Hi, this tutorial covers nonresponse bias. Recall that bias is the tendency for collected data to differ from what is expected in a systematic way. Biased data can offer often favor a specific group of those studied. Biased jeopardizes the accuracy of the data collected. Studies can be carried out and evaluated better if different types of bias can be recognized.
So we're looking at nonresponse bias, also known as participation bias. And it's a type of bias when an individual chosen from the sample cannot be contacted or refuses to cooperate. When nonresponse bias occurs, there is an unrepresentative sample, which is bad.
So let's take a look at an example where we may have some nonresponse bias. So polling company is conducting a study in a certain city and people's attitudes toward their occupation. The company calls random phone numbers each day between the hours of 6 PM and 9 PM.
So this is going to create some nonresponse bias, because those that work during the evening hours will be unable to take part in the study. Because it's 6 to 9, there's going to be some people that just won't be able to be part of it. Many service industry and health care employees work evenings. Since these members of the population may have unique views on their occupation, nonresponse bias has occurred.
So because these people are going to be unable to participate in the study because it's taking place during their work hours, we're going to have some nonresponse bias here. So that's the first example of when nonresponse bias can happen.
Let's look at a second example. Consider a study that examines drug abuse among adults. Many drug users may be unwilling to talk about their views towards drug abuse in light of their own problems. Due to these participation issues, the opinions of drug nonusers would be overrepresented.
So because the segment of the population that uses drugs-- people that would be very important to this study-- might be unwilling to talk about their own use, we're going to have some bias here, again, leaving us with nonusers being overrepresented in the study and users being underrepresented in this study. So again, creating this nonresponse bias, or this participation bias. So that is the tutorial on nonresponse bias. Thanks for watching.
Hi. This tutorial covers response bias. So let's just make sure we have a good working definition of what bias is first. So recall that bias is the tendency for collected data to differ from what is expected in some systematic way. Biased data can often favor a specific group of those studied.
Bias jeopardizes the accuracy of the data collected. Studies can be carried out and evaluated better if the different types of bias can be recognized. So again, we're going to focus on response bias. So response bias is a type of bias that occurs when the response given by someone to a question in a study is not an answer that correctly reflects the respondent's view.
And there are a couple of ways where response bias can creep into a study. The wording of a question, especially if it has some sort of leading wording, will produce some response bias. Sometimes the approach of the researcher can cause some response bias. So if the researcher, in the way that they're asking, is leading the respondent to a specific response, that also can cause response bias. And also, the behavior of the respondent can also cause some response bias. If they're hesitant to answer some sort of question for some reason, that can also-- so if that affects how they answer the question and whether or not that correctly reflects their own view, that can also cause some response bias.
OK, so let's take a look at a couple examples here. So question 1-- so this would be on a survey. So in our town, 25% of car accidents among 18 to 20-year-olds were alcohol-related. Do you support lowering the legal drinking age to 18? And question 2, do you believe that women and men should receive the same pay rate for doing the same job? And just so that we can see that this isn't going to be an anonymous survey, let's suppose this question was asked in a focus group.
So in the first question, you're going to get some response bias there, because the researcher is leading the respondent toward the answer of no, OK? And it's the statement that said, in our town, 25% of alcohol-- or, excuse me, of car accidents among 18 to 20-year-olds were alcohol-related. So that's a pretty negative statistic that maybe would cause people to think, yeah, maybe I shouldn't support lowering the legal drinking age. So there would be an unrepresentative number of "no" answers due to the approach of the researcher and the wording of the question. So again, this would cause the answer not to correctly, possibly, reflect the respondent's view because of the way that thing was worded.
All right, and then in the second question-- remember, the second question was do you believe that women and men should receive the same pay rate for doing the same job? So the respondent would be hesitant to give an answer of "no" because of social desirability concerns. So being opposed to gender equality is sometimes socially undesirable, so perhaps they wouldn't give their own correct view, just because they don't want to be looked at as somebody who doesn't support gender equality.
So the behavior of the respondent in this case would yield an unrepresentative number of "no" answers, again causing response bias. So again, it's important, when you're designing a survey, to try to structure your questions in a way where you won't get response bias. And also, generally having anonymous surveys will also eliminate some of that response bias.
So this has been your tutorial on response bias. Thanks for watching.