Discussão:Pesquisa:Inquérito crack Brasil/Questionário Sócio-demográfico

Questões resolvidas (resolved issues)

Multiple-choice migration options

For the migration questions on this page (section 1), and also on the age-sex page (section 3), I was thinking that a multiple-choice response with the following options would be best:

1 ... Brazilian state (first, alphabetically) ... 26 ... Brazilian state (last, alphabetically) 27 ... Federal district

30 ... Argentina 31 ... Bolivia 32 ... Chile 33 ... Colombia 34 ... Ecuador 35 ... French Guyana 36 ... Guyana 37 ... Paraguay 38 ... Peru 39 ... Suriname 40 ... Uruguay 41 ... Venezuela

50 ... Portugal 51 ... Italy 52 ... Switzerland 53 ... United States 54 ... Japan

60 ... Other country (please specify __________________)

For the state that contains the municipality that the respondent currently resides in, I think 'elsewhere in [state name]' would make sense. For example, for someone who reports knowing someone who used to live in the municipality of Rio de Janeiro, but left the municipality for another part of the state of Rio de Janeiro, the option could read 'elsewhere in the state of Rio de Janeiro.'

For the countries on the list, this document http://idbdocs.iadb.org/wsdocs/getdocument.aspx?docnum=556530 is what persuaded me that there should be categories for Switzerland and Italy --Dfeehan 20h39min de 5 de Julho de 2011 (UTC)

Wow that's a big list... --Solstag 22h03min de 3 de Fevereiro de 2012 (UTC)

Yes, it is! But the idea is not to have the interviewer read the entire list to the respondent for each question; instead, the interviewer asks the where question and gets enough information to put the response into one of these categories. For example, if the respondent says the name of a town where he lived right before, the interviewer might ask to clarify which state the town is in, and then code that on the form. I don't think it makes sense to constantly re-read the entire list of options to the respondent. --128.112.149.247 19h33min de 6 de Fevereiro de 2012 (UTC)

To follow-up on what Dennis said, I think the main purpose of the list is to prevent the interviewers from recording any free-text answers. Thus, as long as it does not take up too much space on the paper, having lots of choices should not introduce many problems that I can think of. --Msalganik 21h13min de 6 de Fevereiro de 2012 (UTC)

This is done now. --Solstag 00h07min de 7 de Fevereiro de 2012 (UTC)

Migrant status

Right now, we can't quite estimate whether or not the respondent moved into the muncipio recently because we are assuming that people who were born in the municipio and who live there at the interview never lived anywhere else. Based on Brazil's census instrument, it seems like the best way to change this would be to have three possible responses to the [SOCIODEM-BORNINCITY} questions:

[SOCIODEM-BORNINCITY]. Você nasceu neste município? a. Sim, and I always lived here b. Sim, but I have lived elsewhere c. Não

Then, [SOCIODEM-BIRTHPLACE] would only be answered by people who respond 'no', and [SOCIODEM-TIMEINCITY] as well as [SOCIODEM-PREVRESIENCE] would be answered by anyone who replied (b) or (c) to [SOCIODEM-BORNINCITY].

Hm, I see, perhaps we are still able to distinguish that. We ask their year of birth and then we ask how long they've been there, and if those don't match we ask where they used to live before going there. But I also see it is much clearer if we ask this other way. What do you think? And is it better to have two alternatives for "Yes" or to have another question? --Solstag 21h58min de 3 de Fevereiro de 2012 (UTC)

Yes, I see what you are saying. The only downside to using the year of birth and the amount of time in the town is that the interviewer then has to make that calculation, and it seems like that's something that they could mess up. (I'm not sure about that, but it doesn't seem impossible.) The reason I put the two 'yes' alternatives instead of a separate question is that the 2010 long-form Census question which asks about migration does so this way. But I'm open to changing it to use another question instead, if we think that would be better. --128.112.149.247 19h35min de 6 de Fevereiro de 2012 (UTC)

If possible I would suggest that we stick with the question *exactly* the way they did it in the Census. I'm pretty sure they field tested before using, and if we use the exact same question we can match our sample to theirs if we want. --Msalganik 21h10min de 6 de Fevereiro de 2012 (UTC)

Fixed, it's like the 2010 Brazilian CENSUS now. --Solstag 22h29min de 6 de Fevereiro de 2012 (UTC)

Other questions about the respondent

Matt and I thought of a few other things that it might be interesting to ask of the respondents, so that we have more background information about them; please let us know what you think. First, we thought it would be interesting to ask whether or not the respondents themselves receive Bolsa Familia. Second, one of our colleagues mentioned that it's possible to ask whether or not the respondent's job pays social security (or something similar to this), as a way of getting at whether or not she works in the formal sector. Third, our colleague also mentioned that the census asks about the material that the respondent's house/roof is made from in order to get a rough measure of wealth. What do you all think? [Dennis 05 Jan 2011]

We have the unemployment question, but we don't think we can afford further increasing the questionnaire. --Solstag 13h54min de 26 de Janeiro de 2011 (UTC)

Q 7. (about work/employment)

I thought we decided that this question was bad because it does not actually measure "unemployment". --Msalganik

Yes, this is not the standard way to ask about employment. We should try "did you look for a job in the last week/month?". Anyway, as far as I remember we proposed this question just to make sense of personal networks related to workplace. I'm not sure about keeping this question in the new questionnaire. [Dimitri 07DEC2010]

I'm not sure. If our objective was to measure workplace network, we don't care if the person was unemployed in the official sense. All we care is whether they have been establishing work relationships. In that sense, the current question "are you currently working" is alright, we could at most add a time window, as in "have your worked at least a full month in the past six months?" and if we're serious that this is interesting, add an extra question like "with how many people are you in contact with because of that job?" --Solstag 02h42min de 8 de Dezembro de 2010 (UTC)

Ok, me and chico and dimitri talked and resolved that since it is less important and we did not even use this question in the last survey analysis, and we need to value our time, we may drop it if necessary. --Solstag 15h28min de 13 de Dezembro de 2010 (UTC)

Neilane thinks this question is useful so we have more insight on our sample. Keeping it with a better phrasing. --Solstag 18h37min de 17 de Dezembro de 2010 (UTC)

I agree with Neilane. I think we added it as a check to make sure that the survey population matched the city population. For example, if the city has 10% unemployment and our sample has 20% unemployment that would suggest a problem. We can also look for comparability on other characteristics (age, color, gender) as well. Finally, with crack use, I think that employment might be related to the number of crack users you know. This would be a good thing to check. --Msalganik 19h26min de 17 de Dezembro de 2010 (UTC)

O que o Matt está dizendo é que a questão sobre emprego poderia ser usada para conferir se a empresa está coletando os dados corretamente. Isso implicaria escrever a questão de uma forma padrão que possa ser comparada a alguma outra pesquisa já existente. É possível isso? Caso contrário, acho melhor deixar como está, pois a segunda preocupação dele, com o número de usuários que se conhece, parece-me melhor refletida na forma atual. --Solstag 08h16min de 20 de Dezembro de 2010 (UTC)

Yes, it is possible to keep the question anyway. If you want to phrase it according to IBGE polls about employment this would be like "did you look for a job in the last week?" If "yes", then the person is unemployed. [Dimitri, 21DEC2010]

This is solved with our current phrasing, which is comparable to census data. --Solstag 13h54min de 26 de Janeiro de 2011 (UTC)

OK, good. --Msalganik 17h40min de 27 de Janeiro de 2011 (UTC)

Sobre as questões 8, 9 e 10

Retirar essas perguntas? --Neilane

Se não me engano, essas perguntas serviriam para avaliar correlações nas respostas do ScaleUp, mas não são necessárias. Talvez o Dimitri saiba explicar melhor. --Solstag 06h24min de 23 de Novembro de 2010 (UTC)

Na minha opiniao poderiamos retirar as perguntas 8, 9 e 10. Lembro vagamente que a ideia era usa-las para checar a precisao das respostas do scale-up, mas agora nao vamos perguntar nem mesmo sobre estudantes em Universidades Federais. Acho melhor deixa-las de fora. [Dimitri 07DEZ2010]

Feito! --Solstag 02h47min de 8 de Dezembro de 2010 (UTC)