Fichier:IQ curve.svg

Current IQ tests typically have standard scores such that the mean score is 100 with each standard deviation from the mean counting for 15 IQ points.^[1] The plot shows, assuming that such scores have a normal distribution, the percentage of people getting a score versus the score itself, from 55 to 145 IQ, that is over a span of six standard deviations. Spans are represented with different colors for each standard deviation above or below the mean. In order to create it, the following Octave code was run:

%standard deviation:
sigma = 15;

%IQ values:
IQ=55.5:144.5;

%Gaussian function:
G=exp(-(IQ-100).^2./(2*sigma^2));
H=exp(-(IQ-100).^2./(2*sigma^2));

%Normalisation in the [0:100] range
G=100*G/sum(G);
H=100*H/sum(H);

%to plot them in different colors
G(1:15)=0;
G(31:45)=0;
G(61:75)=0;

H(16:30)=0;
H(46:60)=0;
H(76:90)=0;

%output to text file:
T=[IQ; G; H];
T = T';
save -ascii 'IQ_curve.dat' T;

A symmetric plot was desired, therefore the domain was shifted by 0.5. It saves the samples in an external file called IQ_curve.dat. To get the plot in different colours, the function is split in two different functions: when one is zero the other one is non-null and vice versa. Plotting G and H will get the effect desired. In order to get the SVG, the following Gnuplot code was used:

set terminal svg
set output "IQ_curve.svg"
set xrange [55.5:144.5]
set key off
set xzeroaxis linetype -1 linewidth 0.5
set yzeroaxis linetype -1 linewidth 0.5
set xtics axis
set ytics axis
plot "IQ_curve.dat" using 1:2 with impulses linewidth 1.5, \
     "IQ_curve.dat" using 1:3 with impulses linewidth 1.5

Then I post-processed the file with Inkscape.

1. ↑ Kaufman, A.S. (2009(2009)). IQ Testing 101. New York (NY) : Springer Publishing. pp. 104-109. ISBN 978-0-8261-0629-2.

6 décembre 2006(2006-12-06)

Alessio Damato

Voir ci-dessous