P. 1
High Range IQ Tests (Big) Flaws

High Range IQ Tests (Big) Flaws

|Views: 664|Likes:
Published by Marco Ripà
High Range IQ Tests (HRTs) are not so reliable above 150-155 SD 15...
Here is my (personal) point of view.
High Range IQ Tests (HRTs) are not so reliable above 150-155 SD 15...
Here is my (personal) point of view.

More info:

Categories:Types, Research
Published by: Marco Ripà on May 30, 2013
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as DOCX, PDF, TXT or read online from Scribd
See more
See less

06/04/2013

pdf

text

original

1

HRTs (Big) Flaws

Marco Ripà (sPIqr Society Founder) e-mail: marcokrt1984@yahoo.it

Standard HRTs (Main Flaws):
Case 1)answers selection criteria:

A solution MUST be:

a) A solution; b) Given a), the best solution (the expected answer have to fit the maximum amount of patterns); c) Given a)+b), it have to be the simplest one.

Case 2)answers selection criteria:

A solution MUST be:

a) A solution; b’) Given a), the answer which is “the most correct”, taking into account the WHOLE test (e.g. Asterix test). c) Given a)+b), it have to be the simplest one. Case 1)+2)You should be able enough (lucky) to thinking like the test creator… this is pretty much easier if you yourself are a test creator.

2

N.B. It is possible that a test creator’s mind/creativity is affected by another HRT (one or more) he took before… e.g. you can find some LS36 ideas which have been reproduced by others in their own spatial test(s) (ditto for numerical items)… Further reading (concerning HRTs’ decreasing norms, e.g. PSN, LS36, 916, etc.): http://en.wikipedia.org/wiki/Hundredth_monkey_effect

Case 1), analysis: The basic rules above define a clear logical ranking: it is WRONG by itself… it has been stated by HRTs’ pioneers. This ranking (by difficulty, etc.) was born together with the first HRTs! Top scorers according to their “best/simplest solutions” rankingHRTs Top Scorers

Case 2), analysis: Errors amplification (DOMINO EFFECT)

3

…Thus, at the present time, HRTs are quite meaningless (you can use a supervised test to measure IQs <150 and you are not measuring something which is strongly related to “geniuses” above 150)… in fact, most of Nobel Prizes’ IQs are below 160. There is more… I am a fast thinker rather than an HRT performer (e.g. http://www.scribd.com/doc/128686857/My-Personal-IQ-Profile)… so, what is my true IQ? And my “blitz” IQ was grown taking a given subtest multiple times as you can see from the link

4

above…

Personal comments:
The ability to learn fasterIs this “mutual knowledge (learning)” IQ related? Considering a given IQ score, fixed… the less tests you have taken, the smarter are you? FALSE! It depends also from: 1) You are a pioneerYou started to take HRTs many years ago. 2) Your mind is similar to the test creator(s) one(s). Thus, IQs>150-155, basing on HRTs (at the present time), are not strongly related to geniusesthey are quite meaningless. Moreover… [Another, different, chapter is the risk of cheating: http://interesjournals.org/ER/pdf/2013/April/Rip%C3%A0%20and%20Morelli.pdf] N.B. You can cheat on a given HRT creating a similar test by yourself, scoring it you’ll “steal” mutual knowledge from the test takers/top z-scorers. The “iter” is as follows: Step 1: write down on a piece of paper the items you aren’t able to solve by yourself; Step 2: create your own IQ test using masked items which overlap the same "problem"; Step 3: score 30+ testee and pick their answers on the same paper (see above)... write down on the right the average IQ of the testee you are scoring. Step 4: now, you are ready for your best IQ achievement ever (on the test you wish to ace).

This (asymmetric info) lead to something like “moral hazard”“Adverse selection” concerning the HRTs.

How big this knowledge systematic error would be? I have made a rough estimation basing on my personal experience, analyzing tests and counting how many items can be affected by a mutual knowledge problem (looking at their norm as well)… I have concluded that, for a score around 190 SD 15 in a standard HRT with a ceiling close to 200, the maximum error could be around 23-30 IQ points (let’s say 1.5-2 SD from the mean). E.g. Bob started to take HTRs in 1992, he took 50 tests from 1992 to 2012, lowest/top scores 158195, average 179… first 4 scores (1992-1995): 169-161-175-184. Bob creates a few, widely accepted, HRTs by himself, from 2002 to 2006. My prediction is that Bob’s real IQ have to be 164+ for sure, but it isn’t 190+ (IMO it would be somewhere around 175).

5

There are some other flaws as well… Fat tails, Flynn effect, wrong norms based on linear fit only (the rarity-IQ distribution is not linear!), etc… The real IQ distribution should be as follows (high curtosis):

You're Reading a Free Preview

Download
scribd
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->