This is a copy of a message sent out under “Decimal places and column headers in MEANS output” but (pace David Marso) I think it warrants a new topic.
Jon
OLAP CUBES gets close to what I want. The FM doesn’t have any reference to having to use pivot tables, but the help menu does (if you scroll far enough down the case studies menu).
If I use ‘epsilon’ instead of ‘earndiff’ it would be almost perfect.
recode incr3 (3=100)(1,2=0)(else=sysmis) into earn12k.
OLAP CUBES earn12k BY edlevel
/CELLS=MEAN COUNT
/CREATE 'earndiff' gac (edlevel (1 3) )
/HIDESMALLCOUNTS COUNT=5
/TITLE='OLAP Cubes on earnings differences for men and women'.
OLAP Cubes on earnings differences for men and women |
edlevel Highest qualification level: Total |
| Mean | N |
earn12k | 32.0463 | 1554 |
OLAP Cubes on earnings differences for men and women |
edlevel Highest qualification level: earndiff |
| Mean | N |
earn12k | -38.9429 | -148 |
Pivot table after processing:
OLAP Cube |
| edlevel Highest qualification level | Mean | N |
earn12k | 1 A-level or above | 54.1463 | 615 |
2 O-level or CSE | 19.9153 | 472 |
3 None | 15.2034 | 467 |
earndiff | -38.9429 | -148 |
Total | 32.0463 | 1554 |
If I use:
OLAP CUBES earn12k BY edlevel
/CELLS=MEAN COUNT (edlevel)
/CREATE 'earndiff' gac (edlevel (1 3) (2 3) (1 2) )
/HIDESMALLCOUNTS COUNT=5
/TITLE='OLAP Cubes on earnings differences for men and women'.
The pivot table comes out:
OLAP Cubes on earnings differences for men and women |
| edlevel Highest qualification level | Mean | N |
earn12k | 1 A-level or above | 54.1463 | 615 |
2 O-level or CSE | 19.9153 | 472 |
3 None | 15.2034 | 467 |
earndiff | -38.9429 | -148 |
3 None - 2 O-level or CSE | -4.7118 | -5 |
2 O-level or CSE - 1 A-level or above | -34.2311 | -143 |
Total | 32.0463 | 1554 |
The label differences and the negative Ns could be confusing: also the Total row needs to be on top, but it’s close to a decent elaboration table with epsilons already calculated. For tutorial purposes I’ll have to go through the pivot table process step by step with screenshots. Still a couple of manual edits needed for column headers and to get decimals down to one place.
OLAP Cubes on earnings differences for men and women |
| sex | % | n=100% |
earn12k | Men | 48.7 | 874 |
Women | 10.5 | 686 |
epsilon | -38.2 | |
Total | 31.9 | 1560 |
Thanks again for pointing me in the right direction.
John
John F Hall (Mr)
[Retired academic survey researcher]
Email: [hidden email]
Website: www.surveyresearch.weebly.com
SPSS start page: www.surveyresearch.weebly.com/spss-without-tears.html
Locked
