Lab: Content Analysis I (Manifest Content)

In this lab we will be doing a content analysis of scholarly journals in sociology and social work and popular magazines and their coverage of our theme topics from 1993 to 2003. We will test the hypothesis that:

Coverage of specific topics goes up and down at the same time for all three sources of information (sociology journals, social work journals, popular magazines).

Therefore the independent variable is the journal or type of journal and the dependent variable is the amount of coverage in a given year.

We will collect our data using electronic bibliographic data bases available though the Lehman Library. We will count the number of citations on your theme in each year for 6 specific publications.

Type	Database	Journal s
Sociology	Sociological Abstracts	1. American Sociological Review 2. American Journal of Sociology
Social Work	Social Service Abstracts	1. Families in Society 2. Social Service Review
Popular	Masterfile Premier	1. People 2. Newsweek

To get to the list of bibliographic data bases available at Lehman click here. If there is any problem with that link, click here

Here are the search terms to be used

For Sociological Abstracts or Social Sevices Abstracts

• Select "Sociological Abstracts" or "Social Service Abstracts" (depending on which journal you are doing). Then click on "advanced search."
• Underneath "Build your Search Strategy" set your search criteria.
• For the first, select "Keyword=(KW=)" or "Descriptors= (DE=) as listed above, then enter the terms above.
• For the second, select "Journal Name=(JN=)" as listed above.
• Where it says "From" type 1993 and where it says "To" type 1993.
• Click on "Search."
• Just above the list of citations, next to a blue arrow, you will see the number of citations (which may be large, small or even 0). Fill this information in the grid below (explanations of abbreviations are in the second graphic. )
• Enter 0 if there are no citations for a year.
• You can then change the year using the box to the left. Fill in the numbers for each year.
• Repeat this procedure for each of the sociology journals. Switch to the Social Service Abstracts data base when appropriate.

Caution #1: Make sure to only count citations for the specified journal. Check a few of the citations to make sure that you are getting accurate results.

Caution #2: Make sure that you use the correct data base. There is some overlap between the two databases, but to get all of the articles from a particular journal you need to use the correct database. Select the data base designated above.

For Ebsco

• Scroll down and click on "Master File Premier."
• Click on "Advanced Search."
• Put the the magazine name (e.g. "People") in the first line and change the first line of Default Fields" to "SO Journal_name.
• Enter your search term exactly as written in the table in the next line.
• For the first "Yr" enter "1993" and for the second enter "1993."
• Click on "Search."
• Looking at the list of citations (which may be short or long) copy the number of citations in your journal each year. Fill this information in the grid below (next page in printed version) (explanations of abbreviations are in the second graphic).
• Enter 0 if there are no citations for a year.
• Click on the "search history" tab, and then "Revise." Change the year. OR click on "search options" and change the year.
• Repeat this procedure for each of the popular magazines. (You can just go to search history and click on revise, then change the name of the magazine).

Caution: Make sure to only count citations for the specified journal. If you use a different method of getting the counts, some listings will include other journals which have the same words in the name (for example, when searching for People you will also get listings for People Management and Teen People Education. For Newsweek you will get some special editions. Do not count these.

Data Entry

When you have completed the grid,

• If you are in internet explorer, you can click here to open the data set.

OR

• go into SPSS (click Start, Program Menu, SPSS for Windows).
• Under "Open an Existing Data Source" select "journals.sav" OR click cancel, then file, open, and find journals.sav on the data set list and open it (in the lab it will be on the Z drive in the SPSS directory).

• Make sure that the screen looks like the blank grid shown somewhere near here (variable names across the top, years down the sides). If it does not, go to the bottom of the screen and select the "Data View" tab.
• Copy your numbers from the paper to the SPSS file (note: you do not enter the variables soctotal, swtotal and poptotal).
• After you have entered your data, take a look at the variables starting with asrn and going to ssrn. These show you how many total articles each of the publications printed each year. Why is this important?
• Once you have entered your counts, you need to calculate the totals and also the rate at which each publication covered your theme. Click on file, open, syntax. Find the file computetotals.sps on the list and open it. This will open a syntax window. (If you are not working in Carman 220 you can click right and select "save target as" here to copy the syntax file to your floppy or hard disk)
• Click run, all. If you look back at the data screen you should now see filled in totals and rates. The totals add up the numbers for each type of publication. The rates are calculating the
  (number of articles on the theme in a year)
  ÷ (total number of articles on any topic in that year)
• At this point, if you wish, you may save the data set to a floppy diskette by selecting File, Save As changing the "Save in" location to A:, typing in the name "journals" and clicking "Save."
• Click on Edit, Options. Select the tab labeleld "Charts." In the box that says "Fill Patterns and Line Styles" select "Cycle through patterns." Click OK.

Creating the Graphs

• Looking at the menu at the top of the SPSS Screen, select Graph, Sequence.
• For "Time Axis Labels" select "year." For "variables" put the names of the six journals.
• Click "OK."
• Repeat using poptotal, soctotal and swtotal.
• Now run the following 2 graphs in exactly the same way: (1) newsrate, peoprate, asrrate, ajsrate, fisrate, ssrrate; (2) swrate, socrate, poprate.
• When you have all 4 graphs in the output window, put your name on the printout by selecting "File, Page Set Up" then Options. Type your name in the header.
• Then print (click on the printer icon OR select File, Print, OK).

When you have the graphs, look at them closely and compare them to each other. Pay attention to the values on the y axis as the graphs may be deceptive. Answer the following question susing a word processor.

IN YOUR OPINION,

Now look at the graphs that use rates and answer the same questions.

Overall:

Here are some titles from these publications for the period immediately prior to the period you studied:

• "The Enclave and the Entrants: Patterns of Ethnic Enterprise in Miami before and after Mariel" American Sociological Review, 1992, 57, 3, June, 411-414
• "A Cross-National Comparison of the Gender Gap in Income" American Journal of Sociology, 1990, 96, 1, July, 69-106
• "When Case Managers Manage the Seriously Mentally Ill: A Role-Contingency Approach" Social Service Review, 1990, 64, 1, Mar, 79-93
• "Religious Denominational Policies on Sexuality" Families in Society, 1992, 73, 5, May, 304-312
• "Pampering the elderly (II)". Newsweek, 11/26/90, Vol. 116 Issue 22, p58, 1p, 1c
• "Princess with a purpose." (cover story) People, 7/16/90, Vol. 34 Issue 2, p64, 8p, 15bw

Think about the titles (and any titles or summaries/abstracts you read). How do the articles in the popular magazines differ from those in the scholarly journals? Do you think that any differences you notice have any implications for the validity of our research design?