http://www.youtube.com/watch?v=HUb71CVWxYU
I've uploaded the 28 multiple choice questions under our classroom resources (as well as our chapter 12 powerpoint) in case you lost it or were absent. Please complete these for homework. Don't forget--this is a big grade!
Here's some help for your homework...
- When considering bias, remember...
- Voluntary Response Bias is always a concern when we have a voluntary response sample; and if we don't have a voluntary response sample, it's impossible to have voluntary response bias. So consider the sampling method to see if this is even a possible bias. Remember, any time we invite people to be surveyed, those with the strongest opinions are more likely to respond and are thus overrepresented.
- Non-Response Bias is when people are given the opportunity to be in a survey, but some do not respond, AND those who don't respond may have a different opinion. If I randomly select 100 people and hand out a lengthy survey, many people won't respond--this survey suffers from non-response bias. Sometimes I give a survey and many are not returned--I can't just hand out more surveys to fix this, as the original non-response bias still exists.
- Undercoverage, on the other hand, is not the fault of the respondents...this is when our survey design fails to cover part of the population. For instance, if I go to one lunch period and hand out a survey I will have overrepresented one of the grade levels at our school, and I have undercovered the other grade levels. This is different from non-response bias because some people weren't even given the opportunity to respond.
- Response Bias is when any part of the survey design influences peoples' responses. This often shows up in the way a question is worded. For example, if I asked, "Considering that AP Statistics is the most applicable mathematics, do you think it should be taught to all students?" The beginning part of the question leads people to answer "Yes," and thus the answer "Yes" will be overrepresented. This suffers from response bias. A better question would ask, "Do you think Statistics should be taught to all students?
- The wording of the question is not the only way we can see response bias. Remember, it's anything in the design that affects a response. If I asked students, face to face, "Do you think I'm an effective teacher?" many people would be inclined to say yes (as they might feel bad or nervous saying no to me, even if they might believe I'm not effective). This design influences responses.
Feel free to email with more questions. I can post any questions I respond to via email on the blog for all of us to use for help.
Enjoy your weekend! Meet up with some friends and do your multiple choice together! See you all Monday for one more week of hard work before break!
No comments:
Post a Comment