I am working on an important (to me) academic project and need to be sure of three things before I invest my time in learning how to use a survey program. I apologize if I could have found the answers in the documentation. I did look, but I am not very technical and so don't know professional jargon well enough to be sure I have an answer.

1. Do academics use LimeSurvey for their research projects?

2. I have to give participants two surveys and I anticipate that many more will take the first than the second, but I can only use surveys by people who completed both. I must be able to filter out the first surveys that are not accompanied by a second. Can I do this?

What would happen if a participant took the surveys on different machines or if they cleaned their comp of cookies and so forth, or started using an IP blocker? I don't anticipate having so many completed second surveys that I could not go through the first list manually and selecting the ones that have a corresponding second survey.

3. Can I finely sort the data? For example, can I sort out those who gave response "d" on question 3 to see what their response was to question 7?

Thank you so much for your help.

Has this has ended up in the wrong forum? If it is, I'm sorry.

Hi,

1. Do academics use LimeSurvey for their research projects?

www.limesurvey.org/forum/can-i-do-this-w...are-using-limesurvey

I must be able to filter out the first surveys that are not accompanied by a second

There are many ways to do that:
1. If you use tokens ( www.limesurvey.org/manual/Survey_participants ) just match the results by that.
2a. You can use end-url to start the second questionnaire if the first is completed.
2b. You can provide the link to the second survey only if the first survey is completed. (e.g. in the "Thank you" message)
3. Export the tokens of the completed first surveys and only invite these to answer the second.

3. Can I finely sort the data?

Well, this is a matter of analysing the data. You can do this in each analysing tool like SPSS, PSPP, EXCEL, or whatever.
LimeSurvey is a data collecting tool.

that I could not go through the first list manually and selecting the ones that have a corresponding second survey.

There are easy to use formulas in EXCEL, SPSS has a function to merge files according to a key, etc.

Has this has ended up in the wrong forum?

Why did you think that?
Because you didn't get an answer for 13 minutes????
Remember: This is the forum of the Community Edition of LimeSurvey. We all are volunteers and do this in our spare time.

Best regards
Joffm

1) Yes.

2) If you use tokens (same for both surveys), you can match the two data sets with the token value. Otherwise, you will need some sort of identifier. You could present the URL for the second survey at the end of the first, passing the token value or custom identifier in that URL.

- www.limesurvey.org/manual/URL_fields/en#...using_GET_parameters

3) LimeSurvey provides some basic statistics with filtering but I'm sure you will be analyzing the data with SPSS, R or another analysis tool.

Thank you so much for your help. I meant I thought I had ended up on the wrong sub-forum. This is because I saw the sub-forum was about installation issues. I didn't know if there was a separate one for people who had yet to install it at all.

Thank you again.

Hi,

I thought I had ended up on the wrong sub-forum

then you are nearly the first one who really cares for the correct sub-forum

Joffm

Tpartner and Joffm gave you all the right answers already.

Regarding the subforum, this would probably have been the best match for your question about "Can I do this with Limesurvey?": www.limesurvey.org/forum/can-i-do-this-with-limesurvey

Hi,


I'm happy to answer your questions:


> 1. Do academics use LimeSurvey for their research projects?

Yes, they do. There are many universities who do research with LimeSurvey. It is also taught in academic summer schools (e.g., next week the Gesis Summer School in Survey Methodology takes place). Here is an (unfortunately somewhat outdated) list:
www.limesurvey.org/about-us/news/1109-as...g-limesurvey-en-gb-4


> 2. I have to give participants two surveys and I anticipate that many more will take the first than the second, but I can only use surveys by people who completed both. I must be able to filter out the first surveys that are not accompanied by a second. Can I do this?

So what's your exact design? How do you recruit participants, do you contact them by email?


> What would happen if a participant took the surveys on different machines or if they cleaned their comp of cookies and so forth, or started using an IP blocker? I don't anticipate having so many completed second surveys that I could not go through the first list manually and selecting the ones that have a corresponding second survey.

If you assign a unique ID to each respondent, it does not matter if they use another browser or machine and you can match cases easily afterwards.


> 3. Can I finely sort the data? For example, can I sort out those who gave response "d" on question 3 to see what their response was to question 7?

You can do some stats within limesurvey but filtering for cases is still quite basic. I'd suggest to export data (e.g., for SPSS/PSPP, R, Stata, or Excel) for more options.

Thanks again for the help you all are giving me. I have launched into reading the manual. In answer to f_funke's question:

>So what's your exact design? How do you recruit participants, do you contact them by email?

I will be recruiting participants through social media. Most of them will not be contacted by email. I would like to make the survey as anonymous as possible to protect the integrity of the survey responses and not have participants fear that I will judge them based on what they say.

They will come to the survey via a link from a website. Once they have completed the survey, they will be returned to the website and can read the materials I have there (I have figured out how to do this). Then, when they are done, they need to take a second survey.

In order to preserve anonymity but still match up the surveys, would like them to register with a username, but no email when they take the first survey which they will then use to unlock the second. However, I don't know if that is possible.

Can I assign a unique ID to each correspondent without getting their email address?

I have a new idea. Could I have people log into my website with only a username, then I give them a token code there which they will use for both surveys? That way it will always be available for them alone on my website so they won't need to keep track of it (I anticipate several weeks between them taking the two surveys). Can I do this using the Dummy Participants feature?

So, your website manages he tokens of the different users?
If the user logs in with his username, he will get his token?
Of course you can do that.
But, isn' it possible that users only answer the second survey?

But that's only how you distribute the tokens and does not affect LimeSurvey.

And Dummy Participants? Why?
I think you will need a relation between you user name and the tokens.

I would do it like this:
Just upload all usernames with dummy emails into the participants table of the first survey.
Generate tokens.
Export this participants table and upload into the second survey.
This exported csv-file you can use in your website to display the tokens.

And a little bit more sophisticated:
Query the LS database (table "lime_tokens_[SurveyID]") to see which users completed the first survey
Either display the link to the first or to the second survey.

Joffm

The tricky part is the recruitment. It makes it difficult to use tokens, which would have been your perfect match.

Now: Why do they have to go to your website to read the information? Couldn't you include this into the survey? Basically have one big surve, separated into 3 parts:

1. Questions from "first survey"
2. Information (as text display questions)
3. Question from "second survey"

Or do you think this would be too long? But I think even a long survey will have a higher finishing rate than 2 surveys with a website visit in between. You will most probably lose quite a few respondents with this "break".

And keep in mind: as you recruit via social media, your relation with these respondents is probably very distant, which also factors into the response rate. If you would have a good, long standing relation with those people, they tend to try harder to finish, even longer surveys. Now if your survey isn't especially interesting to them and there is no real connection between you (or the sponsor of the survey) and the respondents, chances are high that you have a huge drop out rate.

And be prepared for a very low starting rate. Surveys sent out to people tent to have starting rates far below 10%. I have quite some experience with recruitment on automotive websites for brand surveys and only the most popular/cool brands came even close to 5% of starting rate (this means you contact 100 people on the website and invite them to participate and only 5 of those make it even to the first page of the questionnaire. And all those people had a certain connection to the brands, otherwise they wouldn't have visited their website. Brands that don't have a lot of emotional envolvement have even lower starting rates.

And after that you will still have quite a few dropouts, depending of the length and style of your survey.

So be prepared to invite a lot of people.

Of course, this depends a lot on your survey. If the topic of the survey is very interesting for the people you invite you might see far higher rates.