On Mega Admission Standards

Kevin Langdon

(reprinted from Noesis #135)

 

One year ago, in Noesis #125, I presented a table of comparative statistics on high scorers on the LAIT and the Mega Test, introduced by this paragraph:

I have examined much of the available data on the tests we accept. Here are statistics on the highest scores on the LAIT (from Sigma Four #5, January 1980; out of 15,000 testees) and the Mega Test (sixth norming report, May 1989; out of 3920). The [seventh] column shows the number of Mega testees at each level per 15,000, to facilitate comparison. As about 27,000 people have taken the LAIT, I estimate that the current totals are about twice the figures for this early sample.

My original table had only five columns. I have added three more columns to facilitate comparison between the two distributions. Two of the columns added show cumulative totals for the LAIT and the Mega Test.

The mean score for the LAIT sample was 2.84 sigma, approximately one in 495. The mean score for the Mega sample was 3.26 sigma, approximately one in 1735. In order to estimate the number of very high scores to be expected in a Mega sample of the same size as the LAIT sample (15,000) if the tests were of equal difficulty, the cumulative number for each LAIT score was adjusted by 1735/495 = 3.5 (shown in the third added column). Columns 4 and 8 are most directly comparable.
     

LAIT (N=15,000)

Mega Test (N=3,920)

I.Q. Number Cum. X 3.5 Raw Number Per 15K Cum.
176 0 0 0 48 0 0 0
175 2 2 7 47 1 4 4
174 0 2 7 46 1 4 8
173 7 9 32 45 2 8 16
172 15 24 84 44 3 12 28
171 2 26 91 43 6 23 51
170 14 40 140 42 12 46 97
169 28 68 238 41 15 57 154
168 16 84 294 40 7 27 181
167 43 127 445 39 13 50 231
166 53 180 630 38 15 57 288
165 25 205 718 37 18 69 357
164 79 284 994 36 27 103 460


The similarity of the distributions is obvious--and to be expected, given that both samples were predominantly composed of Omni readers. The Mega Test has a slightly higher ceiling than the LAIT. There is a further question regarding the general population percentile corresponding to each score. My norms place the ceiling of the LAIT at 176; my study of Ron's data placed the ceiling of the Mega Test at 178; Ron contends that the ceiling is as much as ten points higher than this. If Ron is right, it follows that the ceiling of the LAIT is much higher than 176.

In my opinion, it's stretching things considerably to claim that these tests are accurate at their ceilings or a couple of points below them. The one-per-million level occurs at 176 on the LAIT and 46 on the Mega. Even if we allow one point for ceiling bumping, fewer than fifteen people have made scores this high on my tests or Ron's. I suggest that we lower our percentile cutoff to 99.9997, one in 300,000, 4.5 sigma, Mega/Titan 43, LAIT 172. Other views are invited.