Mudanças entre as edições de "Usuário Discussão:Abdo/MMMNS/Ampliação da Rede/Questionário"

Edição das 20h03min de 4 de dezembro de 2009

Dúvidas sobre o questionário (Doubts)

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"?) --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.) --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?) --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.) --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"). --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 Resources on the main page, or by checking out the 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.

--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?) --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?) --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?) --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.) --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)

Propostas de alteração do questionário (Modification proposals)

From voip talk between Matt and Ale

Implemented

• ScaleUp goes before summation method
• 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 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

Not yet implemented

• Split ballot / randomizing the question order? If impossible, should the main question about drug use be the first on the ScaleUp survey?
• Add the standard socio-economic level questionnaire? (to determine class A, B, C)
• Check if we also have data on state schools since we removed "municipais"
• Question 8: 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
• Get a lot more concrete information about Vox Populi