Archives of Gerontology and Geriatrics 59 (2014) 577583

Contents lists available at ScienceDirect

Archives of Gerontology and Geriatrics

journal homepage: www.elsevier.com/locate/archger

Inuential factors affecting age-related self-overestimation of

step-over ability: Focusing on frequency of going outdoors and
executive function
Ryota Sakurai a,b,c,*, Yoshinori Fujiwara c, Naoko Sakuma d, Hiroyuki Suzuki c,
Masami Ishihara e, Takahiro Higuchi a, Kuniyasu Imanaka a,*
a

Health Promotion Science, Tokyo Metropolitan University, 1-1 Minami-Osawa, Hachioji-shi, Tokyo 192-0397, Japan
Japan Society for the Promotion of Science, 5-3-1 Kojimachi, Chiyoda-ku, Tokyo 102-0083, Japan
c
Research Team for Social Participation and Community Health, Tokyo Metropolitan Institute of Gerontology, 35-2 Sakae-cho, Itabashi-ku,
Tokyo 173-0015, Japan
d
Research Team for Promoting Independence of the Elderly, Tokyo Metropolitan Institute of Gerontology, 35-2 Sakae-cho, Itabashi-ku, Tokyo 173-0015, Japan
e
Department of Human Sciences, Tokyo Metropolitan University, 1-1 Minami-Osawa, Hachioji-shi, Tokyo 192-0397, Japan
b

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 10 January 2014
Received in revised form 28 July 2014
Accepted 30 July 2014
Available online 8 August 2014

Objective: Self-overestimation of step-over ability among older adults may result in a potential fall risk.
However, the behavioral causal factor(s) of older adults self-overestimation is unclear. We examined
whether older adults overestimation of step-over ability was mediated by an inactive lifestyle and/or
poor executive function.
Methods: A sample of 194 community-dwelling older adults was assigned to either a high (HG, once a
day or more) or low (LG, every 23 days or less) frequency of going outdoors group. Executive function
was determined by the Trail-Making Test (TMT). Both the HG and LG participants performed Step-Over
Tests (SOT) in two ways: self-estimation of step-over ability and the actual step-over task. During the
self-estimation task, participants observed the horizontal bar at a distance of 7 m and predicted the selfestimated maximum height (EH) of successful SOT trials. The motor task was then performed,
determining the actual maximum height (AH) of successful trials.
Results: A total of 36.1% of LG participants failed to successfully perform SOT trials at their EH (i.e.,
overestimation), whereas only 11.3% of HG participants failed. A multiple linear regression analysis
showed that SOT overestimation was associated with an inactive lifestyle (low frequency of going
outdoors) but not with executive function. Analyses of fall experience showed that both executive
function and lifestyle signicantly correlated with SOT overestimation among fallers, whereas only
lifestyle was signicantly correlated among non-fallers.
Conclusion: Our results suggest that an inactive lifestyle is a possible correlate of SOT overestimation
among older adults, while executive function further inuenced overestimation only among fallers.
2014 Elsevier Ireland Ltd. All rights reserved.

Keywords:
Aging
Self-estimation
Stepping-over
Overestimation
Going outdoors
Cognitive function

1. Introduction
Physical ability generally declines as we age (Ahmed, Matsumura, &
Cristian, 2005; Callisaya, Blizzard, Schmidt, McGinley, & Srikanth,
2008; Oberg, Karsznia, & Oberg, 1993). If older adults are unaware of
their age-related physical decline, it is likely that they will overestimate their physical ability (presumably recalling previous conditions

* Corresponding authors at: Tokyo Metropolitan Institute of Gerontology, 35-2

Sakae-cho, Itabashi, Tokyo 173-0015, Japan. Tel.: +81 3 3964 3241x4257;
fax: +81 3 5375 5310.
E-mail addresses: r_sakurai@hotmail.co.jp (R. Sakurai),
imanaka@tmu.ac.jp (K. Imanaka).
http://dx.doi.org/10.1016/j.archger.2014.07.017
0167-4943/ 2014 Elsevier Ireland Ltd. All rights reserved.

of their physical ability when younger). This self-overestimation of

physical ability may be a serious causal factor in movement-related
accidents among older adults. When considering step-over ability,
older adults overestimation of physical ability may lead to tripping on
an obstacle while walking, possibly resulting in a fall.
Regarding older adults self-estimation of step-over ability, we
recently showed that older adults tend to overestimate their stepover ability (Sakurai et al., 2013). Furthermore, the number of
older adult fallers who overestimate step-over ability is almost
double that of non-fallers. This suggests that overestimation of
step-over ability may be a potential risk factor for falling. A similar
nding of overestimation among older adults was evident when
assessing reaching ability (Butler, Lord, & Fitzpatrick, 2011;

578

R. Sakurai et al. / Archives of Gerontology and Geriatrics 59 (2014) 577583

Gabbard, Cacola, & Cordova, 2011; Liu-Ambrose, Ahamed, Graf,

Feldman, & Robinovitch, 2008; Robinovitch & Cronin, 1999).
Although several previous studies have highlighted behavioral
evidence for older adults overestimation of physical ability, the
signicant causal factors affecting older adults overestimation
remain unclear.
Regarding likely causal factors, Liu-Ambrose et al. (2008)
proposed that decits in working memory might affect older
adults self-estimation of physical abilities. The authors showed
that older adults with low working memory capacity (measured by
a cognitive test) tended to overestimate their reaching ability more
than older adults with better working memory performance.
However, that study only included a small sample of fallers, and no
non-fallers. Despite this limitation, it is likely that working
memory capacity, as an element of executive function (which
also comprises inhibitory function and mental exibility: see
Diamond, 2013; Miyake et al., 2000), may be essential in
enhancing older adults self-estimation of physical ability.
Given that poor working memory capacity is related to selfoverestimation of physical ability among older adults, the specic
memory (or awareness) of current physical ability may also be
deteriorated to some extent, which could inuence older adults
overestimations. The specic memory or awareness of physical
ability is probably updated through daily physical activities. For
example, decreased frequency of going outdoors may reduce daily
experiences of re-estimation (and conscious awareness) of agerelated decline in physical ability; this could lead to a lack of
updating specic memory/awareness of physical ability. If this is
the case, an inactive lifestyle may be an inuential factor for both
decits in general working memory (executive function) and
specic memory/awareness of current physical ability (Fujita,
Fujiwara, Chaves, Motohashi, & Shinkai, 2006). Therefore, we
hypothesize that decits in cognitive/executive function, as
mediated by an inactive lifestyle, may lead to older adults
overestimation of physical ability.
Another pertinent question is how closely inactive lifestyles and
cognitive/executive function are related to, or independent from,
each other, and to what extent these respective lifestyle and
cognitive factors inuence older adults overestimation of physical
ability. Therefore, we investigated the relationships between
activity level, executive function, and self-estimation of physical
ability using a step-over task in an attempt to identify inuential
factors of age-related self-overestimation. To this end, healthy
community-dwelling older adults were assigned to two groups with
a HG (go out daily) and LG (every few days or less) groups. These two
groups were tested using our original SOT. The accuracy of selfestimation was evaluated in terms of the difference between EH and
AH during the SOT. Executive function was examined using an
established neuropsychological assessment, the TMT (Demakis,
2004; Sanchez-Cubillo et al., 2009). To examine likely conventional
confounds affecting the frequency of going outdoors, the TMIG Index
of Competence (TMIG-IC) (Koyano, Shibata, Nakazato, Haga, &
Suyama, 1991) and Mini-Mental State Examination (MMSE)
(Folstein, Folstein, & McHugh, 1975) were administered. The
TMIG-IC is a questionnaire consisting of three subscales used to
measure older adults functional capacity: instrumental activities of
daily living (IADL: e.g., using public transportation with no aid),
intellectual activity (e.g., reading newspapers), and sociability (e.g.,
visiting friends). The MMSE, widely used as a tool for assessing
overall cognitive functioning, consists of seven cognitive subdomains/functions: orientation to time, orientation to place, registration of three words, attention and calculation, recall of three words,
language, and visual construction (Tombaugh & McIntyre, 1992).
Based on these measures, we then examined the relationships
between self-estimation of step-over ability, frequency of going
outdoors, and executive function.

2. Methods
2.1. Participants
One hundred ninety-four older adults (mean age = 73.6,
SD = 5.3 years; 78.4% female) participated in the study. Participants were screened for the following exclusion criteria: (i) major
disease or injury (e.g., stroke, heart disease, or injury-related fall)
within 3 months prior to the study; (ii) problems with motor
functions (e.g., use of a walking aid such as a cane); (iii) severe
mental disorders or cognitive impairment (MMSE score <24); and
(iv) vision insufcient to identify the experimental device
(corrected binocular visual acuity <1.0 identied by Landolt ring
chart examination). Written informed consent was obtained from
all participants before examination. The study was conducted in
accordance with the Declaration of Helsinki (1983). The Tokyo
Metropolitan Institute of Gerontology approved the research
protocol.
2.2. Health characteristics and frequency of going outdoors
A physician or physical therapist collected information
regarding participants health conditions and lifestyles through
an interview. Information assessed was age, education, anamnesis
(hypertension, cerebrovascular disorder, cardiac disease, diabetes
mellitus, and arthropathy), and daily practice of going outdoors
(including shopping, taking a walk, going to the hospital, and
gardening; Fujita et al., 2006). The daily practice of going outdoors
was scored as 1 point for once a day or more, 2 points for once every
23 days, and 3 for once a week or less according to Jacobs et al.
(2008); participants were then assigned either a HG (score of 1) or
LG (scores of 2 and 3) group. For the TMIG-IC, rating scores for the
three subscales (IADL, intellectual activity, and social roles) ranged
from 0 to 13, with higher scores indicating greater functional
capacity. Fall experiences over the previous year were also
assessed and rated as falls including an unintentional drop or fall
to the ground, except for bicycle accidents, accidental contact with
environmental structures (e.g., furniture and walls), and sudden
cardiovascular or central nervous system events (Gibson, Andres,
Isaacs, Radebaugh, & Worm-Petersen, 1987).
2.3. Overall cognitive and executive function
We used the MMSE (Folstein et al., 1975) and TMT (SanchezCubillo et al., 2009) to evaluate overall cognitive and executive
function (Diamond, 2013; Miyake et al., 2000), respectively. The
MMSE has a maximum score of 30 points, with higher scores
indicating higher overall cognitive function. The TMT consists of
TMT-A and B (Demakis, 2004; Sanchez-Cubillo et al., 2009). In the
TMT-A, to test simple visual search and motor speed skills
(Gaudino, Geisler, & Squires, 1995; Sanchez-Cubillo et al., 2009),
participants are asked to draw a line with a pencil to connect 25
printed numerals from 1 to 25 in ascending order (Gaudino et al.,
1995; Sanchez-Cubillo et al., 2009). In the TMT-B, to test higherorder cognitive skills such as working memory and mental
exibility (Kortte, Horner, & Windham, 2002; Sanchez-Cubillo
et al., 2009), participants perform a visual-motor task similar to
the TMT-A, except this includes connecting 13 numerical numbers
and 12 Japanese hiragana characters (Hirota et al., 2010) (the
original TMT uses 12 letters) while alternating numbers and letters
in ascending order (i.e., a number followed a hiragana character
and vice versa). Both the TMT-A and B were performed after
ensuring that the participants were fully familiarized with the task
procedures. The elapsed time taken to complete the TMT-A and B
was recorded; the difference in time between the two (D TMT) was
calculated by subtracting TMT-A time from TMT-B time for a

R. Sakurai et al. / Archives of Gerontology and Geriatrics 59 (2014) 577583

measure of executive function (working memory, inhibitory

function and mental exibility; Sanchez-Cubillo et al., 2009).
2.4. SOT
Step-over performance and accuracy of self-estimation was
measured using the SOT (Sakurai et al., 2013). Participants rst
performed the self-estimation test, then the actual step-over task.
For the self-estimation test, a horizontal black wooden bar
(25 mm  25 mm  900 mm) was placed 2 m in front of a white
background wall. Participants were asked to observe the horizontal
bar from a distance of 7 m. The experimenter manually adjusted
the height of the bar from 10 to 80 cm (ascending) or 80 to 10 cm
(descending) at a slow speed. While the bar was moving,
participants were instructed to imagine stepping over the bar
while facing straight ahead, with no restrictions to posture except
in reference to jumping. Participants were asked to respond
verbally when the bar reached the maximum height they
believed they could step over the bar. Participants were allowed
to amend their EH after their verbal response; the experimenter
then manually adjusted the bar height accordingly. No time
restriction was imposed on this estimation. Trials for the selfestimation test were performed 4 times, with 2 ascending and 2
descending trial manipulations for bar height. The mean EH for
the 4 trials was calculated per participant.
Following the self-estimation test, the actual SOT task was
performed. Participants were asked to approach and step over the
bar set at the individual EH. If the participant failed to step over the
bar at the EH (e.g., touching/kicking the bar with the foot/lower
limb), the bar was lowered by 3 cm. If the participant was
successful at his/her original EH, the bar was raised by 3 cm. Each
participant was then asked to step over the bar again at the new
height. This was repeated until the participant either succeeded or
failed with the step-over action; the nal height at which the
participant successfully stepped over the bar across 2 consecutive
trials was recorded as his/her individual AH.
The ability to step over an obstacle generally correlates with
lower limb length; thus, the EH and AH were divided by the length
(distance from the greater trochanter to the ground through the
lateral malleolus) of the lower limb. These ratios were used as
individual representative EH and AH values for subsequent
analyses. The difference (D height) between the EH and AH was

579

then calculated to determine the accuracy, or bias error (i.e.,

underestimation or overestimation), of the self-estimation of stepover ability. The D height was then standardized with the AH in
terms of the following formula: [(EH AH)/AH  100]. The
percentage of participants who failed to step over the bar at the
EH was calculated for both groups.
2.5. Data analysis
Differences in all measurement variables between the HG and
LG groups were assessed with either chi-square tests (for nominal
variables) or multivariate analyses of variance (MANOVA, for
interval variables), except for the EH and AH. A repeated measures
analysis of variance (ANOVA) was performed for both SOT
performance (EH and AH) and frequency group (HG and LG)
factors. The relationships between EH and AH, between the
frequency of going outdoors and cognitive function, and between
the D height and all other measurement variables were examined
in terms of either Pearson product-moment or Spearman rankorder correlation coefcients for the HG and LG groups separately.
A multiple linear regression analysis was then performed to
examine respective associations among measurement variables
with D height on the basis of all the resultant signicant
correlation coefcients, biological covariates (age and gender) of
D height, and functional capacity (TMIG-IC). As Liu-Ambrose et al.
(2008) recently showed that poor cognitive function was related to
self-overestimation of physical abilities among older fallers (not
non-fallers), correlation analyses and a multiple linear regression
analysis were further performed on the measurement variables in
the same manner separately for fallers and non-fallers. All
statistical analyses were performed using IBM SPSS Statistics
Version 20.0 with a signicance level set at p < 0.05.
3. Results
In terms of basic participant characteristics (Table 1), 36 (18.6%)
and 158 (81.4%) participants were classied into the LG and HG
groups, respectively. The male-female ratio, age, and education did
not signicantly differ between the LG and HG groups.
In terms of health characteristics, chi-square tests revealed that
the prevalence rates of cerebrovascular disorder, cardiac disease,
falls within the past year, and diabetes mellitus (with marginal

Table 1
Participant characteristics.
Variables

Full sample (N = 194)

HG (n = 158)

LG (n = 36)

p-Value

Female, n (%)
Age, mean (SD)
Education, mean years (SD)
TMIG-IC, mean (SD)
MMSE, mean (SD)
TMT-A
TMT-B
D TMT
Hypertension, n (%)
Cerebrovascular disorder, n (%)
Cardiac disease, n (%)
Diabetes mellitus, n (%)
Arthropathy, n (%)
Fall experience, n (%)
EH: estimated height, % of LL (SD)
AH: actual height, % of LL (SD)
D height (SD)
D height > 0, n (%)b

152 (78.4)
73.6 (5.3)
13.0 (2.3)
12.2 (1.1)
28.2 (1.7)
42.0 (15.6)
114.7 (49.2)
72.7 (41.9)
70 (36.1)
10 (5.2)
11 (5.7)
13 (6.7)
30 (15.5)
45 (23.2)
66.5 (10.1)
74.1 (8.0)
9.8 (12.0)
31 (16.0)

125 (79.1)
73.3 (6.7)
12.9 (2.3)
12.3 (1.0)
28.2 (1.7)
41.2 (15.4)
114.0 (49.0)
72.7 (41.7)
59 (37.3)
5 (3.1)
6 (3.7)
8 (5.0)
24 (15.1)
31 (19.6)
66.2 (10.2)
74.5 (8.0)
10.9 (12.0)
18 (11.3)

27 (75.0)
74.8 (6.7)
13.4 (2.5)
11.8 (1.4)
28.5 (1.7)
45.1 (16.0)
117.7 (52.7)
72.5 (43.6)
11 (30.6)
5 (13.9)
5 (13.9)
5 (13.9)
6 (16.7)
14 (38.9)
67.9 (9.5)
72.0 (7.6)
4.7 (15.6)
13 (36.1)

0.589
0.125
0.243
0.023
0.317
0.179
0.685
0.983
0.444
0.009
0.018
0.056
0.825
0.013
a
a

p < 0.001
p < 0.001

a
A repeated measures ANOVA was performed. The main effect of SOT performance (F1, 190 = 6.2, p < 0.05) was signicant. The main effect of HG/LG group (F1, 190 = 0.4,
p > 0.1) was not signicant. The interaction between the two factors was signicant (F1, 190 = 4.9, p < 0.05).
b
Percentage of participants who failed to step over the bar at the estimated height (i.e., overestimation). HG and LG refer to high and low frequency of going outdoors
groups, respectively. D TMT indicates the time difference between TMT-B and TMT-A (B A); the D height indicates a difference between estimated height and actual height
[(EH AH)/AH  100].

R. Sakurai et al. / Archives of Gerontology and Geriatrics 59 (2014) 577583

580

100

HG: r = 0.554, p < 0.01

Estimated SOT (% of LL)

Estimated SOT (% of LL)

100

90
80
70
60
50

LG: r = 0.131, n.s

90
80
70
60
50
40

40
40

60
80
Actual SOT (% of LL)

100

40

60

80

100

Actual SOT (% of LL)

Fig. 1. Scatter diagrams of EH and AH for the HG and LG groups. The left scatter diagram (squares) is for the HG group; the diagram on the right (triangles) is for the LG group.
For both of the scatter diagrams, lled markers indicate non-fallers; opened markers indicate fallers. Data points above the diagonal line shows participants who had EH
values higher than their AH values (i.e., overestimation), indicating that the above-diagonal participants failed to step over the bar at the estimated height (EH). Dotted lines
represent the 95% condence interval of the regression line. LL: leg length.

signicance, p = 0.056) were signicantly greater in the LG group

than in the HG group. No other variables showed signicant
differences between the LG and HG groups.
For functional capacity and cognitive function, TMIG-IC scores
were signicantly lower for the LG group than that for the HG
group, whereas the MMSE and TMT scores showed no signicant
differences between the LG and HG groups.
For SOT performance, a 2-way ANOVA revealed a signicant
main effect of estimated SOT (EH) and actual SOT (AH; F1, 190 = 6.2,
p < 0.05), but not for the LG and HG frequency groups (F1, 190 = 0.4,
p > 0.1). There was a signicant interaction (F1, 190 = 4.9, p < 0.05)
between the two factors. The AH was signicantly higher for the
HG than for the LG (p < 0.05) group, whereas EH did not
signicantly differ between the two groups. The mean D height
(bias of self-estimation) indicated that the HG group showed a
signicantly larger underestimation than did the LG group
(p < 0.01). Moreover, 36.1% of the LG participants had a D height
higher than zero (i.e., overestimation, failing to step over the bar at
EH), whereas only 11.3% of the HG group had a value greater than
zero; this difference was signicant (p < 0.01).
Fig. 1 shows the scatter plots depicting the relationships
between EH and AH in the HG and LG older adults. These plots
suggest that EH and AH were strongly correlated in the HG group
(r = 0.554, p < 0.01), but were not signicantly correlated in the LG
group (r = 0.131, p > 0.1); the difference between the two
correlation coefcients was signicant (p < 0.01). This indicated
that EH largely reected AH in the HG group, but this was not the
case for the LG group. This suggests that low frequency of going
outdoors is a possible correlate of SOT overestimation.1
A correlation analysis performed to examine the relationship
between the frequency of going outdoors and cognitive function
showed that none of the cognitive tests (MMSE, r = 0.046,
p = 0.538; TMT-A, r = 0.138, p = 0.058; TMT-B, r = 0.108,
p = 0.188; and D TMT, r = 0.102, p = 0.188) were signicantly
correlated with the frequency of going outdoors. Table 2 shows
the results of the correlation analyses for the D height and
selected variables. Age (r = 0.143, p < 0.05), frequency of going
1
We also examined the relationship between EH and AH separately with each of
the following groups; (1) non-fallers with HG, (2) fallers with HG, (3) non-faller
with LG and (4) faller with LG. A correlation analysis showed that the correlation
coefcients signicantly differ neither for non-fallers with HG (r = 0.566) and fallers
with HG (r = 0.519) nor for non-fallers with LG (r = 0.119) and fallers with LG
(r = 0.111), with a signicant difference in correlation coefcients appearing
between HG and LG, irrespective of fallers and non-fallers. These results suggest
that the low frequency of going outdoors may possibly mediate SOT overestimation
among LG older adults.

outdoors (r = 0.223, p < 0.01), D TMT (r = 0.141, p < 0.05), and

TMIG-IC (r = -0.126, p = 0.079) were all signicantly (or marginally
signicantly) correlated with D height, although these correlations
were quite low.
Based on these correlation analyses, a subsequent multiple
linear regression analysis was performed with the explanatory
variables of gender, age, D TMT, frequency of going outdoors, and
TMIG-IC used to estimate D height. Results (Table 3) revealed that
although the coefcient of determination was somewhat low
(adjusted R2 = 0.215), only the frequency of going outdoors
positively associated D height, but no other variables (age, sex,
TMIG-IC, nor D TMT) signicantly associated with D height.
In further analyses, participants were assigned to two groups,
fallers and non-fallers, based on fall experiences over the previous
year. The relationships between self-overestimation and other
measurement variables were then examined for fallers and nonfallers. A chi-square test showed that the number (proportion) of
participants who overestimated their SOT ability was signicantly
larger (p < 0.01) for the fallers (26.7%) than for the non-fallers
(12.8%), whereas age and the number (proportion) of females did not
differ between the 2 groups (fallers mean age = 70.3 SD = 5.6 years;
82.2% female, non-fallers mean age = 73.7 SD = 5.2 years; 77.1%
female). Correlation analyses for the D height and measurement
variables showed that MMSE (r = 0.356, p < 0.05), TMT-B
(r = 0.323, p < 0.05), and DTMT (r = 0.363, p < 0.05) scores were
all signicantly correlated with D height for fallers but not nonfallers, and the frequency of going outdoors was signicantly
correlated with D height for both fallers (r = 0.282, p < 0.05) and
non-fallers (r = 0.190, p < 0.05). No other variables were signicantly correlated with D height. Table 4 shows the results of
multiple linear regression analyses performed to examine the
Table 2
Results of correlation analysis between age, MMSE, TMT, frequency of going
outdoors, TMIG-IC, and D height.
Variables
Age
MMSE
TMT-A
TMT-B
D TMT
Frequency of going outdoorsa
TMIG-IC

D height
0.143
0.003
0.023
0.113
0.141
0.223
0.126

p-Value
0.046
0.969
0.746
0.117
0.049
0.009
0.079

a
A value of 1 was given to the high frequency of going outdoors group (older
adults who went out once a day or more), and a value of 2 was for the low frequency
of going outdoors group (older adults who went out once every 23 days or less).
D TMT indicates the time difference between TMT-B and TMT-A (B A).

R. Sakurai et al. / Archives of Gerontology and Geriatrics 59 (2014) 577583

Table 3
Results of the multiple linear regression analysis for D height during the SOT.
Variables
a

Gender
Age
D TMT
Frequency of going
outdoorsb
TMIG-IC

Standardized b (95% CI)

p-Value

0.05
0.08
0.11
0.17

( 6.23, 2.93)
( 0.18, 0.55)
( 0.01, 0.08)
(1.20, 10.17)

0.338
0.340
0.148
0.019

0.07

( 0.92, 0.36)

0.911

A value of 1 was for Males, and 2 was for Females.

A value of 1 was given to the high frequency of going outdoors group (older
adults who went out once a day or more), and 2 was given to the low frequency of
going outdoors group (older adults who went out once every 23 days or less).
D TMT indicates the time difference between TMT-B and TMT-A (B A).
b

explanatory variables for estimating D height. Two models were

calculated for fallers, with either TMT-B or DTMT being separately
involved (the p-value was corrected to 0.025, 0.05/2, as a Bonferroni
correction to account for Type 1 error). This was done because a high
correlation coefcient appeared between TMT-B and DTMT
(r = 0.780, p < 0.01), indicating an issue of multicollinearity for
the multiple regression analysis. Results showed that both D TMT
and frequency of going outdoors were positively associated with D
height among fallers, whereas only frequency of going outdoors was
positively associated with D height among non-fallers.
4. Discussion
We observed that LG older adults were less accurate in their
estimates of SOT ability, with a larger proportion of overestimation
among the LG older adults (36.1%) than among the HG older adults
(11.3%). The higher rate of overestimation among the LG older
adults may be mediated by their poorer SOT abilities (as reected
in lower AH values) when compared to the HG group, despite
almost identical self-estimated height (EH) for both LG and HG
older adults (Table 1). This was further highlighted by the
correlation analyses (Fig. 1). These analyses showed a signicantly
high AH-EH correlation for the HG older adults, whereas a nonsignicant low correlation was observed among the LG older
adults. This indicated that the LG older adults self-estimation did
not reect their low actual SOT ability, resulting in overestimation.
In contrast, the HG older adults self-estimation more clearly
reected their actual SOT ability, resulting in less overestimation or
even underestimation. Moreover, the multiple linear regression
analysis (Table 3) showed that only frequency of going outdoors

581

was positively associated with D height (indicating over- and

under-estimation), with no signicant association with cognitive
function. However, comparisons between fallers and non-fallers
showed that D TMT was positively associated with D height for the
fallers but not the non-fallers (Table 4). These ndings suggest that
a lower frequency of going outdoors among the LG group may be
an important behavioral correlate of overestimations of step-over
ability, whereas fallers overestimation may have been mediated
by poor executive function in addition to a lower frequency of
going outdoors.
Previous studies have indicated that going outdoors is essential
for older adults to maintain functional capacity such as IADL
(Ganguli, Fox, Gilby, & Belle, 1996; Gill, Allore, & Guo, 2004; Jacobs
et al., 2008; Kono, Kai, Sakato, & Rubenstein, 2004; Makizako et al.,
2013; Shimada et al., 2010; Simonsick, Guralnik, Volpato, Balfour,
& Fried, 2005). Fujita et al. (2006) showed that older adults who
went out every 23 days or less had signicantly lower physical
function than those who went out once a day or more. This
suggests that the lower amount of physical activity resulting from
a low frequency of going outdoors is a key contributor to
deteriorated physical function among older adults. Indeed,
evidence that more frequent activity/exercise inuences higher
physical function/skills has been suggested by several studies in a
variety of disciplines, including exercise intervention, motor
learning, and rehabilitation (Hasegawa et al., 2005; Jabusch,
Alpers, Kopiez, Vauth, & Altenmuller, 2009; Nakamura, Tanaka,
Yabushita, Sakai, & Shigematsu, 2007). The signicant difference in
AH between the HG and LG groups observed in the present study
may have been mediated by frequency-dependent exercise effects,
such that lower frequency in physical activity (i.e., going outdoors)
causes lower physical abilities.
A comparison of the correlation coefcients between EH and AH
for both HG and LG groups showed a signicantly (p < 0.01) larger
EH-AH correlation coefcient for HG (r = 0.554) than that for LG
(r = 0.131). The larger EH-AH correlation coefcient for HG
indicated that the HG older adults consistently and correctly
estimated their SOT ability depending on their current SOT ability
(AH), whereas the LG older adults estimated their SOT ability less
relying on their current SOT ability. It is likely that the HG group
(practically going outdoors everyday) may have experienced
physical activities well, thus having many opportunities for reexamining their current physical ability, whereas the LG group
(going outdoors every two days or less) may not have fully
experienced physical activities, thus limiting re-examination of
their current physical ability (which was more impaired than that

Table 4
Results of the multiple linear regression analysis for D height among fallers and non-fallers.
Fallers

Non-fallers

Model 1
Variables
Gendera
Age
MMSE
TMT-B
Frequency of going
outdoorsb
TMIG-IC
a

Model 2
Standardized
b (95% CI)
0.19
0.04
0.30
0.25
0.30

( 16.7, 3.17)
( 0.86, 0.64)
( 6.0, 0.17)
( 0.02, 0.23)
(0.62, 17.2)

0.07 ( 4.54, 2.98)

Adjusted R2 = 0.244

p-Value
0.173
0.766
0.051
0.097
0.020*
0.675

Variables

D-TMT

Standardized
b (95% CI)
Standardized
b (95% CI)
0.19
0.07
0.29
0.35
0.35

( 16.4, 2.71)
( 0.87, 0.56)
( 5.7, 0.18)
(0.03, 0.29)
(2.13, 18.3)

0.10 ( 4.75, 2.43)

Adjusted R2 = 0.275

p-Value

p-Value
0.155
0.653
0.051
0.019*
0.015*
0.526

0.04 ( 6.0, 3.60)

0.13 ( 0.08, 0.69)

0.625
0.120

0.17 (0.35, 11.7)

0.022

0.08 ( 3.12, 1.24)

Adjusted R2 = 0.180

0.396

A value of 1 was for male participants; value of 2 for female participants.

A value of 1 was for the high frequency of going outdoors group (older adults who went out once a day or more), and a value of 2 was for the low frequency of going
outdoors group (older adults who went out once every 23 days or less). Multiple linear regression analyses were performed separately with TMT-B (Model 1) and DTMT
(Model 2) given the high correlation coefcient between TMT-B and DTMT.
D TMT was calculated by subtracting TMT-A from TMT-B.
*
Bonferroni correction was applied to avoid Type 1 error, resulting in an alpha of 0.025 (0.05/2).
b

582

R. Sakurai et al. / Archives of Gerontology and Geriatrics 59 (2014) 577583

of HG older adults). This means that the LG group may presumably

have failed to be aware of their current age-related decline in
physical ability. Our results on cognitive function also showed no
signicant difference between the HG and LG groups, suggesting
that LG older adults probably have a good memory of their
previous physical ability in past years. It is therefore likely that the
LG group self-estimated their physical ability according to their
memory of previous conditions of physical ability rather than the
updated knowledge of their current decline in physical ability. This
may have resulted in a relatively high self-estimated SOT height
(EH) as good as the HG group, although the current physical ability
was much more declined for LG than that for HG (as shown in AH:
Table 1). The similar self-estimations (EH) and signicantly lower
AH among the LG older adults compared with those of the HG older
adults caused either overestimation or less underestimation of SOT
ability in LG older adults. In summary, the daily low frequency of
going outdoors in LG older adults may have caused a lack of
updating their awareness of their current physical ability,
therefore relying on their memory of previous conditions of
physical ability in past years while self-estimating their SOT
ability. This may have led to the LGs EH similar to HG despite the
current, more declined physical ability in LG older adults, thus
resulting in overestimation or less underestimation of SOT ability
in LG older adults.
Regarding executive function, it was evident from the separate
analyses with fallers and non-fallers that D TMT was signicantly
associated with an overestimation of SOT ability among fallers but
not non-fallers. This is consistent with results from Liu-Ambrose
et al. (2008), indicating a signicant relationship between an
overestimation of reaching ability and poor cognitive function
(lower working memory) among fallers. We previously showed
that fallers tend to overestimate their SOT ability more than nonfallers (Sakurai et al., 2013). This nding was replicated in the
present study as well. Furthermore, it was evident from the
present study that only fallers showed a signicant relationship
between the overestimation of SOT ability and poor cognitive
function (D TMT). Based on both Liu-Ambrose et al.s study and
the present ndings, poor executive function likely mediates the
self-overestimation of SOT ability among fallers (but non-fallers).
Such an overestimation in older adults could be a potential
precursor of future falling (e.g., recurrent falls), although it is far
from clear whether this is a potential risk factor or whether
additional variable(s) are correlated with both falling and
overestimation of step-over ability. Furthermore, the underlying
mechanisms for how cognitive function mediates self-overestimation of SOT ability (and potential resultant falls) are not
evident. These issues should be further examined in future
studies.
We initially postulated that an inactive lifestyle (e.g., spending
less time outdoors) might negatively inuence both (general)
executive function and (specic) memory/awareness/updating of
current physical abilities; this could result in self-overestimation
of SOT ability. However, our results showed that executive function
(D TMT) was not correlated with the frequency of going outdoors
for either non-fallers (r = 0.110) or fallers (r = 0.105; data not
shown). The likely reason for the lack of relationship between
frequency of going outdoors and cognitive function is because both
faller and non-faller groups of participants were physically and
mentally healthy with high functioning as evidenced by both the
exclusion criteria in this study and the results of TMIG-IC (average
score is 12.3 in non-fallers and 11.9 in fallers). Our results also
showed that among non-fallers only the variable of frequency of
going outdoors was signicantly correlated with self-overestimation of SOT ability, whereas among fallers, both executive function
and frequency of going outdoors signicantly correlated with selfoverestimation of SOT ability. This suggests that for older adult

fallers, general executive function limits the updating of memory/

awareness of physical abilities, and that an inactive lifestyle also
limits this updating ability. For non-fallers, self-overestimation
might only be mediated by an inactive lifestyle. Further studies
will be needed to elucidate likely reasons for these discrepancies
for fallers and non-fallers.
As described above, the correlation analyses showed that the
LG older adults self-estimation did not reect their low actual
SOT ability, whereas the HG older adults self-estimation more
clearly reected their actual SOT ability. However, the results
also showed that LG older adults were more accurate (or
approaching to overestimation) in rating their EH than HG older
adults were, with a larger underestimation for the HG older
adults than that for LG older adults (D height: 10.9 and 4.7,
respectively). A large underestimation of physical ability has
generally appeared in young adults (Robinovitch & Cronin, 1999;
Sakurai et al., 2013). Our previous study has also shown that
young adults tend to more underestimate their step-over ability
than older adults (Sakurai et al., 2013). Despite the typical
underestimation of their step-over ability in young adults, EH
and AH were strongly correlated in young adults but this was not
the case for the older adults, with the degree of correlation
coefcients signicantly different for the young and older adult
groups. According to our previous and present ndings, it is likely
that both young adults and HG older adults may unconsciously
ensure the safety of their motor actions and thus largely
underestimate their physical ability.
Finally, we should address the issue of potential confounding
factors related to inaccurate self-estimation among older adults.
Some psychological factors, such as a fear of falling, might
inuence self-overestimation of step-over ability. However, this
was not likely to be the case in the present study because a fear of
falling should be strongly associated with fall experience. An
increased fear of falling might bias participants self-estimation
toward an underestimation rather than an overestimation of stepover ability. Nevertheless, our previous (Sakurai et al., 2013) and
present ndings showed that the percentage of fallers who
overestimated their SOT ability was almost double than that of
the non-fallers. Therefore, a fear of falling likely does not explain
our results.
Another potential limitation in this study is the use of a crosssectional study. The cross-sectional, rather than longitudinal,
study might preempt proper inferences in explaining the likely
causal relationship between an overestimation of SOT ability and
a low frequency of going outdoors, although our multiple
regression analysis was controlled for several confounding
factors (age, gender, functional capacity, and cognitive function)
affecting the frequency of going outdoors (Cohen-Manseld,
Shmotkin, & Hazan, 2012; Fujita et al., 2006). The present study
also used the two tests, MMSE and TMT, as cognitive assessments. This was because we rstly hypothesized that working
memory was a possible inuential factor for overestimation of
physical ability, according to the previous study (Liu-Ambrose et
al., 2008). In this regard, if we had measured the other functional
domain of executive function (e.g., updating, inhibition function),
it may be found interesting results relating inuential factors
affecting age-related self-overestimation. Therefore, further
longitudinal research will be needed to examine the causal
relationship between lower frequency of going outdoors and agerelated self-overestimation of physical abilities, including
examining widely cognitive function using various assessments.
5. Conclusion
The present study revealed that only the feature of low
frequency of going outdoors was signicantly associated with the

R. Sakurai et al. / Archives of Gerontology and Geriatrics 59 (2014) 577583

overestimation of step-over ability among older adults. This

suggests that the inactive lifestyle is a signicant correlate of agerelated self-overestimation of physical function. For fallers, in
addition to inactive lifestyle, the decits in executive function may
also be a possible correlate of overestimation of step-over ability.
This might result in increasing fall risks, although this should be
further examined in the future studies. Direct causal relationships
between inactive lifestyle (low frequency of going outdoors),
decits in executive function, and overestimation of physical
ability should be examined in relation to the risk of activity-related
accidents, such as falls. This may probably be more effectively
examined with longitudinal than cross-sectional research.

Conict of interest statement

The authors declare that they have no conicts of interest.
Acknowledgments
This study was supported by a Grant-in-Aid for JSPS (Japan
Society for the Promotion of Science) fellows (23-5365). The
authors acknowledge the continued efforts of the members of Unit
412 of the Tokyo Metropolitan Institute of Gerontology. We also
thank Mr. Susumu Ogawa of Chuo University for his help in data
collection.

References
Ahmed, M. S., Matsumura, B., & Cristian, A. (2005). Age-related changes in muscles and
joints. Physical Medicine and Rehabilitation Clinics of North America, 16(1), 1939.
Butler, A. A., Lord, S. R., & Fitzpatrick, R. C. (2011). Reach distance but not judgment
error is associated with falls in older people. Journals of Gerontology. Series A,
Biological Sciences and Medical Sciences, 66(8), 896903.
Callisaya, M. L., Blizzard, L., Schmidt, M. D., McGinley, J. L., & Srikanth, V. K. (2008). Sex
modies the relationship between age and gait: A population-based study of older
adults. Journals of Gerontology. Series A, Biological Sciences and Medical Sciences,
63(2), 165170.
Cohen-Manseld, J., Shmotkin, D., & Hazan, H. (2012). Homebound older persons:
Prevalence, characteristics, and longitudinal predictors. Archives of Gerontology and
Geriatrics, 54(1), 5560.
Demakis, G. J. (2004). Frontal lobe damage and tests of executive processing: A metaanalysis of the category test, stroop test, and trail-making test. Journal of Clinical
and Experimental Neuropsychology, 26(3), 441450.
Diamond, A. (2013). Executive functions. Annual Review of Psychology, 64, 135168.
Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Mini-mental state. A practical
method for grading the cognitive state of patients for the clinician. Journal of
Psychiatric Research, 12(3), 189198.
Fujita, K., Fujiwara, Y., Chaves, P. H., Motohashi, Y., & Shinkai, S. (2006). Frequency of
going outdoors as a good predictors for incident disability of physical function as
well as disability recovery in community-dwelling older adults in rural Japan.
Journal of Epidemiology, 16(6), 261270.
Gabbard, C., Cacola, P., & Cordova, A. (2011). Is there an advanced aging effect on the
ability to mentally represent action? Archives of Gerontology and Geriatrics, 53(2),
206209.
Ganguli, M., Fox, A., Gilby, J., & Belle, S. (1996). Characteristics of rural homebound older
adults: A community-based study. Journal of the American Geriatrics Society, 44(4),
363370.

583

Gaudino, E. A., Geisler, M. W., & Squires, N. K. (1995). Construct validity in the Trail
Making Test: What makes Part B harder? Journal of Clinical and Experimental
Neuropsychology, 17(4), 529535.
Gibson, M., Andres, R., Isaacs, B., Radebaugh, T., & Worm-Petersen, J. (1987). The
prevention of falls in later life. A report of the Kellogg International Work Group on
the Prevention of Falls by the Elderly. Danish Medical Bulletin, 34(Suppl. 4), 124.
Gill, T. M., Allore, H., & Guo, Z. (2004). The deleterious effects of bed rest among
community-living older persons. Journals of Gerontology. Series A, Biological Sciences
and Medical Sciences, 59(7), 755761.
Hasegawa, Y., Yoneda, Y., Okuda, S., Hamada, R., Toyota, A., Gotoh, J., et al. (2005). The
effect of weekends and holidays on stroke outcome in acute stroke units. Cerebrovascular Diseases, 20(5), 325331.
Hirota, C., Watanabe, M., Sun, W., Tanimoto, Y., Kono, R., Takasaki, K., et al. (2010).
Association between the Trail Making Test and physical performance in elderly
Japanese. Geriatrics & Gerontology International, 10(1), 4047.
Jabusch, H. C., Alpers, H., Kopiez, R., Vauth, H., & Altenmuller, E. (2009). The inuence of
practice on the development of motor skills in pianists: A longitudinal study in a
selected motor task. Human Movement Science, 28(1), 7484.
Jacobs, J. M., Cohen, A., Hammerman-Rozenberg, R., Azoulay, D., Maaravi, Y., & Stessman, J. (2008). Going outdoors daily predicts long-term functional and health
benets among ambulatory older people. Journal of Aging and Health, 20(3),
259272.
Kono, A., Kai, I., Sakato, C., & Rubenstein, L. Z. (2004). Frequency of going outdoors: A
predictor of functional and psychosocial change among ambulatory frail elders
living at home. Journals of Gerontology. Series A, Biological Sciences and Medical
Sciences, 59(3), 275280.
Kortte, K. B., Horner, M. D., & Windham, W. K. (2002). The trail making test, Part B: Cognitive
exibility or ability to maintain set? Applied Neuropsychology, 9(2), 106109.
Koyano, W., Shibata, H., Nakazato, K., Haga, H., & Suyama, Y. (1991). Measurement of
competence: Reliability and validity of the TMIG Index of Competence. Archives of
Gerontology and Geriatrics, 13(2), 103116.
Liu-Ambrose, T., Ahamed, Y., Graf, P., Feldman, F., & Robinovitch, S. N. (2008). Older
fallers with poor working memory overestimate their postural limits. Archives of
Physical Medicine and Rehabilitation, 89(7), 13351340.
Makizako, H., Doi, T., Shimada, H., Park, H., Uemura, K., Yoshida, D., et al. (2013).
Relationship between going outdoors daily and activation of the prefrontal cortex
during verbal uency tasks (VFTs) among older adults: A near-infrared spectroscopy study. Archives of Gerontology and Geriatrics, 56(1), 118123.
Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D.
(2000). The unity and diversity of executive functions and their contributions to
complex Frontal Lobe tasks: A latent variable analysis. Cognitive Psychology, 41(1),
49100.
Nakamura, Y., Tanaka, K., Yabushita, N., Sakai, T., & Shigematsu, R. (2007). Effects of
exercise frequency on functional tness in older adult women. Archives of Gerontology and Geriatrics, 44(2), 163173.
Oberg, T., Karsznia, A., & Oberg, K. (1993). Basic gait parameters: Reference data for
normal subjects, 1079 years of age. Journal of Rehabilitation Research and Development, 30(2), 210223.
Robinovitch, S. N., & Cronin, T. (1999). Perception of postural limits in elderly nursing
home and day care participants. Journals of Gerontology. Series A, Biological Sciences
and Medical Sciences, 54(3), B124B130. discussion B131.
Sakurai, R., Fujiwara, Y., Ishihara, M., Higuchi, T., Uchida, H., & Imanaka, K. (2013). Agerelated self-overestimation of step-over ability in healthy older adults and its
relationship to fall risk. BMC Geriatrics, 13, 44.
Sanchez-Cubillo, I., Perianez, J. A., Adrover-Roig, D., Rodriguez-Sanchez, J. M., RiosLago, M., Tirapu, J., et al. (2009). Construct validity of the Trail Making Test: Role
of task-switching, working memory, inhibition/interference control, and visuomotor abilities. Journal of the International Neuropsychological Society, 15(3),
438450.
Shimada, H., Ishizaki, T., Kato, M., Morimoto, A., Tamate, A., Uchiyama, Y., et al. (2010).
How often and how far do frail elderly people need to go outdoors to maintain
functional capacity? Archives of Gerontology and Geriatrics, 50(2), 140146.
Simonsick, E. M., Guralnik, J. M., Volpato, S., Balfour, J., & Fried, L. P. (2005). Just get out
the door! Importance of walking outside the home for maintaining mobility:
Findings from the womens health and aging study Journal of the American
Geriatrics Society, 53(2), 198203.
Tombaugh, T. N., & McIntyre, N. J. (1992). The mini-mental state examination: A
comprehensive review. Journal of the American Geriatrics Society, 40(9), 922935.