<p>One of my favorite blogs, Freakonomics, describes Harvard research that shows male SAT takers guess more than females. This is attributed to the penalty for incorrect answers; women are apparently more inclined to leave a blank answer in that situation.
[quote]
Baldiga’s results might help explain why women often do better in college than their SAT scores would have predicted and raise an important question: Are multiple-choice test scores the best way to fairly “measure aptitude and forecast future achievement”?
<p>Perhaps another reason to support the ACT – no penalty for guessing. :)</p>
<p>Floridadad: according to the study, the authors conclude that the lack of guessing does negatively impact scores, some 70-80 points on a full length SAT.</p>
<p>I’m a male, and I feel more inclined to guess on an answer rather than leave it blank. But the thought of getting points marked off for guessing wrong irked me, so I always second-guessed my answer. With the ACT, however, I knew I should guess and move on, which worked quite well, and so my ACT ended up being better.</p>
<p>Floridadad55, I’m not proposing anything, just reporting the findings. I actually believe in standardized tests as one element of a comprehensive admissions process.</p>
<p>But, if this research holds up, I could see the SAT folks being pressured to remove the guessing penalty.</p>
<p>^^wonder if eliminating that guessing penalty would eliminate the misconception that the ACT is easier than the SAT and, in addition, that many perform better on the ACT…</p>
<p>So if the stats show that Italian Americans do not guess as much as Jewish Americans, does that mean that the test is “biased” against Italian Americans.</p>
<p>Or what if people from the midwest guess less than people from California?.</p>
<p>No test is perfect.</p>
<p>But I have found in life that people with high SAT scores are indeed smarter academically than those that have lower scores. By the way, I myself only got a 1200 (in the days before the writing section). My son, who evidently got his mother’s genes, scored a 2300.</p>
Of course it will. There is a subtractive penalty for wrongly answered questions vis-a-vis no penalty for unanswered questions. Thus does a testtaker’s personality enter in to the score: the more risk-tolerant are apt to answer when lacking certainty, the more cautious to leave it blank. The penalty is such that the elimination of at least one answer choice on a question improves your odds to random or better, but how many testtaker’s are this careful in their evaluation before venturing a wild guess? Moreover, how many are not wildly guessing but, lacking reasoned certainty, are scared to trust to still-useful intuition.The penalty discourages the use of intuition, too.</p>
<p>There really is no reason why this should be a factor in a test and it could easily be removed by dispensing with the guessing penalty as the ACT does.</p>
In order to answer this question, I suppose one could look at ACT scores, where the penalty does not apply, and see if gender-differentiated performance gaps are just as severe there as with SAT scores.</p>
<p>And I agree that the SAT, to an extent, measures intelligence. Those who score low are likely not as smart as those who score high. Of course, 50-80 points shouldn’t be seen as much. Depending on the day, a kid with a 2200 could get a 2100 or a 2260…</p>
<p>Is the abstract implying that females can eliminate wrong answers in some cases and still choose not to answer questions because they are not certain of the correct answers? If it’s not implying that, then what is it implying? That blind guessing actually helps? I don’t see any merit to that. If women choose not to guess in cases in which it is propitious to guess (like when they can eliminate one or more answers), then they need to adopt a new testing strategy because they are working themselves against the odds. Whatever the tendency of a specific sex might be, they can be taught the best strategy for guessing. Eliminating the guessing penalty might be one way of solving the “problem,” but it’s not neccesarily necessary to do that.</p>
<p>I think the suggestion is that guessing is usually a good idea except on those questions where any answer seems equally probable. Therefore, averaged across many individuals, moderately aggressive guessing (though not totally random guessing) will result in higher scores.</p>
<p>A risk-averse individual, female or otherwise, might avoid guessing even if one or two possibilities could be safely eliminated. This risk aversion could result in lower scores, on average.</p>
<p>It’s common sense to guess. If you can safely eliminate one answer than assume your guess isn’t totally blind and more likely to be right than the 1/4 chance you get from blindly guessing for the remaining answers you had after you eliminated the obviously wrong answer it is to your benefit to guess.</p>
<p>Guessing completely at random for five-choice questions results in no benefit, on average, for a 0.25 point penalty (which is why it is 0.25 points, of course). Arguably, the time spent bubbling each guess is a disbenefit. I find that, typically, people have a hard time eliminating answers “for sure”. On the SAT it is unusual to have a math problem ask for a positive number and yet have some negative answers (for example). Or, when you do get these atypical problems, people don’t notice the hints that telegraph which are the wrong answers.</p>
<p>Anyway, the M - F SAT score differences have been about 30 points (3 questions or so in the average score range) for about the last 15 years. This is about 0.25 standard deviations. The M - F ACT score differences have been about 1 point (roughly 2 questions in this score range), or about 0.2 standard deviations. I hesitate to draw any conclusions from this.</p>
<p>The problem with everything related to the SAT is people attempt to draw generic conclusions from individual experiences. In so many words, there is no OVERALL validity behind the recommendation that guessing helps to a higher score. Poor guessing contributes to poor scores. There are cases where guessing could be productive but it is rarely for the people who rely on guessing. Simply put, bad testers do guess. Good testers do not, unless in very specific conditions. </p>
<p>Fwiw, all the theories that support guessing fall short because of a very simple fact. The probabilities are based on a tester being able to eliminate bad answers with certainty. Well, if someone can pick a bad answer, why struggle with picking the correct answer? Where do students guess the most on the SAT? Math or Verbal? Why is verbal the hardest section to increase one’s score? Guess where students lose the most points by following the advice to guess when being able to eliminate 2 or 3 bad answers? Starting to get the point? </p>
<p>All in all, guessing is almost never a good idea. And that is why ETS/ and TCB do not hesitate to recommend it. Think about it. And, if you do not believe, take two official tests and compare your results with usual guessing and without ANY guessing at all.</p>
<p>PS That research based on SAT Subjects has ZERO relevance to the SAT.</p>
<p>In a sort-of related tangent: When the CB removed the guessing penalty on AP Exams for the 2011 testing year, our school’s AP Coordinator said that a main factor behind the decision was the fact that males guessed more than females did. I haven’t found anything from the CB that says this, but take my school’s Coordinator’s word for what it’s worth.</p>
<p>Though…maybe the SAT’s guessing penalty will go away in the near future?</p>