1.2. Differences in achievement not in intelligence
Lynn's (2010a) estimate of IQ was based on the 2006 British PISA (Program for International Student Assessment), an internationally standardized assessment administered to 15 year olds in schools, that found higher scores for students in northern Italy when compared to students in the south. PISA tests, however, were developed to measure achievement and not intelligence. In fact, the aim of PISA is to measure "how far students near the end of compulsory education have acquired some of the knowledge and skills that are essential for full participation in society"....
Nevertheless, Lynn (2010a) uses achievement tests as "proxies for intelligence" (p. 95) adopting the logic that educational attainment and intelligence are highly correlated (from r=0.5 to r=1.0) across nations (Lynn & Meisenberg, 2010; Lynn & Mikk, 2007). However, in his studies it is not clear what kind of IQ tests have been used, and the other factors affecting achievement such as school quality, sociocultural level, and so on, are not controlled.
1.3. Correlation relationships discussed as causality relationships
It is widely known and accepted that a correlation coefficient describes the degree of relationship between two variables. However, two variables may correlate highly, but they may be different from each other. It is also possible that changes in the variables being studied are influenced by some other unobserved variable. Finally, correlation does not assume causality.
Against such universally shared methodological rules, Lynn (2010a) discusses association among variables as if they are equivalent or in a simple unilinear causal relationship.
1.5. Measuring intelligence using unvalidated tests
In his more recent paper, Lynn (2010b) reports further evidence of the lower IQs of southern Italians. The first is the report of an intelligence test given to a sample of 50,000 individuals who self-administered the test over the internet on www.sitozero.it. This is a commercial site with an inadequate description of the psychological tests used, with a considerable amount of advertisements and without any control of scientific and methodological issues. We do not consider these non-scientific data to be suitable for making assumptions about IQs.
1.6. Intelligence scores and Flynn effect
Lynn (2010b) uses data from several studies on Raven's test (Pruneti, 1985; Pruneti, Fenu, Freschi, & Rota, 1996; Tesi & Young, 1962) and Cattell Culture Fair test (Buj, 1981; Pace & Sprini, 1998). None of the studies used the same age groups and none were aimed at comparing IQs across regions of Italy.
Moreover, Lynn (2010b) did not consider the calculation of IQs made by the authors, but rather he recalculated the IQ scores in light of the well known and controversial (Colom, Lluis-Font & Andrés-Pueyo, 2005) Flynn effect (2007), described as a general increase of intelligence scores over the world in the last 50 years. So, for instance, an IQ of 99 collected in 1960, was increased by 4 points considering the Flynn effect = 4 of the Italian IQ in the years 1960-79.
Such procedure is questionable, as also Hagan, Drogin, and Guilmette (2008) pointed out. Indeed, different studies demonstrated that the Flynn effect is concentrated in the lower half of the normal distribution or in undeveloped countries (Colom et al., 2005), whereas a possible stagnation of IQ scores in developed ones is currently under debate (Teasdale & Owen, 2005; 2008).
2. New evidence against the north-south differences in IQs
With the aim to contribute to the study of regional differences in IQs, we obtained two new sources of evidence based on the direct assessment of IQs in children of different Italian regions, using measures of intelligence that do not contain highly academic content.
Despite the minor differences between the studies, our results demonstrate quite clearly that raw scores [on Raven's Coloured Progressive Matrices] of children from Sicily are not lower than those [of children from the North and Central-South] reported by Cornoldi et al. (2010). On the contrary, they are sometimes higher. This result could be related to the fact that the children in our group were tested in group sessions, while children in Italian standardization scores (Belacchi et al., 2008) were tested both in group and individual administration. Belacchi et al. (2008), indeed, found mean raw scores significantly higher in group sessions administration than in individual administration. Moreover, the children in our group were selected for other research purposes, and did not include children with socio-cultural disadvantage or other type of behavioral or cognitive problems. The more extensive sample reported by Cornoldi et al. (2010), on the contrary, was collected with the aim of building norms, and it likely includes a more diverse sample of children coming from different urban and suburban areas, and showing different socio-cultural levels.
Naglieri et al. (submitted for publication) studied the differences between the psychometric qualities of the CAS [Cognitive Assessment System] for the Italian and US standardization samples. Although the goal of that study was not to make regional comparisons, they did report that there were no significant differences (F(1, 806)=2.19, p=.11) between the average CAS-Italian Full Scale standard scores (set at a mean of 100 and standard deviation of 15) for students from the northern (M=100.5; SD=13.2), central (M=101.2; SD=11.9), and southern (M=103.1; SD=11.6) regions of Italy. The mean standard scores for the students in the north were only slightly lower than the mean for those in the south (effect size=.21). These results suggest that a test of intelligence that measures basic neuropsychological processes, and does not include academically laden verbal and quantitative tests, yields small differences between the regional groups. These findings also amplify the importance of measuring intelligence directly when comparing groups and argue against using reading, math and science test scores as "proxies for intelligence" (Lynn, 2010a).
5. General conclusions
Our examination of intelligence test score differences between the north and south of Italy led to results that are very different from those reached by Lynn (2010a). Our results demonstrate that by using intelligence tests to assess differences in ability rather than using achievement scores as a proxy for intelligence, children from the south of Italy did not earn lower scores than those from the north of Italy. Rather, they were even higher in Raven's CPM. However, we see no advantage in claiming that children in the south are "more intelligent" than children in the north, because these groups are different on a number of variables (e.g., environmental factors, educational influences, composition of the samples) that influence differences in test scores. We also disagree with Lynn's genetically-centered explanation of intelligence which denotes a fixed conception not only about intelligence but also about learning.
D'Amico et al. "Differences in achievement not in intelligence in the north and south of Italy: Comments on Lynn (2010a, 2010b)". Learning and Individual Differences, 2012.