This course will cover the basic elements of designing and evaluating questionnaires. We will review the process of responding to questions, challenges and options for asking questions about behavioral frequencies, practical techniques for evaluating questions, mode specific questionnaire characteristics, and review methods of standardized and conversational interviewing.
Asking Sensitive Questions
Loading...
來自 University of Michigan 的課程
社会调查的问卷设计
248 個評分
從本節課中
Unit 3: Asking Factual Questions
Facts and quasi facts; Memory and recall; Asking sensitive questions; Mode, privacy and confidentiality
與講師見面
Frederick Conrad, Ph.D.Research Professor, Survey Methodology
Institute for Social Research
Frauke Kreuter, Ph.D.Professor, Joint Program in Survey Methodology
Adjunct Research Professor, Institute for Social Research
The next three segments are about asking sensitive questions.
There are very many techniques in the literature on asking sensitive questions.
Not all of them can be covered here.
But, the most important ones I will introduce to you.
And in doing so, I want to acknowledge the work that Roger Tourangeau and
Tin Yan did, not only to assemble material here for the Joint Program in
Servey Methodology when we teach this
questionnaire, but also in the terrific overview
piece that is part of the questionnaire material that you have on your website.
Learning outcomes for these segments are the following.
We wanr to teach you what makes a question sensitive.
What the consequences are of asking sensitive questions in surveys.
And what we can do to minimize these consequences.
And of course, what conclusions you can draw
from them for your own questionnaire.
So let's start with this set of items here
and think for yourself, "Which of these questions do you think are sensitive?"
And think about why.
Now you will notice, some of them are about
behaviors you do that you would not want to reveal to me.
Some of them are things that you think one should do, but don't.
There's sort of a mixture here on these and we will see in
the course statistics that not everybody answered these questions in the same way,
actually not answered but evaluated them, with respect to their sensitivity.
And the last item on this list has the biggest variance because having
a library card or not is not all that sensitive anymore, depending on
the context you are in. Because in many settings, having a library card
is obsolete because literature and things
of that nature are nowadays available online.
So you already see, personal perception, the own state you
are in, might drive what you think is sensitive.
There are a couple of other ways to
structure sensitivity and we'll walk you through them.
There's three main types for sensitive questions.
The first one is, you can consider
a question as just being about private information.
So income and other identifiers are part of that category.
Illegal behaviors and socially stigmatizing behavior.
Let's go back to the income example.
People do not like to be asked about their income, regardless of what the income is.
With illegal behaviors, there might be a threat considered
in revealing that behavior. And the social stigmatizing behavior,
this is a set of things where respondent think that the interviewer or the
researcher has a particular behavior in mind and they fail
to match that behavior.
So there's a mixture of issues here.
One is, can the information be used in a legal way?
Is the information personal?
And does the information relate to the image of the respondent?
In general, there are different levels of sensitivity.
A good thing for you is to ask yourself, "Is the question threatening to ask,
regardless of the respondent answer?", as I mentioned for the incoming example.
Or is this question threatening to answer
because the respondents are afraid of the consequences this might have?
Now what are the consequences of
asking sensitive questions in your questionnaire.
There are three consequences that you might
want to be aware of when constructing a new survey.
The first one is, if you include a sensitive
question and the respondents do know about it
ahead of time, you might have a higher
unit nonresponse rate than you would have otherwise.
People are less likely to participate in a survey with sensitive questions.
If they do participate and you hit that segment, you will have item nonresponse.
People will, if asked, not tell you that information that they consider sensitive.
And a good example, again, is the income question.
Across many surveys there's about a quarter of the respondents,
who do not reveal the answer to that particular question.
And finally there's misreporting.
People do tend to want to give an answer to
a survey question, but they might just edit their answers.
This can lead to over reporting of social desirable behavior
and to under reporting of socially undesirable behaviors and attitudes.
So why do people edit their answers?
There's a norm out there, for many questions, on how they should be answered.
"Did you vote at the last election?"
Well, good citizens do vote, so people feel
that social norm of being a voter and
therefore might edit their answer even so they
know or remember that they did not vote.
People should not use drugs is another example of that nature.
People choose to give the norm answer, because they
want to avoid embarrassment or because of privacy and confidentiality concerns.
And finally, you should keep in mind, people are more
likely to edit and misreport if they have something to hide,
if the topic is more sensitive than less sensitive,
and if there are bystanders, other people in the
room that would listen and also hear the answer.
The last piece gives already an idea
of other factors than can influence the reporting
and the amount of editing people do when they answer a sensitive question.
The mode of data collection, the presence of an interviewer
in the data collection, makes a big difference on the
amount of misreporting to sensitive questions. And I will show
you in the subsequent segments a few examples of that.
If you remove the interviewer, if it's self administered and you
have bystanders in the survey, you get more reports on sensitive behavior.
But they are also techniques, the wording of the question itself,
that you can use in order to improve reporting to those items.
And that's what we're going to talk about in the next segment.
