Is there a way to find which language the survey was answered in?

[zvig]

I have a form which I translated into another language. It is of value to me to know which language was used to fill the form. The interest is to compare or separate people who answered in different languages, so I can see if French respondent and British respondent, for example, can be pooled together into one data set, or should be analyzed separately.

I know that data download let’s me choose in which language to export the data (which is a very useful feature), but that does not tell me what language was used initially. I suppose I can make two copies, each one in a different language, and then combine the exported data, but that is an unnecessary hassle, that can also lead to human error (duplicate or missed cases, columns of data going to the wrong place, etc. - whatever humans can mess up with cut and paste.

I should mention that I hope to deploy a web form as a web survey - each response is from an different, anonymous, respondent, unknown to me. The language can suggest their country of origin.

Thanks,

Zvi.

[Kal_Lam]

@zvig, you could use this approach outlined by @wroos:

Can you tell the language that the data was collected in for multi-language forms?

Welcome @jturner, not directly, as far as I know. Also be aware that a user can change the language any time, even during the interview, so the same filled form can include multiple languages. Beside adding a question where the user enters her/his language used, e.g. as select_one LanguageUsed, there are workarounds: You can trace e.g. in a calculate variable the language-based choice_label of one select_one question, with a calculation: jr:choice-name(${ChoiceVar}, ‘${ChoiceVar}’) the local…

We also however have this feature request in the community:

Separation of the dataset by the submitted language

sometimes when we have multilingual forms, we have some choices different in the content than the others, I think it will be great to add some features to download the dataset as per the language submitted and then to do separate analysis for each submitted language.

You could VOTE for it to make it a reality!

[zvig]

Okay, thanks. I suppose I can add a language question "what language do you use to answer the form? and give choices for the translations languages. This, I think, is a less sophisticated version of @wroos proposes.

As for the suggestion - which I did vote for - it’s not exactly what I wanted.

I want a single dataset with a language variable, not several datasets each in a different language. Two (or more) datasets would actually make life more difficult - because in each set the variables names will be in a different language (at least in some export settings).

[wroos]

Hello @zvig,
I would like to mention again, that the same submitted form could have been done switching between the available languages. The user can change the (available) language at any place during working on a form.

A workaround to “see” the language used at a select question would be to calculate and store the
jr:choice-name.
Another option could be to clone the form and deploy/use different fixed (single) language versions/links.

@zvig

because in each set the variables names will be in a different language (at least in some export settings).

Hint: Variable and choice names are not language specific, only their labels. Furthermore you can chose the language (for the labels!) of an export.