Anonymous wrote:
Anonymous wrote:I used to do CDC phone surveys as a college student in the late 80s. What a nightmare — some people are not nice! But yes, they are 100% legit and provide important health information.
By picking the numbers randomly, the researchers try to ensure that the surveys avoid any bias that might come from asking people to reach out and answer a survey. (Eg, the “type of person willing to helpfully answer a survey” could be correlated with things like general “behavior that’s helpful to a group,” which here might make it look like more people are in favor of vaccines than a completely random sample.) Don’t know how it works these days, but because of the randomness factor they tried very hard to follow through once a number was randomly chosen. If we part-time employees couldn’t get it done after several tries, we kicked it up to management for even more attempts.
Today you get the bias of people who have landlines. And answer them. I don't think the venn diagram is great intersected with people who have kids and will answer questions on vax.
DP but I mean, yes, clearly random digit dialing has bias problems. It does include cell phones now.
The researchers may combine random digit dialing with other methods of sampling or the use of a panel. They will also be checking their responses and perhaps over sampling for groups that are underrepresented. And they will also go on the back end and try to adjust their data for response biases they think they can find.
It’s really hard to do! A young, but important science.