<p>
[quote]
Because the OP specified that if you hadn't taken the SAT yet then you could still answer the question
[/quote]
Uh, no I didn't. </p>
<p>Anyways, this is supposed to be a random poll. There is no bias, because I'm trying to find the association of income and gpa to sat scores. The data will have scattered points, hopefully positive, and I can then conclude to an equation, etc. Obviously there will be outliers and my residual plot will vary, this is important, so outliers are just as important. Thank you for your cooperation and time, hopefully more will respond.</p>
<p>No offense, but we all understand what you're TRYING to do.</p>
<p>It is not a random sample though. A random sample would be if you went through a phone book and randomly picked phone numbers (even that is not fair because not all people own phones, especially with more people relying on cell phones for their daily calls).</p>
<p>Furthermore, using members from <a href="http://www.collegeconfidential.com%5B/url%5D">www.collegeconfidential.com</a> is a biased sample. First, you are only taking a sample from high schoolers with internet access. Secondly, and most importantly, you are taking a sample from the most elite and interested college-bound students. Not saying that all collegeconfidential kids are elites but they typically are students with a higher level of thinking or interest.</p>
<p>Basically, you're graph is going to have unrealistically high data. For instance, where a 2000 may be all the way in 93rd national percentile, it may only be the 50th percentile among members on this site.</p>