The Official September 2014 Study Group Forum

[Gray]

.

[bound]

Ahhhhh okay I get it. I just always thought that the percentile was what gave the scaled score its weight. Anyway, thanks for clearing that up.

Lol no problem. I really hope I'm not completely wrong... Someone please validate my explanation!?

You're correct I just didn't have the time to explain it further than "it is predetermined" hahah

[Gray]

.

[blacklungz]

Here's my thinking. I was a statistics minor for what it's worth.

Think of your Reported Score and Percentile Rank as the same thing. Those are predetermined. 170 has usually been 97th percentile, 160 is 80th percentile and the median test takers (50th percentile) score at about 151 or 152. However, Raw Score is not predetermined.

Hypothetical: If a certain test is unusually easy and 20% of all test takers only miss 10 questions (extremely unlikely, I know, but just go with me here), then all those people will not be getting a 170 or above (since we usually think of -10 as sufficient for a 170). They would be getting the predetermined Reported Score and above, which in this case (since we made up 20%) would be 160 and above.

I'm almost positive the "curve" has to be determined after the test is administered. Yes, they do administer experimentals but anything can happen when those sections are actually accounting for a student's score.

[chimera]

I need the new/updated medians ASAP.

[Gray]

.

[blacklungz]

Hypothetical: If a certain test is unusually easy and 20% of all test takers only miss 10 questions (extremely unlikely, I know, but just go with me here), then all those people will not be getting a 170 or above (since we usually think of -10 as sufficient for a 170). They would be getting the predetermined Reported Score and above, which in this case (since we made up 20%) would be 160 and above.

In your hypothetical, if the scale was set, in advance, to -10=170, and 20% of people got 10 questions wrong, then 20% of people who took that test would get a 170. That's how it works.

However, the percentile rank would be changed, because 20% at or above 170 probably significant enough to shift the percentiles a bit.

I don't think so. It looks like LSAC has already released it's predetermined Scaled Score and the corresponding Percentile Rank for the June 2014 to February 2015 cycle. It was posted as a bulletin on my TestMasters portal. So they do not change the Scaled Score and corresponding Percentile Rank; those are completely predetermined from three years prior (meaning one test cannot change this). Mathematically, 20% cannot get a 170+ if the Percentile Rank for that score is 97th percentile. In my hypothetical, those people would in fact be receiving only a 160+ and I believe my argument still stands. If you can show me that they change this, then I'll take back what I said.

[mornincounselor]

Post removed.

[Learn_Live_Hope]

I wouldn't expend an ounce of energy or a minute of the day thinking about curves. It's 6 weeks till the LSAT !!

little more than six

[Gray]

.

[ilikebaseball]

I wouldn't expend an ounce of energy or a minute of the day thinking about curves. It's 6 weeks till the LSAT !!

its just fun to talk about. Especially since its more likely to be generous than not. It may boost some people's confidence a tad

[flash21]

that feel when you do an LR section, and BR it, and end up with the exact same score.

BRB, never improving in LR after months of drilling

[longhorn65]

I wouldn't expend an ounce of energy or a minute of the day thinking about curves. It's 6 weeks till the LSAT !!

little more than six

43 days 17 hours and about 59....make that 58 min depending on geographic location you have less or more.

[blacklungz]

I don't think so. It looks like LSAC has already released it's predetermined Scaled Score and the corresponding Percentile Rank for the June 2014 to February 2015 cycle.

But the percentile is based on the last three years of tests, not the individual test, as per LSAC's website...???

(Check out the previous page of this thread for further explanation... I think we've been over this)

In short, they are really good at scaling the test so that the score distributions/percentiles are always about right, but if they eff that up, the percentile is NOT tied to the scaled score and it WILL change.

The raw score required for a certain scaled score does not change after the test has been written, unless an item is removed from scoring.

I wouldn't expend an ounce of energy or a minute of the day thinking about curves. It's 6 weeks till the LSAT !!

Sighhh... this is probably TCR....I'm done.

"A percentile rank is also reported for each LSAT score, reflecting the percentage of candidates scoring below your reported test score. The percentile for a score is based on the distribution of scores for the three-year period prior to the year in which the score is reported. Examine your report for further details."

It's towards the bottom of this website: http://www.lsac.org/jd/lsat/your-score. I can also screenshot the TestMasters bulletin they posted if you really want. It literally says this new Reported Score/Percentile Rank will apply to all tests administered between June 2014 and February 2015.

Btw, this is how I take breaks from studying for the LSAT. I can't seem to evade it no matter what I do.

[Gray]

.

[bound]

I don't think so. It looks like LSAC has already released it's predetermined Scaled Score and the corresponding Percentile Rank for the June 2014 to February 2015 cycle.

But the percentile is based on the last three years of tests, not the individual test, as per LSAC's website...???

(Check out the previous page of this thread for further explanation... I think we've been over this)

In short, they are really good at scaling the test so that the score distributions/percentiles are always about right, but if they eff that up, the percentile is NOT tied to the scaled score and it WILL change.

The raw score required for a certain scaled score does not change after the test has been written, unless an item is removed from scoring.

I wouldn't expend an ounce of energy or a minute of the day thinking about curves. It's 6 weeks till the LSAT !!

Sighhh... this is probably TCR....I'm done.

"A percentile rank is also reported for each LSAT score, reflecting the percentage of candidates scoring below your reported test score. The percentile for a score is based on the distribution of scores for the three-year period prior to the year in which the score is reported. Examine your report for further details."

It's towards the bottom of this website: http://www.lsac.org/jd/lsat/your-score. I can also screenshot the TestMasters bulletin they posted if you really want. It literally says this new Reported Score/Percentile Rank will apply to all tests administered between June 2014 and February 2015.

Btw, this is how I take breaks from studying for the LSAT. I can't seem to evade it no matter what I do.

I'm honestly not understanding what your position is on this.

[blacklungz]

I don't think so. It looks like LSAC has already released it's predetermined Scaled Score and the corresponding Percentile Rank for the June 2014 to February 2015 cycle.

But the percentile is based on the last three years of tests, not the individual test, as per LSAC's website...???

(Check out the previous page of this thread for further explanation... I think we've been over this)

In short, they are really good at scaling the test so that the score distributions/percentiles are always about right, but if they eff that up, the percentile is NOT tied to the scaled score and it WILL change.

The raw score required for a certain scaled score does not change after the test has been written, unless an item is removed from scoring.

I wouldn't expend an ounce of energy or a minute of the day thinking about curves. It's 6 weeks till the LSAT !!

Sighhh... this is probably TCR....I'm done.

"A percentile rank is also reported for each LSAT score, reflecting the percentage of candidates scoring below your reported test score. The percentile for a score is based on the distribution of scores for the three-year period prior to the year in which the score is reported. Examine your report for further details."

It's towards the bottom of this website: http://www.lsac.org/jd/lsat/your-score. I can also screenshot the TestMasters bulletin they posted if you really want. It literally says this new Reported Score/Percentile Rank will apply to all tests administered between June 2014 and February 2015.

Btw, this is how I take breaks from studying for the LSAT. I can't seem to evade it no matter what I do.

Dude I literally quoted that exact LSAC thing on the previous page. Bolded explains everything I'm trying to say.

Well then you're supporting what I'm trying to say. LSAC wants to see a normal distribution with the majority at 151 or 152 no matter how easy/hard the test is. If 20% of test takers get -10, you can be sure that the curve will be tight as hell.

Check this out to see how it works: http://en.wikipedia.org/wiki/Normal_distribution

[Gray]

.

[blacklungz]

I don't think so. It looks like LSAC has already released it's predetermined Scaled Score and the corresponding Percentile Rank for the June 2014 to February 2015 cycle.

But the percentile is based on the last three years of tests, not the individual test, as per LSAC's website...???

(Check out the previous page of this thread for further explanation... I think we've been over this)

In short, they are really good at scaling the test so that the score distributions/percentiles are always about right, but if they eff that up, the percentile is NOT tied to the scaled score and it WILL change.

The raw score required for a certain scaled score does not change after the test has been written, unless an item is removed from scoring.

I wouldn't expend an ounce of energy or a minute of the day thinking about curves. It's 6 weeks till the LSAT !!

Sighhh... this is probably TCR....I'm done.

"A percentile rank is also reported for each LSAT score, reflecting the percentage of candidates scoring below your reported test score. The percentile for a score is based on the distribution of scores for the three-year period prior to the year in which the score is reported. Examine your report for further details."

It's towards the bottom of this website: http://www.lsac.org/jd/lsat/your-score. I can also screenshot the TestMasters bulletin they posted if you really want. It literally says this new Reported Score/Percentile Rank will apply to all tests administered between June 2014 and February 2015.

Btw, this is how I take breaks from studying for the LSAT. I can't seem to evade it no matter what I do.

I'm honestly not understanding what your position is on this.

I'm saying that if a Reported Score and a Percentile Rank is predetermined, it's almost next to impossible to have a predetermined curve (amount you can miss). They can certainly have a rough estimate, but if they adhere to the strict Score/Percentile relationship, they need to have the flexibility to adjust the curve after they see the results.

[blacklungz]

Well then you're supporting what I'm trying to say. LSAC wants to see a normal distribution with the majority at 151 or 152 no matter how easy/hard the test is. If 20% of test takers get -10, you can be sure that the curve will be tight as hell.

Check this out to see how it works: http://en.wikipedia.org/wiki/Normal_distribution

No, no, I do understand how normal distribution works. The thing is that if 20% of test takers get -10, LSAC has fucked up IN ADVANCE, because the test was way too easy/the scale was all off.

Technically, they could have not seen a problem earlier and ran into one on the actual test those sections were administered (again, highly unlikely, but possible). And like I said above to someone else, they can certainly have a range in which they think the curve will land, but the actual figure is determined after they see the results. If it wasn't, why not release that information as well in addition to the Scaled Score/Percentile Rank that they already release a year in advance?

[Gray]

.

[Gray]

.

[valen]

I wouldn't expend an ounce of energy or a minute of the day thinking about curves. It's 6 weeks till the LSAT !!

^^ yep

[schmelling]

.

[P.J.Fry]

I wouldn't expend an ounce of energy or a minute of the day thinking about curves. It's 6 weeks till the LSAT !!

its just fun to talk about. Especially since its more likely to be generous than not. It may boost some people's confidence a tad

You can't look at a "more generous" curve in isolation for a confidence boost. The implication is a harder test. The only way you do better is by getting more questions correct than others whether the test is very easy or very hard.