Question

please help me with this question :c I will give you a shiny medal! :)


Suppose you are conducting a survey regarding violence in a youth hockey leagues. You obtain a cluster sample of 21 hockey teams within a youth hockey league and sample all hockey players in the randomly selected hockey team. The survey is administered by the coaches.

I'm confused about this because I personally would make a cluster sample since we will be selecting ALL individuals within a randomly selected collection or group of individuals? or am I thinking this wrong? What can be done to solve this?

Which of these best describes the bias in the​ survey?

Sampling bias

Response bias

Nonresponse bias

Undercoverage bias

Answer 1

@jim_thompson5910 Could help explain this for me? I'm sorry for bothering you.

Answer 2

I'm thinking it might be sampling? But I don't understand why it would be, because I think the clustering was something I would do...

Answer 3

hmm let me think

Answer 4

oki doki :>

Answer 5

Does it say what the survey is about? or what questions are on the survey?

Answer 6

oh it says it's regarding violence, but that's a bit vague

Answer 7

it's administered by coaches and its a cluster sample of 21 hockey teams within a youth hockey league that were chosen by random I think.

Answer 8

The question confused me a lot and I don't see what could be a flaw:
Sampling bias (I was thinking this, but I dont think it could be it).
Response bias
Nonresponse bias
Undercoverage bias

Answer 9

One moment, I might have found something

Answer 10

:D yaya

Answer 11

Ok so when you do a cluster sample, you'll randomly select 21 clusters. In each cluster, you'll sample all of the team members. The other clusters are ignored. As you can probably guess, there's a big problem here: the other players are ignored and the clusters selected are probably very homogeneous

homogeneous means "similar"

Sure there will be star players or outliers, but for the most part, teams tend to be cohesive units that work well together. They tend to be the same. They think the same way, they have the same methods, workouts, practices, drills, etc. The coach is the one that helps them all have this same mindset. If everyone was doing their own thing, then it would be a very bad team.

As a general rule, it is a bad idea to use a cluster sample when the clusters are homogeneous. You will only focus on one subset of the population and ignore the rest

Answer 12

So basically what I'm saying is that cluster sampling with homogeneous clusters leads to `Undercoverage bias`

Answer 13

oh so it wont be sampling or response bias? Undercoverage bias is when one proportion of one segment of the population is lower in a sample than it is in the population

Answer 14

Is that why it would be undercoverage?

Answer 15

undercoverage bias is explained here

http://onlinestatbook.com/2/research_design/sampling.html

they point to an example where people without phones weren't included in the sample

Answer 16

it's not response bias because they don't mention anything about leading or loaded questions

Answer 17

hmm sampling bias seems to be the umbrella term for many other types of bias

Answer 18

sampling bias is more general while undercoverage is the actual type of bias I'm talking about

Answer 19

oh I see, I put undercoverage and I got it wrong, :/ :c the program said it was response bias..

Answer 20

I think the program is wrong.

Answer 21

Maybe it's response because the hockey players would want to look good in the eyes of their coaches? I don't know.. :/

Answer 22

it could be right. I'm trying to think of why

Answer 23

I'm thinking the fact that the coaches are in charge is what leads to a conflict of interest. So it's not being done by an impartial/objective third party

Answer 24

So for instance, if the survey asks about violent head injuries and the coach doesn't want the hockey team to change, then they would probably skew the results somehow. That's just a guess really

Answer 25

oh I see.. hmm

Answer 26

I would think the coaches would sample incorrectly, but since the coaches are sampling then maybe so..

Answer 27

that could be it too

Answer 28

but I think the remedy would be to have some outside person/party do the sampling (and not the coaches)

Answer 29

Ohh! I see, I think thats it!

Answer 30

An impartial party would be the solution! 8D

Answer 31

Thanks so much! :D