<p>I have a question about null hypotheses:</p>
<p>In question 6, part a) of last year’s statistics FRQ, it states that a consumer organization is concerned that a car company is overstating the fuel efficiency of their car model and wants to investigate. The company states it as 27 mpg.</p>
<p>What are the null hypothesis and alternate hypothesis?</p>
<p>I thought the null hypothesis was x ≥ 27 and x < 27 is the alternate hypothesis.
(where x is the actual efficiency of the car model).
However, the solution given by the college board says that
the null hypothesis is x = 27 and the alternate is x < 27.
However, the consumers are not concerned if the company understates the efficiency, so wouldn’t my assertions also be correct?</p>
<p>For reference:
<a href=“Supporting Students from Day One to Exam Day – AP Central | College Board”>Supporting Students from Day One to Exam Day – AP Central | College Board;
<a href=“Supporting Students from Day One to Exam Day – AP Central | College Board”>Supporting Students from Day One to Exam Day – AP Central | College Board;
<p>NO NO NO NO NO!!!
You have to ALWAYS set the null equal to something, not </=, not >/=, not >, not <. IT HAS TO BE =. The alternative can be anything. keep that in mind.</p>
<p>Also, if it’s like. The population will be happy if the car’s efficiency is 36 mpg or higher. The null is at 36 mpg. You’ll test the alternative hypothesis and set it to H(alternative) < 36.
And for clarification. I just noticed i made a mistake up there. The alternative can’t be everything. Alternative hypothesis can only be <, >, or does not equal (the equal sign with a cross on it) Sorry about the confusion!</p>
<p>Oh, that’s terrible… my CliffsAP book says that null hypothesis can be ≥ or <= as well as =; it uses this notation in many examples.</p>
<p>WOAH! I must double check now.</p>
<p>I just checked my text book, Barron’s book, and AMSCO’s book. all says H(0) = the hypothesized value,</p>
<p>the alternative hypothesis H(a) > hypothesiszed value, or <hypothesized value, or DOES NOT EQUAL to hypothesized value</p>
<p>Also, I actually took the 2009 AP STATS FRQ as my final exam last week. my friend got points off for what you did, and my teacher used to be AP grader. I think it’d be safe just to stick with = sign.</p>
<p>CliffsAP is teaching you the wrong way, even though I believe statisticians do do that for null and alternative hypotheses. </p>
<p>However, like AznCo0Co0 said, null is always =, whereas alternative is “not equal”, <, or >. Just do it this way for the AP exam.</p>
<p>thanks guys
darn, the FRQ scoring guidelines are really really harsh
even if i can get a perfect on the MC, which i probably can, i still am worried that i may not get a 5…</p>
<p>Actually, I am confused about question 4 as well… When do you use the t-distribution, as opposed to the z-distribution? The sample size is 50 and 50 > 30 so you should be able to use it, right? Is it because the populations may not be normal?</p>
<p>I hate the AP Stats FRQ grading. It’s almost impossible to get a 4 on a question. I’m aiming for one 4, three 3’s, one 2, and on the 6th question, a 3. That’ll be a easy 5 then.</p>
<p>This concept is called directionality. Wikipedia actually has a fairly good article on the topic: [Null</a> hypothesis - Wikipedia, the free encyclopedia](<a href=“Null hypothesis - Wikipedia”>Null hypothesis - Wikipedia)</p>
<p>In general, the effective null hypothesis (=) is what the AP graders will expect.</p>