Business Statisticas Assignment: Lab | Variable Selection.

profiledeef
LabGuide_VariableSelectionwiththeGSSDataExplorer.docx

Lab Guide: Variable Selection with the GSS Data Explorer

In this lab, you will use the GSS Data Explorer to guide your selection of predictor and outcome variables of interest to you (from within the GSS Datasets). You will analyze these variables in future labs. Follow this lab guide and the lab report template closely.

1. Go to the GSS Data Explorer and set up an account. If you already have an account, verify that you can log in. They send an email confirmation that you need to use to confirm your account.

a. https://gssdataexplorer.norc.org/

2. Create a project.

a. Title your project: YourLastName Firstname_Variables of Interest.

b. The following options may not show up in the same order shown here. If you save your project but still need to adjust some settings, just go to your project on the My GSS page, click settings (the gear), and you will be able to finish adjusting the settings.

c. Set file format extract type to SPSS.

d. Set the privacy settings of your project to public.

e. Use the Copy Link option to copy the link to your project and paste it into your lab report. If copy link does not show up (as in the following screenshot), you can manually copy the link. Make sure that the final part of the link (the numbers) copies over. If it doesn’t copy, manually enter the numbers at the end of the link.

3. Details and goals for your variable selections. (The next section has step by step instructions for locating variables in the GSS Data Explorer).

a. Select sets of variables fitting the following criteria. Each of these variable sets corresponds to an analysis pathway that we will explore in the next labs. Your variable sets should include variables that are interesting to you and that you expect to be related for some reason. If there are variables that you are very interested in analyzing but they don’t fit into the any of the following sets of criteria, let me know (via email, phone or text). Whenever possible, I will do my best to put your preferred variables in an appropriate format for you to use in the later labs.

i. Set A: Nominal predictor, interval/ratio outcome

ii. Set B: Nominal predictor, nominal outcome

iii. Set C: Interval/ratio predictor, Nominal or interval/ratio predictor, interval/ratio outcome (3 variables: 2 predictors, 1 outcome)

b. For each variable set, be ready to describe

i. Why you chose these variables.

ii. The relationship you expect to see among the variables in the set.

iii. The response categories or range of each variable.

iv. The link to each Variable Detail Page.

c. Follow the Lab Report template to report your variable selections, hypotheses, and other requested information. Use the template to help you organize your information while you are doing your variable selections.

d. How do I pick a subject area for my variables?

i. Ideally, you should select variables (especially outcome variables) that are related to the key trend area about which you did your library research.

ii. If you want to change or refine your area, that is fine, but you may need to do a small amount of library research on the new topic for a later lab (e.g. to identify 2-3 relevant research articles).

e. How do I know if a variable is a predictor or an outcome variable?

i. Most variables could be predictor or outcome variables, but some variables can only be predictors because they are associated with stable demographic characteristics. For example, Respondent’s Sex and Age can only be predictors. If there are issues with your variable selections, you will have time to change them before the analysis stage (in a later lab).

4. How to find variables and the information associated with them.

a. From in My GSS, click on your project.

b. Click the plus sign in the Variables box to add variables to your project. Year, id, and Ballot are variables that are always present by default and do not count as variable selections.

c. In the Search Data area, select Range of Years, then Select Specific Years, and set it to a single survey year. You can select any survey year. The year you select will affect the variables that are available.

Range of Years:

Select Specific Years:

Select your year of interest. You can choose any year you like. The year (survey wave) you choose will affect the variables available. Many variables are available in multiple wave, but some variables are only in one wave.

Filter your results by Module/Subject to help you more easily identify and explore relevant variables. Modules/subjects correspond generally to the Key Trends used in the last lab.

The Module/Subject view is helpful for exploring the variables available in areas of interest to you. Select module/s of interest to you to see the specific variables related to that subject. The numbers in parentheses refer to the number of variables related to that subject.

If you are looking for a specific variable or want to do more targeted exploration, search by keyword. If you don’t select any modules, it will search all the modules by default.

Example of a Keyword search. I was interested in Labor force status.

Example of a Module/Subject search. I was looking for variables related to the Children module, but the selected survey wave/year (2014) did not actually include variables from that module. In this situation, either select another module or select a survey wave that includes the Children subject area.

In order to figure out the appropriate survey year for a module, reset Filter by Specific Years to show all years, and search the module again. If you hover over the green block, it will show you the survey wave for the variable. In the following example, these variables are only in the 1990 wave.

To see the information about a variable, click on the variable name, and you will be taken to its Variable Detail page. You will use this information in your Lab Report. In the following image, the labeled areas identify:

A. The variable name (for use in areas of your report where you need to identify the variable name> Thus, the variable name in this example is: Being parent now harder (NOT parharder).

B. The survey question associated with the variable.

C. The survey wave/s (years) that the variable is present in.

D. The response options for the variable. When reporting response options, report the Labels (NOT the codes).

E. If you want to save the variable for possible use later, add it to your cart. It is easy to delete variables later, if necessary.

You are now ready to select variables of interest to you!

Follow the Lab Report template to report your variable selections, hypotheses, and other requested information. Use the template to help you organize your information while you are doing your variable selections.

Examples of formatting and organization for the tables in the Lab Report.

Use the example below for your reference when you are completing the lab report. The Lab Report Template has template tables and formal guidance for completing the tables.