
doing
them all.
A number of students have asked if they need to include the results for
the total sample in their tables for the multiple bivariate
assignment. The answer is YES.
Where do you get these? If you look at the bottom of any of your
cross-tabulation tables you will see them. These may vary
slightly
from table to table, however, because of missing data. Another
way to
get them, for the whole sample, is to type the name of the column
variable in your table in the ROW box on SDA with nothing in the column
box. This will get you the frequencies.