Accident Analysis and Prevention 80 (2015) 16

Contents lists available at ScienceDirect

Accident Analysis and Prevention

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

The development, factor structure and psychometric properties of

driving self-regulation scales for older adults: Has self-regulation
evolved in the last 15 years?
Ides Y. Wong a,b, * , Simon S. Smith b , Karen A. Sullivan b,c
a

School of Population Health, The University of Queensland, Queensland Centre for Mental Health Research, The Park Centre for Mental Health, Australia
Centre for Accident Research and Road Safety Queensland (CARRS-Q) & Institute of Health and Biomedical Innovation (IHBI), Queensland University of
Technology, Australia
c
Clinical Neuropsychology Research Group, School of Psychology and Counselling & Institute of Health and Biomedical Innovation, Queensland University of
Technology, Australia
b

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 6 November 2014
Received in revised form 1 March 2015
Accepted 27 March 2015
Available online xxx

The term driving self-regulation is typically used to describe the practice of drivers who avoid driving in
situations that they regard as unsafe because of perceived physical impairment. Older adults report using
this strategy to improve safety while retaining mobility. Self-regulation is typically assessed using the
driving avoidance items from the driving habits questionnaire (DHQ) and the driver mobility
questionnaire (DMQ-A). However, the psychometric properties of these measures are not well
understood. Using data from 277 older drivers, exploratory factor analysis was used to test the
homogeneity of three driving self-regulation scales: the DHQ, DMQ-A, and an extended DMQ-A. Good
internal consistency for each of the scales was identied (all as  .9). A one factor solution was identied
for two of the measures (DHQ, DMQ-A) and a two factor solution accounting for over 70% of the score
variance was identied for the third measure. The two factors assessed situations that may be avoided
while driving because of the external (e.g., weather-related) or internal (e.g., passenger-related)
driving environments, respectively. The ndings suggest that the interpretation of an overall summated
scale score, or single-item interpretations, may not be appropriate. Instead, driving self-regulation may
be a multifaceted construct comprised of distinct dimensions that have not been identied previously
but can be reliably measured. These data have implications for our understanding of driving selfregulation by older adults and the way in which this behavior is measured.
2015 Elsevier Ltd. All rights reserved.

Keywords:
Driving
Older adults
Self-regulation scale
Factor analysis
Self-regulation

1. Introduction
The number of older adults that rely on driving for transport is
projected to rapidly increase over coming decades (Organisation
for Economic Co-operation and Development, 2009). This demographic change has important implications for many communities,
one of which is how to sustain the mobility and activity of older
persons. Self-regulation of driving is one individually-tailored
strategy that has been suggested as a means of ensuring safety
whilst maintaining mobility (Dickerson et al., 2007).

* Corresponding author at: Queensland Centre for Mental Health Research, The
Park Centre for Mental Health, Wacol, Queensland 4076, Australia.
Tel.: +61 7 31387698.
E-mail address: i.wong@qcmhr.uq.edu.au (I.Y. Wong).
http://dx.doi.org/10.1016/j.aap.2015.03.035
0001-4575/ 2015 Elsevier Ltd. All rights reserved.

Self-regulation involves voluntarily restricting ones driving to

avoid situations that one considers unsafe. These situations can
include driving at night, in the peak hours, or in poor weather
conditions (Baldock et al., 2006; Molnar and Eby, 2008; Sullivan
et al., 2011). This restriction of driving is assumed to be a response
to functional limitations (such as reduced vision or cognitive
ability) that the driver recognizes as increasing their risk while
driving. Appropriate use of self-regulation may assist older drivers
to reduce their crash risk and safely maintain mobility, thereby
avoiding the negative outcomes, perceived or otherwise, that may
be attributed to premature driving cessation (Edwards et al., 2009).
To date, most studies of older adults driving self-regulation
have operationalized this construct using questionnaires. The
driving habits questionnaire (DHQ) (Owsley et al., 1999), and a
modied form of this questionnaire known as the avoidance scale
of the driving mobility questionnaire (DMQ-A) (Baldock et al.,
2006) are the most commonly utilised. The history of the DHQ and

I.Y. Wong et al. / Accident Analysis and Prevention 80 (2015) 16

Fig. 1. the evolution of items and response formats to assess driving self-regulation by older adults, highlighting three variants, the driving habits questionnaire (DHQ), the
driving mobility questionnaire avoidance scale (DMQ-A), and the extended DMQ-A.

DMQ-A and their relation to each other is shown in Fig. 1. This

gure shows a large degree of overlap in the item content of these
scales (the DMQ-A adds one item to the eight items from the DHQ).
The DMQ-A adjustment of the DHQ response scale allows
participants to indicate that they are avoiding situations for
reasons other than their vision. In addition to these two forms of
the measure, several extended versions have been created by
adding more situations, including those that have been nominated
by older drivers themselves. For example, a recent study by
Sullivan et al. (2011) described an additional 15 such situations.
The use of multiple scales (or item combinations) to assess driving
self-regulation by older adults may account for some of the
inconsistencies in the literature regarding this strategy (see Wong
et al., 2014) for a recent review of the varied use of this measure to
assess older drivers self-regulation). It may also raise questions
about how this concept should be assessed and whether the items
should be extended or updated as proposed (Sullivan et al., 2011).
The psychometric properties of the DHQ, DMQ-A and its
variants are difcult to establish. Owsley et al. (1999, p. 211)
reported that the DHQs domain 4 items had a 2-week testretest
reliability of .60 (.44.74); however, it is unclear how these data
should be interpreted. Besides these test-retest reliability data for
the DHQ, to our knowledge no further psychometric data
(e.g., item-correlations; factor analysis) have been reported for
these measures.
Even if scale level data were available, the past use of scores
from these tests shows further variation. Most studies operationalise self-regulation as the overall scale score (summated or
averaged) of the avoidance items, and, less commonly, it has been
measured using a single-item extracted (and/or modied) from the
9-item DMQ-A. The use of summed, or averaged, item scores
presumes that self-regulation is a unidimensional construct, with a
single underlying factor, although this has yet to be tested. It is
possible, instead, that driving self-regulation is multi-faceted. For
example, the original DHQ items focussed specically on driving
situations that older adults avoid due to visual problems; however,
additional items proposed by subsequent research including those
nominated by older adults themselves (e.g., avoiding driving with
passengers) may suggest other reasons for driving self-regulation.
The driving situations that older adults avoid because of perceived
cognitive limitations, for example, could include high trafc roads,
peak hour driving, or situations that involve navigational
challenges (e.g., roundabouts). These situations may be different
from those that are avoided because of visual acuity problems (e.g.,
night time driving, driving into the sun). Driving self-regulation by

older adults could also include the use of other compensatory

strategies (i.e., not just reducing the amount or location of driving,
but using alternative transport or riding with others), which would
support the idea that situations are avoided for reasons other than
vision. It is also possible that the current pool of driving avoidance
items including those that have been recently proposed contain
items that are statistically redundant, yet could contribute to
response biases (for example, due to fatigue or response set). An
investigation of the DMQ-A factor structure would further our
understanding of the construct/s underpinning older adults'
driving self-regulation.
The purpose of this study was to explore the concept of driving
self-regulation by older adults, and focus on its measurement.
Specically, the study aimed to investigate the internal consistency
and factor structure of three self-regulation item sets: 1) the
original 8-item DHQ, 2) the 9-item DMQ-A, and 3) the extended
21-item DMQ-A.
2. Method
2.1. Participants
The participants were 277 older adults (63% female, age range:
6592 years, Mage = 71.64; SD = 5.87) recruited from the community in response to newspaper and email advertisements, iers
distributed via a range of organisations (e.g., Country Womens
Association, Council On The Aging and Health Clinics). Eligible
participants were current drivers and aged 65 years or older.
Potential participants were told that the research was interested in
their opinions on the driving experience and the transportation
needs of older drivers. The response format of the questionnaire
include postal survey (n = 108; Mage = 72.04; SD = 7.21) and
web-based
questionnaire
(Mage = 70.95;
SD = 5.85).
The
demographic characteristics of participants appear in Table 1.
2.2. Materials
2.2.1. Extended DMQ-A
An extended version of the avoidance items of Baldock et al.s
(2006) DMQ was used to assess participants self-regulation.
Twelve new items from the set generated by Sullivan et al. (2011)
were added after the items from the DHQ/DMQ-A. These 12 items
were selected because they described situations that have also
been used in previous studies of older drivers' self-regulation
(Braitman and Williams, 2011; Horswill et al., 2011; Kostyniuk and

I.Y. Wong et al. / Accident Analysis and Prevention 80 (2015) 16

Table 1
Demographic characteristics of participants (n = 277).
Sample characteristic

M (SD)

Age (M/SD)
Years licensed
1120
years
2130 years
3140 years
4150 years
Over 50 years

71.36 (6.38)

Advanced driver training

No
Yes
H/s wk spent driving
Employment (n/%)
Not employed/no voluntary work
Not employed/voluntary work
Retired
Part time
Full time
Residential location (n/%)
Major city
Inner regional
Outer regional
Remote
Very remote
Medical conditions (n/%)
None
1
>2

Table 2
Means and standard deviations of scale items and the total scale score.
n

1
2
21
117
134

0.4
0.7
7.6
43
48.4

236
41

85.2
14.8

8.98 (8.17)
3
20
78.7
29
6

1.1
7.2
79
10.5
2.2

95
1.6
64
6
1

34.3
38.3
23.1
2.2
0.4

129
104
44

46.6
37.5
15.9

Residential location was identied by the accessibility/remoteness index of

Australia (ARIA; ABS, 2011). The medical conditions that were self-reported by
participants as potentially impacting their driving included: vision problems
(27.8%), arthritis (18.4%), heard condition (15.2%), previous heart attack (3.2%),

Molnar, 2008; Meng and Siren, 2012; Vance et al., 2006; Windsor
et al., 2008) (see Table 2 for a list of the items). Participants were
asked to rate on a 5-point Likert scale from 1 never to 5 always,
the extent to which they avoid driving in 21 potentially risky
driving situations.
2.3. Procedure
Ethical clearance for this project was provided by the
institutional review board (see Acknowledgements). Under the
approved protocol, informed consent was inferred from the return
of a completed questionnaire. Questionnaires were presented in
16-point font. The extended DMQ-A was administered as part of a
larger xed-order battery that was prepared and mailed to
interested participants. Questionnaires were returned in a replied
paid envelope. Participants also had the option of completing the
questionnaire using a web based electronic survey.
2.4. Statistical analysis
Scale reliability was assessed using a measure of internal
consistency. Cronbachs alpha (a) is a standard index of this test
property. Although debated, the commonly cited minimal acceptable value of alpha is .7 (Nunnally, 1978; Tavakol and Dennick,
2011). In a recent systematic review that assessed the internal
consistency of selected questionnaires, the following descriptors
were associated with alpha values of .5, through .8 as: poor,
marginal, acceptable and good, respectively (Spector et al., 2012),
and these descriptors were applied in this study. Exploratory factor
analysis was used to determine the factor structure of the original
8-item DHQ, the 9-item of the DMQ-A, and the 21-item extended
DMQ-A.

Extended DMQ-A items

SD

1.At night in the rainb

2.In the raina
3.At nighta
4.Peak houra
5.In foggy conditionsc
6.High trafc roadsa
7.Long distance drivingc
8.Freewaysa
9.At the start/end of school timesc
10.Tunnelsc
11.Other people's carc
12.Unfamiliar roadsc
13.Lane changesc
14.Roadworksc
15.Parallel parkinga
16.Right turnsa
17.Roundaboutsc
18.With passengers (children)c
19.When alonea
20.With passengers (adult)c
21.Familiar roadsc

2.49
2.35
2.09
2.01
1.89
1.88
1.80
1.76
1.76
1.73
1.64
1.63
1.56
1.56
1.50
1.45
1.42
1.40
1.30
1.28
1.25

1.31
1.39
1.24
1.15
.99
1.20
1.17
1.18
1.07
1.16
.89
.90
.93
.93
.87
.90
.89
.83
.69
.68
.75

Total scale score

1.70

.77

n = 277. Avoidance items were rated on a scale ranging from 1 (never avoid) to 5
(always avoid). Higher scores indicate greater avoidance.
a
DHQ/DMQ-A items.
b
The added DMQ-A item.
c
Items that were added to the DMQ-A (extended DMQ-A).

3. Results
3.1. Degree of driving self-regulation and scale reliability
Table 2 demonstrates that, overall, participants reported
relatively low avoidance of the driving situations. Approximately
17% (n = 46) of the sample indicated that they never avoided any
of the 21 driving situations. The highest avoidance rating for an
individual item was driving at night in the rain (M  2.5), where
2 and 3 on this scale indicate rarely and sometimes,
respectively. The lowest avoidance rating was on familiar roads
(M = 1.25; SD = .77). Analysis with Cronbachs alpha revealed good
internal consistency for the 8- (a = .90), 9- (a = .91), and 21-item
sets (a = .96).
3.2. Factor structure of the 8-item DHQ and the 9-item DMQ-A
The KaiserMeyerOlkin values of the original 8-item DHQ and
the 9-item DMQ-A were .89 and .90, respectively, and both values
exceeded the recommended value of .6 (Kaiser, 1974). Bartletts test
of sphericity (Bartlett, 1954) for the correlation matrix for each of

Table 3
Scale characteristics of the DHQ and DMQ-A.

Item

DHQ
Communalities

DMQ-A
Communalities

In the rain
When alone
Parallel parking
Right turns
Freeways
High trafc roads
Peak hour
At night
At night in the rain
Scale reliability (a)

.660
.368
.372
.685
.736
.809
.719
.687

.90

.696
.337
.376
.682
.707
.789
.710
.695
.691
.91

n = 277. DHQ, driving habits questionnaire (Owsley et al., 1999); DMQ-A, driving
mobility questionnaire avoidance (Baldock et al., 2006). Note that the DMQ-A
includes the 8 items of the DHQ, plus 1 additional item: at night in the rain.

I.Y. Wong et al. / Accident Analysis and Prevention 80 (2015) 16

Table 4
Pattern and structure matrix for principal components analysis with Oblimin rotation of two factor solution of 14 driving self-regulation items selected from the 21-item
extended DMQ-A.a
Item

At night in the rain

In the rain
At night
Peak hour
In foggy conditions
Long distance driving
Freeways
Tunnels
Other peoples car
High trafc roads
With passengers (children)
When alone
With passengers (adult)
Familiar roads

Pattern coefcients

Structure coefcients

Factor 1

Factor 2

0.87
0.80
0.75
0.79
0.88
0.83
0.74
0.84
0.78
0.78
0.14
0.04
0.04
0.03

0.05
0.05
0.15
0.10
0.11
0.06
0.21
0.15
0.06
0.18
0.73
0.93
0.90
0.90

Factor 1

Factor 2

Communalities

0.84
0.83
0.82
0.83
0.82
0.86
0.84
0.77
0.75
0.87
0.49
0.41
0.47
0.40

0.37
0.44
0.51
0.47
0.31
0.46
0.56
0.26
0.31
0.56
0.79
0.92
0.92
0.89

0.71
0.69
0.69
0.70
0.69
0.75
0.73
0.61
0.56
0.64
0.84
0.85
0.78
0.77

n =277.
a
Six cross loading items were dropped from the nal model, and parallel parking was excluded because of low communality (.36) on Factor 1. Major loadings for each item
are bolded. With the exception of driving when alone, all items for Factor 2 were items added to the DMQ-A. Factor 1, external driving environment; Factor 2, internal
driving environment.

the scales reached statistical signicance (p < .001), supporting the

factorability of both scales. A principal component analysis (PCA)
of the 8-item DHQ and the 9-item DMQ-A revealed a one-factor
solution, which accounted for 62.95% and 63.15% of the variance,
respectively. Table 3 displays the communalities of the items
across the 8-item DHQ and the 9-item DMQ-A. For both versions,
the communalities of all items exceeded .65 with the exception of
driving when alone and parallel parking, where the communalities for both items ranged from .34 and .39 (see Table 3).
3.3. Factor structure of the 21-item extended DMQ-A
The KaiserMeyerOlkin value of the 21-item extended DMQ-A
was .94. Bartletts test of sphericity (Bartlett, 1954) reached
statistical signicance (p < .001), supporting the factorability of the
correlation matrix. Inspection of the correlation matrix of the 21item extended DMQ-A revealed that the items were moderately
intercorrelated (all rs > 0.3). However, several inter-item correlations exceeded 0.8, indicating a potential problem with multicollinearity and singularity. All 21 items were entered into a PCA.
The PCA of the extended DMQ-A revealed three components with
eigenvalues exceeding 1, explaining 58.16%, 9.85% and 5.65% of the
variance, respectively. Inspection of the screeplot revealed a clear
inection point after the second component. The two-factor
solution was further supported by the results of Parallel Analysis,
which showed that there were only two components with
eigenvalues exceeding the corresponding criterion values (PCA
eigenvalue 12.21, 2.07, 1.19; Parallel Analysis value 1.54, 1.43, 1.36)
for a randomly generated data matrix of the same size
(21 variables  277 respondents). The component matrix revealed
that there were two, cross-loading items on component 31,
suggesting a two-factor solution may be more parsimonious.
A second exploratory factor analysis using the 21 items was
undertaken and two factors were extracted. This two-factor
solution explained a total of 68% of the variance, with Factor
1 contributing 58.16% and Factor 2 contributing 9.85% of the total.
The component correlation matrix revealed strong positive
correlations between the two factors (r = .54), supporting the
use of the direct oblimin rotation. However, the pattern matrix
revealed that several cross-loading items. As per the method of

1
Parallel parking had a loading of .54 on factor 3; and .3 on Factor 1; avoiding
right turns had loadings of .4 .5 on all three factors.

Tabachnick and Fidell (2001), items with cross-loading above

.32 were deleted from subsequent analyses (n = 6). These items
related to avoiding driving: at the start/end of school times, on
unfamiliar roads, or when lane changes, roadworks, right
turns, or roundabouts might be encountered. The parallel
parking item was also deleted from further analysis because of
low communality (<.5) on its corresponding factor (i.e., Factor 1),
leaving 14 items.
A third analysis with the 14 remaining items produced a
two-factor solution that accounted for 71.5% of the variance
(Factor 1: 57.45%, Factor 2: 14.07%). The Cronbachs alpha for
each of the factors was .95 and .90, respectively indicating good
internal consistency for the two subscales. There was a
moderate and positive correlation between the two factors
(r = .48). All items, on both factors had communalities >.5.
Driving in another persons car had the lowest communality (.56,
see Table 4). Most of the items that loaded onto Factor
1 represented driving situations that are inuenced by
environmental conditions or factors that are arguably external
to the car. On this basis, we labeled this factor the external
driving environment. With the exception of familiar roads, the
variables that loaded onto Factor 2 were primarily related to
factors that are internal to the car. On this basis, we labeled this
factor as the internal driving environment.
4. Discussion
This study investigated the psychometric properties of the
DHQ, DMQ-A and a recently extended DMQ-A; three variants of
an item set that have been used to assess older adults driving
self-regulation. These data were gathered to assist researchers in
their selection and interpretation of these scales by providing
data on their psychometric properties and to explore the nature
of this construct. This study documented the development and
evolution of an item set that was initially developed to assess
driving avoidance due to visual problems (the DHQ), to one that
was updated using recent driver-input from a community sample
and now includes situations that may be avoided for other
reasons.
All three scales were internally consistent (a > .9). The 8-item
DHQ and 9-item DMQ-A both yielded a one factor solution
suggesting that the items assess a common theme. However, the
low factor loadings of two items that assessed driving avoidance
when alone and when parallel parking, respectively, indicate

I.Y. Wong et al. / Accident Analysis and Prevention 80 (2015) 16

that these items did not t well with the other items within the
scale. This suggests that these two scales provide a reliable
measure of driving self-regulation by older adults, but when the
total scale score is used it may not adequately characterise
responses in these two areas. In other words, these items may be
assessing domains that the DHQ or DMQ-A summary score does
not fully capture.
After a series of preliminary analyses, a factor analysis of
14 items from the extended DMQ-A yielded a two factor solution,
measuring driving avoidance due to concerns about the external
and internal driving environments, respectively. The identication
of an internal driving environment factor could account for the
problem of low loading of the item driving when alone identied
in the shorter versions. With more items assessing this construct, a
separate factor emerged. This nding suggests that older adults
may also self-regulate their driving for reasons related to the
internal environment of their car, rather than, as has often been
construed, as a result of physical impairments, such as failing
vision. Although it is not possible to conclude from this study why
older adults might avoid driving based on factors related to the
internal (in-vehicle) environment (e.g., the number and nature of
passengers), it is possible that this factor assess concerns about
distraction, cognitive stimulation (either adding to or sharing the
load), or a fear of being evaluated by others. In addition, the clear
separation of this factor from the external driving environment,
which more closely relates to traditional interpretations of this
concept, is notable. In this study, the external driving environment included items that may involve a visual element (such as
night time driving) and items assessing trafc-dependent situations (such as driving during peak hour or on freeways), raising
the possibility that the concept relates to perceptions about the
external environment, broadly dened. Such perceptions may
represent a higher order consideration for older drivers that has
until now been difcult to identify.
There are several study limitations. First, the results of this
study indicate that, on average, this sample of older drivers did not
frequently avoid the nominated driving situations. The nding of
low avoidance, on average, is consistent with previous reports
(Gwyther and Holland, 2012; Horswill et al., 2011; Kostyniuk and
Molnar, 2008; Sullivan et al., 2011). Low driving avoidance could
reect: (a) a sampling bias, whereby participants who were not
self-regulating their driving were the ones that volunteered for the
study; (b) a response bias, whereby older drivers did not admit to
self-regulation; or, (c) that self-regulation is not a commonly
employed strategy in this sample. Second, the two-factor solution
that emerged from the 14 extended DMQ-A item had two
conceptual anomalies. Whilst most of the items that loaded onto
each factor could be interpreted, the item driving other peoples
cars would seem to be related to the in-vehicle environment rather
than the external environment, and driving on familiar roads
would seem to be related to the external rather than internal
environment. Statistically, the item driving other peoples cars had
the lowest communality on its corresponding factor (i.e., Factor 1),
and the familiar roads item had the lowest mean of the items, and
was highly correlated with another item on this factor (e.g., when
alone, r = .84). Further scale analysis may clarify the contribution of
these items. Third, this study used a measure of self-reported
driving self-regulation, whereas it is now feasible to measure this
construct objectively (Molnar et al., 2013). However, we suggest
that there are still important applications for questionnaire data on
driving self-regulation, especially if these measures can be rened
and this study makes an important contribution to this achieving
this objective.
The results of this study may guide researchers in their
selection of measures and their interpretation of scores from selfreported driving self-regulation scales. A review of current practice

reveals that various approaches to the use of these data, ranging

from the use of an overall scale score, through item level
interpretations of questions from existing scales, to the use of
new items with unknown psychometric properties. In general, the
use of an overall scale score is encouraged over the use of item level
interpretations because this improves measurement reliability and
as this study shows all three measures, including the two subscales
had good internal consistency. Further, the development and use of
new items that are not tested together with others risks atomising
the problem, and as this analysis has shown, if these items are
added to an existing scale, it can reduce the functionality of the
scale overall. Seven of the new items from the extended DMQ-A did
not improve the psychometric properties of the overall scale or
subscales. However, as the results of the factor analyses show some
caution in the use of overall scores is also warranted. If the DHQ,
DMQ-A, or extended DMQ-A overall score is interpreted as a
measure of driving self-regulation (i.e., it is interpreted as a single
construct), it could be masking important variation of participants
self-regulatory driving. Using either scale in this way may not
adequately capture older adults self-regulation because: (a) the
items primarily assess the way in which the external environment
affects self-regulation, yet the analysis of the extended DMQ-A
suggests that there may also be internal driving environment
factors that explain some driving avoidance; and (b) the scales
themselves, whilst reliable, include some relatively poorly tting
items. At a conceptual level, the presence of a two-factor structure
of the recently extended DMQ-A item pool suggests that older
drivers decisions to self-regulate may be multifaceted, and the
cautious use and interpretation of the 14-item version of it is
encouraged in future research. Further analysis of this measure and
its subscales is needed to fully characterise the psychometric
properties of this measure. Nevertheless, if the two factor structure
is replicated and further invested in the interpretation of the factor
scores, it could lead to differential advice for a driver whose selfregulation is primarily in response to internal rather than external
environmental factors and vice verse.
This study represents the rst step in validating and improving
a measure of driver self-regulation. The next steps should follow a
scale development process such as that described by Myers et al.
(2008). Conrmatory factor analysis should be used to further
evaluate the factor structure that this study has identied, and
Rasch analysis could be applied to explore items that could be
redundant. Further, the periodic but systematic review of items
will be necessary to ensure the relevance and representativeness of
the items in relation to the underlying constructs. It is possible that
our conceptualisation of driving self-regulation by older adults
needs to evolve from that which was captured by the DHQ and
DMQ-A. This research suggests that older adults avoid driving in
specic situations for different reasons. Future studies should
investigate whether the driving self-regulation by older adults is
multi-factorial as suggested, and if so, rene our understanding of
these factors and this safety strategy.

Acknowledgements
The Human Research Ethics Committee of Queensland
University of Technology (QUT-HREC#1000000311) approved
this research. This project was granted an occupational workplace
health and safety clearance. This study was supported by a
research grant awarded to Ides Wong, Dr. Simon Smith,
Dr. Patricia Obst and Dr. Kirsteen Titchener by the NRMA-ACT
Road Safety Trust and the Queensland Injury Prevention Council.
The authors wish to thank Council on the Aging, the University of
3rd Age and the University of Canberra for their contribution to
this research project.

I.Y. Wong et al. / Accident Analysis and Prevention 80 (2015) 16

References
Baldock, M.R.J., Mathias, J.L., McLean, J., Berndt, A., 2006. Self-regulation of driving
and older drivers functional abilities. Clin. Gerontol. 30 (1), 5370. doi:http://
dx.doi.org/10.1300/J018v30n01_05.
Bartlett, M.S., 1954. A note on the multiplying factors for various x2 approximations.
J. R. Stat. Soc. Ser. B 296298.
Braitman, K.A., Williams, A.F., 2011. Changes in self-regulatory driving among older
drivers over time. Trafc Inj. Prev. 12 (6), 568575. doi:http://dx.doi.org/
10.1080/15389588.2011.616249.
Dickerson, A.E., Molnar, L.J., Eby, D.W., Adler, G., Bdard, M., Berg-Weger, M.,
Kerschner, H., 2007. Transportation and aging: a research agenda for advancing
safe mobility. Gerontologist 47 (5), 578590.
Edwards, J.D., Lunsman, M., Perkins, M., Rebok, G.W., Roth, D.L., 2009. Driving
cessation and health trajectories in older adults. J. Gerontol. Ser. A 64 (12),
12901295.
Gwyther, H., Holland, C., 2012. The effect of age, gender and attitudes on selfregulation in driving. Accid. Anal. Prev. 45, 1928. doi:http://dx.doi.org/10.1016/
j.aap.2011.11.022.
Horswill, M.S., Anstey, K.J., Hatherly, C., Wood, J.M., Pachana, N.A., 2011. Older
drivers insight into their hazard perception ability. Accid. Anal. Prev. 43 (6),
21212127. doi:http://dx.doi.org/10.1016/j.aap.2011.05.035.
Kaiser, H.F., 1974. An index of factorial simplicity. Psychometrika 39 (1), 3136.
Kostyniuk, L.P., Molnar, L.J., 2008. Self-regulatory driving practices among older
adults: health, age and sex effects. Accid. Anal. Prev. 40 (4), 15761580. doi:
http://dx.doi.org/10.1016/j.aap.2008.04.005.
Meng, A., Siren, A., 2012. Cognitive problems, self-rated changes in driving skills,
driving-related discomfort and self-regulation of driving in old drivers. Accid.
Anal. Prev. 49, 322329. doi:http://dx.doi.org/10.1016/j.aap.2012.01.023.
Molnar, L.J., Charlton, J.L., Eby, D.W., Bogard, S.E., Langford, J., Koppel, S., Man-SonHing, M., 2013. Self-regulation of driving by older adults: comparison of selfreport and objective driving data. Transport. Res. Part F: Trafc Psychol. Behav.
20, 2938.

Molnar, L.J., Eby, D.W., 2008. The relationship between self-regulation and drivingrelated abilities in older drivers: an exploratory study. Trafc Inj. Prev. 9 (4),
314319.
Myers, A.M., Paradis, J.A., Blanchard, R.A., 2008. Conceptualizing and measuring
condence in older drivers: development of the day and night driving comfort
scales. Arch. Phys. Med. Rehabil. 89 (4), 630640.
Nunnally, J.C., 1978. Psychometric Theory, second ed. McGraw-Hill, New York.
Organisation for Economic Co-operation Development, 2009. Ageing Societies in
OECD Factbook 2009: Economic, Environmental and Social Satistics. OECD
Publishing.
Owsley, C., Stalvey, B., Wells, J., Sloane, M.E., 1999. Older drivers and cataract: driving
habits and crash risk. J. Gerontol. Ser. A 54 (4), M203M211.
Spector, A., Orrell, M., Schepers, A., Shanahan, N., 2012. A systematic review of
knowledge of dementiaoutcome measures. Age. Res. Rev. 11 (1), 6777.
Sullivan, K.A., Smith, S.S., Horswill, M.S., Lurie-Beck, J.K., 2011. Older adults safety
perceptions of driving situations: towards a new driving self-regulation scale.
Accid. Anal. Prev. 43 (3), 10031009. doi:http://dx.doi.org/10.1016/j.
aap.2010.11.031.
Tabachnick, B.G., Fidell, L.S., 2001. Using multivariate statistics, fourth ed. Allyn and
Bacon, Boston.
Tavakol, M., Dennick, R., 2011. Making sense of Cronbachs alpha. Int. J. Med. Educ. 2,
5355.
Vance, D.E., Roenker, D.L., Cissell, G.M., Edwards, J.D., Wadley, V.G., Ball, K.K., 2006.
Predictors of driving exposure and avoidance in a eld study of older drivers
from the state of Maryland. Accid. Anal. Prev. 38 (4), 823831. doi:http://dx.doi.
org/10.1016/j.aap.2006.02.008.
Windsor, T.D., Anstey, K.J., Walker, J.G., 2008. bility perceptions, perceived control,
and risk avoidance among male and female older drivers. J. Gerontol. Ser. B 2,
P75P83. doi:http://dx.doi.org/10.1093/geronb/63.2.
Wong, I.Y., Smith, S.S., Sullivan, K.A., Allan, A.C., 2014. Toward the multilevel older
Persons transportation and road safety model: a new perspective on the role
of demographic functional, and psychosocial factors. J. Gerontol. Ser. B
gbu099.