How to deal with mistakes in the variable codes (between languages)

Question

Hello :)

I am already running my questionnaire and I realized I made a mistake. My questionnaire is in two languages: English and Italian. I found out that in the question about gender I have switched the responses codes, that means in the English version I have the following answer options: 1.female, 2.male, 3.other; and in Italian I have them listed like: 1.male, 2.female, 3.other, that means male and female are reversed in the two languages. Since in could be a problem to analyze the data with this code mistake, I wanted to ask if there is a solution to the problem (maybe if I can correct the Italian codes, if I have already collected data).

Thank you very much

Answer 1

Not a problem at all, you just have to write a simple recode in the Tool you are using. For example in STATA:

recode variable (1=2)(2=1) if Language == ITA

I dont know if you can do it directly in SoSci

Comments

I agree that solving this issue post-hoc in the analysis would be the way to go. Anything else involves a severe risk of messing the data up.

---

Do you have an idea about how I could do it in SPSS?

---

IF (language = "ita") newVar = 2 - oldVar.
IF (language = "eng") newVar = oldVar.
EXECUTE.

---

Thank you! Sorry, but I am a beginner in using SPSS, do I have to paste the code in the syntax? And do I need to create a new variable before? (as in the code there is a newVar and an oldVar). Sorry, but I don't get if I have to take the code exactly how it is or if I should change/do something before

---

The IF works like a COMPUTE. It will create the new variable eventually.

You may also take a look into the SPSS reference manual, looking for the LOOP command. It allows you to do the same command for a whole set of variables (of one question).

---

That might be also personal preference, but if you are new to data cleaning/data transformation, you might want to start with STATA (if your organization pays for it) or R (which is for free). For both of those tools you find way more help and information online compared to SPSS. Also R might be the better tool to get to know in the long run.

Note: That is probably subjective and has nothing to do with SoSci. Just some input from someone who has to deal with those tools on a daily basis :)

---

Well, it's a good point that the material available about SPSS online is scarce. Howwever, getting started with R (which is an awesome tool) is quite a challenge. I would not advice to try that on your own during the analysis phase of the thesis ;)

Just took me two hours today to remove a secondary class from an R variable in order to make the tidy rbind via bind_rows() work ;)