<p>Here is a question for you. I think it is probably fairly easy for anyone well versed in statistics, which is NOT me. What do you think? </p>
<p>The Question:
</p>
<p>But you really need to read the entire question with data provided in the following link. Submissions were closed as of yesterday (so I wouldn’t be accused of asking for help). Answer will be posted in a week.</p>
<p>If anyone graphs tuition increases, they should also compare what’s going on with federal subsidies. Every time the Federal govt increased its student subsidies, tuition went up.</p>
<p>Not going to invest much time in an exam question after just finishing finals, but can’t you easily solve this with a standard normal table? If the excess values follow the std normal distribution, we can calculate the z-score for an excess of zero and then consult the table for our answer.</p>
<p>“noimagination” (I object to your username!) That was pretty much my approach, and I got the same answer. You wrote, “I guess” but you did the calculation, right? I think it worked out to 10.89 years. As I said, I think its an easy problem if you know statistics.</p>
<p>Also, if the data are real, it is kind of interesting. When inflation is high, colleges hold back on their increases. They still raise tuition, but are more likely to keep their increases less than the inflation rate. But when inflation is low, they hike tuition, more than the low inflation rate.</p>
Yes, I did the calculation. Wasn’t sure if the question was supposed to be that simple.
BLS only has tuition CPI data back to 1979. Looking at 12-month percent change, only 1979 and 1980 (both years with very high inflation) show higher inflation for all items than for college tuition/fees. Either 8-9 of the years from 1964-78 also met that condition or the question was based on different (or invented) data.</p>
<p>These are the BLS series I used:
CUUR0000SEEB01
CUUR0000SA0</p>
<p>without burning any brain cells on this, the one fact that caught my eye was r = 0.5. that would make r2 = 0.25. That’s not a very significant correlation.</p>
<p>wow… I just noticed that either their answer is worded incorrectly (meaning their answer is actually wrong) or my answer was wrong. </p>
<p>They wrote:
</p>
<p>“every 10.89 years” over 45 years would mean it happens 4 times in 45 years. I thought the answer was 10.89 (rounded up to 11) times in 45 years, or 24.2%% of the time.</p>
<p>^ Their answer cannot be correct because they give two inconsistent responses: “every 10.89 years over those 45 years” is not “25 percent of the time”. I believe we did this right, and the magazine editors were just a bit careless with their language.</p>
I guess it wasn’t very hard after all…</p>