Attention: Jeff Bloem/Michelle Pratt: Follow up on my earlier question.

INTERNATIONAL
1

Dear Jeff Bloem/Michelle Pratt:

Thanks very much for your quick response to my question regarding linking husbands and wives in the Liberian 1974 and 2008 Liberian IPUMS Census samples for the purpose of studying intra-/inter-ethnic marriages. I am somewhat a bit more familiar with the IPUMS’ :Attach characteristics method" than the SPLOC method that you refer to in your response; although I would eventually like to learn both methods. The SPRULE – for linking husbands and wives/spouses sounds even more direct; although I have not as yet experimented with that variable.

I have few more questions about the “attach characteristics” method for linking spouses. When I get into attaching characteristics, should I check spouse? Should I also check sex, [spouse’sex]? since a spouse can be of either sex? I am ambivalent/unsure about checking these two variables.

I also have few other questions: In my extra #1 for 1974, I have only 39 variables; are there more variables than those? I also have had some problems with the variables Year and EthnicLr. Using SPSS, I recoded the variable Year into two categories: 1 for 1974 and 2 for 2008. However, when I crosstabulate ETHNICLR by ETHNICLR_SP, controlling for the recoded variable Year, only year 2008 appears. Does the variable ETHNICLR exists in both the 1974 and 2008 Liberian samples with the same name? Every indication seems to suggest that, since an X is under both years for this variable.

When I renew my Winzip for decompressing zipped files, and resubmit my Extract # 2, I

will try my hand at linking husbands and wives using SPSS, and let you look at my result/output. I tried to download the free Wingzip, but, I got a message that it does not work on my Dell computer. I am told to contact the publisher of Wingzip.

Finally, is there a direct e-mial address through which I can communicate with you and/or Michelle Pratt? I had a bit of difficulty using the forms containing your answers, which would be the ideal way, by just clicking Reply. But, if it is not possible to communicate directly with you via a direct e-mail, then I will figure out how to do so using the “Question and Answer” form. Thanks so much for your help so far. I greatly appreciate it.

Konia T. Kollehlon

konia1@verizon.net

2

I’ll try to address each of your questions one at a time.

Regarding the attach characteristics tool: Yes, you can certainly include the spouses sex in the attach characteristics tool, if you think this will be important for your research. As you suggest this will identify the sex of the spouse.

Regarding the number of variables: In IPUMS International, you can select the specific variables you want to include from each sample. Specifically, there are slightly more than 39 variables available for the Liberia 1974 sample.

Regarding the YEAR and ETHNICLR variables: The variable ETHNICLR variable is available for both 2008 and 1974 Liberian samples. For future reference, information about the availability of variables can be found on the availability tab. The issue is the family interrelationship variables (which the attach characteristics tool relies on to construct the “_sp” variables) are only available for 2008. This is why you only see observations from 2008 in your cross-tabulation.

Regarding decompression software: We tend to recommend 7zip. It seems to work best on most Windows computers.

Regarding directly contacting IPUMS User Support: You can always email ipums@umn.edu to directly contact one of us.