Sociology 428 Assignment #1

Sociology 428: Research in Gender Inequality

Assignment #1 Due February 16 9:00AM

Use the "Boolean" search in LexisNexis to count the annual number of "working mother" stories in the New York Times from January 1,1981 to December 31, 2008. Here's the search string we will start with:
(work! OR employ! OR career) w/2 mother!

Plot the counts on a spreadsheet and draw a fitted curve (a polynomial of order 2). Compare the overall trend with the trend found for the Washington Post articles. Is the NYTimes trend also curvilinear, peaking at a similar time?

The total number of articles is too many to code manually. We might be able to do more coding if we used just a sample of these articles. But would a sample show the same trend as the entire population?
How would the trend results above in #1 be any different if we sampled just one four week period each year? Use your two assigned 4-week periods and repeat step #1 above for each of the two sampled 4-week periods. Are the two trend curves basically similar to the total curve? If you added the two 4-week periods together into an eight-week sample?

The computer search counts many articles that don't really mention working mothers (e.g., "He worked with Mother Teresa"). Human coding can eliminate some of these errors. Is the human coding worth the extra effort?
Repeat step 2 above for each of the two monthly counts, but count only articles that genuinely mention working mothers. Use the "extended list" results from LexisNexis to identify erroneous articles. Occasionally you may need to look at the article itself.
List any cases that you found marginal or difficult to classify as actually mentioning working mothers. What was your final judgment and why?
What proportion of all the articles identified by the Boolean search are erroneous? Plot the two combined 4-week revised counts of correct articles (i.e. an eight-week count). Does the trend look any different than the uncorrected trend?

Suggest an alternative search string that you suspect might do a better job of identifying articles that mention working mothers. Do you think your alternative would have fewer incorrect hits? Would your alternative search string identify more articles about working mothers that were skipped by our first definition?

return to:

Last updated January 26, 2009