2200 words

The IQ-ists are at it again. This time, PP is claiming that his little tests he created are on an absolute scale—meaning that they have a true 0 point. This has been the Achilles heel of psychometry for many decades. But abstract concepts don’t have true 0 points, and this is why “cognitive measurement” isn’t possible. I will conceptually analyze PP’s arguments for his “spatial intelligence test” and his “verbal intelligence test” and show that they aren’t on absolute scales. I will then use the IQ-ists favorite measurement—temperature (one they try to claim is like IQ)—and show the folly in his reasoning on claiming that these tests are on an absolute scale. I will then discuss the real reasons for score disparities and relate them to social class and one’s life experiences and the argue that the score results reflect merely environmental variables.

Fixed reference points and absolute scales

There are no fixed reference points for “IQ” like there are for temperature. IQ-ists have claimed for decades that temperature is like IQ while thermometers are like IQ tests (Nash, 1990). But I have shown the confused thinking of hereditarians on this issue. An absolute scale requires a fixed reference point or a true 0 point which can be objectively established. Physical quantities like distance, weight, and temperature have natural objective 0 points which can serve as fixed reference points. But nonphysical or abstract concepts lack inherent or universally agreed-upon 0 points which can serve as consistent reference points. So only physical quantities can truly be measured in an absolute scale, since they possess natural 0 points which provide a foundation for measurement.

If “spatial intelligence” is a unitary and objectively measureable cognitive trait, then all individuals’ spatial abilities should consistently align across various tasks. But individuals often exhibit significant variablity in their performance across spatial tasks, excelling in one aspect and not others. This variablity suggests that “spatial intelligence” isn’t a unitary concept. So the concept of a single, unitary, measurable “spatial intelligence” is questionable.

If the test is on an absolute scale for measuring “spatial intelligence”, then the scores obtained directly reflect the inherent “spatial intelligence” of individuals, without being influenced by factors like puzzle complexity, practice, or other variables. The scores are influenced by factors like puzzle complexity and practice effects (like doing similar things in the past). Since the scores are influenced by various factors, then it’s not on an absolute scale.

If a measurement is on an absolute scale, then it should produce consistent results across different contexts and scenarios, reflecting a stable and underlying trait. But cognitive abilities can be influenced by various external factors like stress, fatigue, motivation, and test-taking conditions. These external factors can lead to fluctuations in performance which aren’t indicative of the “trait” that’s attempting to be measured. It’s merely reflective of the circumstances of the moment one took the test in. So the concept of an absolute scale for measuring cognitive abilities fails to account for the impact of external variables which can introduce variability and inaccuracies in the “measurement.” This argument undermines the claim that this—or any test—is on an absolute scale, since motivation, stress and other socio-cognitive factors, like Richardson (2002: 287-288) notes:

the basic source of variation in IQ test scores is not entirely (or even mainly) cognitive, and what is cognitive is not general or unitary. It arises from a nexus of sociocognitive-affective factors determining individuals’ relative preparedness for the demands of the IQ test. These factors include (a) the extent to which people of different social classes and cultures have acquired a specific form of intelligence (or forms of knowledge and reasoning); (b) related variation in ‘academic orientation’ and ‘self-efficacy beliefs’; and (c) related variation in test anxiety, self-confidence, and so on, which affect performance in testing situations irrespective of actual ability.

Such factors, which influence test scores, merely show what one was exposed to in their lives, under my DEC framework. Socio-cognitive factors related to social class could introduce bias, since people from different backgrounds are exposed to different information, have unequal access to information and test prep, along with familiarity with item content. Thus, we can then look at these scores as mere social class surrogates.

If test scores are influenced by stress, anxiety, fatigue, motivation, familiarity, non-cognitive factors, and socio-cognitive factors due to social class, then the concept of an absolute scale for measuring cognitive abilities may not hold true. I have established that test scores can indeed be influenced by myriad external factors. So given that these factors affect test scores and undermine the assumption of an absolute scale, the concept of measuring cognitive ability on such a scale is challenged (don’t forget the irreducibility arguments). Further, the argument that “spatial intelligence” is not measurable on an absolute scale due to its nonphysical nature aligns with this perspective, which further supports the idea that the concept of an absolute scale isn’t applicable in these contexts. Thus, the implications for testing are profound, and so score differences are due to social class and one’s life experiences, nor any kind of “genotypic IQ” (which is an oxymoron).

Regarding vocabulary, this is influenced by the home environment—the types of words one is exposed to as they grow up (and can therefore also be integrated into the DEC). Kids from lower SES families here fewer words at home and in their neighborhoods (low SES children hear 30 million fewer words than higher SES children) (Brito, 2017). We know that word usage is the strongest determinant of child vocabulary growth, and that less educated parents use fewer words with less complex syntax (Perkins, Finegood, and Swain, 2013). The language quality that is addressed to children also matters (Golinkoff et al, 2023). We can then liken this to the Vygotskian More Knowledgeable Other (MKO). An MKO would have the knowledge that their dependent doesn’t. But if the MKO in this instance isn’t educated or low income, then they will use fewer words and they then will have this feature in their home. Such tests merely show what one was exposed to in their lives, not any underlying unitary “thing” like the IQ-ists claim.

Increasing both the amount and diversity of language within the home can positively influence language development, regardless of SES. Repeated exposure to words and phrases increases the child’s opportunity to learn and remember (McGregor, Sheng, & Ball, 2007). The complexity of grammar, the responsiveness of language to the child, and the use of questions all aid language development (Bornstein, Tamis-LeMonda, Hahn, & Haynes, 2008; Huttenlocher, Waterfall, Vasilyeva, Vevea, & Hedges, 2010). Besides frequency of language input, how caregivers communicate with children also affects children’s language skills. Children from higher SES families experience more gestures by their care-givers during parent–child interactions; these SES differences predict vocabulary differences at 54 months of age (Rowe & Goldin-Meadow, 2009). Parent–child interactions provide a context for language exposure and mold the child’s language development. Specific characteristics of the caregiver, including affect, responsiveness, and sensitivity predict children’s early and later language skills (Murray & Hornbaker, 1997; Tamis-LeMonda, Bornstein, Baumwell, & Melstein Damast, 1996). Maternal sensitivity partially explains links between SES and both children’s receptive and expressive language skills at age 3 years (Raviv, Kessenich, & Morrison, 2004). These differences also appear across culture (Mistry, Biesanz, Chien, Howes, & Benner, 2008). Maternal supportiveness partially explained the link between SES and language outcomes at 3 years of age, for both immigrant and native families in the United States. (Brito, 2017: 3-4)

The issue of temperature

This can be illustrated using the IQ-ists favorite (real) measurement—temperature. The Kelvin scale avoids the issues in the first argument. In the Kelvin scale, temperature is measured in relation to absolutel 0 (the point where molecular motion theoretically stops). It doesn’t involve factors like variability in measurement techniques, practice effects, or individual differences. The Kelvin scale has a consistent reference point—absolute 0—which provides a consistent and fixed baseline for temperature measurement. The values in the Kelvin scale are directly tied to a true 0 point.

There are no external influences on the measurement of temperature (beyond that which influences the mercury in the thermometer to move up or down),  like the type of thermometer used or one’s familiarity with temperature measurement. External factors like these aren’t relevant to the Kelvin scale, unlike puzzle complexity and practice effects on the spatial abilities test.

Finally, temperature values on the Kelvin scale are universally applicable, which means that a specific temperature corresponds to the same level of molecular motion regardless of who performs the measurement, or what measurement instrument is used. So the Kelvin temperature scale doesn’t have the same issues as PP’s little “spatial intelligence” test. It has a clear and consistent measurement framework, where values directly represent the underlying physical phenomenon of molecular motion without being influenced by external factors or individual differences. When you think about actual, established measurements like temperature and then try to relate them to IQ, then the folly of “mental measurement” reveals itself.

Now, having said all of this, I can draw a parralel between the argument against an absolute scale for cognitive abilities and the concept of temperature.

Temperature measurements, while influenced by external factors (since this is what makes the mercury travel up or down in the thermometer) like atmospheric pressure and humidity, still have an absolute 0 point in the Kelvin scale which represents a complete absence of thermal energy. Unlike “spatial intelligence”, temperature has a fixed reference point which served as an objective 0 point, which allows it to be measured on an absolute scale. The external factors influencing temperature measurement are fundamentally different from the factors which influence one’s performance on a test, since they don’t introduce subjective variations in the same manner. So while temperature is influenced by external factors, it’s measurement is fundamentally different from nonphysical concepts due to the presence of an objective 0 point and the presence and distinct nature of influencing factors. This is put wonderfully by Nash (1990: 131):

First, the idea that the temperature scale is an interval scale is a myth and, second, a scale zero can be established for an intelligence scale by the same method of extrapolation used in defining absolute zero temperature. In this manner Eysenck (p. 16) concludes, ‘if the measurement of temperature is scientific (and who would doubt that it is?) then so is that of intelligence.’ It should hardly be necessary to point out that all of this is special pleading of the most unabashed sort. In order to measure temperature three requirements are necessary: (i) a scale, (ii) some thermometric property of an object and, (iii) fixed points of reference. Zero temperature is defined theoretically and successive interval points are fixed by the physical properties of material objects. As Byerly (p. 379) notes, that ‘the length of a column of mercury is a thermometric property presupposes a lawful relationship between the order of length and the temperature order under certain conditions.’ It is precisely this lawful relationship which does not exist between the normative IQ scale and any property of intelligence. The most obvious problem with the theory of IQ measurement is that although a scale of items held to test ‘intelligence’ can be constructed there are no fixed points of reference. If the ice point of water at one atmosphere fixes 276.16 K, what fixes 140 points of IQ? Fellows of the Royal Society? Ordinal scales are perfectly adequate for certain measurements, Moh’s scale of scratch hardness consists of ten fixed points, from talc to diamond, and is good enough for certain practical purposes. IQ scales (like attainment test scales) are ordinal scales, but this is not really to the point, for whatever the nature of the scale it could not provide evidence for the property IQ or, therefore, that IQ has been measured.

Conclusion

It’s quite obvious that IQ-ists have no leg to stand on, which is why they need to claim that their tests are on absolute scales even when it leads to an absurd conclusion. The fact that test performance is influenced by myriad non-cognitive traits due to one’s social class (Richardson, 2002) shows that these—and all tests—take place in certain cultural contexts, meaning that all tests are culture-bound, as argued by Cole (2004) with his West African Binet argument.

The fact of the matter is, “mental measurement” is impossible, and all these tests do is show the proximity to a certain kind of class-specific knowledge, not any kind of general cognitive “strength”. Taking a Vygotskian perspective on this issue will allow us to see how and why people score differently from each other, and it comes down to their home environment and what they learn in their lives.

Nevertheless, the claims from IQ-ists that they have a specified measured object, object of measurement and measurement unit for IQ or that their tests have a true 0 point are absurd, since these things are properties of physical objects, not non-physical, mental ones. The Vygotskian perspective will allow use to understand score variances between individuals and groups, as I have argued before. We don’t need to claim that there is an absolute scale for cognitive assessment nor do we need to claim that mental measurement is possible for this to be a truism. So, yet again, PP’s argument fails.