SSCI202 Workshop 3: Exploring the 2012 AuSSA Data

2012 AuSSA data

The Workshop 3 introduces the 2012 Australian Survey of Social Attitudes (AuSSA). AuSSA is a longitudinal survey about social attitudes, beliefs and opinions of Australians. It is a biennial survey that began in 2003. It has been the main source of data for studies about how Australians think and feel about their lives as well as about how they have changed over time. AuSSA is also the Australian component of the International Social Survey Programme (ISSP). The ISSP is a cross-national collaboration on surveys covering important topics for social science research (If you want to know more about it, visit here). In addition, the ISSP chooses a special topic each year, which is called as module, and repeats that topic from time to time. The module topic for 2012 was Family and Changing Gender Roles. This is the most recent ISSP data on this topic, which was surveyed four times in 1988, 1994, 2002, and 2012.
In this workshop, we will use this dataset. The dataset is extracted from the 2012 ISSP data but includes only Australian respondents. This 2012 AuSSA is one of the datasets that you will use for the final survey report. If you want to know more about the 2012 AuSSA, visit https://www.acspri.org.au/aussa/2012.

How to open the 2012 AuSSA in SPSS

The dataset is not a public dataset. I get the permission of using this dataset just for educational purposes. Therefore, use this dataset only for this unit.

Go to the course iLearn page and find The 2012 AuSSA SPSS Data File under the section of Datasets. Download this file. The downloaded data file should be aussa2012.sav. Also, download and look at A Simple Codebook of the 2012 AuSSA and A Detailed Codebook of the 2012 AuSSA which all provide useful information on variables and their values.

<Figure 1>

Figure 1: <Figure 1>

Open SPSS and follow the below steps.

  1. Go to File > Open > Data.
<Figure 2>

Figure 2: <Figure 2>

  1. Go to the folder in which the downloaded file is. Choose the downloaded file (aussa2012.sav). Click Open.

Note: If you are using SPSS via AppStream, you need to upload your SPSS data file from your local computer. Follow the instruction of How to upload your SPSS work files to AppStream.


<Figure 3>

Figure 3: <Figure 3>

Exploring the 2012 AuSSA

Click Variable View at the bottom. You will see the information of all the variables. Note that the level of measurement for all variables are not correctly specified: they are all Scales. Thus, you NEED TO ASSIGN an appropriate level of measurement for any variables you are analysing before you start an analysis.

<Figure 4>

Figure 4: <Figure 4>

For example, suppose that we are analysing a variable, fepresch. To obtain detailed information on this variable, open “A Detailed Codebook of the 2012 AuSSA” using Adobe Acrobat Reader. Click the icon of Bookmark in the left pane.

<Figure 5>

Figure 5: <Figure 5>

You will see the variable list in the left pane. Click the variable you want to study. In this example the variable is fepresch. Then, the pdf will show the page which explains the variable of your choice.

<Figure 6>

Figure 6: <Figure 6>

The codebook shows the questionnaire and response options, which helps you to figure out the level of measurement for variables.

<Figure 7>

Figure 7: <Figure 7>

<Figure 7> shows information on fepresch. It asks respondents the extent to which they agree or disagree with the statement that a preschool child is likely to suffer if his or her mother works. Although you see nine response options, we will take into account only five options (Strongly agree; Agree; Neither agree nor disagree; Disagree; Strongly disagree) as valid options. The other two options (Can’t choose; No answer) will be counted as system missing values, which will be explained in the next section. Because the five response options can be rank-ordered but do not have the same distance between them, fepresch is an ordinal variable. Change the level of measurement for fepresch. If you don’t remember how to do that, see Entering political orientation variable (ordinal variable).

I recommend you spend some time in looking around variables in the 2012 AuSSA. The dataset includes so many variables relating to gender and family issues. You may find many variables that inspire your intellectual curiosity.

Missing values

Missing or invalid values are generally too common to ignore. Survey respondents may refuse to answer certain questions, may not know the answer, or may answer in an unexpected format. If you do not filter or identify these incorrect responses, your analysis may not provide accurate results.

For numeric data, empty data fields or fields containing invalid entries are converted to system-missing, which is identifiable by a single period(.). But more common practice is that some numerical values are treated as missing values. For example, in <Figure 7>, 8 denotes Can’t choose, and 9 represents No answer. These responses are typical types of missing values. In Variable View, click a small square in the column of Missing for fepresch (see <Figure 8>). You will see that 8 and 9 are treated as missing values in the popped-up window (see <Figure 9>). To go back to the previous Variable View tab, click OK in the popped-up window.

<Figure 8>

Figure 8: <Figure 8>

<Figure 9>

Figure 9: <Figure 9>

The reason why we treat some values as missing may be important to your analysis. For example, you may find it useful to distinguish between those who refused to answer and those who didn’t answer since it was not applicable to them.

When analysing variables, SPSS always take into account of missing values. For example, look at the frequency table of fepresch in <Figure 10>. You will see the frequency of two missing categories (8 = Can’t choose; 9 = No answer) in the bottom. Note that the Percent in the table is based on all responses including missing categories, while the Valid Percent is based on only valid responses which exclude missing values. Your next question might be which percent (more broadly, which statistics) you should report. The norm is to report the valid percent unless the non-response rates (the percent of missing values) are too much high. This is because a high level of non-response rate could provide an inaccurate estimation. In <Figure 10>, the non-response rate is 3.5%, which is very small and negligible.

<Figure 10>

Figure 10: <Figure 10>

Workshop Activity 3.1: Exploring variables in the 2012 AuSSA

  1. Download A Detailed Codebook of the 2012 AuSSA from iLearn and explore variables in this dataset. Find three variables that may interest you. Read the descriptions of these three variables in the codebook and decide appropriate levels of measurement for them. In Variable View, assign appropriate levels of measurement for the three variables. If you are not sure how to assign levels of measurements, see How to enter data. Then, make the frequency tables of your three variables.

Note: External students should post their answers to these four questions on the iLearn. This activity will contribute to your workshop participation marks.



NOTE: After you get the outcome of statistical analysis in the Output window, DO NOT close the Output window to go back to Data Editor. Instead, use the icon of Switch windows on the top menu of AppStream. For more details, see Switch windows. If you follow this, all your statistical outcomes will be stored in the Output window. Then, you will be able to export them to other documents. The last section of this workshop will introduce how to export your statistical outcomes in a web report format.


Descriptive Statistics

This section introduces three SPSS commands for computing descriptive statistics: Frequencies, Descriptives and Explore. fepresch (ordinal variable) and age (continuous variable) are used for this task. Please assign correct levels of measurements for these two variables first.

Computing descriptive statistics using Frequencies

Go to Analyze > Descriptive Statistics > Frequencies. In the window of Frequencies, select variables for which you want to compute descriptive statistics (in this example, they are fepresch and age) and move them into the pane of Variable(s). Then, click Statistics.


Note: If you see variable labels instead of variable names as in <Figure 11>, right-click at the left variable pane. Choose Display Variable Names. Variable labels will be changed into variable names. Also, choose Sort Alphabetically. Then, variables will be listed in an alphabetical order. For more details, see the second step of Making a frequency table.


<Figure 11>

Figure 11: <Figure 11>

In the window of Frequencies: Statistics, 1) tick Mean, Median and Mode in the section of Central Tendency, 2) tick Std. deviation, Variance and Range in the section of Dispersion, and 3) tick Skewness in the section of Distribution. Click Continue at the bottom. You will be back to the previous window. Click OK at the bottom.

<Figure 12>

Figure 12: <Figure 12>

In the output window, you will see the descriptive statistics of both fepresch and age (see <Figure 13>). In the week 4 lecture, I explained how to read SPSS descriptive statistics tables. If you have troubles in reading this table, review the week 4 lecture slides again. I want to ask you which measure of central tendency and variability is most relevant for fepresch and age.

<Figure 13>

Figure 13: <Figure 13>

Computing descriptive statistics using Descriptive

Go to Analyze > Descriptive Statistics > Descriptive. In the window of Descriptives, select variables for which you want to compute descriptive statistics (in this example, they are fepresch and age) and move them into the pane of Variable(s). Click Options.

<Figure 14>

Figure 14: <Figure 14>

In the window of Descriptives: Options, 1) tick Mean at the top, 2) tick Std. deviation, Variance and Range in the section of Dispersion and 3) tick Skewness in the section of Distribution. Click Continue at the bottom. You will be back to the previous window. Click OK at the bottom.

<Figure 15>

Figure 15: <Figure 15>

In the output window, you will see the descriptive statistics of both fepresch and age (see <Figure 16>). As you notice, the output does not show the median of variables. This is why I always like using Frequency over Descriptive for getting descriptive statistics.

<Figure 16>

Figure 16: <Figure 16>

Computing descriptive statistics using Explore

Go to Analyze > Descriptive Statistics > Explore. In the window of Explore, select variables for which you want to compute descriptive statistics (in this example, they are fepresch and age) and move them into the pane of Dependent List. Click Plots.

<Figure 17>

Figure 17: <Figure 17>

In the window of Explore: Plots, tick only Histogram in the section of Descriptive. Click Continue. You will be back to the previous window. Click OK at the bottom.

<Figure 18>

Figure 18: <Figure 18>

In the output window, you will see the descriptive statistics of both fepresch and age (see <Figure 19>). The output shows almost all descriptive statistics. Moreover, the visualisation of variables (e.g., histograms and box plots) is also shown as well. Therefore, Explore command is a very useful way to see the distribution of variables.

<Figure 19>

Figure 19: <Figure 19>

Workshop Activity 3.2: Descriptive statistics

  1. Compute measures of central tendency and variability for respondents’ years of schooling (educyrs). 1) Based on your SPSS output, is the distribution of educyrs symmetrical or skewed? 2) Which measures of central tendency and variability are most appropriate to summarise the distribution of educyrs? And explain why.

Note: External students should post their answers to these four questions on the iLearn. This activity will contribute to your workshop participation marks.


Comparing descriptive statistics between groups using Explore

Another advantage of using Explore is that it allows you to compare descriptive statistics between groups. Suppose that we want to compare the age of respondents between men and women. Thus, we need to compute the measures of central tendency and variablity for men and women respectively so that we can compare these measures between them.

Go to Analyze > Descriptive Statistics > Explore. In the window of Explore, 1) select variables for which you want to compute descriptive statistics (in this example, it is age) and move it to the pane of Dependent List. 2) select a group variable (in this example, it is sex) by which descriptive statistics are compared and move it to the pane of Factor List. Click Plots. As you did in <Figure 3-18>, tick only Histogram in the window of Explore: Plots. Click Continue. You will be back to the previous window. Click OK at the bottom.

<Figure 20>

Figure 20: <Figure 20>

In the output window, you will see the descriptive statistics for both males and females. (see <Figure 3-21>). Also, you will see the box plot of age by gender, which is very helpful for comparing the distribution between men and women (see <Figure 22>). If you are not sure how to read box plots, review the week 4 lecture slides.

<Figure 21>

Figure 21: <Figure 21>

<Figure 22>

Figure 22: <Figure 22>

Workshop Activity 3.3: Comparing descriptive statistics

  1. Suppose that you are interested in investigating whether men and women spend similar hours doing household chores. Use the variable rhhwork and sex for this investigation. rhhwork is a variable showing the hours spent on household chores. sex is a variable of gender. Compare the central tendency of rhhwork between men and women. Based on the comparison, do you think men and women spend similar hours on household chores?

Note: External students should post their answers to these four questions on the iLearn. This activity will contribute to your workshop participation marks.


Creating a web report of your analysis

Now, I would like to introduce how to save all your output in the same document. It is a convenient way to record your analytic work. I assume that you didn’t close the SPSS Output window so far. Thus, your output window should include all the outcomes of the workshop 3.

In the window of Output, go to File > Save As….

<Figure 23>

Figure 23: <Figure 23>

In the window of Save Output As, 1) choose the folder of Temporary Files in the section of Look in, 2) type file names, 3) choose SPSS Web Report (.htm) in Save as type, and 4) click Save. Your file will be saved in the folder of Temporary Files.

<Figure 24>

Figure 24: <Figure 24>

Next, you need to download this file from AppStream. Click the icon of My Files in the AppStream navigation bar. Go to Temporary Files. You will see the saved file (in this example, it is Workshop3.htm). Click the name of this file. It will be downloaded.

<Figure 25>

Figure 25: <Figure 25>

Open the downloaded file in your local computer. You will see your saved outputs in your web browser (See <Figure 26>. Use the navigation pane to find a specific output. Also, you can copy any outputs in this report and paste them to other documents.

<Figure 26>

Figure 26: <Figure 26>

Workshop Activity 3.4: Creating a web report

  1. Create a web report of your analysis so far. And show it to your tutor. External students should post their web report file in the External Forum.

Note: External students should post their answers to these four questions on the iLearn. This activity will contribute to your workshop participation marks.


Last updated on 25 August, 2019 by Dr Hang Young Lee(hangyoung.lee@mq.edu.au)