AP Stats

<p>lol yea i was about to write homeless but i didn't wanna make it sound like people without high school diplomas are all homeless. i changed my wording to make it say that they might not be as wealthy and therefore may not be able to afford telephones.</p>

<p>What did yall get for the Confidence Interval for #6? I wasn't sure whether to use n=9 or n=18. I used 9 since thats the number of differences if that makes sense.</p>

<p>It's really 9 and on the calculator it was actually 8.7 or something but since I only had the table to work with, I just chose the t value at df = 9</p>

<p>I just did a T Interval on my calculator and used -13 I think as X bar and the standard deviation of the differences of the means which equaled 4.854 and n as 9 and got an interval of (-16.73,-9.269)</p>

<p>Mine was (-16.6, -9.43) actually. 2-sample T interval which is just x1-x2 +/- t* (s1^2/n1 + s2^2/n2)^1/2</p>

<p>df was 8.537 exactly</p>

<p>Did you have to do it that way or could you do it with a T-Interval with the differences as the sample numbers.</p>

<p>Not sure, I was actually stunned with the question at first since we really never did much practice two-sample t intervals so I have no idea right about this point. I'm not even completely certain if I'm right.</p>

<p>I concur with Meng, I used a two sample t-interval for unpaired means. I can see why you would have done it paired, but I don't think it was appropriate, not 100% sure though.</p>

<p>Although, after getting the problem from AP Central, I am now not sure which is the proper method because I think they may be match pairs even though the sample is not the same.</p>

<p>my stats teacher said that it must be a 2-samp T-int. if you did matched pair, you are wrong b/c you are not sure if each variable is independent of each other.</p>

<p>I did just a T-int with the differences :( i hope i get partial credit.</p>

<p>oooops, would i get even a 1/ if i did everything right except lol i made it a z ?...</p>

<p>you could have done a one sample t-interval with the differences. I subtracted the mean of one list from the mean of the other an got -13, which is the same thing that the two sample people got, so i am pretty sure i did it right. a bunch of people did it that way. you wouldnt say it was a matched pairs though.</p>

<p>lol for the bias ones, your guys answers were pretty...out there. I just put how the time of the day can cause bias, because people who have a diploma are more likely to have jobs and be out of the house at work, thus not being able to answer the phone and underestimating the number of people with high school diplomas while overestimating the people without because they could likely be home and able to answer the phones.</p>

<p>I think that bias would have been appropriate also.</p>

<p>I think the 1 sample t test would have been appropriate, but I really will have to ask my stats teacher on monday. I'll post what he says.</p>

<p>i said that it was lack of privacy, those who did not have a high school diploma may feel embaressed and would not want to report it, making the proportion of those who do not have a high school diploma be lower than the true population</p>

<p>Of course there was bias involved. If there was a sample taken by telephone numbers EVERYONE WITHOUT A TELEPHONE would not be represented! People without telephones tend to be poorer. People that are poorer tend to have low education. Therefore the proportion of adults that do not have a high school diploma would be UNDERESTIMATED! Tell me if I am wrong, but was this question not one of the main backbones of AP Statistics. All telephone surveys have nonresponse!</p>

<p>Nonresponse is what I put down at first. But then I didn't feel like it was the best to answer the second part of the question referring to the adults with did not have diplomas.</p>

<p>So I settled with what KingFalcon used as well - response bias as some would be ashamed to admit that they did not have diplomas. Therefore, the proportion of adults without diplomas would be lower than it really is.</p>

<p>Does anyone have all of the questions worked out correctly that they could share?</p>