8.0 The KEYVARS Variables

The KEYVARS Area of Interest contains many variables that have already been created for you. If you are used to creating your own variables, look at this list of variables, first, to see if one exists that has already been created for you and that you can use. Often the 'coding' used to create these variables is given for you in the documentation. If you find a variable that you could use but minor changes need to be made, you may be able to modify the coding, adapt a variable to your use, and ease the process of creating a variable. If you become familiar with this list of variables, you can save yourself time and effort.

Like with the COMMON variable we'll begin by looking at what the NLSY79 User's Guide says about the KEYVARS Area of Interest (Appendix C: NLSY79 Areas of Interest). You can navigate to this quote by starting in the Contents window, Documents subdirectory, NLSY79 User's Guide: Double-click on it to open Text/Cdbk window and go to Appendix C: NLSY79 Areas of Interest.

Name: KEYVARS - Created key variables Description: This area of interest contains the *KEY* variables as well as other summary variables created by the Center for Human Resource Research from raw data. They encompass a wide range of employment, military, education, income, poverty status, local labor market unemployment rate, and various other geographic-related variables. Sources of Data: The variables on this file are all created with the exception of the 'Type of Residence' variables, which are derived from the Household Interview Forms. Related Areas: This area of interest contains the primary longitudinal (all main file) created variables. Other constructed variables can be found in JOBINFO, the supplemental FERTILE area of interest, and the special NLSY79 geocode, workhistory, and child data files.

NLSY79 User's Guide (Appendix C: NLSY79 Areas of Interest)

From this documentation you get a description of the contents and an understanding of the range of topics the list of variables cover; how or from what the variables were created; and, when appropriate, related areas of interest. Use this range of topics to explore the variables in the list. Look for the areas of interest that will aid in your research.

Let's open the KEYVARS area of interest. In the Contents window make sure the index opened is the Area of Interest index (double-click on it to open it in the middle frame). Type 'k', 'e', 'y', to move the highlight bar to KEYVARS. Double-click on KEYVARS to open the list of variables in the Variables window.

When you open KEYVARS, a listing of the variables assigned to this group is displayed. Notice some of the general characteristics of the group first. Look at the column headings that depict all the variables. The variables in the group that have been created for you have the '*Created' indicator in the Question Name column. The variables that are a result of a question ask of a respondent have the question's name in this column. The first variable in the list 'R0188000 TYPE OF RESIDENCE R IS LIVING IN 79 INT' has the question name S24Q25. Many following it were created for you. The KEYVARS list contains many of the created variables you will use and is a good place to search first before you begin to create your own variables.

You could also find this variable using the Qname Index [Contents window, Index = Qname]. Find and open the S24Q group by double-clicking on it. In the Variables window the S24Q list will appear. Search/scroll down through the list using the Question Name column as a guide until you find the question name S24Q25. Notice the variable's 'Name' is R0188000. You have located the same variable with different information and using a different index. Remember, the indexes on the CD are just access points to the variables in the NLS.

You may sort the variables by clicking on any column heading. The default sort order is by variable Name. The variable's Name is what you will use in SAS, SPSS, Stata Dictionary when you analyse the data. The variable's Name is derived from the Reference Number minus the decimal point (.).

Figure 7 KEYVARS Variables in the Variables Window

Suppose you want to look at those variables like R0188000 TYPE OF RESIDENCE R IS LIVING IN 79 INT in the figure above. You could sort by Description to see how many variables have a similar Description.

Sort these variables, now, by Description! (Click on the column heading Description.) Now you must locate the variable R0188000. Scroll down to where its description begins. (For this example, look for the first few occurrences of the word 'TYPE OF ...' and look for the Question Name 'S24Q25' in that group.

8.1 TYPE OF RESIDENCE

Notice some of the general characteristics of this 'group' of questions as highlighted in Figure 8 below. You can glean several bits of information about the variable:

1. The description of each begins 'TYPE OF RESIDENCE R IS LIVING IN'. From the description you can often glean words that were used in the question. Familiarity with the words or concepts used in the question can enhance the use of another index "Any Word in Context" and may enrich the use of the written documentation in this case the NLSY79 User Guide.
2. The Survey Years column reflects a change in the frequency the question is asked: annually from 1979-1994; every two years thereafter, 1996, 1998. (The reason for this is the respondents were in the middle of their lives and there was less potential for change in them from year to year. As a consequence, a new 'youth' survey, NLSY97, could be funded and put in the field in the inactive years.) The Survey Years are consecutive and the same as in the Description except for the last three years (94, 96, 98). (The 'survey year' and 'INT' were dropped in 1994 and following active years. There is no difference in the questions other than they sort in a different order due to the omission of 'yy INT')
3. The Question Name column reveals that the question was asked under several different 'names' and in different sections of the survey. (HHI = Household Information; S = Screener or Section; D = Deck; Q = other.) You may want to investigate if the questions differed in the respective sections, or if the answer categories changed, and/or if the sponsor of the section changed.
4. You can highlight this group of variables or just a subset of them and work with global tagging functions. See below.

Figure 8 R0188000 TYPE OF RESIDENCE R IS LIVING IN 79 INT

If you wanted to include this block of variables in a tagset for later extraction, you have to 'tag' (or 'check') the variables in the group. You can tag them individually by left-clicking on the 'checkbox' of each or you can highlight a group of them and tag that group all at once.

To highlight a group of variables: left-click on the first occurrence of the variable, hold down the <Shift> key, and left-click on the last variable in the group. You are now ready to work with a block of variables.

Suppose you wish to begin building a file of variables that you will use in an extract of variables. To build the file you must 'tag' the variables you wish to include in to the 'tagset'.

To view the tagging commands: place the cursor over one of the highlighted variables in the group and right-click. The pop-up tagging menu should appear, you can select one of these functions by moving the cursor to it and left-clicking on it. In the figure below, the 'Tag-Selected F3' option was chosen. Remember, you always want to include the variable 'R0000100 IDENTIFICATION CODE 79 INT' (Location = Area of Interest, COMMOM) in your tagset so that you can identify individual cases when the data are extracted corresponding to the variables in the tagset.

8.2 Poverty Variables

Now suppose you want to find out something about the respondent's 'poverty'. In the KEYVARS variables list, the only columns that will tell about 'poverty' are the Description and possibly the Question Name columns.

Question Name. Click on the heading "Question Name" and the data will sort by that field. When you do, investigate how the variables sort in the group. The ones that have been created for you sort to the top, those that are the result of a question asked, sort by the name. In the case of variables resulting from a question, notice the groupings of question names, for example, look at the question names "POVSTATUS" or "POVLEVEL". Read the Description of the variables to get more information about the individual variable. The Description contains additional information in the words "FAMILY POVERTY...". And what happened to the longitudinality of this variable for the years 1979 - 1992, if available?

Figure 9 "POVSTATUS" and "POVLEVEL" Variables

Sort this variables list, now, by Description (left-click on Description) and scroll to find the "FAMILY POVERTY..." variables. You should see the information displayed in Figure 10 below.

Description. The variables whose Description begins "FAMILY POVERTY..." were created from 1979 - 1992. From 1993 - 1998 they were the result of the questions POVLEVEL and POVSTATUS. You may want to investigate these variables in detail as to how they were asked, what were the response categories, etc. You can get some of this information from the Codebook page for each of these variables. Double-click on the POVLEVEL or POVSTATUS variable you want to look at and it will appear in the Text/Cdbk window. Look at the others, too. You should notice after looking at a few codebook pages that one question for POVLEVEL and one for POVSTATUS was asked in each survey year between 1993 - 1998.

Figure 10 "FAMILY POVERTY …IN…" Variable

If you looked at the codebook for one of these POVERTY variables you would see a hyperlink to "Documentation Links: Codebook Supplement-Appendix 2". If you followed this link you would see two choices for further information:

• Variable Creation: Total Net Family Income
• Family Poverty Status and Family Poverty Level Variables

"Variable Creation: Total Net Family Income" contains only the 'raw' coding for the "Total Net Family Income" variable (no textual explanation). "Family Poverty Status and Family Poverty Level Variables" contains a textual explanation of the variables and Poverty Guideline tables.

8.3 Process of Finding Variable Details

In order to find out about a variable in detail, you may have to attempt a series of steps to get the complete perspective. The index, the variable list, description of the variable, the codebook, and the documentation, if available, all contain information about the variable. Below is a summary of the steps to attempt in order to view all perspectives of a variable.

Contents window, open Index of choice (double-click). Look at the groupings in the Index. These contain the variables. Open a group in the Index by double-clicking on it: Variables window. Look at the individual variables in the list. Sort them by Description and study the textual description of like sorts. Sort them by Question Name, see how the sorts change, and examine the Question Names of like sorts. Double-click on a variable to view its details in the Text/Cdbk window. Study the codebook display and observe at the bottom if there is more information hyperlinked in "Documentation Links". If so, link to it. Try the NLSY79 User's Guide on the CD to find out if more information on the scope, genesis, etc. is available. Look first at Chapter 4: Topical Guide to the NLSY79.

In the next section we will look at some additional KEYVARS variables in detail using the process described above. The additional KEYVARS variables are discussed without accompanying images. You should now be able to use the software to find the accompanying screens yourself. Reminders are given during the discussion.