[SOLVED] Secondary Data Analysis Lab Exercise

Secondary Data Analysis Lab Exercise: Descriptive Statistics and Simple Graphs

In this lab, you will learn how to import a dataset into SPSS and generate descriptive statistics. These will include frequencies, percentages, and measures of central tendency (mean, median, and mode).

Make sure you have the GSS18.sav file open.

This first part is a “practice” where you are walked through how to use SPSS.

Finding Frequencies, Percentages, Measures of Central Tendency, Measures of Dispersion

1. Navigate through the menus – Analyze > Descriptive Statistics > Frequencies

2. This should open a new “Frequencies” window where you can select your variables

a. For this lab, let’s work with: WRKSTAT (Labor Force Status), DEGREE, AGE

3. To select your variables you must move them from the left box into the right box (either double-click on the variable name or select it and use the arrow between them)

4. After selecting your variables, make sure that the “Display Frequency Tables” option at the bottom of the window is checked.

a. Let’s explore some of the other frequency functions in SPSS

i. Click the “Statistics” button in the “Frequencies” window

1. A new window should open and you will see a lot of different options. Let’s pay particular attention to the “Central Tendency” box. Let’s check: Mean, Median, and Mode.

2. Click “Continue” to return to the “Frequencies” window

ii. Click the “Format” button in the “Frequencies” window

1. This function allows to choose different ways to order your frequency table. For now, we can leave these options alone, but it’s good to know these options exist for future reference.

2. Click “Continue” to return to the “Frequencies” window

iii. Now back at the “Frequencies” window, click “OK”

5. SPSS will now open an “Output” window and will generate the requested tables and statistics. Note that percentage and cumulative percentage are both automatically included in the frequency table.

NOTE: Percentage is the percent of the total N encompassed by any given value. Cumulative percentage is a cumulative summation of all percentages as you move from one value to the next higher value. Once the highest value is reached, it should sum to 100%. This is useful for knowing what percentage falls below a particular value.

Now it is your job to interpret this output!!

6. The output only provides numerical output. In order to make sense of the information, you will need to reference back to the variable information to see what labels match with what numerical values.

7. You will note in your output that SPSS has provided you with all the requested statistics for each of your variables. But are they all meaningful or useful? NO! YOU MUST BE SMARTER THAN THE COMPUTER!!! Remember, each measure of central tendency only works with variables with certain levels of measurement. While SPSS does have a way to mark the level of measurement for variables, they may not always be labelled correctly and regardless of the label, SPSS will still run calculations on any numerical value.

a. WRKSTAT is a categorical/nominal variable. At this level of measurement, frequencies and percentages are meaningful as is the mode. All other calculated statistics for this variable in the output is meaningless. You cannot have a mean or median for a categorical/nominal level variable…but SPSS will give it to you if you ask for it because everything is represented by a number. Be careful!!

b. DEGREE is an ordinal level variable. At this level of measurement, frequencies, mode, and median are meaningful, but mean here doesn’t really work. You’ll occasionally see a mean calculated for some ordinal variables, but that is not a universal.

c. AGE is a ratio level variable. For this level of measurement, all statistics are possible and technically meaningful. However, it is likely that in an interval/ratio level variable, values will not repeat themselves very often and so frequencies and mode are not particularly useful statistics.

Creating Graphs

In the video, I showed you how to also get simple graphs with the descriptives but if you want to just go in and make standalone graphs, the process is pretty easy

8. To do graphs, at the top you select Graphs and then select Chart Builder. Click okay on the pop-up window.

9. At the bottom left, you’ll see “Choose from:”

a. Pick the type of graph you want. Pie/Polar or Bar are classics for nominal/ordinal data. If you’re using an interval/ratio variable, Histogram is great.

b. You may see in the box that there are many types of each graph. Click and drag the first one to the spot that tells you to drag it there.

10. Now find your variable in the box on the top left and drag into that blue, rectangle below the graph shape it’s showing.

a. Two important notes: for pie or bar charts, on the left you’ll see the word “Count” in a spot that has a downward facing triangle. Click that downward facing triangle and switch it to Percentage (?)

11. Now click okay

Final step: exporting your output so you can have the tables/graphs in a Word document so you can use them.

12. On the Output window (where you can see the tables and graphs), click File and then Export.

13. In the middle you’ll see a section that says File Name and on the far left a blue box that says Browse. Click Browse

14. In the pop up, at the top you’ll see “Look in:” and should change that to desktop

15. Give it whatever name you want. I recommend leaving the file type as a Word Document as any word processor can open it. Click Save. And then click Okay. You’ve now exported those tables and graphs into a usable file type to incorporate into papers and presentations.

Now it’s your turn to pick some variables. You can just scroll through the dataset or pick some by searching around in the GSS Data Explorer that’s linked on the module.

Once we have finished the steps with the sample variables, you need to pick different variables (one for each of the three levels of measurement) on your own and then do the following:

1. What variables did you choose? What is their level of measurement?

a.

2. For the nominal or ordinal variable(s):

a. A frequency table for each variable

b. The meaningful measure of central tendency (nominal: mode; ordinal: mode and median)

3. For the interval or ratio variable(s):

a. The summary table that has all measures of central tendency: mean, median, and mode

4. A useful graph for each item

a. So you will put three graphs here!