Exibir código-fonte para Usuário Discussão:Abdo/MMMNS/Ampliação da Rede/Questionário

<center>'''[[Usuário:Abdo/MMMNS/Tasks|Tasks related to this project]]'''</center>


==Open questions and proposals==

===Split ballot / randomizing the question order? If impossible, should the main question about drug use be the first on the ScaleUp survey?===
*I think split ballot is not possible because they (vox populi) don´t use doing this. It seems better asking first about drug use as a filter. --Dimitri
**I don't get it. A filter for what? And is it the best thing to use the main question as a filter? Perhaps it should wait for the point where there is going to be less noise, for example, far from the start so that the respondent has already understood the procedure but close enough to it so that he is not tired. --[[Usuário:Abdo|Abdo]] 07h49min de 5 de janeiro de 2010 (UTC)
***You´re right, Ale. I was a bit confused about the idea of filtering and made a mistake. Also, maybe we could re-order the questions about IDUs, MSMs and SWs to the end of ScaleUp questions. I´m not quite sure yet because it seems that in other surveys researchers used to put first those questions. -- Dimitri 12JAN2010
*<s>'''Update:''' Since we're now considering going with Ibope, we'll try to randomize within ScaleUp questions and, at the level of blocks, the ScaleUp and Summation blocks.</s> --[[Usuário:Abdo|Abdo]] 22h11min de 14 de janeiro de 2010 (UTC)
*'''Update:''' We've closed the deal with Vox. Dimitri will talk to them to see if we can trust them to manually randomize the two middle question blocks. --[[Usuário:Abdo|Abdo]] 08h13min de 8 de abril de 2010 (UTC)

===Move drug use questions to bottom===
We moved the drug use questions to the end of the questionnaire, because they're sensitive, but agreed that we should hear more from Maeve on this. Should they be asked earlier because of fatigue? Perhaps between the Scale Up and Summation questions? --[[Usuário:Abdo|Abdo]] 22h08min de 14 de janeiro de 2010 (UTC)
:Seems like we've settled on keeping this as it is, right? --[[Usuário:Abdo|Abdo]] 08h13min de 8 de abril de 2010 (UTC)

===New question at end of survey for interviewer===
We have 3 questions for the interviewer at the end of the survey
(http://wiki.stoa.usp.br/Usu%C3%A1rio:Abdo/MMMNS/Amplia%C3%A7%C3%
A3o_da_Rede/Question%C3%A1rio).  Do you think we can add one more:
"Which do you think were probably more accurate: responses to the
scale-up questions, responses to the summation questions, about the
same."  I think this will give us just one more piece of information
that we can use to compare the two methods. -- Matt

:I think it is OK but we need to talk to Vox and ask them about the estimation of time spent during application of questionaires - I´m not sure about that anymore because we added new questions and I don´t know how long the questionnaires will take. -- Dimitri

===Question 5 (How long have you lived in Curitiba?)===
("Em 01 de Julho de 2009, fez quanto tempo que você morava, sem interrupção, na cidade de Curitiba?
   1. Até 4 anos
   2. De 5 a 9 anos
   3. 10 anos ou mais
")

I thought we decided to make this continuous, in other words "How
long have you lived in Curitiba?"  We can always collapse it to these
groups and I think it is easier for the respondent if they don't have to
use these categories which are a bit unnatural.  One thing is that we
need to decide what kind of answer we want if they lived in Curitiba
until they were 20 years old, moved to Rio, and then moved back to
Curitiba 2 years ago.  Do we want them to say 2 years or 22 years? -- Matt

:I think the idea is to ask only for how long they lived in Curitiba
(i.e., continuously) because if we want to know about their history of
events (migration) then this would be another different question (in
fact we would need at least 2 new questions: for how long do you live
in curitiba + how old were you when you departed the last/first time. -- Dimitri

::If the purpose of this question is to check that people who have lived in Curitiba for longer seem to know more people in Curitiba then I think we want to ask "How many years of your life have you spent living in Curitiba?"  What is the purpose of this question? -- Matt

::: The usual question in the Brazilian Census is "Há quanto tempo (anos) você reside sem interrupção em Curitiba?" ["For how long (in years) do you live ininterruptly in Curitiba?". I think this is enough to get the information we want (the longer stay the bigger degree) -- Dimitri 08APR2010 10:50hs

:::: But does the question in the Brazilian Census intends to reflect the number of people a person might know in curitiba, or is it more related to migration patterns? I tend to agree with Matt, someone who has lived in Curitiba during his youth, then leaves the city and later returns, is likely to reconnect with his past acquaintances. If we want to have an idea of the number of people somebody knows, then asking "how many years of your life" makes more sense. Even though we might lack systematic evidence supporting this idea that people returning to Curitiba reconnect, I think it is easier to justify than the opposite assumption. --[[Usuário:Abdo|Abdo]] 22h32min de 10 de abril de 2010 (UTC)

===Add question about children===
Since we just removed 7 questions (questions 8-14 (Você frequentou nos últimos 2 anos) I would like to propose that we add two questions:

* How many children do you have?

* How old are they?


The reason to add these question is that they can help us to check the responses to the scale-up questions:
Quantas pessoas você conhece que vivem em Curitiba e são estudantes de 5ª. a 8ª. Séries de escolas públicas?________
Quantas pessoas você conhece que vivem em Curitiba e são estudantes de 5ª. a 8ª. Séries de escolas particulares?________
Quantas pessoas você conhece que vivem em Curitiba e são estudantes do ensino médio de escolas particulares?________
Quantas pessoas você conhece que vivem em Curitiba e estudam em Universidades Federais?__________
Quantas meninas você conhece que vivem em Curitiba e são menores de 5 anos?__________
Quantos meninos você conhece que vivem em Curitiba e são menores de 5 anos?__________  

For example, people with child under 5 years old should be more likely to know children under 5 years old; people with children in grades 5 to 8 should be more likely to know children in grades 5 to 8; and so on.  If we see these patterns in the data we can have more confidence that people are answering these questions accurately.

Dimitri, I assume that these are standard questions in demography.  Is there some national demography survey that you can copy the questions from?

If anyone has other ideas for demographic questions that we can compare to network size questions, I think it might be worth it.

:: Yes, this is a very good idea. In the Brazilian Census the questions about children are directed only to women but I think we should also ask men and change a little bit the usual census questions. Actually we don´t need the history of birth events. Definetively we could add one question proposed by Matt, "Quantos filhos você tem?" [How many children do you have?], but I´m not quite sure about the second one because we would need to ask more questions if the respondent had more children. I think it will be enough to know the respondent has children and the age of the youngest (we might estimate the age of the eldest comparing the younger´s and respondent´s age). So, my suggestion is to ask a simple question like "Qual é a data de nascimento do seu filho mais novo?" [What´s the date of birth of your youngest child?"] -- Dimitri 08APR2010 11:15am

==Closed issues==

===Item 4. cor (color)===

É normal não ter uma opção "não quero responder"? (Is it usual to not have an option "don't want to answer"?) --[[Usuário:Abdo|Abdo]] 21h35min de 26 de novembro de 2009 (UTC)

:Engracado vc perguntar isso porque em geral nao se coloca essa alternativa "nao quero responder", a nao ser em casos muito especificos. Mas perguntas gerais, como essas, nao ha porque colocar esta alternativa. Eh normal proceder deste modo. (Dimitri)

::Ok, não é um problema. (Ok, not an issue.) --[[Usuário:Abdo|Abdo]] 17h00min de 1 de dezembro de 2009 (UTC)

===Item 7. escola (school)===

Aí cabe escola em qualquer nível, até pós-graduação? Escola de línguas também? De esportes? (Does that include school of any level, even post-graduate? Language schools too? Sports?) --[[Usuário:Abdo|Abdo]] 21h35min de 26 de novembro de 2009 (UTC)

:Neste caso a escola eh fundamentalmente a de formacao basica ate a universidade, mas penso que no nosso caso nao ha problema se pegarmos "qualquer tipo de escola" pois a ideia aqui eh ver se o entrevistado possui uma rede mais movimentada do que se ele nao frequentasse qualquer tipo de atividade fora de casa. (Dimitri)

::Legal, de fato faz sentido. (Alright, it does make sense.) --[[Usuário:Abdo|Abdo]] 17h00min de 1 de dezembro de 2009 (UTC)

===Item 10+11 uso de dorgas (drug use)===

Ainda me confunde o formato ideal para perguntar disso e compararmos com outras pesquisas como PCAP. Na 10 o tipo das respostas difere daquela pesquisa. Na 11 haviam alternativas: "com que frequencia vc usa drogas" ou "você consideraria uma vez por semana uso frequente". (I'm still cofused about the ideal format to ask about this and compare to other research like PCAP. On 10 the answers differ from the one in that poll. On 11 there were alternatives to the current "what do you consider frequent": "with what frequency do you use" or "would you consider frequent to use once a week"). --[[Usuário:Abdo|Abdo]] 21h35min de 26 de novembro de 2009 (UTC)

:I'm not sure that I understand exactly what is happening here, but I think we want the question and answer choice to be EXACTLY the same as the PCAP. -- Matt

:I think we only want to ask question 11 if they report using drugs in question 10.  Is that correct? I don't speak Portuguese so I'm not sure if this is already accounted for. -- Matt

:Agora eu estou confuso porque pensei que a questao 10 e 11 fossem exatamente as mesmas utilizadas na PCAP! (Now, I am confused because I believed I had used exactly the same questions 10-11 as in the PCAP) (Dimitri)

::Ale, do you think the current question wording for question 10 does not match the PCAP? (Matt)

:::I think they're different. You can check the questions from the PCAP by looking at the PDF linked from the [[Usuário:Abdo/MMMNS#Artigos|Resources on the main page]], or by checking out the [[Usuário:Abdo/MMMNS/Ampliação_da_Rede#Perguntas_coerentes_para_extrapola.C3.A7.C3.A3o|suggested questions]] I previously included on the wiki.
:::I think they differ enough so that it would be hard to consider a valid comparison. In particular, the PCAP question only refers to "frequency" in one alternative, which in turn makes question 11 more clear.
:::Also, the PCAP has two questions, there's specific mention of "powder cocaine" and "injected cocaine". Those are the categories the PCAP actually asks about. I'm not sure if it's desirable to change this to "general drug use" in order to make our comparison, as it is now.
:::About question 11, what we want to know is how people answering PCAP understand drug usage in the terms of the RDS study, that is, an average of once a week during the last six months. We're using the ScaleUp as a "miniature PACP" to figure that out.
:::So I'm uncertain whether the current phrasing is the best alternative to get this information. As I wrote above, we could ask this in two other ways. The "would you consider frequent to use on average once a week" is more connected to what we want to know, and the "with what frequency do you use" asks for more concrete information. However, both alternatives give us less information as it only tells us about one specific point in the scale. Thinking of it now, the first alternative is enough for current considerations, the second is not. But even the first may fail us if we come up with other ideas about how to use this data.
:::In short, I understand we must at least conform question 10 to the PCAP. We can leave it as a general question but the answers must look like the PCAP. We can leave question 11 as it is.
:::--[[Usuário:Abdo|Abdo]] 17h00min de 1 de dezembro de 2009 (UTC)

::::OK, I looked at the PCAP (following the link Ale listed above) and the question we are using is NOT the same as the PCAP.  I agree with Ale that we should follow the PCAP exactly.  Therefore, I think we should replace question 10 with the two PCAP question (one about "powder cocaine" and one about "injected cocaine") so that we can exactly match the estimated heavy drug use rate from 2004 to the one is 2009.  That way we can see if using the exact same method, we detect a change in drug use frequency.  To see the exact wording see page 168 of the PCAP report that Ale linked to. -- Matt

::::I'm still not sure why we need question 11 (and I can't read Ale's description on this page: http://wiki.stoa.usp.br/Usu%C3%A1rio:Abdo/MMMNS/Amplia%C3%A7%C3%A3o_da_Rede).  I think that there are a couple of estimates that we want to make with this survey.  First, we want to *exactly* match the definition and methods from the 2004 PCAP.  That will allow us to make a direct comparison to see if the overall rate of drug use seems to have increased in Curitiba.  Second, we want to make the scale-up estimate which uses the RDS definition of heavy drug user (25 times in the last 6 months).  Those are the two estimates that I think we need to make.  A third estimate that we might want to make is a direct estimate of the prevalence of heavy drug users based on the RDS study definition.  In other words, I think we might want to add the question that is used to screen people into the RDS study.  -- Matt

:So, we will use exactly as the two questions in the PCAP (P29 and P30 about use frequency of powder and injected cocaine). I think the confusion started when we were in Curitiba and talked about making compatible the RDS+NSU+PCAP as a way to measure the transmission errors in Curitiba and then extrapolate it to (or correct?) the PCAP estimates. I remember it wasn't an easy question but I don't remember the final decision and why we decided it. However, I believe Matt made it quite clear now: we will use the PCAP questions to make the estimates compatible between the 2004 survey and ours for 2009; and we will keep the drug use definition as in RDS to investigate the transmission errors. Fair enough, we can add questions P29 and P30 and keep the question 11 as it is (Ale has already corrected it) -- Dimitri

::::Also, I don't see any mention in the survey that we are going to keep the responses confidential.  There is a lot of text about this in the PCAP right before they ask the heavy drug use questions
:::::"Vamos fazer agora mais algumas perguntas antes de encerrarmos, e gostaria de repetir que nenhuma entrevista será analisada individualmente, mas	sempre pelo conjunto de respostas, garantindo a confidencialidade das respostas. As próximas perguntas podem ser consideradas de caráter íntimo, mas	é importante que suas respostas sejam muito sinceras. Para nem mesmo eu tenha acesso as suas respostas, gostaria que as marcasse nesta folha e colocasse na urna. Caso tenha alguma dúvida, estarei à disposição para possíveis esclarecimentos."
::::I don't know exactly what this says, but I think we probably need something like it, perhaps right before question 10.  Obviously, I'm not the best person to write this. -- Matt

:::::You're right, Matt. However the PCAP is a huge questionnaire and has so many different parts that at any new starting part they have to remind the respondent that that is a confidential questionnaire and people can be sure of their privacy. In our case, the questionnaire is straightforward and small and in addition we have the so-called "Termo de Consentimento Livre e Esclarecido"(TCLE) where the respondent is informed thoroughly about the survey in the very beginning of the interview, especially about the confidentiality of all information she provides. -- Dimitri

::::::OK, has the "Termo de Consentimento Livre e Esclarecido"(TCLE) been written yet?  Can we post that (or have Ale post that) so that everyone can see it?  -- Matt

===Item 12 uso de drogas pelos contatos (contacts drug use)===

Não é esta a questão mais importante? Não deveria dar algum tipo de ênfase a ela, ou isso só piora a resposta? (Isn't this the most important question? Shouldn't there be some kind of emphasis on that, or would it be worse?) --[[Usuário:Abdo|Abdo]] 21h35min de 26 de novembro de 2009 (UTC)

:Na realidade esta questao se repete no final do terceiro modulo (sobre as populacoes desconhecidas). Creio que eh importante ela manter o equilibrio em relacao as duas precedentes. Nao seria bom criar alguma enfase neste caso. (Dimitri)

::Bem então. Mas no terceiro módulo deveria ser a mesma pergunta, com a condição de "em média uma vez por semana nos últimos seis meses", não? (Fair enough. But i the third module the question should have the same condition about "on average once a week for the last six months", not?) --[[Usuário:Abdo|Abdo]] 17h00min de 1 de dezembro de 2009 (UTC)

:Ver resposta mais abaixo junto ao questionamento do Matt -- Dimitri

===Item 17 - família do parceiro (partner's family)===

Para aqueles que não tem esposo(a) ou namorado(a), o que responder nessa questão? Zero? (For those who don't have husband/wife or boyfriend/girlfriend, what's the answer here? Zero?) --Neilane

:Sim. A pessoa deve ser avisada desde o inicio que ela deve contar apenas para as categorias nas quais ela tem contatos pessoais (Yes, each respondent should be instructed to count only people in the categories of relations he/she has a personal contact with) (Dimitri)

::Isso já está claro no questionário ou precisa de alguma modificação? (Ok, so is this clear in the questionnaire or do we need to include it?) --[[Usuário:Abdo|Abdo]] 17h00min de 1 de dezembro de 2009 (UTC)

::: Acho que seria bom incluir mais uma nota explicativa no inicio do modulo sobre categorias de relacao nas redes pessoais -- Dimitri

===Item 49. How many professors do you know?===

Previously there was some concern that the number of professors was not counted correctly because there was something like 2.5 university students for each university professor.  Do we now have the correct number of professors?  If not, maybe we should not include this question.  -- Matt

:Yes, it seems that the category of professors has a lot of double counting. I agree, and we should not included it. (Dimitri)

::Removi a pergunta do questionário. (Following this discussion, the question has been removed.) --[[Usuário:Abdo|Abdo]] 17h00min de 1 de dezembro de 2009 (UTC)

===Item 12 and item 35===
I don't speak Portuguese, but are question 12 and question 35 the same?  Google translate says they are.  -- Matt

:: They are "almost" the same but they have a fundamental difference: the question 12 asks ego's knowledge about ANY drug user (as defined in the RDS), i.e., it does not specify any kind of "boundary" or frame of reference (time or space); the question 35 asks specifically ego's knowledge about drug users LIVING in Curitiba. Of course, we could change it all and add the specificity of place (Curitiba) in the question 12 and delete the 35. What do you think? I like the idea of preserving both because we could check answers in two different moments (before and after questions about personal network - this might be useful in the future, but I'm not quite sure, that's just a guess). -- Dimitri

:::Interesting.  I did not notice these differences.  I'm still not sure about the purpose of question 12.  I think that the question about "How many heavy drug users do you know in Curitiba?" should be in the same place as the similar questions so I think it makes sense to keep question 35.  If we don't have a good reason for question 12, then I think we should remove it.  One possible reason to ask both 12 and 35 would be to ask the same question two times to measure consistency (as you said).  However, If we did that, then I think we would really need to ask the /exact/ same question both times (including using the same prompt before the question: "Finalmente, agora vou perguntar quantas pessoas você conhece de vista ou pelo nome E que também conhecem você de vista ou pelo nome, e com quem você entrou em contato seja pessoalmente ou por telefone, correspondência ou e-mail, nos últimos 2 anos que fazem parte de grupos específicos").  Also, I think the issue of response consistency is somewhat beyond the scope of this study.  To keep things simple I would suggest not having two of the same questions on this survey. -- Matt

===Questions for Vox Populi===
1) How much will it cost for GPS locations of respondents?  If I understood correctly they could give us this information for extra money.  Do we know how much this would cost? -- Matt

:They didn't informed us yet but I guess it is cheap. (Dimitri)

2) Will they give us information about how long each interview takes including possibly how long each question takes?  This is useful information and it should be possible if all the data is being collected on a computer (CASI). -- Matt

:Yes, this is easy for them. I'll ask Vox about that. (Dimitri)

===Implemented From voip talk between Matt and Ale===
*ScaleUp goes before summation method.
*:The reason for this is twofold: first because this study as a whole is more concerned with the ScaleUp than the summation method, but second and more important because the summation method involves "exclusive counting", which is a mentally complex procedure that in order to avoid confusion should only be introduced to the respondent after all the usual counting questions from the ScaleUp method. --[[Usuário:Abdo|Abdo]] 07h48min de 5 de janeiro de 2010 (UTC)
*Summation method should make it a lot more clear that each contact falls into only one category and repeat thoroughly (at each level - "don't count the people that ) that he should be counted only once also within that category (give concrete examples)
*Add a question about level of education
(Question numbers below refer to [http://wiki.stoa.usp.br/wiki/Usuário:Abdo/MMMNS/Ampliação_da_Rede/Questionário&oldid=11053 an old version of the questionnaire])
*Question 10: Agree to replace with the two PCAP questions
*Question 11: Replace with direct estimation using the exact RDS drug use definition (inclusion criteria)
*Question 12+35: Supress 12, fix 35 to use RDS definition
*Question 45: Supress "municipais" because people won't know the difference between municipal and state schools?
*Question 7+9: Combine and make it all "last 2 years" since those are social network related questions

===Add the standard socio-economic level questionnaire? (to determine class A, B, C)===
*This I don´t think is a good idea. It´s always a problem asking questions about socioeconomic status in Brazil, and in fact we don´t need that if we are already asking about education. In Brazil these are strongly related variables. --Dimitri
**Well, maybe we could ask those at the end of the questionnaire, if there is enough time. What exactly is problematic with that? Respondents don't like when you ask? Our approval by the ethics committee doesn't extend to this kind of question? --[[Usuário:Abdo|Abdo]] 07h49min de 5 de janeiro de 2010 (UTC)
***This is considered a sensitive question and usually people either refuse to respond or give an incorrect/innacurate answer. Since education level is highly correlated with income in Brazil maybe we could save space/time in the questionnaire. -- Dimitri 12JAN2010 

===Check if we also have data on state schools since we removed "municipais"===
*Yes, that´s ok. The implementation of "public schools" as a summation of municipal and state schools is ok and we have data about that. --Dimitri
**Great! :) --[[Usuário:Abdo|Abdo]] 07h49min de 5 de janeiro de 2010 (UTC)
*** Actually I think this is not done yet. I don´t  remember the percentage of students at state schools in Curitiba and I think the reason using municipal school is that students at state school is too high (more than 6% of total population while students at municipal schools are only 0.43%). My experience in Belo Horizonte is that most people knows perfectly well who´s in the municipal school because these are a kind of "elite" schools - they´re just few enough to be noted. Sure, Curitiba might be the opposite but still the percentage of students at state schools is too high. -- Dimitri 12JAN2010
**** Dimitri and I talked, and he remembered it was desirable for the populations to be within some range, and public schools attendance in general is out with about 6% of the population, but other than that it is better to keep it general. --[[Usuário:Abdo|Abdo]] 19h51min de 13 de janeiro de 2010 (UTC)

===Question 8<nowiki>:</nowiki> Should be "last 2 years" in case it's a social network related question or moved up in case it's a socio-demographic question===
*I don´t know but I think it is not a good idea changing this. We have already some questions in the summation part asking about ego´s contacts in the workplace. So, this question is more important for sociodemographic profile purposes (we should recall that job market in Brazil is really volatile and people change around a lot. So it is important to make sure that people are working right now, and how this fact would affect their personal network sizes). --Dimitri
*:Again I don't get it. If it is important because of the effect on personal network size, then it should be "last two years" because later we always ask about their networks "in the last two years". If it is only a social-demographic question, then it should not say "last two years" to let us compare to data from other surveys. Does that make sense? --[[Usuário:Abdo|Abdo]] 07h49min de 5 de janeiro de 2010 (UTC)
**I think is different when I ask if you´re working right now than if you worked in the last 2 years. If you´re working right now it says not only about your active personal network but also about your income and living conditions. Besides we could compare people that are not working right now and still have contacts with old friends at work, and people that is working but have fewer contacts at work. -- Dimitri 12JAN2010
*** I talked to Dimitri and we agreed that it is better to treat it as a socio-demographic question. Using this as a network question faces the problem of knowing people from work done before 2 years ago but still being in touch with them. So having worked in the last two years has no direct relationship to having work contacts. Unless the question was "how many people do you know from work you did in the last two years", but that still has the problem of leaving out people you also knew from work in previous years. So anyway this would be messy as a network question. Maybe there's a magical formula that fixes both issues, but for now it's better as it is. --[[Usuário:Abdo|Abdo]] 19h51min de 13 de janeiro de 2010 (UTC)

===Do we want from Vox Populi the name frequency distribution of the respondents, to compare with the frequencies from the data we have from death records?===
*This is a very good point, and I think we could ask them for that name frequency distribution. --Dimitri
**That's another thing to ask voxpopuli, and we probably want to have it explicit in the contract. --[[Usuário:Abdo|Abdo]] 07h49min de 5 de janeiro de 2010 (UTC)

===I think this question (is really 7 questions):===

 9. Quais das seguintes alternativas você frequentou nos últimos 2 anos?<br/>
   1. Escolas
   2. Igrejas
   3. Associações de voluntariado
   4. Associações de lazer e clubes
   5. Associações culturais
   6. Associações de classe
   7. Partidos políticos 

Is really 7 questions.  I would rewrite as:

 Quais das seguintes alternativas você frequentou nos últimos 2 anos?<br/>
   9. Escolas: Sim () Não ()
   10. Igrejas: Sim () Não ()
   11. Associações de voluntariado: Sim () Não ()
   12. Associações de lazer e clubes: Sim () Não ()
   13. Associações culturais: Sim () Não ()
   14. Associações de classe: Sim () Não ()
   15. Partidos políticos: Sim () Não ()

*Yes! That´s great! -- Dimitri 12JAN2010
**Done. --[[Usuário:Abdo|Abdo]] 18h48min de 13 de janeiro de 2010 (UTC)

===Don't ask head of household===
<pre>
One thing I forgot to mention was that I don't think we want to 
interview the head of the household.  In the US this is usually 
considered to be the man, and we don't want most of our sample to be 
men.  Also, we don't want to interview whoever answers the door.

I think the most common method for selecting adults from a household is 
the "next birthday" method (sometimes called the last birthday method).  
For example, the interviewer asks "How many adults live in this 
household?"  Then, "Of those [X] adults, which of them has the next 
birthday?"  (or last birthday?)  Then you interview that person.  The 
"birthday" method allows you to sample from the adults living in the 
household, and avoids the biases involved in interviewing the person who 
answered the door or the head of the household.

Here's a paper that reviews methods for sampling within households: 
http://dl.dropbox.com/u/1437602/sample_within_household.pdf
and here is a paper that uses the next birthday method: 
http://dl.dropbox.com/u/1437602/next_birthday.pdf

It was nice talking to everyone today,
Matt
</pre>

:According to Dimitri they'll randomly sample among valid respondents and return up to 3 times to the household. Closed. --[[Usuário:Abdo|Abdo]] 08h13min de 8 de abril de 2010 (UTC)

===Get a lot more concrete information about Vox Populi (what are the relevant dates, where are the formal proposals, do they use PDAs or paper, etc)===
*Yes, I agree with that. Some of those questions we know the answers: the relevant dates depends on firming the contract (this depends on bureocracy between FIOTEC and Vox Populi, and I think only in the end of January we´ll have a definitive answer). Anyway, according to Vox Populi as soon as they got the final contract they can start fieldwork a week later and have the results in other 2 or 3 weeks; I got the formal proposals in pdf (but these are in Portuguese). I don´t think is a good idea making them available to the public in the wiki. What do you think?; About the last point, if they use PDAs or not I really don´t know and I´ll need to ask them, but I believe they will use PDAs. --Dimitri
**Ok, but please ask them about the PDAs and randomized question orders (using PDAs that should be easy and usual), and send us voxpopuli's pdfs in portuguese by e-mail.--[[Usuário:Abdo|Abdo]] 07h49min de 5 de janeiro de 2010 (UTC)
*Update: Since VoxPopuli does not use PDAs, we're considering going with IBOPE because PDAs are relevant to have the precise response time of each question and to allow random question order. --[[Usuário:Abdo|Abdo]] 22h12min de 14 de janeiro de 2010 (UTC)
*We've closed the deal with Vox. No PDAs, might randomize question blocks, will give us the addresses for geolocation. --[[Usuário:Abdo|Abdo]] 08h13min de 8 de abril de 2010 (UTC)

===Questions from the General Social Survey to be answered by the interviewer===
*Matt suggested we include two questions from the GSS: "No geral, qual foi a atitude do entrevistado diante da entrevista?" and "A compreensão das perguntas pelo entrevistado foi..." I already translated and included them within "Respondidas pelo entrevistador" in "Questões adicionais", but comments are welcome. --[[Usuário:Abdo|Abdo]] 22h05min de 14 de janeiro de 2010 (UTC)
**These quesions are important given the study [http://www.asanet.org/images/journals/docs/pdf/asr/Aug09ASRFeature.pdf in this paper] which concludes that answers to them might be surprisingly correlated to personal network size. --[[Usuário:Abdo|Abdo]] 22h47min de 14 de janeiro de 2010 (UTC)
***Questions translated by me and included as a task for the interviewer "after concluding the interview". --[[Usuário:Abdo|Abdo]] 08h13min de 8 de abril de 2010 (UTC)