Francisco, 

There's no single question, but the accepted best practice is to use something that is a) topic-relevant (i.e., related to the topic of the survey), and b) easy to answer (i.e., not too hard, and not too sensitive). The California Health Interview Survey, or first substantive question (after screener questions) is a question on general health that reads:

"These next questions are about your health.

Would you say that in general your health is excellent, very good, good, fair, or poor?"

EXCELLENT .............................................................1

VERY GOOD ............................................................2

GOOD .......................................................................3

FAIR ..........................................................................4

POOR........................................................................5

REFUSED ............................................................... -7

DON’T KNOW......................................................... -8

See more at http://healthpolicy.ucla.edu/chis/design/Pages/questionnairesEnglish.aspx 

There are other things to consider, too, though. 

First, the interviewer will have already developed some rapport in order to get the interview started, and what they do in those initial few exchanges before the survey officially "starts" is probably more important than what the first survey question is. They should not read scripted text, but be friendly and listen to the respondent's concerns about the survey and tailor their their responses. 

Second, there may be reasons that you have to ask other questions before the substantive questions in your survey. For example, you may want to collect household or personal demographic info to use in skip logic within the survey, or to use in weighting, and you want to get that info early (for the skip logic you need it for the interview, and for weighting, you want to get it early in case the respondent doesn't complete the interview. 

So, in conclusion, I would encourage you to assess the entire approach and recruitment process and interviewer training more than worrying about what specific question you ask first. By the time the respondent says "OK" to the interview and has started it, the worst of your worries are over. Getting to yes from the knock on the door is the hardest part. 

Francisco, 

There's no single question, but the accepted best practice is to use something that is a) topic-relevant (i.e., related to the topic of the survey), and b) easy to answer (i.e., not too hard, and not too sensitive). The California Health Interview Survey, or first substantive question (after screener questions) is a question on general health that reads:

"These next questions are about your health.

Would you say that in general your health is excellent, very good, good, fair, or poor?"

EXCELLENT .............................................................1

VERY GOOD ............................................................2

GOOD .......................................................................3

FAIR ..........................................................................4

POOR........................................................................5

REFUSED ............................................................... -7

DON’T KNOW......................................................... -8

See more at http://healthpolicy.ucla.edu/chis/design/Pages/questionnairesEnglish.aspx 

There are other things to consider, too, though. 

First, the interviewer will have already developed some rapport in order to get the interview started, and what they do in those initial few exchanges before the survey officially "starts" is probably more important than what the first survey question is. They should not read scripted text, but be friendly and listen to the respondent's concerns about the survey and tailor their their responses. 

Second, there may be reasons that you have to ask other questions before the substantive questions in your survey. For example, you may want to collect household or personal demographic info to use in skip logic within the survey, or to use in weighting, and you want to get that info early (for the skip logic you need it for the interview, and for weighting, you want to get it early in case the respondent doesn't complete the interview. 

So, in conclusion, I would encourage you to assess the entire approach and recruitment process and interviewer training more than worrying about what specific question you ask first. By the time the respondent says "OK" to the interview and has started it, the worst of your worries are over. Getting to yes from the knock on the door is the hardest part. 

Thank you for the answer. I am consciousof the crucial importance of the entire approach. However, in the step of this project I need a specific question to start the questionnary that is quasi designed already. In this case, The California Health Interview Survey is an useful example.

I certainly agree with Matt's suggestion about starting with something that is easy to answer. For this reason, some people like to start survey interviews with a few background  and demographic questions, as he also suggests.

Indeed, Matt has excellent advice (this is his area of expertise)!

I have found in my own work that starting off a survey with the "how do you rate your health" can elicit different responses than if you ask that at the end of the survey. Of course, it depends on the survey, and what is the construct you are focusing on. I like the demographic questions to ease into the survey...

It would be useful to include a background opening question that creates an awareness of what health and 'good health' means to that demographic. So an overall general question such as, 'When you hear the term 'good health' what does that mean to you?'

I realise that this is not standardised as a first question but it does encourage them to start thinking about their health before you commence the survey.

Another option would be to provide a short defintion of 'good health' and ask if they agree with this statement and why.