Hi

I'd to adjust my dataset by Education using SPSS Rake Weights.

Here's the unweighted data for education level:

Education level Freq. Percentage

Primary 0 0%

Lower-Sec. 50 21%

Higher-Sec. 83 34%

Tertiary 110 45%

The population data for education level is:

Education level Freq. Percentage

Primary 2000 8%

Lower-Sec. 6200 26%

Higher-Sec. 7000 29%

Tertiary 9100 37%

Can someone tell me how to fill in the "Categories and Control Totals or Fractions" under Rake Weights to Control Totals for my case? It's confused because there are no respondents who belong to Primary level in my dataset.

Many thanks.

James

If your dataset has no cases with education = Primary, then no case weighting scheme can produce a positive percentage, and if you were to decide to add, say, one case and assign it to primary, the resulting weight on that case would be quite large and would distort your results.

You might want to just eliminate the primary category. You don't need to renormalize the remaining percentages to sum to 100%, since RAKE will do that automatically. Just leave that value out of the education list categories.

Hi Jon,

Thank you for your prompt reply!

Based on your suggestion, if 1=Primary, 2=Lower-Sec., 3=Higher-Sec., 4=Tertiary, then what I need to fill in the Categories and Control Totals or Fractions for Education category in the RAKE window would be 2 .26 3 .29 4 .37. Am I right?

James

That's right. Cases in categories that don't have a specification get a weight of sysmis, and they will be automatically excluded from procedures when weighting is on. RAKE automatically turns on the weight when it is run.