ArticlePDF Available

The Flynn effect: Smarter not faster



Inspection time (IT) and Peabody Picture Vocabulary Test (PPVT) scores from 75 school children aged 6–13 years in 2001 were compared with the performances of 70 children aged 6–13 years who had attended the same primary school in 1981 [J. Exp. Child Psychol. 40 (1985) 1.]. ITs for the 2001 sample were measured with the same four-field tachistoscope and identical computer-based procedures followed by Wilson in 1981. The 2001 sample completed two versions of PPVT concurrently: PPVT (1965, Form A) as used in 1981 and PPVT-III (1997, Form IIIA). Mean ITs from both samples, 20 years apart, were essentially the same (123±87 and 116±71 ms in 1981 and 2001, respectively). There was, therefore, no evidence that speed of processing had improved. Correlations between IT and raw PPVT scores were significant (P<.01) for both 1981 (r=−.43) and 2001 (r=−.31). Within the 2001 sample, concurrent PPVT scores correlated .68; however, means revealed a significant Flynn effect. Thus, scores for the 2001 cohort on the earlier PPVT were higher (M standardized IQ 118.52±16.62) than the recently restandardized PPVT-III (113.97±12.23), although, compared in terms of the most recent standardization sample, the 2001 cohort was equivalent to the 1981 sample (113.66±16.72). The Flynn effect has nothing to do with speed of processing as measured by IT, despite the effect being strongest for ability tests that earn bonus scores for quick performance. Because IT correlates with IQ but appears to be stable across 20 years, whereas IQ is not, IT may have promise as a useful biological marker for an important component of cognitive decline during old age.
The Flynn effect: Smarter not faster
Ted Nettelbeck
*, Carlene Wilson
Department of Psychology, University of Adelaide, Adelaide, South Australia 5005, Australia
CSIRO, Adelaide, South Australia, Australia
Received 30 September 2002; received in revised form 13 June 2003; accepted 30 June 2003
Inspection time (IT) and Peabody Picture Vocabulary Test (PPVT) scores from 75 school children aged 613
years in 2001 were compared with the performances of 70 children aged 6 13 years who had attended the same
primary school in 1981 [J. Exp. Child Psychol. 40 (1985) 1.]. ITs for the 2001 sample were measured with the same
four-field tachistoscope and identical computer-based procedures followed by Wilson in 1981. The 2001 sample
completed two versions of PPVT concurrently: PPVT (1965, Form A) as used in 1981 and PPVT-III (1997, Form
IIIA). Mean ITs from both samples, 20 years apart, were essentially the same (123 F87 and 116 F71 ms in 1981
and 2001, respectively). There was, therefore, no evidence that speed of processing had improved. Correlations
between IT and raw PPVT scores were significant ( P< .01) for both 1981 (r=.43) and 2001 (r=.31). Within
the 2001 sample, concurrent PPVT scores correlated .68; however, means revealed a significant Flynn effect. Thus,
scores for the 2001 cohort on the earlier PPVT were higher (Mstandardized IQ 118.52 F16.62) than the recently
restandardized PPVT-III (113.97 F12.23), although, compared in terms of the most recent standardization sample,
the 2001 cohort was equivalent to the 1981 sample (113.66 F16.72). The Flynn effect has nothing to do with speed
of processing as measured by IT, despite the effect being strongest for ability tests that earn bonus scores for quick
performance. Because IT correlates with IQ but appears to be stable across 20 years, whereas IQ is not, IT may have
promise as a useful biological marker for an important component of cognitive decline during old age.
D2004 Elsevier Inc. All rights reserved.
Keywords: Flynn effect; Inspection time; Peabody Picture Vocabulary Test
1. Introduction
This study drew on two lines of inquiry: a steady, continuing, long-term, and worldwide
improvement in IQ and the theoretical contribution of inspection time (IT), envisaged as a measure
0160-2896/$ - see front matter D2004 Elsevier Inc. All rights reserved.
* Corresponding author. Tel.: +61-8-8303-5738; fax: +61-8-8303-3770.
E-mail address: (T. Nettelbeck).
Intelligence 32 (2004) 85 93
of speed of processing, to an understanding of the nature of individual differences in intelligence. The
critical proposition on which the study was based was that comparisons between ITs from cohorts of
children separated by 20 years, made by replicating earlier research procedures, provided the means for
testing whether rising IQ (the ‘‘Flynn effect’’) was the consequence of or accompanied by faster
Flynn (1999) has clearly documented widespread rises in mean IQs from substantial samples from
some 20 nations representing western European/North American cultures or technologies. These
increases in mean IQ, apparently without changes in variance, are presumed to be caused by
environmental influences as yet unidentified. There is considerable interest in Flynn’s finding (Neisser,
1998) and ongoing debate about a range of explanations, generally covering improved physical health,
nutrition and well-being, and extensive educational and technological changes within the countries
involved across the 20th century. IQ has risen, despite evidence that differences in IQ are substantially
influenced by genetic variation (Plomin & Petrill, 1997) and that individual IQ is generally not
susceptible to improvement (Spitz, 1999). Moreover, improved IQ appears to represent gains in
problem-solving abilities more than straightforward knowledge acquisition, because the largest effects
involve tests designed to measure nonverbal reasoning and abstract problem-solving, like Raven’s
progressive matrices and the Performance subtest from the Wechsler scales. Of immediate relevance to
the current study, most of these tests carry bonus points for quicker responding. By Flynn’s account,
average gains of about 1 IQ point every 3 years have probably been occurring since the IQ test was
invented. Yet, almost no one believes that human genotypic intelligence has improved significantly
during the course of the 20th century. Although the Flynn effect has thus far been demonstrated
predominantly for young (male) adults (which seems to rule out earlier maturation in more recent
generations as an explanation), at least one study, by Tasbihsazan, Nettelbeck and Kirby (1997), has
demonstrated an improved Mental Development Index of 18 points across 25 years among infants aged
1827 months based on concurrent Bayley’s (1969, 1993) comparisons. This finding is not plausibly
attributable to improved education.
It is important to note that these IQ improvements are cross sectional, derived from the achieve-
ments of different cohorts on test content that has not changed or changed little across generations.
There is no suggestion that individual IQs have improved longitudinally. Within generations, IQ scores
remain good predictors of academic, work performance, and other life achievements (Jensen, 1998),
and there is now general consensus that individual differences in IQ reflect a substantial genetic
component (Neisser et al., 1996). Nonetheless, rising IQs can only be substantially explained
environmentally and they therefore challenge the construct validity of the tests as measures of
fundamental, inborn cognitive abilities. The Flynn effect implies that abstract reasoning abilities,
previously held by many to reflect basic capacities, are influenced by as yet undetermined
environmental circumstances.
Considerable speed-based research has found that speed measures correlate with IQ (Nettelbeck,
1998, 2001).IT(Vickers, Nettelbeck, & Willson, 1972) measures individual differences in threshold to
detect the location (left or right) of the shorter of two vertical lines displayed in a briefly exposed target
figure. The threshold measure is essentially a critical stimulus onset asynchrony (CSOA), defined as the
minimum delay required, between the onset of the target figure and the subsequent onset of a backward
masking figure so as to achieve predetermined high accuracy. IT correlates at about .5 with nonverbal
IQ (Deary & Stough, 1996; Grudnick & Kranzler, 2001; Kranzler & Jensen, 1989; Nettelbeck, 1987).
The basis of this correlation has not been clearly identified, but there are strong grounds for supposing
T. Nettelbeck, C. Wilson / Intelligence 32 (2004) 85–9386
that it reflects more than application of higher-order ‘‘intelligent’’ strategies (Deary, 2000; Nettelbeck,
2001). It may involve IT’s sensitivity to the psychometric construct ‘‘speediness’(Burns & Nettelbeck,
2003), defined by Horn and Noll (1997) as speed under relatively undemanding circumstances.
However, speediness is unlikely to provide a sufficient explanation for ITIQ correlation, because
Burns and Nettelbeck (2003) also found that IT shared variance with a higher-order, orthogonal general
factor and recent research has raised the possibilities that IT is sensitive to attentional capacities (Hutton,
Wilding, & Hudson, 1997) and fluid abilities (Osmon & Jackson, 2002). Although it is unlikely that a
single kind of mental speed could account for individual differences in IQ (Roberts & Stankov, 1999),
Carroll (1993) has allowed that some fundamental aspect of processing speed could underpin the higher-
order general ability factor that distinguishes his model for human intelligence from similar ‘‘multiple
intelligences’’ models (Horn & Noll, 1997).
The current study set out to replicate with a current sample measures of vocabulary and IT
initially made by Wilson in 1981 as part of a cross-sequential investigation of childhood
developmental changes in processing speed (this work was published by Nettelbeck & Wilson,
1985). To this end, primary school children were recruited from the same school that was involved
in 1981. This school had continued to serve the same catchment area, as 20 years previously, from
upper middle-class socioeconomic suburbs.
As for the earlier study, vocabulary achievement was a
proxy for IQ and estimated with the same test. IT was measured using exactly the same apparatus
and procedures followed in the earlier study. If rising IQ is accompanied by improved processing
speed as is implied by theories that have drawn heavily on the ‘‘fast is smart’’ assumption common
to western European cultures (Brand, 1996; Eysenck, 1987; Jensen, 1998), then this would be
revealed by comparison of IT measures made now with those recorded 20 years ago. On the other
hand, if IQ was shown to improve but IT had not, this would rule out speed of processing as an
explanation for improving IQ.
2. Method
2.1. Participants
Seventy-five school children (36 boys, 39 girls) aged 6 13 years took part, with the permission of
their parents. They attended the same school as the 70 children (38 boys, 32 girls) also aged 6 13
years in the 1981 sample. As had been the case in 1981, all had normal or corrected-to-normal
vision. Following Wilson in 1981, this was a sample of convenience, aiming to draw about 10
children, approximately balanced for gender, from each of seven consecutive grade levels. Response
rates were high, with many more children volunteering than were required. Those participating were
determined according to the children’s availability and teachers’ convenience at the time of testing
(see Table 1).
The Australian census categorizes postcodes in capital cities within quintiles for socioeconomic strata, defined by
household income and other indices of relative social advantage. In 1981 and 2001, the postal districts encompassed by the
school’s catchment area were in the highest quintile.
Grade levels in 2001 were different from those in 1981. Today’s children aged 6– 13 years were located in Grades 1 7,
whereas in 1981 these age groups were in Grades 2–8.
T. Nettelbeck, C. Wilson / Intelligence 32 (2004) 85–93 87
2.2. Materials and apparatus
Vocabulary was measured with the Peabody Picture Vocabulary Test (PPVT) using both the original
version (Dunn, 1965; Form A) and the most recent PPVT-III (Dunn & Dunn, 1997; Form IIIA). IT was
measured using the same Gerbrands tachistoscope, previously modified in-house to provide four fields
so that the four stimulus cards could be set undisturbed throughout the study, following appropriate
initial alignments. These cards displayed in turn an initial visual fixation cue, the two alternative targets
with the shorter vertical line to left or right, and the backward masking figure. The tachistoscope was set
at the same field luminances, employing the same stimulus cards, the same sequence for lighting the
fields, and the same presentation technique used by Wilson in 1981. Thus, the two vertical lines in the
target were 24 and 34 mm, were 10 mm apart, and aligned at the top by a horizontal line. The backward
mask that subsequently overlaid each target display had both vertical bars 44 mm long and 5 mm wide,
centered at 10 mm apart. The software that controlled onset and offset of the four fields and recorded
responses as correct or not was the same program used by Wilson in 1981. It was an early version of the
Parameter Estimation by Sequential Testing (PEST) program (Taylor & Creelman, 1967), an adaptive
staircase algorithm that estimated the CSOA with an associated probability of 85% correct responding.
The response keypad was that used by Wilson in 1981. Further details for these pieces of equipment are
to be found in Nettelbeck and Wilson (1985) (Study 3).
2.3. Procedures
All children were tested individually, first completing both versions of PPVT concurrently at a single
session, with f10 min break between. Order of completion was balanced across children. A single
estimate of IT was made at a second session, following exactly the same instructions used by Wilson in
Table 1
Means FS.D.s for grade levels, chronological ages, standardized PPVT scores, and ITs from children in 1981 and 2001
Grade level nAge (years-months) Gender (M/F) PPVT
1 10 7-4 F0-6 6/4 101 F19 231 F171
2 10 7-10 F0-3 5/5 116 F19 133 F45
3 10 8-8 F0-2 5/5 116 F9 125 F48
4 10 9-10 F0-5 4/6 115 F18 115 F46
5 10 10-11 F0-4 6/4 117 F14 101 F40
6 10 11-11 F0-3 6/4 107 F770F42
7 10 13-2 F0-3 6/4 123 F22 86 F24
1 14 6-9 F0-5 7/7 116 F11 151 F64
2 12 7-9 F0-5 6/6 113 F16 135 F120
3 10 8-8 F0-2 4/6 122 F12 132 F85
4 10 9-9 F0-2 5/5 110 F8114F53
5 10 10-8 F0-4 4/6 115 F11 91 F29
6 10 11-7 F0-2 5/5 109 F11 96 F28
7 9 12-9 F0-5 5/4 113 F14 74 F18
Original PPVT norms were applied in 1981, while PPVT-III norms were applied in 2001.
T. Nettelbeck, C. Wilson / Intelligence 32 (2004) 85–9388
1981. These emphasized accuracy of responding, not speed. The same practice routines and staircase
algorithms were used. Nettelbeck and Wilson (1985) provided a full description of these.
3. Results
As can be seen from Table 1, the age sample in 2001 (overall M= 9-5 F2-4 years-months) closely
matched the 1981 cohort (9-11 F2-6 years-months) [t(143) = 1.49, P> .05, two-tailed]. The numbers of
children and gender balances within age levels were similar for both cohorts. There were 48% boys in
2001 compared with 54% in 1981. Thus, the two samples drawn from the same school 20 years apart
were age and gender equivalent for comparison purposes.
Distributions of standardized PPVT scores within the current and the 1981 sample, based on age
norms applying in 1981 and 2001, were also essentially the same. The overall mean in 1981 (initial
PPVT age norms) was 113.66 F16.72 compared with 113.97 F12.23 in 2001 (PPVT-III age norms).
The difference was not statistically significant (t< 1.0), and correcting scores according to the extent to
which norms for both versions of PPVT had become obsolete at time of testing (see Flynn, 1987) did
not change this outcome. Nonetheless, although samples were equivalent for verbal achievement for
their respective times, the current sample demonstrated a clear Flynn effect. Thus, although concurrent
scores for the 2001 children with original and recent versions of PPVT were highly correlated
[r(73)=.68, P< .01, two-tailed], these children were significantly advantaged on the early PPVT
(M= 118.52 F16.62) compared with PPVT-III (M= 113.97 F12.23) [t(74) = 3.22, P< .01, two-tailed].
This within-subjects outcome was confirmed by between-subjects analysis, comparing the earlier
version PPVT scores from the 2001 sample with the earlier version PPVT scores from the 1981 sample
[t(143) = 1.76, P< .05, one-tailed]. This rise of almost 5 points across 20 years was lower than but
consistent with 20 years’ improvement in Wechsler Verbal IQ (7 points), estimated by comparing WISC
with WISC-R standardization samples (Wechsler, 1949, 1974). (The improvement for word knowledge
was also smaller than the Flynn effect of f8 Performance IQ points across 20 years embedded in the
WISC/WISC-R norms.)
Despite the Flynn effect for vocabulary achievement, Table 1 demonstrates that there was no evidence
of improvement in IT from 1981 (overall M= 123 F87 ms) to 2001 (M=116F71 ms). Of course, this
conclusion amounts to accepting a null hypothesis, but the effect size of only about 0.09 would require
more than 2000 cases in both cohorts to achieve a=.05 (two-tailed) at power = 0.80. Overall, the 1981
and 2001 distributions were remarkably similar, being positively skewed to the same extent (2.26 and
2.94, respectively) around the same medians (103 and 102 ms) and with similar minima (18 and 33 ms)
and maxima (450 and 500 ms). Two-way ANOVA found no cohort (1981 vs. 2001) effect
[F(1,131) < 1.0]; as expected, there was a highly significant age effect [ F(6,131) = 6.20, P< .001] but
no Cohort Age interaction [ F(6,131) = 1.24, P>.05]. Visual inspection of the IT distributions across
age and cohorts confirmed that the longer mean IT in the youngest 1981 subsample, compared with
2001, was the consequence of three children aged 6/7 years whose low PPVT scores and long IT
estimates made them outliers. Nonetheless, correlations within both cohorts between raw PPVT scores
and IT (1981) and raw PPVT-III scores and IT (2001) were significant and similar: 1981
r(68) = .43 F.24 (95% confidence limits), P< .01; 2001 r(73) = .31 F.23, P< .05. These coeffi-
cients did not differ significantly (z= 1.07, ns). Both outcomes were, of course, confounded by strong
age effects.
T. Nettelbeck, C. Wilson / Intelligence 32 (2004) 85–93 89
4. Discussion
The current study aimed to replicate Wilson’s 1981 study (Nettelbeck & Wilson, 1985) and succeeded
to a remarkable degree. Word knowledge outcomes derived from 1981 and 2001 norms provided an
excellent match. As predicted by Flynn’s observations, concurrent testing of word knowledge in the
2001 sample, using both current and earlier Peabody test versions and norms, found that word
knowledge had risen significantly by about 5 points across 20 years. This result was statistically
significant and consistent with restandardizations of the Wechsler scales across this period, which have
found a Flynn effect of 7 Verbal IQ points (cf. about 8 points for nonverbal aspects). However, most
importantly, estimates of IT were essentially the same from both cohorts, results for 1981 and 2001
demonstrating the expected significant age and IQ effects to the same extent. The monotonic reduction in
mean IT with age was marked and not consistent with Anderson’s theory that IT does not change with
development (Anderson, 1992; Anderson, Reid, & Nelson, 2001).
In other words, whereas average IQs of 613-year-olds are known to have risen appreciably across 20
years, including on a pencil-and-paper marker test for ‘‘speediness’’ (the Coding subscale from
Wechsler), IT was not subject to cohort improvement. IT, which as expected correlated significantly
with word knowledge scores within both the 1981 and the 2001 cohort, did not change at all on average.
Thus, based on current results, IQ gains are not explicable in terms of improved processing speed, as
operationalized by IT. This result suggests, moreover, that IT measures some fundamental aspect of
mental functioning that is relevant to understanding of human intelligence, which is not influenced by
whatever environmental circumstances are responsible for rising IQ scores. What this function is,
however, is not clear. Debate continues around what psychological processes are tapped by IT
(Nettelbeck, 2001), and the nature of processing speed appears to be complex (Roberts & Stankov,
1999). Moreover, it is necessary, though difficult, to replicate this result. The samples in both the original
1981 study and the 2001 replication were small, with only about 10 children in each age group. Insofar
as a major assumption underpinning current conclusions is that the 1981 and 2001 cohorts were
demographically equivalent, it would have been desirable to confirm this more precisely, e.g., by
recording parents’ educational levels, occupations, and salaries. Without more evidence to the contrary
than the broad census data available here, it is always possible that the increased word knowledge found
was the consequence of idiosyncratic improvement to the socioeconomic circumstances of the children
involved, beyond those that speculation has linked with the Flynn effect (Neisser, 1998). It is also
important that future research of this issue should test the stability of IT across time for a much wider
range of ages than the 613 years included here.
If the current result is confirmed, two future directions for research are suggested. The present result
for IT begs the question as to whether different kinds of processing speed, similarly known to correlate
with IQ although not necessarily appreciably with each other (Kranzler & Jensen, 1991), are stable or
improve across time. For example, existing large data sets derived for parameters of Decision time and
Movement time (Jensen, 1987) might be used to explore this question by replicating these earlier
procedures with current samples.
A second suggestion is that IT may provide a useful biomarker for cognitive aging. The causes of
aging are not known. Nonetheless, there is considerable evidence to support a conclusion that normal
aging beyond the mid-1960s is, on average, accompanied by cognitive decline that, although different
abilities change at different rates, is largely attributable to slowing of information processing speed
(Deary, 2000; Salthouse, 1996; Schaie, 1994). Although gradually slowing processing speed and
T. Nettelbeck, C. Wilson / Intelligence 32 (2004) 85–9390
declining cognitive abilities may be ongoing beyond early adult years, considerable evidence points to
relative stability before the sixth decade but a marked shift in rate beyond. However, there are marked
individual differences in the onset and progress of age-related changes, so that chronological age is an
unreliable marker for functional aging. Obviously, individual differences in functional aging have
implications of considerable practical importance for those involved, and reliable biomarkers, capable of
predicting functional change as a consequence of aging, particularly any accelerated decline in cognitive
integrity, would therefore be extremely useful (Stern & Carstensen, 2000).
Much aging research has relied on the Digit Symbol test (Wechsler) to measure processing speed
(Salthouse, 1996). Although longitudinal studies leave little doubt that aging effects are real (Salthouse,
2000; Schaie, 1994), it is possible that effect sizes from cross-sectional designs have been exaggerated
by the Flynn effect, which would tend to favor younger cohorts on Digit Symbol. Thus, a measure of
processing speed, known to be stable across age cohorts, would be desirable for researching age-related
cognitive decline.
At the very least, IT appears to tap ‘‘speediness’’ (Nettelbeck, 1994) and may also be relevant to
general ability (Burns & Nettelbeck, 2003). Moreover, on current evidence, IT is stable across
generations, at least among children, unlike conventional psychometric tests. We suggest that these
attributes, together with its noninvasive measurement procedure, may make IT an attractive prospect as a
biomarker to monitor cognitive changes that accompany aging, providing that it meets other essential
criteria. These should include sensitivity to cognitive change within a short period, predicting important
life changes ahead of time (e.g., decline in workplace competence or the onset of driving difficulties) and
predicting longevity. It is also desirable that a biomarker should be measurable in other species
(McClearn, 1997) and therefore available for animal modeling of functional aging. IT is certainly
sensitive to cross-sectional age comparisons among elderly people (Nettelbeck, 1987) and to age-related
differences in cognitive functioning (Nettelbeck & Rabbitt, 1992; see also Deary, 2000, pp. 244– 246 for
a relevant reanalysis of these data), but nothing is known currently about whether IT is subject to
longitudinal slowing. In principle, the discrimination required to estimate IT is simple and should be
achievable with animals; however, IT’s utility as a lead biomarker, capable of predicting accelerated
decline in cognitive integrity, would be a more immediate priority for future research.
The University of Adelaide’s Small Research Grant Scheme supported this research. We thank Dr.
Greg Evans for his valuable assistance when repairing and calibrating the tachistoscope. We gratefully
acknowledge the help and cooperation extended by the principal, staff, parents, and children at
Mitcham Primary School; without their generous participation, this research would not have been
Anderson, M. (1992). Intelligence and development: A cognitive theory. Oxford: Blackwell.
Anderson, M., Reid, C., & Nelson, J. (2001). Developmental changes in inspection time; What a difference a year makes.
Intelligence,29, 475486.
Bayley, N. (1969). Bayley scales of infant development: Birth to two years. New York: The Psychological Corporation.
T. Nettelbeck, C. Wilson / Intelligence 32 (2004) 85–93 91
Bayley, N. (1993). Bayley scales of infant development: Manual (2nd ed.). San Antonio, TX: The Psychological Corporation.
Brand, C. R. (1996). The g factor. Chichester: Wiley.
Burns, N., & Nettelbeck, T. (2003). Inspection time in the structure of cognitive abilities: Where does IT fit? Intelligence,31,
Carroll, J. B. (1993). Human cognitive abilities: A survey of factor analytic studies. Cambridge, UK: Cambridge University
Deary, I. J. (2000). Looking down on human intelligence: From psychophysics to the brain. Oxford: Oxford University Press.
Deary, I. J., & Stough, C. (1996). Inspection time and intelligence: Achievements, prospects and problems. American Psychol-
ogist,51, 599608.
Dunn, L. M. (1965). Peabody Picture Vocabulary Test: Manual. Circle Pines, MN: American Guidance Service.
Dunn, L. M., & Dunn, L. M. (1997). Examiner’s manual for the Peabody Picture Vocabulary Test (3rd ed.). Circle Pines, MN:
American Guidance Service.
Eysenck, H. J. (1987). Speed of information processing, reaction time, and the theory of intelligence. In P. A. Vernon (Ed.),
Speed of information processing and intelligence ( pp. 21 67). Norwood, NJ: Ablex.
Flynn, J. R. (1987). Massive IQ gains in 14 nations: What IQ tests really measure. Psychological Bulletin,101, 171–191.
Flynn, J. R. (1999). Searching for justice: The discovery of IQ gains over time. American Psychologist,54, 520.
Grudnick, J. L., & Kranzler, J. H. (2001). Meta-analysis of the relationship between intelligence and inspection time. Intelli-
gence,29, 523535.
Horn, J. L., & Noll, J. (1997). Human cognitive abilities: Gf-Gc theory. In D. P. Flanagan, J. L. Genshaft, & P. L. Harrison
(Eds.), Contemporary intellectual assessment: Theories, tests, and issues ( pp. 53 91). New York: Guilford Press.
Hutton, U., Wilding, J., & Hudson, R. (1997). The role of attention in the relationship between inspection time and IQ in
children. Intelligence,24, 445460.
Jensen, A. R. (1987). Individual differences in the Hick paradigm. In P. A. Vernon (Ed.), Speed of information processing and
intelligence (pp. 101175). Norwood, NJ: Ablex.
Jensen, A. R. (1998). The g factor: The science of mental ability. Westport, CT: Praeger.
Kranzler, J. H., & Jensen, A. R. (1989). Inspection time and intelligence: A meta-analysis. Intelligence,13, 329– 347.
Kranzler, J. H., & Jensen, A. R. (1991). The nature of psychometric g: Unitary process or a number of independent processes?
Intelligence,15, 397422.
McClearn, G. E. (1997). Biomarkers of age and aging. Experimental Gerontology,32, 87– 94.
Neisser, U. (Ed.) (1998). The rising curve: Long-term gains in IQ and related measures. Washington, DC: American Psycho-
logical Association.
Neisser, U., Boodoo, G., Bouchard, T. J., Boykin, A. W., Brody, N., Ceci, S. J., Halpern, D. F., Loehlin, J. C., Perloff, R.,
Sternberg, R. J., & Urbina, S. (1996). Intelligence: Knowns and unknowns. American Psychologist,51, 77 101.
Nettelbeck, T. (1987). Inspection time and intelligence. In P. A. Vernon (Ed.), Speed of information processing and intelligence
(pp. 295346). Norwood, NJ: Ablex.
Nettelbeck, T. (1994). Speediness. In R. J. Sternberg (Ed.), Encyclopedia of human intelligence ( pp. 1014 1019). New York:
Nettelbeck, T. (1998). Jensen’s chronometric research: Neither simple nor sufficient but a good place to start. Intelligence,26,
Nettelbeck, T. (2001). Correlation between inspection time and psychometric abilities: A personal perspective. Intelligence,29,
Nettelbeck, T., & Rabbitt, P. M. A. (1992). Aging, cognitive performance, and mental speed. Intelligence,16, 189 205.
Nettelbeck, T., & Wilson, C. (1985). A cross-sequential analysis of developmental differences in speed of visual information
processing. Journal of Experimental Child Psychology,40, 1– 22.
Osmon, D. C., & Jackson, R. (2002). Inspection time and IQ: Fluid or perceptual aspects of intelligence? Intelligence,30,
119– 127.
Plomin, R., & Petrill, S. (1997). Genetics and intelligence: What’s new? Intelligence,24, 53 77.
Roberts, R. D., & Stankov, L. (1999). Individual differences in speed of mental processing and human cognitive abilities:
Toward a taxonomic model. Learning and Individual Differences,11, 1 120.
Salthouse, T. A. (1996). The processing-speed theory of adult age differences in cognition. Psychological Review,103,
Salthouse, T. A. (2000). Aging and measures of processing speed. Biological Psychology,54, 35–54.
T. Nettelbeck, C. Wilson / Intelligence 32 (2004) 85–9392
Schaie, K. W. (1994). The course of adult intellectual development. American Psychologist,49, 304–313.
Spitz, H. H. (1999). Beleaguered Pygmalion: A history of the controversy over claims that teacher expectancy raises intelli-
gence. Intelligence,27, 199 234.
Stern, P. C., & Carstensen L. L. (Eds.) (2000). The aging mind: Opportunities in cognitive research. Washington, D.C.:
National Research Council, National Academy Press. 271 pp.
Tasbihsazan, R., Nettelbeck, T., & Kirby, N. (1997). Increasing mental development index in Australian children: A compa-
rative study of two versions of the Bayley mental scale. Australian Psychologist,32, 120 125.
Taylor, M. M., & Creelman, C. D. (1967). PEST: Efficient estimate in probability functions. Journal of Acoustical Society of
America,4, 782787.
Vickers, D., Nettelbeck, T., & Willson, R. J. (1972). Perceptual indices of performance: The measurement of ‘‘inspection time’’
and ‘‘noise’’ in the visual system. Perception,1, 263–295.
Wechsler, D. (1949). Wechsler Intelligence Scale for Children: Manual. New York: The Psychological Corporation.
Wechsler, D. (1974). Manual: Wechsler Intelligence Scale for Children-Revised. New York: The Psychological Corporation.
T. Nettelbeck, C. Wilson / Intelligence 32 (2004) 85–93 93
... We observed no meaningful effect changes in processing speed (Gs) and reading and writing (Grw). Past research about Flynn effects for processing speed did not yield evidence for meaningful changes in Australian (Nettelbeck & Wilson, 2004) and Estonian measurement invariant test score changes (Must & Must, 2018), thus conforming to our findings for Gs. To our knowledge, we are the first to report evidence about potential Flynn effects for Grw which did not show any meaningful test score changes in our investigation. ...
... We observed no meaningful effect changes in processing speed (Gs) and reading and writing (Grw). Past research about Flynn effects for processing speed did not yield evidence for meaningful changes in Australian (Nettelbeck & Wilson, 2004) and Estonian measurement invariant test score changes (Must & Must, 2018), thus conforming to our findings for Gs. To our knowledge, we are the first to report evidence about potential Flynn effects for Grw which did not show any meaningful test score changes in our investigation. ...
Generational IQ test score changes (the Flynn effect) were globally positive over large parts of the 20th century. However, accumulating evidence of recent studies shows a rather inconsistent pattern in past decades. Patterns of recently observed test score changes appeared to be markedly different in strength and even signs between countries and domains. Because of between-study design differences and data availability in terms of differing IQ domains, it is so far unclear if these inconsistencies represent a consequence of differences in Flynn effect trajectories between countries, covered time-spans, or investigated IQ domains. Here, we present data from 36 largely population-representative Germanophone standardization samples from 12 well-established psychometric tests (17 subtests) of 10 stratum II domains from 1996 to 2018, thus providing a comprehensive assessment of domain-specific changes according to the Cattell-Horn-Carroll intelligence model. Examination of both raw score and measurement-invariant latent mean changes yielded positive (comprehension-knowledge, learning-efficiency, domain-specific knowledge), negative (working memory capacity), stagnating (processing speed, reading and writing), and ambiguous (fluid reasoning, reaction and decision speed, quantitative knowledge, visual processing) stratum II Flynn effects. This means that in the present sample, the Flynn effect is surprisingly differentiated on domain level and does not conform to the frequently observed IQ test score gains in crystallized and fluid intelligence. These findings could be attributed to either (i) a so far undetected domain-specificity of the Flynn effect due to an unavailability of test data beyond crystallized and fluid domains or (ii) a symptom for an impending stagnation of the Flynn effect.
... For example, mean score differences between the PPVT-4 and the PPVT-III were only 1.7 points for ages 2-4 years and 0.2 for ages 5-6 years (Dunn & Dunn, 2007). These score differences are much lower, and that disparity considers any variation due to the Flynn effect (Nettelbeck & Wilson, 2004) and any actual item differences from one version to another. Even mean score differences between the PPVT-4 and an unrelated test of language (CELF-4) found only a 3.4-point difference (Dunn & Dunn, 2007). ...
... No psychometricians beyond the test publisher have defined an exact value to determine practical significance in equivalency score differences and/or effect size, thus the publisher's assumption that 0.20 or less denotes equivalency may be viewed as too liberal. When examining the Flynn Effect (Nettelbeck & Wilson, 2004), a threepoint difference per decade may be considered as practically significant (Sattler, 2018); therefore, the 2.98 score difference found in the current study may also be considered as practically significant. Whether the PBT and TBT/ Qi scores were significantly different is disputable for practical significance, while it is clear that the differences are statistically significant. ...
Pearson now uses a technology‐based testing platform, Q‐Interactive, to administer tests previously available in paper versions. The same norms are used for both versions; Pearson's in‐house equivalency studies indicated that both versions are equated. The goal of the current study is to independently evaluate equivalency findings. For the current study, equivalency was measured using the three‐part test set forth by American Psychological Association in 1986. First, the researchers examined rank order similarity; then, they examined mean score similarity; and finally, they examined score‐distribution similarity. One of these equivalency standards (rank order similarity) was not met, and one other standard is debatable (mean score similarity); therefore, the authors noted concerns about the use of Peabody Picture Vocabulary Test, Fourth Edition Q‐Interactive for preschoolers. New normative data should be collected.
... More generally, gains on fluid intelligence have been stronger than gains on crystallized intelligence, at least in most of the times and places for which we have informative data. Some abilities included in the g nexus did not show any evidence of Flynn effects, including reaction time (e.g., Nettelbeck & Wilson, 2004), digit span (Gignac, 2015;Woodley of Menie & Fernandes, 2015), and ability-based emotional intelligence (Pietschnig & Gittler, 2017). ...
The aim of the study is to estimate the most recent trends of intelligence world-wide. We find that the most recent studies report mainly positive Flynn effects in economically less developed countries, but trivial and frequently negative Flynn effects in the economically most advanced countries. This is confirmed by an analysis of 48 countries in the 2000–2018 PISA tests, showing that high pre-existing IQ and school achievement are the best predictors of declining test scores. IQ gaps between countries are still large (e.g., 19 IQ points in PISA between East Asia and South Asia) but are diminishing world-wide. We predict that these trends, observed in adolescents today, will reduce cognitive gaps between the working-age populations of countries and world regions during coming decades. As is predicted by the well-established relationship between intelligence and economic growth, there is already evidence that the ongoing cognitive convergence is paralleled by global economic convergence. These developments raise questions as to how long this cognitive and economic convergence will continue, whether it will eliminate cognitive and economic gaps between countries entirely, and whether a condition with high levels of cognitive ability and economic prosperity is sustainable long-term.
... Lastly to determine the role visual development alone, played in threshold time for object/shape identification without a motor component we utilized the visual Inspection Time (IT) task, where improvements in IT performance has previously been used to predict future abilities on cognitive tasks such as perceptual speed, verbal IQ and working memory (Nettelbeck and Young, 1990;Nettelbeck and Wilson, 2004;Brown and Crewther, 2017;Ebaid et al., 2017). A modified Inspection Time (IT) task, was based on the version of Vickers et al. (1972) and adapted by Brown and Crewther (2017) for children aged 7-11 years. ...
Full-text available
CITATION Alhamdan AA, Murphy MJ and Crewther SG (2022) Age-related decrease in motor contribution to multisensory reaction times in primary school children. Front. Hum. Neurosci. 16:967081. Traditional measurement of multisensory facilitation in tasks such as speeded motor reaction tasks (MRT) consistently show age-related improvement during early childhood. However, the extent to which motor function increases with age and hence contribute to multisensory motor reaction times in young children has seldom been examined. Thus, we aimed to investigate the contribution of motor development to measures of multisensory (auditory, visual, and audiovisual) and visuomotor processing tasks in three young school age groups of children (n = 69) aged (5−6, n = 21; 7−8, n = 25.; 9−10 n = 18 years). We also aimed to determine whether age-related sensory threshold times for purely visual inspection time (IT) tasks improved significantly with age. Bayesian results showed decisive evidence for age-group differences in multisensory MRT and visuo-motor processing tasks, though the evidence showed that threshold time for visual identification IT performance was only slower in the youngest age group children (5−6) compared to older groups. Bayesian correlations between performance on the multisensory MRT and visuo-motor processing tasks indicated moderate to decisive evidence in favor of the alternative hypothesis (BF 10 = 4.71 to 91.346), though not with the threshold IT (BF 10 < 1.35). This suggests that visual sensory system development in children older than 6 years makes a less significant contribution to the measure of multisensory facilitation, compared to motor development. In addition to this main finding, multisensory facilitation of MRT within race-model predictions was only found in the oldest group of children (9−10), supporting previous suggestions that multisensory integration is likely to continue into late childhood/ early adolescence at least.
... Cognitive aging is also a reflection of the historical context and the numerous life-course influences. In that context it is not surprising that many studies have shown that later born cohorts tend, on average, to perform better on several cognitive abilities which is often referred to as Flynn effects [11][12][13][14][15][16][17][18]. But, do these birth cohort differences manifest themselves also regarding proportions of individuals showing cognitive decline, stability, and gain? ...
... For receptive vocabulary, the Spanish version of the Test de Vocabulario en Imágenes Peabody -Peabody Picture Vocabulary Test -PPVT was administered (Dunn & Dunn, 1981) at times 1 and 2. The Peabody Test is straightforward to administer. One of its main advantages is that children are merely required to indicate the picture that matches the spoken word (Nettelbeck & Wilson, 2004 The block building from the British Scales Abilities (BAS) was administered at time 1 as a measure of general cognitive performance. This task has shown to be an effective way to measure intelligence , and it has been considered a good predictor of later academic achievement (Casey et al., 2008;Wolfgang et al., 2003). ...
... However, these norms have not been updated for three to four decades. Thus, cohort effects related to changes in music-listening and in cognitive variables known to be related to musical abilities may make these norms less valid for current use (Nettelbeck and Wilson, 2004). More recent test batteries include the Montreal Battery of Evaluation of Musical Abilities (MBEMA; Peretz et al., 2013), which was administered to a large sample of Canadian and Chinese children aged 6-8. ...
Full-text available
Measuring musical abilities in childhood can be challenging. When music training and maturation occur simultaneously, it is difficult to separate the effects of specific experience from age-based changes in cognitive and motor abilities. The goal of this study was to develop age-equivalent scores for two measures of musical ability that could be reliably used with school-aged children (7–13) with and without musical training. The children's Rhythm Synchronization Task (c-RST) and the children's Melody Discrimination Task (c-MDT) were adapted from adult tasks developed and used in our laboratories. The c-RST is a motor task in which children listen and then try to synchronize their taps with the notes of a woodblock rhythm while it plays twice in a row. The c-MDT is a perceptual task in which the child listens to two melodies and decides if the second was the same or different. We administered these tasks to 213 children in music camps (musicians, n = 130) and science camps (non-musicians, n = 83). We also measured children's paced tapping, non-paced tapping, and phonemic discrimination as baseline motor and auditory abilities We estimated internal-consistency reliability for both tasks, and compared children's performance to results from studies with adults. As expected, musically trained children outperformed those without music lessons, scores decreased as difficulty increased, and older children performed the best. Using non-musicians as a reference group, we generated a set of age-based z-scores, and used them to predict task performance with additional years of training. Years of lessons significantly predicted performance on both tasks, over and above the effect of age. We also assessed the relation between musician's scores on music tasks, baseline tasks, auditory working memory, and non-verbal reasoning. Unexpectedly, musician children outperformed non-musicians in two of three baseline tasks. The c-RST and c-MDT fill an important need for researchers interested in evaluating the impact of musical training in longitudinal studies, those interested in comparing the efficacy of different training methods, and for those assessing the impact of training on non-musical cognitive abilities such as language processing.
Full-text available
Η Περιβαλλοντική Εκπαίδευση (ΠΕ) λαμβάνει χώρα σε πολλά δημοτικά σχολεία. Κατά τη διάρκεια της καραντίνας και εν μέσω πανδημίας του Κορωνοϊού, έπρεπε τα περιβαλλοντικά προγράμματα που είχαν ξεκινήσει στην αρχή της σχολικής χρονιάς να αλλάξουν μορφή και να ενσωματώσουν σχέδια εργασίας εξ αποστάσεως ΠΕ. Τα οφέλη τόσο της ΠΕ, όσο και της εξ αποστάσεως ΠΕ, είναι πολλαπλά σε ένα παιδί αλλά και στην κοινωνία γενικότερα, αφού η αλλαγή στάσεων, αντιλήψεων και συμπεριφορών των μαθητών (Ράπτης, όπως αναφέρεται στη Σκαναβή, 2004) έχει βραχυπρόθεσμα και μακροπρόθεσμα αποτελέσματα. Επίσης, προωθείται η υπεύθυνη περιβαλλοντική συμπεριφορά και αναπτύσσεται στους μαθητές η δεξιότητα να συμμετέχουν ενεργά στην λήψη αποφάσεων. Μαθητές της Ε’ τάξης Δημοτικού Σχολείου της Νίκαιας, πήραν μέρος όλη την περίοδο του εγκλεισμού στο σπίτι σε καινοτόμες δράσεις και έδειχναν υπερβάλλων ζήλο στη συμμετοχή τους. Environmental Education (EP) takes place in many primary schools. During the quarantine and in the midst of its pandemic Covid 19, needed the environmental programs that had started at the beginning of the school year to change form and incorporate distance work plans IP. The benefitsof both EP and and distance EP, are multiple in a child but also in society in general, since the change of attitudes, perceptions and behaviors of students (Raptis, as mentioned in Skanavi, 2004) have in short and long-term results. Responsible environmental behavior is also promoted and the skill is developed in the students to be actively involved in decision-making. Elementary school students of Nikaia, took part throughout his period inclusion at home in innovative actions and showed excessive eagerness in their participation.
Inspection time tasks assess the ability to make a simple visual discrimination, typically in milliseconds. Typically, IT stimuli consists of a pi-shaped figure, in which subjects select the side with the significantly longer leg. To prevent storage in iconic memory, a backward mask is then introduced. However, some participants have reported that the mask may cause the shorter leg to appear to lengthen, creating a possible strategy that facilitates performance. As a result, alternative stimuli/masks have been developed; however, these alternative stimuli may be processed differently. This study assessed the cognitive correlates and stability of an alternative stimuli/mask. Results indicated that processing of the stimuli was influenced by an interaction between the complexity of the stimuli and the number of times it was presented. Specifically, the alternative stimulus/mask produced slower processing, particularly at the time of a second administration; however, it contributed an important and unique relationship with speeded, general intelligence.
This book is about intelligence and cognitive development. The choice of a conjunction is deliberate. There are many books about intelligence and many more about cognitive development. However, there are no books that are about intelligence and development. This is because intelligence and development are regarded as merely different ways of talking about the same thing. If we are interested in intelligence, we talk about the steady state structure of cognition; and if we are interested in development, we talk about how this structure changes. So, typically, psychologists theorize about intelligence, or they theorize about its development. This book is about intelligence and cognitive development because I think these terms refer to quite distinct cognitive properties. The theory [presented in this book] will attempt to specify the mechanisms that underlie individual differences and their relationship with those that underlie cognitive development. Along the way it will have to elucidate the relationship between intelligence and knowledge, accommodating both biological and cognitive conceptions. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Data from 14 nations reveal IQ gains ranging from 5 to 25 points in a single generation. Some of the largest gains occur on culturally reduced tests and tests of fluid intelligence. The Norwegian data show that a nation can make significant gains on a culturally reduced test while suffering losses on other tests. The Dutch data prove the existence of unknown environmental factors so potent that they account for 15 of the 20 points gained. The hypothesis that best fits the results is that IQ tests do not measure intelligence but rather a correlate with a weak causal link to intelligence. This hypothesis can also explain differential trends on various mental tests, such as the combination of IQ gains and Scholastic Aptitude Test losses in the United States.
Conference Paper
Humane-egalitarian ideals, whose aims are group justice and reducing environmental inequality and privilege, must be rested against reality: as revealed by psychology and other social sciences. Four issues are addressed: the equation between IQ and intelligence, whether group potential is determined by a group's mean Ie, whether the Black-White re gap is genetic, and the meritocratic thesis that genes for Ie,will become highly correlated with class. Massive Ie gains ol er time test the IQ-intelligence equation, reveal groups who achieve far beyond their mean les, and falsify prominent arguments for a genetic racial le gap. Class re trends suggest America is not evolving toward a meritocracy, brit ct core refutation of that thesis is needed and supplied. Finally, the viability of humane ideals is assessed against a worst-case scenario.
More than 25 years of research suggests that the measure inspection time (IT) does capture low-level aspects of cognitive functioning that contribute to human intelligence. However, recent evidence does not support earlier claims that IT estimates the speed of a single mechanism like “sampling input” or “apprehension.” Rather, together with other tasks that employ pattern backward masking to limit the duration for which information is available for processing, IT is probably sensitive both to focused attentional capacities to detect organization and change under severe time constraints and to decision processes, ongoing beyond mask onset, that monitor responding. Among normal young adults, IT is correlated with the broad psychometric factor Gs (“speediness”). This mediates correlation with general intelligence. In this group, IT is not correlated with Gf. However, whether this outcome generalizes to samples of persons with an intellectual disability, to young children, or to elderly persons is not yet known. Psychological processes underpinning IT are currently only speculatively defined, but it should prove possible to unravel these by experimentation. To this end, backward masking procedures are arguably more theoretically tractable than reaction time tasks because they reduce the impact of higher-level cognitive strategies on performance. On this basis, IT may hold promise as a means for developing partial explanations for intelligence in psychological terms. However, whether this is realized depends on identifying the psychological functions that support IT.
Original and revised editions of the Bayley Scales of Infant Development were administered to 97 healthy infants aged 18 to 27 months. Concurrent scores on the Mental Development Index (MDI) were consistent with an average increase of approximately 18 points over the past 25 years. MDI gains varied from 4 to 35 points and were apparent for both genders and for infants from different cultural backgrounds, but they were marginally higher for females than for males, and significantly higher for Australian infants in comparison to infants from non-English-speaking families. No significant effect was found for age. These gains for Australian infants were higher than those reported for an American infant population. Although the gains observed in all levels of intelligence ranged from −1 to 2.2 standard deviations, there was significantly more improvement for higher scoring infants than for infants in the normal range. This finding was, therefore, similar to IQ gains which have been reported for older children and adults. The results suggest that norms for the first version of the Bayley Scales of Infant Development, and probably for all other infant developmental scales that have not been updated, now have questionable research and clinical utility.