Authors
Clive Seale

Pub Date: December 2011
Pages: 648

Click here for more information.
Clive Seale
20 Statistical reasoning: causal arguments and multivariate analysis
Clive Seale

1. Choose any article or book that reports the results of a qualitative research study. What sort of causal propositions does the author assume to be true? What sort of causal arguments are contained in the text? How could these be tested in quantitative data analysis? What would the independent and dependent variables be?

2. Table 20.3 in the book and the discussion that accompanies it suggests that hospice care may cause people to want euthanasia. What plausible objections might there be to this causal argument? How could they be tested in further research? How could qualitative research be used to investigate this proposition?

3. Take Table 20.4 to be a zero-order table. Draw hypothetical conditional or first-order tables that you might expect to find by entering the test variable of income, measured as low or high. The pairs of tables should, in turn, illustrate:

    (a) the existence of a spurious or intervening relationship between social class and home ownership;
    (b) replication of the original relationship;
    (c) specification of the original relationship;
    (d) suppression of a stronger relationship.

Table 20.4. Table showing relationships between social class and home ownership (column %)

                                              Social class                           

Home ownership

Lower

Middle

Upper

Owner

20

30

50

Private, rented

30

40

30

Council, rented

50

30

20

                             p <0.01, Q = 0.6                                     

4. Use SPSS to analyse a statistical data set. Chapter 17 has web links to data archives, from which such data sets can be downloaded. Use data transformations (e.g., Compute or Recode) as well as univariate, bivariate and multivariate statistical procedures to investigate the data. Try to pursue an argument, hypothesis or research question in your analysis, considering counter-arguments to yours and dealing with these, where possible, with further data analysis.