IQ: can intelligence really be measured?

The pur­pose of the intel­li­gence quo­tient (IQ) is to estim­ate a per­son’s intel­li­gence, using a tool made up of a series of ques­tions: the IQ test. Taken in con­junc­tion with a psy­cho­lo­gist, the test pro­duces a score which can be com­pared with the aver­age for the pop­u­la­tion in ques­tion, set at 100. But to what extent can IQ tests be used to assess intel­li­gence? Jacques Grégoire, Emer­it­us Pro­fess­or of Psy­cho­logy and Psy­cho­met­rics at the Cath­ol­ic Uni­ver­sity of Louv­ain, looks at some of the main mis­con­cep­tions sur­round­ing IQ.

The intel­li­gence quo­tient (IQ) is some­times con­sidered to be an object­ive meas­ure of intel­li­gence, but this is not the case. Intel­li­gence is a qual­ity that can­not be seen, so how can it be meas­ured? From the begin­ning of the 20^th Cen­tury onwards, research­ers refined their know­ledge of the struc­ture of intel­li­gence. This has led to the devel­op­ment of major sets of tests to assess vari­ous facets: verbal, visu­al-spa­tial, work­ing memory, etc. Cur­rent tests estim­ate an indi­vidu­al’s intel­lec­tu­al level based on the res­ults obtained in the vari­ous exercises.

But it should be noted that this is only an estim­ate. Firstly, the score is a rel­at­ive meas­ure: it defines where the indi­vidu­al stands in rela­tion to the aver­age for the group to which they belong, which is set at 100. What’s more, any meas­ure – in psy­cho­logy, as in any oth­er field – is sub­ject to error. This is why IQ should always be giv­en with a con­fid­ence inter­val, for example of plus or minus five points.

What’s more, an IQ test is sub­ject to many uncer­tain­ties: the con­di­tions under which the test was con­duc­ted, the object­ive sought, the delib­er­ate or inad­vert­ent errors made, etc. As a res­ult, it is imper­at­ive that the res­ult obtained be accom­pan­ied by an ana­lys­is by a trained prac­ti­tion­er. Without inter­pret­a­tion, an IQ score means nothing.

IQ is con­struc­ted in such a way that, with­in an inter­val of plus or minus one stand­ard devi­ation around the mean (from 85 to 115), we find 68% of the pop­u­la­tion. Con­sequently, hav­ing a score below 100 is cer­tainly not a sign of intel­lec­tu­al disability.

There is, how­ever, a stand­ard whereby a score of two stand­ard devi­ations below the mean (less than 70) is con­sidered to be a sign of men­tal dis­ab­il­ity. But this score alone is not enough to make a dia­gnos­is. It must always be accom­pan­ied by oth­er tests or exam­in­a­tions, to provide a more detailed analysis.

Espe­cially as the res­ult of an IQ test can be influ­enced by many factors. For example, fatigue, drug use or a psy­chi­at­ric con­di­tion can affect an indi­vidu­al’s per­form­ance. This means that IQ tests are not always reliable.

An IQ test is first and fore­most a clin­ic­al tool, used to meet a spe­cif­ic object­ive: to dia­gnose men­tal dis­ab­il­ity, to assess the cereb­ral con­sequences of a cra­ni­al trauma, to detect a deteri­or­a­tion in cog­nit­ive fac­ulties due to age­ing, and so on. The ques­tion must there­fore be of a gen­er­al nature: what is the prob­lem to be solved? The IQ test may – or may not – help to achieve this goal, but it is not an end in itself. Some people want to con­sult a pro­fes­sion­al to find out their IQ, often in the hope of obtain­ing an excep­tion­ally high score. But what’s the point? Even in a pro­fes­sion­al or edu­ca­tion­al envir­on­ment, IQ is not enough to pre­dict per­form­ance. Oth­er indic­at­ors may be more rel­ev­ant, such as res­ults in exam­in­a­tions, com­pet­i­tions or tech­nic­al tests. The IQ test should not be used as an abso­lute ref­er­ence to clas­si­fy indi­vidu­als; that is not its purpose.

Just over forty years ago, IQ tests showed gender dif­fer­ences: girls per­formed less well on visuo-spa­tial func­tions (abil­ity to visu­al­ise in 3D an object rep­res­en­ted in 2D) and bet­ter on verbal tasks. Today, these dis­par­it­ies have com­pletely dis­ap­peared. This means that there is no sci­entif­ic data to sup­port a gender-spe­cif­ic ori­ent­a­tion. On the oth­er hand, there is one area in which girls gen­er­ally obtain bet­ter res­ults than boys: per­cept­ive speed, which reflects the abil­ity to spot small dif­fer­ences. This was true over forty years ago, and it remains true today.

The pur­pose of an IQ test is always to com­pare an indi­vidu­al with the group to which they belong, for example their com­pat­ri­ots. On the oth­er hand, it makes no sense to com­pare two dif­fer­ent pop­u­la­tions: there is cur­rently no tool that can do this. In the past, some ini­ti­at­ives aimed to devel­op cul­ture-inde­pend­ent tests whose res­ults would not depend on the coun­try. But this is impossible because intel­li­gence can­not devel­op out­side of culture.

This close link is some­times con­firmed in unex­pec­ted ways. Take the fol­low­ing task: hav­ing to remem­ber a sequence of num­bers and recon­struct it in reverse order. The per­fect example of an exer­cise free from any cul­tur­al influ­ence, isn’t it? Well, a pre­vi­ous study showed that one coun­try per­formed sig­ni­fic­antly worse than oth­er West­ern coun­tries: Lithuania. And why was this? In the Lithuani­an lan­guage, most words for num­bers have two or three syl­lables. And work­ing memory stor­age depends on the length of the words to be remembered, which is great­er here than in oth­er lan­guages. So, there are many examples of ques­tions that may seem uni­ver­sal, but which in real­ity hide major dis­par­it­ies between coun­tries. This is why any rel­ev­ant test needs to be adap­ted to each cul­ture, which requires a sig­ni­fic­ant amount of work.

Intel­li­gence, while partly innate, can only devel­op in a favour­able con­text. School, fam­ily and social envir­on­ment all play a fun­da­ment­al role. As a res­ult, IQ can­not be meas­ured without ref­er­ence to edu­ca­tion or the social and fam­ily con­text, because intel­li­gence is linked to these. This may seem unfair, but it is a real­ity that can be seen in oth­er areas: the child of two top sports­men and women will more eas­ily devel­op ath­let­ic skills, and the same is true of the chil­dren of pro­fes­sion­al musicians.

Thanks to com­plex stat­ist­ic­al mod­els, it is pos­sible to make com­par­is­ons between eras. And from the 1940s onwards, there has been a fairly clear increase in aver­age scores in West­ern coun­tries over the dec­ades. This is known as the Flynn effect. But this trend largely slowed down in sev­er­al developed coun­tries as the 2000s approached, to the point of stag­na­tion; not regres­sion as some stud­ies with meth­od­o­lo­gic­al weak­nesses claim.

How can this phe­nomen­on be explained? There is no abso­lute cer­tainty, but it is reas­on­able to assume that the growth in IQ was encour­aged by the improve­ment in liv­ing con­di­tions after the Second World War: lower mor­tal­ity rates and child­hood ill­nesses, more school­ing, high­er liv­ing stand­ards, etc. And in recent years, we may be approach­ing a kind of lim­it in terms of aver­age intel­li­gence. Don’t human beings have lim­its in every field?

For example, at the 1896 Olympic Games, the win­ner of the 100 m race took 12 seconds. Since then, this mark has been pro­gress­ively improved, drop­ping below 10 s. But will a human being ever man­age to run the 100 m in under 7 or 5 s? There’s bound to be a lim­it that can­not be exceeded.

Bastien Contreras