- Welcome
- Important notice
- About the author
- About this site
- Site guide + Search box
- Dedications
- Acknowledgments
- My personal pantheon (of the great and the good in survey research)
- Recent and planned activities
- Textbooks for Research Methods and Data Analysis
- 1: Survey Analysis Workshop (SPSS)
- 1a: Statistical concepts and methods
- 1b: Teaching with Survey Data
- 1c: Developing research projects using survey data
- 1d: Workshop and presentations for ASSESS (SPSS users in Europe)
- 2: Survey Research Practice
- 2a: Survey Research Methodology, Practice and Training
- 2b: Major survey series
- 3: Subjective Social Indicators (Quality of Life)
- 4: Survey Unit, Social Science Research Council (UK)
- 5a: Polytechnic of North London (1976-1992)
- 5b: Survey Research Unit (1978-1992)
- Village life in Normandy
- Contact
- Origins of the British Crime Survey
- British Crime Survey
3.2 Three (or more) variables
[Page last updated 20 May 2019]
Introducing a third variable. Controlling for test variables. Elaboration.
Logical model is X → Y . T (the effect of X on Y controlling for T) where:
Y = Dependent variable
X = Independent variable
T = Test variable(s)
Introducing a third variable. Controlling for test variables. Elaboration.
Logical model is X → Y . T (the effect of X on Y controlling for T) where:
Y = Dependent variable
X = Independent variable
T = Test variable(s)
3.2.1 Elaboration
For a brief statistical explanation see: 3.2.1 Elaboration (extract from Statistical Notes for the course)
3.2.1.1 Earnings differences – Elaboration [Revised 2 July 2021]
Using data from the 1989 British Social Attitudes Survey this detailed work-through produces zero- and 1st- order contingency tables to see what happens to differences in earnings from paid work (dependent variable) between men and women (independent variable) when controlling for selected test variables .
It uses SPSS command: CROSSTABS <depvar> BY <indvar> BY <testvar>
for some tables but also demonstrates how CTABLES produces neater and cleaner output. Epsilon (percentage point difference) is calculated at each level.
New tutorials replicate this elaboration exercise on data from the British Social Attitudes Survey 2009-2014
Data safeguarding requirements mean that no actual data can be uploaded to this site from the British Social Attitudes Survey (BSAS) . There is a list of SPSS files available from UKDS on BSAS filenames. Using this file, click on the survey you want, click on Use Data, click on Access data Alternatively, check the series list of available files and click the Access link to request downloads of individual source files direct from UKDS. See: Downloading British Social Attitudes Survey (BSAS) data from the UK Data Service
As well as the source files for each survey year, the list also includes two cumulative files and a panel study:
British Social Attitudes Survey, 1983-1989
British Social Attitudes Survey, 1983-1991; Cumulative File
British Social Attitudes Survey Panel Study, 1983-1986
The metadata in the original source files are not mutually compatible: they need to be converted before two or more files can be combined. British Social Attitudes: Making files from different years compatible contains direct hyperlinks links to the source files at UKDS and an explanation of how to convert and combine them.
New General Data Protection Regulations (GDPR) mean that hyperlinks to SPSS *.sav files in the following tutorials will cause type 404 errors. The links have been disabled and all three tutorials are currently being rewritten to point users to the correct place.
They include hyperlinks to an earlier version of a cumulative SPSS "mother" *.sav file I created for all waves of the British Social Attitudes 1983-2014 which, because of GDPR issues, I have removed from my site. Users must now download the data for each survey direct from UKDS, but major differences in metadata mean that the files are incompatible. (See: British Social Attitudes 1983 onwards: Cumulative SPSS file 1983 - 2017)
To get round the problem, Jon Peck (retired Senior Software Engineer, IBM-SPSS) made a brilliant suggestion: create a special version of the mother file containing no cases. Users can now use bsa1983to2017zerocases.sav to create mutually compatible files using the SPSS command: APPLY DICTIONARY. Any combination of two or more converted files can then be generated using SPSS command ADD FILES.
See: British Social Attitudes: Making files from different years compatible and
British Social Attitudes 1983 onwards : Cumulative SPSS file
The following tutorials are still accessible, but are currently being re-written without links to data files.
3.2.1.1 Earnings differences – Elaboration [revised 2 July 2021]
3.2.1.2 Elaboration 2 (Income differences BSA 2009 - 2014)
Selection and extraction of variables from mother file; selecting years 2009 to 2014 only; tidy up metadata; derived variables; check contents of file; frequency counts.
3.2.1.3 Elaboration 3 (Income differences 2009 – 2014_CROSSTABS)
Zero order 2-way contingency tables
3.2.1.4 Elaboration 4 (Income differences 2009 – 2014_CTABLES)
1st, 2nd etc order (multi-way) contingency tables
New tutorials using BSAS 2009 [May 2019].
Replication of 3.2.1.1 Earnings differences – Elaboration
Three new modules:
Stage 1: 3.2.1.5 Earnings differences 2009: Download and check file
Stage 2: 3.2.1.6_[earnings_differences_2009]_extracting_and_saving_selected_variables
Stage 3: 3.2.1.7 Earnings differences 2009: Elaboration
To see if earnings differences have changed over time, further tutorials will replicate exercises using data from all waves 1983 to 2017..
3.2.1.8 Earnings differences 2017: Preparing the data