Top Law Schools

LSAC reports that ~68% of a person's LSAT scores should fall in the range of +/- 2.6 points. Therefore, the score band should be considered.

I didn't major in statistics, but I'm pretty certain that Joe Average's score band of 151+/- is considerably different than Joe Pro's 175+/-. The reasoning being that an individual that doesn't do as well is generally guessing more, which causes considerable scatter in their final scores. Someone who is doing nearly perfect probably isn't guessing on hardly any. So for Joe Average there may be as many as 30 questions that were a guess. For Joe Pro there may be only 1 or 2 that he had to rush a guess on.

It seems like Joe Pro's scores will cluster much more tightly together. In general, measurement error goes down when randomness goes down. People who guess less engage in less random behavior.

What do people think?

Your mistake is in thinking +/- 2.6 points is the equivalent amount of questions answered correctly for a 150 or a 170 scorer.

Scores are approx. normalized and thus 2 points at the 170 level is only one or two questions different, while 2 points at the 150 level is many, many more questions.

Nom Sawyer wrote:Your mistake is in thinking +/- 2.6 points is the equivalent amount of questions answered correctly for a 150 or a 170 scorer.

Scores are approx. normalized and thus 2 points at the 170 level is only one or two questions different, while 2 points at the 150 level is many, many more questions.

Ah, good point.