<p>I am sort of getting sick of all the people who are SO keen on advising you to omit rather than guess on an SAT question. Here is the reasoning for my attack on such a practice:</p>
<li><p>Firstly, if you don’t even guess, then you know you have that question wrong, which seems like giving up to me.</p></li>
<li><p>There are five possible choices (=.20 chance by pure guessing), and they subtract .25 for each wrong answer.
At first look, this seems to make sense: of course you would guess if there is only a .2 chance that you would get it right and they take off that much. However, this is not the case–</p></li>
</ol>
<p>Say you guess on 5 questions; on average, you will get 1 out of those 5 correct, and the other 4 incorrect. This means that you will go up 1 point for the correct answer and down 1 point (4 x .25) for the incorrect answers. So, even just blind guessing can’t hurt you (on average). Usually, you can AT LEAST narrow the choices down by 1, so this leaves you up by .25 over omitting.</p>
<p>That’s my advice, take it or not, it doesn’t matter to me :)</p>
<p>From a statistical standpoint, it would make sense to guess if you could eliminate one answer, and that you would come out ahead if you could eliminate two or more. But alas, the uncertainty of life. Guessing randomly on five questions with one choice eliminated could result in you getting all five right. Or all five wrong. Or any combination of right and wrong. If you ran this trial an infinite amount of times, then yes, the average would indicate that you would break even. But we're only taking the test once - we could get all five right. More likely, we'll get at least three wrong.</p>
<p>The people on CC tend to the brightest of the bright, with most students trying to improve from a 700, or at least a 600. In the ozone layer of test scores, every wrong answer is punished heavily. For students in the upper range of scores, the benefit of getting another point might not be worth the risk of losing a point.</p>
<p>I've gotten numerous 50/50 questions wrong. Statistically speaking, I should come out wayyy ahead if I guessed on every 50/50, no? But even still, there's only a 1/2 chance of getting the answer right. I'm an unlucky man. Again, statistically speaking, you are 100% correct. But we live not in a perfectly mathematical world. So guess at your own discretion. To paraphrase what Stephen Colbert said, we need to rely less on facts and go with our gut. :)</p>
<p>ecthelion, you give the same advice that Princeton Review does. But I have to agree with tdn.</p>
<p>While from a statistical standpoint, it seems to work out, real life is not as tidy.</p>
<p>On practice exams, I always scored lower when I guessed than when I omitted. Luck is involved when you're guessing, the statistical probability will not always be the same as what actually happens. Call it anecdotal evidence (after all, it is), but I've advised friends who've asked, to omit rather than guess, and their scores have always improved.</p>