Inappropriate June 2012 Curve Forum

[RodionRaskolnikov]

I'm pretty much just going to sum up my feelings, most of which have been mentioned before, and hope the trolls leave well enough alone.

1) The extra space threw me off. Doing my initial diagram on the left page then having the visual disconnect from the left to the right page took extra time, and I feel had a negative impact on my performance.

2) I recognize the majority of people probably praised their various gods that they had extra space.

3) If you think of an experimental section as an actual experiment, the one variable that is being tested is the only thing that should change (the questions). Everything else is a control and should remain the same test to test. If not, then you can't be sure that the difference from one experiment to another is based entirely on the questions themselves. Just Scientific Method 101 here.

4) The fact that something else changed COULD be a valid concern for the LSAC. Do I think they're going to take it into consideration? Maybe a little. Do I think I'm going to get an extra 3 point curve in my favor? Hell no.

Number 3 is what I was originally going for. I'm glad I'm not the only one who thinks this way.

[RodionRaskolnikov]

The mathematical models that LSAC uses are based on IRT (Item Response Theory). When items are pretested, IRT is used to obtain estimates of parameters such item difficulty, guessability, and discriminating power (ability of item to distinguish between more/less able test takers). LSAC uses something known as item parameter calibration to obtain accurate characteristics of these parameters.

When items are actually used on a test (operational), the responses are also used for improving parameter estimation. However, if there is a noticeable difference ("drift") between pretested items and operational items, the responses cannot be pooled in order to improve parameter estimation. This difference is known as parameter drift.

In the case that LSAC finds questions that do not fit the expected distribution, LSAC will adjust the scale (one of the methods they use for adjusting the conversion scale is throwing out questions that do not follow the expected distribution). LSAC is very good at what they do, and it is very unlikely that the extra space would have caused enough people's responses to deviate from the expected distribution. This isn't as if LSAC just assumed giving extra space would produce no change. Rather, they use mathematical models for detecting such a difference and adjusting to it. I'm almost certain that the extra space won't produce a more lenient curve though.

bingo, it is about the performance of majority people and weather this performance fits the expectation of LSAC or not. Maybe majority people performed better than the expectation because of the extra space, so the curve will be tighter

That is a real possibility. I hope it doesn't happen though

[bobbyh1919]

Unless you had to actually flip a page over to go back and forth between the rules and the questions, this should not have any impact. Some people sit next to students who have a terrible cough. Some take the test in the middle of a thunderstorm, some take it next to a construction site (me), some have rude proctors. Part of success on the LSAT is doing well under circumstances you didn't plan on. Given all of the possibilities, I would say some extra space in the LG section is pretty low on the list of concerns.

[maxmartin]

The mathematical models that LSAC uses are based on IRT (Item Response Theory). When items are pretested, IRT is used to obtain estimates of parameters such item difficulty, guessability, and discriminating power (ability of item to distinguish between more/less able test takers). LSAC uses something known as item parameter calibration to obtain accurate characteristics of these parameters.

When items are actually used on a test (operational), the responses are also used for improving parameter estimation. However, if there is a noticeable difference ("drift") between pretested items and operational items, the responses cannot be pooled in order to improve parameter estimation. This difference is known as parameter drift.

In the case that LSAC finds questions that do not fit the expected distribution, LSAC will adjust the scale (one of the methods they use for adjusting the conversion scale is throwing out questions that do not follow the expected distribution). LSAC is very good at what they do, and it is very unlikely that the extra space would have caused enough people's responses to deviate from the expected distribution. This isn't as if LSAC just assumed giving extra space would produce no change. Rather, they use mathematical models for detecting such a difference and adjusting to it. I'm almost certain that the extra space won't produce a more lenient curve though.

bingo, it is about the performance of majority people and weather this performance fits the expectation of LSAC or not. Maybe majority people performed better than the expectation because of the extra space, so the curve will be tighter

That is a real possibility. I hope it doesn't happen though

well then wish majority people did worse than expected distribution on LG

[RodionRaskolnikov]

The mathematical models that LSAC uses are based on IRT (Item Response Theory). When items are pretested, IRT is used to obtain estimates of parameters such item difficulty, guessability, and discriminating power (ability of item to distinguish between more/less able test takers). LSAC uses something known as item parameter calibration to obtain accurate characteristics of these parameters.

When items are actually used on a test (operational), the responses are also used for improving parameter estimation. However, if there is a noticeable difference ("drift") between pretested items and operational items, the responses cannot be pooled in order to improve parameter estimation. This difference is known as parameter drift.

In the case that LSAC finds questions that do not fit the expected distribution, LSAC will adjust the scale (one of the methods they use for adjusting the conversion scale is throwing out questions that do not follow the expected distribution). LSAC is very good at what they do, and it is very unlikely that the extra space would have caused enough people's responses to deviate from the expected distribution. This isn't as if LSAC just assumed giving extra space would produce no change. Rather, they use mathematical models for detecting such a difference and adjusting to it. I'm almost certain that the extra space won't produce a more lenient curve though.

bingo, it is about the performance of majority people and weather this performance fits the expectation of LSAC or not. Maybe majority people performed better than the expectation because of the extra space, so the curve will be tighter

That is a real possibility. I hope it doesn't happen though

well then wish majority people did worse than expected distribution on LG

If TLS forums are representative of how LG went, then I think the majority of people did worse than expected. But, from what people seem to be saying, good test takers got swindled and bad test takers got a boost. Since there are far more bad test takers than good, I think TLS isn't representative and the majority might have done a tad bit better.

[bobbyh1919]

Good test takers are ones who are well versed with the material and have fine tuned their skills to the point where they can tune out distractions and minor deviations (and yes, this is absolutely a minor deviation) and succeed regardless.

[maxmartin]

Good test takers are ones who are well versed with the material and have fine tuned their skills to the point where they can tune out distractions and minor deviations (and yes, this is absolutely a minor deviation) and succeed regardless.

in the end it won't be how good the good test takers did on LG, but about how well the majority did on LG and compared with LSAC expectations and majority people definitely are not good test takers.

[fronkman]

Rodion can you please explain how good text takers got swindled but bad test takers got a boost from extra space?

[bdeebs]

Good test takers are ones who are well versed with the material and have fine tuned their skills to the point where they can tune out distractions and minor deviations (and yes, this is absolutely a minor deviation) and succeed regardless.

Yeah, I'm pretty sure by "good test taker" he meant "good logic games theorist but bad test taker".

[Mal Reynolds]

Would you guys just indulge OP's every thought until score day? Let this boy cook.

Oh and OP, no matter what the curve is, attributing it to extra space will be wrong.

[RodionRaskolnikov]

Rodion can you please explain how good text takers got swindled but bad test takers got a boost from extra space?

Sure. I'l use a golf example. Suppose there are two golfers: Tiger Woods and some random person who barely played golf. When they go to play golf, Tiger Woods has enough expertise and practice to see nearly every minute detail of the game. Is the wind too fast for 45 degree shot? Is the ground too steep to put the ball a foot to the left? And so on. The person who barely played doesn't see those details. He/She perhaps just sees the ball, the club, and the hole far off in the distance. When they hit the ball, both get a hole in one. But, Tiger Woods has seen all the intricacies that went into it. He took the wind, ground level, crowd noise, type of club, club weight, amount of sunlight, and everything that might affect his swing into consideration and by dealing with all those he got the hole in one. But, the guy who barely played golf just swung and hit the ball. He didn't even know any of the things that went into it. He simply swung and hit a hole in one. Now, if one of the things Tiger Woods takes into consideration unexpectedly changes, Tiger Woods will notice it far more than the guy who barely plays golf. It will alter his gameplay, might even fluster him, while the latter guy doesn't flinch. The same goes with good LSAT test takers who see far more little intricacies in the test than a normal test taker. They're the ones who notice the key words, the phrasing, and the extra space much more than the guy who just goes in to take the LSAT but is relatively average.

[Mal Reynolds]

Tiger Woods wouldn't be flustered by an immaterial change to a course. Well he would now actually but that's because he sucks. Good test takers will still get the same score they should have gotten on this test. If you did poorly it won't be because of the extra space. You just aren't as good as you thought you were.

[RodionRaskolnikov]

Tiger Woods wouldn't be flustered by an immaterial change to a course. Well he would now actually but that's because he sucks. Good test takers will still get the same score they should have gotten on this test. If you did poorly it won't be because of the extra space. You just aren't as good as you thought you were.

There are a lot of statements there, but little evidence. I think the extra space to a good test taker is analogous to a sudden increase in crowd noise, maybe a scream, right as TW swings to hit the ball.

[094320]

..

[Mal Reynolds]

Tiger Woods wouldn't be flustered by an immaterial change to a course. Well he would now actually but that's because he sucks. Good test takers will still get the same score they should have gotten on this test. If you did poorly it won't be because of the extra space. You just aren't as good as you thought you were.

There are a lot of statements there, but little evidence. I think the extra space to a good test taker is analogous to a sudden increase in crowd noise, maybe a scream, right as TW swings to hit the ball.

Lol no. It's more analogous to allowing a player to keep an extra club in their bag. Why am I using your crappy analogy?

[RodionRaskolnikov]

I think the golf analogy is pretty poor.

If by good you mean you learned how to apply methods rotely without actually having logic/analytical skills, then sure, you might mess up. But if you actually have the skills the game is trying to test, it should've made no difference whatsoever.

What is the LSAT? Isn't it the same basic logic tested through similar questions (same save for the wording/names)? A difference between 170 and 177 is the "application methods" you mentioned.

[RodionRaskolnikov]

Tiger Woods wouldn't be flustered by an immaterial change to a course. Well he would now actually but that's because he sucks. Good test takers will still get the same score they should have gotten on this test. If you did poorly it won't be because of the extra space. You just aren't as good as you thought you were.

There are a lot of statements there, but little evidence. I think the extra space to a good test taker is analogous to a sudden increase in crowd noise, maybe a scream, right as TW swings to hit the ball.

Lol no. It's more analogous to allowing a player to keep an extra club in their bag. Why am I using your crappy analogy?

I don't know why you're even in this thread. You're not providing any insight.

[094320]

..

[Mal Reynolds]

Bro dog, I'm saying your analogy is terrible. It's more insight than you have ever provided in your walls o' text. The extra space is beneficial or negligible. Nothing more. If you got rattled by this test you aren't as good of a tester as you think you are.

[VasaVasori]

.

[094320]

..

[fronkman]

Rod I also disagree with your golf analogy and would like you to consider mine. At a well known tournament like the masters lets say a pin placement is used on a par 3 that has never been used before in tournament play. And for this hypothetical lets assume it is neither objectively harder or easier than a typical pin placement. I think that yes some good experienced golfers may be flustered and perform poorly. But I think most good golfers will fall back on their practice and experience and perform better on aggregate than inexperienced underpreppared golfers.

[bobbyh1919]

Rod I also disagree with your golf analogy and would like you to consider mine. At a well known tournament like the masters lets say a pin placement is used on a par 3 that has never been used before in tournament play. And for this hypothetical lets assume it is neither objectively harder or easier than a typical pin placement. I think that yes some good experienced golfers may be flustered and perform poorly. But I think most good golfers will fall back on their practice and experience and perform better on aggregate than inexperienced underpreppared golfers.

This is a much more appropriate analogy. If your skill set is so limited that a change in the amount of space can weaken it, you should worry more about improving that skill set and less about the random nuances that may appear on test day.

[tmon]

Just stop using analogies, bro. They really don't work for you.

[RodionRaskolnikov]

Rod I also disagree with your golf analogy and would like you to consider mine. At a well known tournament like the masters lets say a pin placement is used on a par 3 that has never been used before in tournament play. And for this hypothetical lets assume it is neither objectively harder or easier than a typical pin placement. I think that yes some good experienced golfers may be flustered and perform poorly. But I think most good golfers will fall back on their practice and experience and perform better on aggregate than inexperienced underpreppared golfers.

Good analogy. I get what you're saying. I think I might have been thrown off by the amount of test takers that were thrown off here on TLS so it seemed that a lot of the good test takers got swindled but the ones that usually troll here weren't affected. But I think you're right.

Originally though I was going for an adjustment to the curve being necessitated by the mere change in the layout, regardless of its effect, since that's what you'd do in a scientific experiment. Then it got changed to the curve should be higher and then people started arguing over that. Oh well. I just hope I did well enough. If not, there's always October.