<p>So the formula above is an approximation. Usually I say that that formula comes with +/- of about five points, which is an awful lot. I have two theories about how you can go further.</p>
<p>This is an important disclaimer. The formula itself is based on nothing other than anecdotal evidence. It is literally just me asking my friends about how their LSAT and SAT scores compare. It is based on like fifteen data points, which is just pathetic and self-selected. However, it seems to work pretty well.</p>
<p>This is an even more important disclaimer. The two theories which I am about to state are based on NO EVIDENCE WHATSOEVER. Not even anecdotal evidence. So really, really, don't believe this. This is just me theorizing. On top of that, the two theories sometimes even contradict. So really, seriously, don't take this too seriously.</p>
<p>Theory #1: The +/- 5 is determined by your relative studying. If you study equally hard for both tests, you'll end up at about +0. Roughly. If you study much harder for the LSAT, you'll end up at about +4ish. If you studied hard for the SAT and really blew off the LSAT, you'll end up at -4.</p>
<p>I think this theory is false. The most important reason I think it's false is that I see just as many (-)'s as (+)'s according to this formula, and I really don't think very many people study LESS for the LSAT than the SAT. So this formula would normally make a lot of sense, but I think it's wrong.</p>
<p>What it might be, however, is just that it's mis-centered. Maybe equal studying will net you like a -3 or something. That would be a pretty simple correction.</p>
<p>Theory #2 is not mutually exclusive with theory #1. Again, none of this is backed up by any evidence -- not even anecdotal evidence. So please don't take this as gospel. I am saying this over and over again in the hopes that I am being clear.</p>
<p>Theory #2 requires that you take a diagnostic. It is usually most reasonable to expect 10-15 points of improvement from that score. Obviously the higher you start, the less improvement you can expect. (For example, you cannot start off at 174 and expect 10 points of improvement.) But let's start with 10-15 as our estimate. This will give you a five-point range. Our original formula gives you a ten-point estimate.</p>
<p>Let's take a kid who scored a 1560 on the SAT and a 159 on his first diagnostic. We then have two predictions: 171-180, and 169-174. You can see that these two estimates overlap by three points: 171-174. This, then, would be his best estimate. Now pretend that kid got a 1560 on the SAT but a is 151 diagnostic. The ranges no longer overlap: 171-180 and 161-166. You'd still use the SAT-based prediction, but now you'd choose the very tail end of it (that is, the 171.)</p>
<p>I cannot stress enough that this is literally just me sitting around making stuff up. I haven't even checked this anecdotally. But that's my first thought.</p>