Journal of Gerontology: PSYCHOLOGICAL SCIENCES

2008, Vol. 63B, No. 4, P235P240

Copyright 2008 by The Gerontological Society of America

Age and Ability Affect Practice Gains in

Longitudinal Studies of Cognitive Change
Patrick Rabbitt,1 Mary Lunn,2 Danny Wong,2 and Mark Cobain3
1

Department of Experimental Psychology, University of Oxford, England, and University of West Australia
2
Department of Statistics, University of Oxford, England.
3
Department of Research and Development, Unilever PLC, London, England.

Key Words: Cognitive gainLongitudinal studyPractice gains.

MAIN goal for cognitive gerontology is to study

individual differences in trajectories of cognitive change
in old age and so to identify factors that accelerate and retard
changes in mental abilities. This requires longitudinal designs
in which participants repeatedly take the same, or very similar,
cognitive tests. A new problem then arises because practice
gains that are due to repeated testing can disguise declines
associated with increasing age and frailty. This has long been
recognized as a theoretical possibility (Schaie, 1965; Schaie,
Labouvie, & Barrett, 1973) and has been shown to be a problem
in the assessment of testretest reliability of clinical diagnostic
measures (Beglinger, 2005; Dikmen, Heaton, Grant, & Temkin,
1999; Falleti, 2006; Kulik, Chen-Lin, Kulik, & Bangert, 1984;
Mitrushima & Satz, 1991). However, it is only recently that the
development of new statistical models has made it possible to
demonstrate formally that when testretest intervals are as long
as 4 to 7 years, practice gains are large enough to cause serious
underestimations of the true rates of declines caused by increasing age, pathology, and frailty (Ferrer, Salthouse, McArdle,
Stewart, & Schwartz, 2004; Ferrer, Salthouse, Stewart, &
Schwartz, 2004; Rabbitt, Diggle, Holland, & McInnes, 2004;
Rabbitt, Diggle, Smith, Holland, & McInnes, 2001). Because it
is clear that practice gains occur in longitudinal studies, a further
possibility is that they may vary between individuals. This
seems probable because many brief laboratory experiments have
found that elderly people learn new material more slowly (for
reviews, see Craik & Jennings, 1992; Kausler, 1990; and
Rabbitt, 2002). Another possibility, also suggested by comparisons on brief training studies, is that as intervals between
successive testing become very long, older participants will
forget more than young participants and so appear to benefit less
from repeated testing. In either case, the acceleration of agerelated cognitive decline would be overestimated.
There is also evidence that practice gains may vary with
individual differences in general fluid intelligence (gF). Rabbitt,

Bithell, Perdicou, Stollery, and Moore (2007 submitted)

practiced 93 volunteers aged from 61 to 82 years on eight
different tests of verbal learning, spatial learning, motor
learning, and information processing speed. These researchers
found that initial scores and subsequent rates of improvement
were uncorrelated between different cognitive tests but were
significantly predicted by unadjusted scores on three different
tests of gF, namely, the Heim 1970 AH4-1 and AH4-2 and the
Cattell and Cattell (1960) Culture Fair test. Other studies have
also shown that, irrespective of their ages, individuals with
higher intelligence test scores learn new material more rapidly
and remember it better than do those with lower intelligence
test scores (e.g., Rabbitt & Anderson, 2006). This made it
useful for us to check whether there are indeed marked individual differences in practice gains during longitudinal studies
and whether these gains systematically differ with an individuals age between 60 and 82 years and with an individuals
intelligence test scores on entry to a longitudinal study.
Apart from these methodological issues, individual differences in practice gains are in themselves substantively interesting as a further index of the age-related cognitive changes that
longitudinal studies purport to assess. To confidently extrapolate from age-related changes observed in laboratory studies to
changes in the ability to manage everyday life, we must
recognize reciprocities between what people do and what
happens to them and, even in old age, how and by how much
their experiences alter their abilities.
In order to estimate and compare practice gains accurately, we
also need to take account of some other frequently neglected
factors. One is self-selection on entry. Volunteers for demanding longitudinal studies are unusually healthy, able, and
highly motivated members of their age groups (Lachman,
Lachman, & Taylor, 1982). This elite bias may be unavoidable
but, because volunteers are typically recruited in successive
waves over many years, it is at least possible to check whether

P235

Downloaded from http://psychsocgerontology.oxfordjournals.org/ at Galter Health Sciences Library on June 26, 2015

During a 20-year longitudinal study, 5,842 participants aged 49 to 93 years significantly improved over two to four
successive experiences of the Heim AH4-1 intelligence test (first published in 1970), even with between-test
intervals of 4 years and longer. After we considered significant attrition by death and dropout and the effects of
gender, socioeconomic advantage, and recruitment cohort, we found that participants with high intelligence test
scores showed greater improvement than did those with lower intelligence test scores. Practice gains also reduced
with age, even after we took into consideration the individual differences in intelligence test scores. This
emphasizes the methodological point that neglect of individual differences in improvement during longitudinal
studies underestimates age-related changes in younger and more able participants and the theoretical point that,
like all experiences during everyday life, participation in longitudinal studies alters the ability of aging humans to
cope with cognitive demands to different extents according to their baseline abilities.

P236

RABBITT ET AL.

METHODS

Participants, Procedure, and Materials

Researchers recruited a sample of 5,842 volunteers, that is,
2,615 residents of Greater Manchester and 3,277 residents of
Newcastle-upon-Tyne, United Kingdom, by appeals on local
media and by word of mouth. The 1,711 men were between the
ages of 49 and 93 years (M 65.6, SD 7.7), and the 4,131
women were between 49 and 92 years (M 64.4, SD 7.8) All
traveled independently to the Department of Psychology at
either the University of Manchester or the University of
Newcastle, where they completed batteries of cognitive tasks in
quiet rooms supervised by two experienced testers. Participants
were each reimbursed expenses of 5 (UK) per session. A
search by HM Registry Office UK identified all 2,342 deaths
between 1983 and the close of the census on July 31, 2004.
Between 1983, when the study began, and July of 2003, when it
ended, there were 3,204 dropouts, of whom 1,208 also died
before the 2004 census. Because many dropouts could only be
identified by failures to return for further testing, dates of
dropout are recorded as the last session attended. The remaining
1,996 dropout participants did not drop out before July of 2003
and also survived the July 2004 census.
Earlier analyses by some of us and our colleagues, namely,
Rabbitt, Diggle, Holland, McInnes, Bent, and colleagues
(2004), examined only Newcastle residents and found strong
practice effects on the Heim (1970) AH4-1 intelligence test,
contrasting with relatively slight, though still significant, effects
on verbal learning tasks and vocabulary tests. As AH4-1 scores
show the greatest practice gains, they were selected as being the
most sensitive indices of possible individual differences in
practice effects. The AH4-1 intelligence test consists of 64
problems with equal numbers of logic problems, verbal
comparisons, arithmetic problems, and number series. After

introductory practice on one question from each of these categories, participants answer as many problems as possible within
10 minutes. Scores are the percentages of correct answers.
The results analyzed are from one Heim (1970) AH4-1 group
intelligence test, which was included in a test battery that
was repeated at 4-year intervals. Results for Manchester and
Newcastle are closely similar and so we combine them.
Previous analyses of data from the Newcastle sample by our
earlier group (Rabbitt, Diggle, Holland, McInnes, Bent, et al.,
2004) found that participants levels of performance on cognitive
tests markedly vary with their levels of socioeconomic advantage
(SEA) as categorized by reference to the UK Office of
Population Census and Surveys Classification of Occupational
Categories (1980). Categories are SEA C1 (n 261), made up of
professionals such as doctors, lawyers, senior managers, and
academics; SEA C2, (n 1,854), made up of other professionals
such as schoolteachers, pharmacists, and junior managers; SEA
C3N (n 2,064), made up of skilled nonmanual workers such as
secretaries; SEA C3M (n 771), made up of skilled manual
workers such as craftsmen, joiners, fitters, and machinists; SEA
C4, (n 433), made up of nonskilled nonmanual workers such as
clerical assistants and storekeepers; and SEA C5 (n 40), made
up of nonskilled manual workers such as laborers, cleaners, and
janitors. The remaining 427 did not record occupational data.
Our group found that cognitive test scores also significantly
differ with gender: Men scored higher on tests of gF and women
higher on tests of verbal memory and learning. Recruitment
cohorts also differ significantly in test scores and levels of
socioeconomic advantage. Accordingly, we also entered occupational category and gender into the analyses.
We found that cognitive test scores, and rates of change,
markedly varied between subsets of individuals who completed
the study between 1983 and 2003, those who died during the
course of the study, those who dropped out during the study but
survived beyond a census of deaths completed by HM Registry
of Births, Marriages and Deaths, Stockport UK in July of 2004,
and those who dropped out during the study but subsequently
also died before the 2004 census (Rabbitt, Lunn, & Wong,
2005). As the incidence and timing of deaths and dropouts
might affect the sizes of practice effects, we found necessary to
test for interactions between practice and deathdropout status.
Accordingly, we divided participants into 11 subgroups
according to their histories of survival, dropout, or dropout
followed by death with respect to the time points of the four
quadrennial test sessions at which the Heim (1970) AH4-1 test
was administered. These groups were as follows.
Group C completed the study and survived the 2004 census
of deaths (n 1,510); Group D1 died between Test Session 1
and Test Session 2 (n 365); Group D2 died between Test
Session 2 and Test Session 3 (n 409); Group D3 died between
Test Session 3 and Test Session 4 (n 246); Group D4 died
between Test Session 4 and the 2004 census of deaths (n 116);
Group WD1 withdrew before Test Session 2 and subsequently
died (n 745); Group WD2 withdrew before Test Session 3 and
subsequently died (n 354); Group WD3 withdrew before Test
Session 4 and subsequently died (n 109); Group W1 withdrew
before Test Session 2 and survived beyond the 2004 census (n
1,013); Group W2 withdrew before Test Session 3 and survived
beyond the 2004 census (n 595); and Group W3 withdrew
before Test Session 4 and survived the 2004 census (n 388).

Downloaded from http://psychsocgerontology.oxfordjournals.org/ at Galter Health Sciences Library on June 26, 2015

apparent differences in trajectories of change and in practice

effects are affected by demographic factors such as gender, age,
differences between recruitment cohorts, differences in levels
of socioeconomic advantage, and geographical locations of
residence. A more difficult problem is to allow for selective
attrition that occurs because older, frailer, and less able
individuals die and withdraw earlier than do others. The longer
studies continue, the more elite and atypical of their age groups
the survivors become (e.g., Lachman, Lachman, & Taylor;
Rabbitt, 2002; Rabbitt, Lunn, & Wong, 2005; Rabbitt, Watson,
Donlan, Bent, & McInnes, 1994a, 1994b; Rabbitt, Wong, &
Lunn, in press; Schaie et al., 1973). Because true trajectories of
change cannot be estimated unless deaths and dropouts are taken
into consideration, these must also be included in the analysis.
Data collected during the University of Manchester longitudinal study of cognitive change in healthy old age, described in
detail elsewhere (Rabbitt, Diggle, Holland, McInnes, Bent, et al.,
2004), allowed us to examine individual differences in improvements that were due to practice during two to four successive experiences of the Heim (1970) AH4-1 intelligence test,
administered to participants at 4-year intervals over total periods
of 8 to 16 years, after the effects of recruitment cohort, city of
residence, gender, and socioeconomic advantage and selective
attrition by death and dropout had been taken into consideration.

AGE, ABILITY, AND LONGITUDINAL PRACTICE GAINS

RESULTS

Table 1. Estimated Parameters for the Mean AH4-1

Percentage Score
Parameter

Estimate

SE

Intercept
Age 70
(Age 70)2

53.04
0.76
0.01

1.08
0.06
0.004

48.24
12.83
2.50

,.0001
,.0001
.0126

Practice effect
Practice 2 vs practice 1
Practice 3
Practice 4

3.94
6.45
8.37

0.31
0.53
0.92

12.81
12.25
9.12

,.0001
,.0001
,.0001

1.57

0.65

2.42

.0157

9.05
6.58
7.96
11.33
17.07
2.47

1.28
0.70
0.96
1.24
3.45
1.16

7.37
9.37
8.26
9.17
4.95
2.12

,.0001
,.0001
,.0001
,.0001
,.0001
.0337

0.68
1.83
3.00
9.27
0.98
3.59

0.71
1.13
1.00
2.84
1.45
0.96

0.96
1.62
3.02
3.26
0.67
3.73

.3384
.1056
.0026
.0011
.5001
.0002

8.91
5.70
2.87
3.14
9.71
7.69
6.20
9.37
4.35
2.06

1.71
1.33
1.67
3.47
1.02
1.53
3.91
0.85
1.00
1.23

5.22
4.29
1.72
0.91
9.47
5.02
1.59
11.08
4.34
1.68

,.0001
,.0001
.0864
.3646
,.0001
,.0001
.1128
,.0001
,.0001
.0937

0.15
0.32
0.40
0.10

0.05
0.09
0.13
0.04

3.01
3.57
3.05
2.27

.0026
.0004
.0023
.0234

Gender
Female vs male
Socioeconomic status
C1 vs C3(N)
C2
C3(M)
C4
C5
Missing
Entry year
1986 vs 1985
1987
1988
1990
1991
1992
Pattern
D1 vs C
D2
D3
D4
WD1
WD2
WD3
W1
W2
W3
Interactions

Effects of Age on Improvement With Practice

The results of a linear mixed effects pattern-mixture model
are shown in Table 1. Age is centered at 70 years, because this
is close to the mean age of the sample. The linear effect of age
is highly significant and the significant quadratic age term
indicates that rates of decline in participantsAH4-1 scores
accelerate as they grow older. The significant difference for
gender occurs because average scores for men are 1.57% higher
than average scores for women. There are also highly
significant effects of socioeconomic advantage and of recruitment year. There are highly significant differences between
groups with different histories of survival, death, and dropout
over successive time points during the study. Finally, after
variance associated with all of these other factors has been
taken into consideration, practice gains are robustly significant.
Participants score significantly higher on their second, third, or
fourth experience than they do on their first experience of the
AH4-1 test. The key finding of interest is the significant
interaction between the effects of age and the amount of
practice gains between Sessions 1 and 2, Sessions 1 and 3, and
Sessions 1 and 4. These occur because, on all these transitions,
the older participants benefit less from practice than the
younger participants do. Another new finding is the Age 3

(Age
(Age
(Age
(Age

70)
70)
70)
70)

3
3
3
3

Practice 2
Practice 3
Practice 4
Gender

Note: The estimated parameters use a pattern-mixture model with an

11-level pattern of groups experiencing survival, dropout, death, and dropout
followed by death. Note that age is centered at 70 years, which is approximately the mean age for the sample. AH4-1 Heim test of general fluid intelligence. Covariance and residual parameters are as follows: rA, 13.33; rB,
0.39; qAB, 0.15; re, 5.54. For the socioeconomic status information, the category descriptions for levels of socioeconomic advantage are given in the text,
in the Participants, Procedure, and Materials section. For the pattern information, the 11 subgroups are also described in that section.

Gender interaction that occurs because men experience more

rapid age-related declines than do women. No other interactions
are significant, so there is no evidence that the amount of
practice gain is affected by proximity to death or to dropout.

Effects of Intelligence on Improvement With Practice

To examine the effects of individual differences in gF (AH4-2
scores) on practice gains, an obvious procedure was to fit
a linear term in AH4-2 and examine the interactions. However,
in the present case this was clearly inappropriate, and the fit of

Downloaded from http://psychsocgerontology.oxfordjournals.org/ at Galter Health Sciences Library on June 26, 2015

Because our study aim was to examine, independently, the

effects of age and of general intellectual ability on practice
effects, we found it necessary to have a different measure of
ability than scores on the AH4-1, for which practice data were
analyzed. This was available because the Heim (1970) AH4-2,
a nonverbal test of general fluid ability, had been administered
to all participants on each testing occasion cotemporaneously
with the AH4-1. Within the entire sample, the correlation
between age-unadjusted percentage correct of AH4-1 and AH4-2
scores is r .84, so AH4-2 scores are a good independent
measure of gF and, specifically, of individual differences in
ability on the AH4-1. The AH4-2 intelligence test also consists
of 64 problems. Each consists of a series of five line drawings
with a missing element that must be supplied from among five
provided alternatives. Solutions of series require mental rotation, addition, and subtraction or irregular shapes and recognition of logical progressions. After introductory practice on one
question from each of these categories, participants attempt to
solve as many problems as possible within 10 minutes. Scores
used are percentages of correct answers.
We conducted our analysis by using a linear mixed effects
model with fixed effects including age, demographic factors
such as gender, socioeconomic advantage, city of residence, and
recruitment cohort, AH4-2 score, practice session, and death and
withdrawal group. The random effects were an intercept for each
individual and an individual age effect, so that variance between
individuals increased with age. A key point of the analysis was
the introduction of the group effect, allowing the groups from C
to W3 to influence the regression coefficients of the usual fixed
effects. This is a so-called pattern-mixture model. If the
regression on the groups, which could include interaction terms,
is significant, then this model allows the time and type of
dropout to influence the effects of other factors (Little, 1993).

P237

RABBITT ET AL.

P238

Table 2. Estimated Parameters for the Mean AH4-1

Percentage Score: Final Model
Estimate

SE

Intercept
Age 70
Low AH4-2 vs middle AH4-2
High AH4-2

52.68
0.52
21.67
20.92

0.99
0.05
1.17
3.16

53.07
10.37
18.58
6.61

,.0001
,.0001
,.0001
,.0001

3.11
5.14
7.04

0.31
0.52
0.84

10.12
10.12
8.35

,.0001
,.0001
,.0001

0.51

0.60

0.85

.3981

7.64
6.29
6.16
8.03
9.68
0.19

1.18
0.64
0.89
1.15
3.20
1.07

6.47
9.77
6.94
7.00
3.02
0.17

,.0001
,.0001
,.0001
,.0001
.0025
.8626

0.95
1.69
2.95
7.08
0.25
2.09

0.65
1.04
0.91
2.61
1.33
0.89

1.45
1.63
3.24
2.71
0.18
2.36

.1460
.1034
.0012
.0067
.8541
.0184

7.36
5.22
2.05
3.03
8.66
6.57
6.69
8.12
3.92
2.01

1.57
1.22
1.53
3.16
0.94
1.41
3.57
0.78
0.92
1.12

4.68
4.29
1.34
0.96
9.21
4.67
1.87
10.44
4.27
1.79

,.0001
,.0001
.1800
.3369
,.0001
,.0001
.0614
,.0001
,.0001
.0735

0.25
0.53
0.69
0.11
0.40
0.04
3.93
5.30
7.66
3.72
4.40
4.30

0.03
0.05
0.07
0.04
0.14
0.26
1.42
2.72
4.04
1.79
3.05
4.39

7.59
11.44
9.53
2.52
2.88
0.14
2.77
1.95
1.89
2.08
1.44
0.98

,.0001
,.0001
,.0001
.0117
.0040
.8855
.0057
.0516
.0582
.0377
.1500
.3272

Practice effect
Practice 2 vs practice 1
Practice 3
Practice 4
Gender
Female vs male
Socioeconomic status
C1 vs C3(N)
C2
C3(M)
C4
C5
Missing
Entry year
1986 vs 1985
1987
1988
1990
1991
1992
Pattern
D1 vs C
D2
D3
D4
WD1
WD2
WD3
W1
W2
W3
Interactions
(Age 70) 3 Practice 2
(Age 70) 3 Practice 3
(Age 70) 3 Practice 4
(Age 70) 3 Gender
(Age 70) 3 Low AH4-2
(Age 70) 3 High AH4-2
Low AH4-2 3 Practice 2
Low AH4-2 3 Practice 3
Low AH4-2 3 Practice 4
High AH4-2 3 Practice 2
High AH4-2 3 Practice 3
High AH4-2 3 Practice 4

Note: The estimated parameters use a pattern-mixture model with an 11level pattern of groups experiencing survival, dropout, death, and dropout followed by death. Three levels of AH4-2 scores (AH4-1 and AH4-2 are Heim
tests of general fluid intelligence) are also coded. Note that age is centered at
70 years, which is, approximately, the mean for this sample. Covariance and
residual parameters are as follows: rA, 12.08; rB, 0.37; qAB, 0.16; re, 5.56.
For the socioeconomic status information, the category descriptions for levels
of socioeconomic advantage are given in the text, in the Participants, Procedure, and Materials section. For the pattern information, the 11 subgroups are
also described in that section.

Downloaded from http://psychsocgerontology.oxfordjournals.org/ at Galter Health Sciences Library on June 26, 2015

Parameter

this model was not good because there is little practice effect at
the lowest and highest scoring individuals. We also considered
a linear quadratic term in AH4-2. Again the fit to the data was
not good. At this stage it became apparent that (a) the
relationship is complicated and (b) the precise form of this
complex relationship is of substantive interest in interpreting the
pattern of individual differences in practice effects. The form of
the interaction reveals that estimates of true practice effects
may be miscomputed both because of ceiling effects for the
most able and floor effects for the least able. In other words,
the relationship between practice gains and AH4-2 scores is
represented by an inverted U function, and this feature of the
data is best revealed by subdivision into three groups. Given
a requirement for simplicity of interpretation, we should look for
two change points to give lowest, middling, and highest score
groups. This selected model gave a good fit to the data, and the
interaction was significant. This procedure would clearly have
been inappropriate had it been intended be used as a prediction
for individual trajectories. This is not the case. It is only intended
to test for overall differences between ability groups. The results
of fitting this final model are shown in Table 2.
Because of the strong correlation between AH4-1 and AH4-2
scores, we expected the overall effects of level of ability to be
highly significant. After we took level of ability into
consideration, we found that the linear but not the quadratic
effects of age (centered at the average age of 70 years) and the
effects of socioeconomic advantage, recruitment year, and
survival, death, and withdrawal history were similar to those
found in the first analysis. An interesting further detail is that
the effect of gender was now abolished. We interpret this as
evidence that all of the variance in AH4-1 scores that is
associated with differences between men and women was now
picked up by scores on another, different, well-validated
nonverbal test of gFthe AH4-2. In other words, differences
in AH4-1 test scores between men and women are not due to
any factor, for which gender is a proxy, other than differences
in the particular kind of general fluid mental ability that is
measured by the AH4-2 and AH4-1 tests. The significant Age 3
Gender interaction replicates that found in the first analysis,
with women showing less age-related decline in AH4-1 scores.
The Age 3 Practice interactions found in the first analysis
still remain significant, even after the effects of differences in
AH4-2 scores have been taken into consideration. In other
words, greater age significantly reduces practice gains, even
after individual differences in general fluid mental ability have
been taken into account. We interpret this as new evidence that
not all of the age-related cognitive changes that lead to declines
in efficiency of learning (practice effects) can be explained by
differences in gF.
The significant interaction between level of ability and
practice gains is a new finding, but the complex nature of this
relationship requires interpretation. For both the low-ability
group and the high-ability group, there are negative interactions
with practice. That is, both the low-ability and the high-ability
groups benefit less from practice than does the medium-ability
group. For the low-ability group this interaction is significant
for comparisons between Session 1 and each of Sessions 2, 3,
and 4. For the high-ability group the interaction is significant
for the comparison between Session 1 and Session 2 but not for
the comparisons between Sessions 1 and 3 or Sessions 1 and 4.

AGE, ABILITY, AND LONGITUDINAL PRACTICE GAINS

DISCUSSION
The present analyses of data from the combined Manchester
and Newcastle samples replicate the main findings of earlier
analyses of data from the Newcastle longitudinal samples by
Rabbitt, Diggle, Holland, McInnes, Bent, and colleagues
(2004) and the Manchester and Newcastle sample by Rabbitt,
Lunn, and Wong (2005). There are marked effects of socioeconomic advantage and of year of entry to the study. Men
score higher than women but women experience less rapid
age-related decline. This is of interest because it probably
reflects the fact that, in Western industrialized societies, women
live longer than men and so experience slower biological
changes and retain their competence to a later calendar age, as
distinct from biological age. A further new detail in exploration
of gender differences is that differences between men and
women on one test of gF, the AH4-1, disappear when scores
on a different, highly correlated test of gF, that is, the AH4-2,
are taken into consideration. Thus we have no evidence that
factors other than levels of gF are responsible for gender differences in performance of intelligence tests.
The significant effects of survival, death, and dropout history
replicate earlier effects reported by Rabbitt, Lunn, and Wong
(2005) and Rabbitt, Lunn, and Wong (2007, submitted).
Performance declines with approach to either death or dropout
and the amounts of declines preceding death and dropout are
closely similar; comparing equivalent time points of dropout,
we see that the effects of approaching dropout are greater when
dropout is shortly followed by death than when dropout is
followed by survival. This empirically shows that not only the
effects of age-related decline but also the effects of other factors
such as health or socioeconomic advantage, gender differences,
and practiceall of which influence rates of changes in
cognitive performance over timeare miscalculated unless
both the occurrence and timing of both deaths and dropouts are
also logged and taken into consideration.
The main point of the present analyses is that, even after we
take into account the effects of initial selection and selective
attrition of a sample, the marked practice effects during
a prolonged longitudinal study found by some of us in earlier
research (Rabbitt, Diggle, Holland, & McInnes, 2004) are
replicated on a different and very much larger sample of
participants (i.e., the Manchester and Newcastle group rather
than just the Newcastle cohort of the University of Manchester
longitudinal study). As in our earlier analyses (Rabbitt, Diggle,
Holland, & McInnes, 2004), significant practice improvements
are found even when intervals between successive presentations

of the task are as long as 4 years. Indeed, the present analyses

show that even the oldest participants show gains with
quadrennial experiences of the AH4-1 test over periods of 8
and of 12 years.
The new questions asked by these analyses were whether
practice gains during a longitudinal study differ between
individuals of different ages and different levels of ability. The
first analysis shows that younger participants gain more from
practice than do relatively older participants. The second
analysis shows that, even after variance associated with age
differences has been taken into account, higher scores on one
test of general intellectual ability, the AH4-2 intelligence test,
are associated with significant increments in practice gains on
another, the AH4-1. The further finding that age differences in
practice remain significant even after effects of differences in
unadjusted intelligence test scores have been considered makes
the additional new point that the reduction in practice effects
with increasing age cannot entirely be attributed to the fact that
peoples levels of gF decline as they grow older. This is
inconsistent with recent speculations that age-related differences in all cognitive skills can be parsimoniously treated as
differences in gF (e.g., Anderson, 1992, Deary 2000). The
particular age-related cognitive changes that reduce practice
gains in this longitudinal study are not entirely captured by
scores on a well-validated test of gF.
Because the form of the interaction between AH4-2 test score
group and practice gains is complex, it requires interpretation.
Both the high-ability and the low-ability groups show less
practice gains than does the medium-ability group. Many
participants in the high-ability group scored the maximum
possible, and many others had such high scores that there was
little scope for improvement. Thus ceiling effects offer
a sufficient explanation as to why high-ability participants show
less improvement. Within the range of scores in which ceiling
effects are no longer possible so that practice gains can appear,
the significant advantage for the middle-ability group over the
low-ability group shows that practice gains do increase with gF.
We must conclude that all previous analyses of longitudinal
data in which only mean values for practice effects have been
calculated have, because of this, significantly underestimated
the true amounts of age-related declines for younger and more
able participants. For older and less able participants, who show
significantly less improvement with practice, estimated rates of
age-related declines have been closer to their true values.
This finding highlights a nontrivial issue because, unless it is
implemented in further analyses, we cannot confidently address
some theoretically interesting and socially important questions
such as whether more and less able individuals experience
different rates of cognitive decline and whether data show
generation cohort effects. For example, do young participants in
longitudinal studies, who have benefited from recent improvements in socioeconomic conditions and in medical care,
experience less rapid cognitive decline than earlier generations
who have, historically, been disadvantaged in these respects?
To find reliable answers to these questions, we must measure
and consider individual differences in practice effects.
Perhaps the most general point that these analyses make is
that individuals cognitive abilities are altered by their
involvement in prolonged longitudinal studies, just as by their
involvements in all other aspects of their everyday lives. The

Downloaded from http://psychsocgerontology.oxfordjournals.org/ at Galter Health Sciences Library on June 26, 2015

There is also a significant interaction between age and low

level of ability relative to age and either medium or high ability.
The age-related decline of 0.52 per year of age overstates
the amount of decline for those individuals at the bottom level
of the ability range. It is possible that it is the inclusion of
this interaction that results in the quadratic effect in which age
becomes no longer significant in this model. As in the first
analysis described earlier, the absence of any interaction
between practice effects and the occurrence or timing of deaths
or withdrawal means that there is no evidence that approach to
death or dropout affects the amount of practice gains after
differences in ability have been taken into consideration.

P239

P240

RABBITT ET AL.

REFERENCES
Anderson, M. (1992). Intelligence and development: A cognitive theory.
Oxford. Blackwell.
Beglinger, A. ( 2005). Practice effects and the use of alternate forms in
serial neuropsychological testing. Archives of Clinical Neuropsychology, 20, 512529.
Craik, F. I. M., & Jennings, J. M. (1992). Human memory. In F. I. M. Craik &
T. Salthouse (Eds.), The handbook of aging and cognition (pp. 51110).
Hillsdale, NJ: Erlbaum.
Deary, I. J. (2000). Looking down on human intelligence. Oxford, England:
Oxford University Press.
Diggle, P. J., Liang, K. Y., & Zeger, S. L. (1994). Analysis of longitudinal
data. Oxford, England: Oxford University Press.
Dikmen, S. S., Heaton, R. K., Grant, J., & Temkin, N. R. (1999). Test
retest reliability and practice effects of Expanded HalsteadReitan
Neuropsychological Test Battery. Journal of the International Neuropsychological Society, 5, 346356.
Falleti, K. (2006). Practice effects associated with the repeated assessment
of cognitive function using the Cog State Battery at 10 min, one week,
and one month testre-test intervals. Neuropsychology, Development
and Cognition: Journal of Clinical and Experimental Neuropsychology,
28A, 10951112.
Ferrer, E., Salthouse, T., McArdle, J., Stewart, W. F., & Schwartz, B.
(2005). Estimating retest effects in longitudinal assessments of cognitive functioning in adults between 18 and 60 years of age. Developmental Psychology, 40, 813822.
Ferrer, E., Salthouse, T., Stewart, W., & Schwartz, B. (2004). Modeling age
and re-test processes in longitudinal studies of cognitive abilities.
Psychology and Aging, 19, 243259.
Heim, A. W. (1970). The AH4 group test of general intelligence. Windsor,
England: NFER-Nelson.
Lachman, R., Lachman, J. L., & Taylor, D. W. (1982). Reallocation of
mental resources over the productive lifespan: Assumptions and task

analyses. In F. I. M. Craik & S. Trehub (Eds.), Aging and the cognitive

process (pp. 304350). New York: Plenum Press.
Kausler, D. H. (1990). Experimental psychology, cognition and human
aging. New York: Springer-Verlag.
Kulik, J. A, Chen-Lin, C., Kulik, R. L., & Bangert, L. (1984). Effects of
practice on aptitude and achievement test scores. American Educational
Research Journal, 21, 435447.
Little, R. J. A. (1993). Pattern-mixture models for multi-variate incomplete
data. Journal of the American Statistical Association, 88, 125134.
Little, T. D., Lindenberger, U., & Maier, H. (2000). Selectivity and
generalizability in longitudinal research: On the effects of continuers
and dropouts. In T. D. Little, K. U. Schanabel, & J. Baumert, (Eds.),
Modelling longitudinal and multilevel data: Practical issues, applied
approaches and specific examples (pp. 187200). Mahwah, NJ:
Erlbaum.
Mitrushima, M., & Satz, P. (1991). Effect of repeated administration of
a neuropsychological test battery in the elderly. Journal of Clinical
Psychology, 47, 790801.
Nagi, M. H., & Stockwell, E. G. (1973). Socioeconomic differentials in
mortality by cause of death. Health Services Reports, 88, 449456.
Office of Population Censuses and Surveys. (1980). Classification of
occupations 1980. London, England: Her Majestys Stationary Office.
Rabbitt, P. (2002). Cognitive gerontology. In H. Pashler & J. Wixted (Eds.),
Stevens handbook of experimental psychology (Vol 4., pp. 401453).
New York: Wiley.
Rabbitt, P., Diggle, P., Holland, F., & McInnes, L. (2004). Practice and
drop-out effects during a 17-year longitudinal study of cognitive aging.
Journal of Gerontology: Psychological Sciences, 59B, P84P97.
Rabbitt, P., Diggle, P., Smith, D., Holland, F., & McInnes, L. (2001).
Identifying and separating the effects of practice and of cognitive
ageing during a large longitudinal study of elderly community residents.
Neuropsychologia, 39, 532543.
Rabbitt, P., Lunn, M., & Wong, D. (2005). Neglect of dropout
underestimates effects of death in longitudinal studies. Journal of
Gerontology: Psychological Sciences, 60B, P106P109.
Rabbitt, P., Watson, P., Donlan, C., Bent, N., & McInnes, L. (1994a).
Subject attrition in a longitudinal study of cognitive performance in
community resident elderly people. In B. J. Vellas, J. L. Albarende, &
P. J. Garry (Eds.), Facts and research in gerontology: Epidemiology
and aging (pp. 2934). Paris: Serdi.
Rabbitt, P. M., Watson, P., Donlan, C., Bent, L., & McInnes, L. (1994b).
Subject attrition in a longitudinal study of cognitive performance in
community-based elderly people. Facts and Research in Gerontology,
14, 203207.
Rabbitt, P. M. A., & Anderson, M. (2006). The lacunae of loss? Aging and
the differentiation of cognitive abilities. In E. Bialystok & F. I. M. Craik
(Eds.), Lifespan cognition: Mechanisms of change (pp. 331343).
Rabbitt, P. M. A., Diggle, P., Holland, F., McInnes, L., Bent, N., Abson,
V., et al. (2004). The University of Manchester longitudinal study of
cognition in normal healthy old age, 1983 through 2003. Aging,
Neuropsychology and Cognition, 11, 245279.
Schaie, K. W. (1965). A general model for the study of developmental
problems. Psychological Bulletin, 64, 92107.
Schaie, K. W., Labouvie, G. V., & Barrett, T. J. (1973). Selective attrition
effects in a fourteen-year study of adult intelligence. Gerontology, 28,
328334.
Received November 30, 2005
Accepted January 2, 2008
Decision Editor: Thomas M. Hess, PhD

Downloaded from http://psychsocgerontology.oxfordjournals.org/ at Galter Health Sciences Library on June 26, 2015

variety and the particular nature of our experiences significantly

alter our abilities to cope with demands made by our everyday
lives or, indeed, by psychological tests. People do not simply
and inexorably decline in mental ability as they grow old. Their
everyday experiences may also degrade or enhance their ability
to cope with the demands that their lives make upon them. We
may speculate that, as well as learning to cope with particular
life demands and situations, people also gradually learn to adapt
to and cope with the changes in their mental abilities brought
about by neurophysiological aging. It is striking that, even in
elderly individuals, very brief episodes of practice, even as brief
as 10 minutes on first encounter with the simple problems in
the AH4-1, can bring about domain-specific improvements
that last for 4 to 7 years (see Rabbitt, Diggle, Holland, McInnes,
Bent, et al., 2004). It therefore seems less interesting to continue to explore changes in individuals scores on particular
cognitive tests than to study interactions between their baseline
levels of ability and their life experiences, and so the levels
of competence they can achieve and so also the extent to which
they can maintain, in old age, their ability to cope with their
lives and with any laboratory experiments into which they may
be inveigled.