Present and briefly report frequency tables (no need for bar charts) for your variables and the confidence interval you calculated for the dependent variable
An Analysis of the British Social Attitudes Survey (2018) For this assignment you will select ONE of the attitudinal variables in the dataset as your dependent (outcome) variable together with a small selection of explanatory variables to carry out an analysis exploring variation in your chosen dependent variable among the British population in 2018
- In addition to these instructions please also read the document ‘about the bsa2018_iqm dataset’ also available on the Blackboard site under ‘Assessment’
- Your analysis and the report must follow the guidelines provided below. Pay careful attention to cover all the requirements specified for the SPSS analysis and for the structure and content of the report.
- All the techniques needed for the data analysis for this assignment were covered in the SPSS practical exercises using the BSA2018– please consult these where required for reminders.
Detailed Instructions Step 1. Define the analysis
- Take time to study the variables in the dataset – The document ‘About the BSA2018_IQM dataset’
includes a full list of these. Note the first variables in the list are characteristics about the respondent with the main section of attitudinal variables (from which you pick your dependent variable) starting from number 55 (view on the monarchy). Select any ONE of these attitudinal measures of interest to use as the dependent variable in your analysis.
Note: While the variable labels give a sort summary of what each variable measures you are advised you may need to go back to the original questionnaire to see exactly what was asked
· Your research objective is to carry out an analysis of your chosen dependent variable using TWO main explanatory variables of your choice (and ONE control variable) to investigate variation in your dependent variable. You may choose your explanatory variables from any in the dataset.
- It is helpful to specify this as a clear research question or research objective which will guide your analysis. This might be defined in fairly broad terms as befitting an exploratory analysis e.g if
interested in public attitudes to capital punishment your objective could be ‘an investigation of factors associated with support for capital punishment’ but you might adopt a more focused question perhaps designed to test some theory (and any related hypothesis) from the literature e.g. ‘that support for capital punishment varies by education level’. You might have more than one hypothesis.
- For this assignment you should limit yourself to just TWO main explanatory variables (one additional variable may be used as a control variable – see stage 4 of the analysis below). In selecting your explanatory variables for the analysis it is very important that you provide a clear reasoning for their choice (with some reference to relevant literature).
Note that in some cases recoded versions of original variables are provided to make them easier to use in crosstabs e.g. the dataset includes a recoded version of the original variables for age and education but you may find it helpful to carry out your own recodes.
Step 2. Carry out the Analysis
N.B. Remember for all your analysis of the 2018 BSA make sure you have applied the weighting variable (‘WtFactor’) to correct for sampling selection bias and non–response (see the SPSS practical 3 for a reminder of how to do this)
- Use appropriate methods to describe the univariate distributions of your dependent and explanatory variables separately.
- For your dependent variable calculate a 95% confidence interval around one of the percentage estimates (e.g. ‘% in favour of capital punishment’) using the method covered in the course (week 4).
- Explore the relationship between the dependent variable and each of your two explanatory variables separately with crosstabs (carrying out any variable recoding as appropriate*) and include a Chi– square and Cramers V test of association for each table.
- Finally generate ONE three–way table where you take one of the crosstabs already produced (in stage 3) and introduce a control variable (again include a Chi–square and Cramers V test). For your control variable you may choose either a variable already used in stage 3 or a new variable from the dataset (you are advised to ensure the variable you use as a control has only a few categories to keep your 3–way table to a manageable size).
* Remember that using variables with many answer categories in crosstabs may result in tables that are hard to read and interpret. You may also find that sample sizes get very small for some of your categories (note it is the number of cases on which percentages are based that is important here).
Percentages based on small totals of less than 50 cases should be treated with caution). Some recoding to group categories together can help with this issue (a technique covered in the workshops). As mentioned some of the variables in this dataset have already been recoded in this way but you might need or wish to do further recoding.
Step 3. Write up the analysis in a report
You should write up Part 1 of the assignment using the following report structure (please note the mark scheme and tips on reporting provided at the end of the document)
- Introduce the research question/research objective (and any hypotheses) that you will be investigating with the secondary analysis with some reference to relevant literature
- You should also introduce and briefly give key details about the BSA dataset you will be using
- Explain the selection of variables used in the analysis (and describe any use of recoding) making reference to relevant literature to support your choices.
- Present and briefly report frequency tables (no need for bar charts) for your variables and the confidence interval you calculated for the dependent variable