LSAT Test Construction

[Broncos15]

Hello,

i had two questions regarding test construction

1. What criteria does LSAC use whether or not to alert test takers of any changes in composition prior to the test; .... For instance LSAC notified test takers before the test of the new comparative passages in RC in 2007 ....but did not announce they would throw in old Logic Games from the 90's/early 2000's on recent LSATs ( the Feb 2014 exam which had a circular game and the June 2014 pattern)...(I am interested in this but was more curious as to the bigger picture as to what principles/ criteria are used in making these decisions)

I understand that at a certain point LSAC needs to alert tests takers of any changes to keep it a "standardized exam"since LSAT scores are valid for 5 years....so at least in theory someone who got a 170 on the first LSAT with comparative reading demonstrates the same type of performance as someone who scored a 170 on the exam prior to this switch.

2. How do you feel the is constructed in terms of balance?.....Are all four sections equal in terms of difficulty or do they offset....for instance could one LSAT a relatively harder Reading Comp and then inserted a relatively easier Games section to compensate and then another LSAT have the vice versa ( harder LG, easier RC) ( again, any relative difficulty changes would be minimal to keep it standard)..i know within each section their is question difficulty variance but was curios about how sections as a whole compared to each ther

[Nikki Siclunov]

Hi Broncos15,

The introduction of a comparative reading section in 2007 was a substantive design change to the test, so they obviously had to notify us. Circular or pattern games may seem unusual, but they’ve been around for awhile and don’t change the composition of the test. Hence, no prior warning was necessary. Obviously, the test is curved, which should take care of any abnormally high or low difficulty levels and ensure that scores are comparable across different administrations.

To answer your second question, while they do attempt to equalize the difficulty within each section (balancing out easy games with difficult ones), it's hard to say if the same always holds true at the test level. Common sense would say yes: if you want to keep difficulty at a constant, a particularly easy games section should theoretically be balanced out by a more difficult RC section, for instance. If all the sections were relatively easy, the curve would need a significant readjustment, which is highly unusual. Overall, variances in difficulty are minimal and should not be a cause for concern.

Thanks!