rowmean help

[Vanessa Ojeda]

Hello,

Based on a scale from research, the researchers recoded their likert scale that measured microaggressions into a dichotomous (0=no, 1=yes) variable. There are 10 set of items that make up the construct "microaggressions". Question: with other sets of variables that create constructs, I usually create a rowmean and then run t-tests/anovas on that construct and demographic variables. However, would I create a rowmean if the set of variables are dichotomous?

Also, is there a user-friendly accessible book out there that outlines data cleaning, etc. in STATA that you would recommend? I've taken several stats classes, but when it comes to making sure I am checking for assumptions, errors, etc. it would be nice to have a checklist to go through in STATA to make sure I have my bases covered. Thank you!

[Clyde Schechter]

-egen, rowmean()- does not care whether the numbers it is averaging are a dichotomy. You can just do it in the usual way. And you can interpret the rowmean as the proportion of items that received a positive response in that observation. One thing to be careful about here is the problem of missing responses. Using -rowmean()-, in effect, treats the missing response as if it were equal to the average of the non-missing responses in the same observation. Sometimes that's appropriate (and then it is called ipsative mean imputation), and sometimes it's not. It depends on the meanings of the variables and how you plan to use the combined construct. There is also the question of whether the items cohere sufficiently into a construct to warrant averaging them, but, again, this has nothing to do with whether they are dichotomous.

Concerning your second question, there are several such books, and you can find most of them in the Bookstore at stata.com. While I have my own preferences, I think it would make more sense for you to look at the options available and pick one that you think will be most compatible with your style. These books differ on the content matter of the examples they use to demonstrate approaches--and you are probably going to relate better to one that features examples in your own field. Also some are more advanced than others. And there is just the matter of whether you find the author's writing style clear. So I would review the previews and see what looks best for you.

[Vanessa Ojeda]

Clyde Schechter I realize I never said "Thank You." I really appreciate your help and your thoughtful response to my questions!