# Exercise 3.5

1. Suggest a hypothesis concerning the relationship between a woman’s education and the amount of time she spent in part-time employment (rather than full-time) when caring for children.
2. How would you expect the relationship between the two variables to differ for the UK and Denmark?

### Procedure:

• Create a subset of countries that includes the UK and Denmark
• Select only working (work_sta=1) women (gender = 0).
• Compute the Pearson correlation between years of schooling (eduyrs) and time spent in part-time employment when caring for children G123 (ptmhmcc - 'Total time part-time work rather than full-time caring for children'), separately for the UK and Denmark.

### Questions

1. Are the correlations statistically significant and are they negative or positive?
2. In which country is the correlation stronger?
3. Discuss these results in light of your hypothesis.
Nesstar
• Open the dataset Family, Gender and Work
• Click the ‘Analysis’ tab
• Click the icon for weighting, and select ‘dweight’
• Click the icon for subset, and subset the data to working women in the UK (gender = 0, work_sta = 1 and country = GB)
• Click the variables ‘Years of full-time education completed’ and ‘Total time part-time work rather than full-time caring for children’ and select ‘Add to correlation’
• Change the subset to working women in Denmark and run the analysis
SPSS syntax
*You need to have a copy of SPSS installed on your computer, and you should download and use the dataset Family, Gender and Work.
*Open SPSS by clicking on the appropriate link.
*Open the ESS data by clicking ‘File’, ‘Open’, and ‘Data’ on the SPSS menu bar before you select the folder and the data set.
*Open a new syntax window by clicking ‘File’, ‘New’, and ‘Syntax’ on the SPSS menu bar.
*You can copy the syntax below and paste it into the syntax window in SPSS.
*Execute the syntax using the 'Run' option on the menu bar.

*Commands must always end with a dot.

*The following command causes the cases to be weighted by the design weight variable 'dweight'.

WEIGHT BY dweight.

*Create filter variable - only include working women from the United Kingdom and Denmark.

USE ALL.
COMPUTE filter_\$=gender = 0 & work_sta = 1 & (cntry ='DK' | cntry ='GB').
VARIABLE LABEL filter_\$ 'gender = 0 & work_sta = 1 & (cntry =DK or cntry =GB) (FILTER)'.
VALUE LABELS filter_\$ 0 'Not Selected' 1 'Selected'.
FORMAT filter_\$ (f1.0).
FILTER BY filter_\$.
EXECUTE.

*SPLIT FILE splits the active dataset into subgroups that can be analysed separately.

SORT CASES BY cntry.
SPLIT FILE, LAYERED BY cntry.

CORRELATIONS
/VARIABLES=eduyrs ptmhmcc
/PRINT=TWOTAIL NOSIG
/MISSING=PAIRWISE.

*Turn off the split file and weight, and select all cases.

SPLIT FILE OFF.
WEIGHT OFF.
USE ALL.
Sample solution

### Problem

1. Hypothesis: A negative correlation is expected between women’s education and the length of time they spent in part-time (rather than full-time) employment when caring for children. The hypothesis derives from the fact that better educated women are likely to earn more than less educated women. Hence, working part-time is more ‘costly’ for them. Better educated women can more easily enter professional and other career-oriented fields where part-time employment is not always possible.
2. Hypothesis: We would expect the relationship discussed above to be stronger in the UK, a society best characterised by the market-dependence regime.