This article was downloaded by: [University of California Santa Cruz]On: 10 October 2014, At: 13:05Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Aging, Neuropsychology, and Cognition:A Journal on Normal and DysfunctionalDevelopmentPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/nanc20

Executive Functioning and Gambling:Performance on the Trail Making Test isAssociated with Gambling Problems inOlder Adult GamblersWilliam von Hippel a , Lily Ng b , Laura Abbot a , Samantha Caldwella , Georgia Gill a & Kym Powell aa School of Psychology , University of Queensland , St Lucia,Australiab Department of Psychology , University of New South WalesSydney , AustraliaPublished online: 19 Oct 2009.

To cite this article: William von Hippel , Lily Ng , Laura Abbot , Samantha Caldwell , GeorgiaGill & Kym Powell (2009) Executive Functioning and Gambling: Performance on the Trail MakingTest is Associated with Gambling Problems in Older Adult Gamblers, Aging, Neuropsychology,and Cognition: A Journal on Normal and Dysfunctional Development, 16:6, 654-670, DOI:10.1080/13825580902871018

To link to this article: http://dx.doi.org/10.1080/13825580902871018

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoever orhowsoever caused arising directly or indirectly in connection with, in relation to or arisingout of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

Aging, Neuropsychology, and Cognition, 16: 654–670, 2009http://www.psypress.com/ancISSN: 1382-5585/05 print; 1744-4128 onlineDOI: 10.1080/13825580902871018

© 2009 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business

NANC1382-5585/051744-4128Aging, Neuropsychology, and Cognition, Vol. 1, No. 1, April 2009: pp. 1–23Aging, Neuropsychology, and Cognition

Executive Functioning and Gambling: Performance on the Trail Making Test is Associated with Gambling Problems in Older Adult GamblersExecutive Functioning and GamblingWilliam von Hippel et al.

WILLIAM VON HIPPEL1, LILY NG

2, LAURA ABBOT1, SAMANTHA CALDWELL

1, GEORGIA GILL

1, AND KYM POWELL1

1School of Psychology, University of Queensland, St Lucia, Australia, and 2Department of Psychology, University of New South Wales Sydney, Australia

ABSTRACT

Rates of gambling problems in older adults have risen with increased accessibility ofgambling venues. One possible contributor to problem gambling among older adults isdecreased self-control brought about by diminished executive functioning. Consistentwith this possibility, Study 1 revealed that older adults recruited from gambling venuesreported greater gambling problems if they also experienced deficits in executivefunctioning, measured via the Trail Making Test. Study 2 replicated this finding anddemonstrated that problem gambling is associated with increased depression amongolder adults, mediated by increased financial distress. These studies provide supportfor the hypothesis that older adult gamblers who have executive functioning problemsare also likely to have gambling problems.

Keywords: Executive functioning; Mental set shifting; Inhibition; Problem gambling;Self-control.

INTRODUCTION

Over the past 30 years the percentage of adults over the age of 65 who gamblehas risen dramatically – more than doubling in some locales – and rates of prob-lem gambling have risen as well, particularly in populations with easy access tocasinos and other forms of betting (Petry, 2005). It is unclear from these trends,however, why gambling problems are increasing among older adults. It could

Address correspondence to: William von Hippel, School of Psychology, University of Queensland,St Lucia, QLD 4072, Australia. E-mail: billvh@psy.uq.edu.au

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

EXECUTIVE FUNCTIONING AND GAMBLING 655

be the case that the increased accessibility of gambling, in combination with theavailability of leisure time and expendable income, leads some older adults todevelop gambling problems. Alternatively, it is possible that gambling prob-lems in older adults are at least partially the result of age-related atrophy of thefrontal lobes leading to diminished executive functioning.

In late adulthood the total volume and weight of the brain graduallyreduces as neurons and the connections between neurons shrink in size,thereby making them less effective (Tang, Whitman, Lopez, & Baloh, 2001).One area of the brain that shows substantial atrophy with age is the prefrontalcortex (Raz & Rodrigue, 2006). Because the prefrontal cortex is the corticalregion primarily responsible for executive functioning, which includes con-trol of thought and action (Banfield, Wyland, Macrae, & Heatherton, 2004;Koechlin, Ody, & Kouneiher, 2003; Miller & Cohen, 2001), atrophy of thisregion of the brain can lead to deficits in self-control (Raz & Rodrigue,2006; Stuss & Knight, 2002; West, 1996). For older adults who gamble,losses in the ability to control their behavior may lead to excessive or patho-logical gambling (see Potenza et al., 2003).

Consistent with such a possibility, executive dysfunction has beendocumented among young adult pathological gamblers (Goudriaan, Oosterlaan,de Beurs, & van den Brink, 2006), suggesting that executive control may playa role in curtailing excessive gambling. Poor executive control has also beenassociated with a more impulsive decision making style, in which peoplewith poorer executive functioning show greater discounting of delayedrewards (Hinson, James, & Whitney, 2003). This finding suggests that exec-utive control may be important in a variety of decision making tasks inwhich people must weigh current costs and benefits against future or distantconsequences. Gambling is one such task, in which the current desire togamble must be balanced against future costs should the likely outcome ofmonetary losses emerge. Because deficits in executive control are a commonconsequence of normal adult aging (Hasher, Zacks, & May, 1999; West,1996), many older adults who gamble are likely to experience losses in theirability to control their gambling behavior. That is, as a consequence ofnormal age-relate declines in executive functioning, older adults might findit increasingly difficult to control their urge to gamble, and might find them-selves betting more money or more frequently than they should.

Although there is no direct evidence to support such a possibility, exec-utive decline in late adulthood has been shown to predict losses of self-controlin other domains (von Hippel, 2007). For example, age-related deficits inexecutive control have been implicated in age-related increases in off-targetverbosity (Pushkar et al., 2000), socially inappropriate behavior ( Henry, vonHippel, & Baynes, in press; von Hippel & Dunlop, 2005), and stereotypingand prejudice (Henry et al., in press; Stewart, von Hippel, & Radvansky,2009; von Hippel, Silver, & Lynch, 2000). Indeed, individual differences in

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

656 WILLIAM VON HIPPEL ET AL.

executive control have also been linked to socially inappropriate behavioramong young adults (von Hippel & Gonsalkorale, 2005; see also von Hippelet al., 2000), suggesting that executive functioning can play an importantrole in self-control independent of general cognitive decline.

It is important to note, however, that poor executive functioning isunlikely to predict gambling problems among all older adults. Rather, execu-tive control should only predict gambling problems among those individualswho have an urge to gamble. Those older adults who do not gamble or whohave very little interest in gambling are unlikely to show a relationshipbetween executive dysfunction and problem gambling, as they are unlikelyto have a desire to gamble that must be restrained. So although most olderadults show declining executive functioning, these declines in executivefunctioning are expected to cause gambling problems only in those olderadults who are active in gambling. Thus, the primary hypothesis of the cur-rent research is that older adults who gamble will show increasing gamblingproblems to the degree that they also show decreasing executive functioning.

Two studies were conducted to test this hypothesis. In both studies olderadults were recruited from gambling establishments to ensure that all partici-pants engaged in at least occasional gambling activities. Participants werethen given a measure of executive functioning (described below) and a self-report measure of gambling problems. If executive decline leads to gamblingproblems, then an association should emerge in this sample of older adultgamblers between a measure of executive functioning and problem gambling.The goal of Study 1 was to establish whether such an association exists. Thegoal of Study 2 was to assess the reliability of such an association, and also toassess whether self-reported gambling problems in this sample are related tothe types of financial and mental health problems typically associated withproblem gambling. In combination, these studies were intended to provideevidence that among older adults who gamble, poorer executive functioningis associated with an increased likelihood of experiencing gambling problemsand the attendant difficulties that such problems entail. Such a pattern of find-ings would be a first step in establishing that executive decline actuallycauses gambling problems among older adult gamblers.

Both of these studies relied on a measure of executive functioningknown as the Trail Making Test (TMT; Army Individual Test Battery,1944). Part A of the TMT requires participants to draw a line from thenumbers 1 to 25 in consecutive order. Part B requires participants to draw aline from 1 to A to 2 to B, etc., until 24 lines have been drawn. The primarymeasure of executive functioning in the TMT is the increase in time taken tocomplete Part B compared to Part A, as Part B requires participants to inhibitthe tendency to move from number to number, or letter to letter, as they mustrepeatedly switch back and forth between numbers and letters. Differentialperformance on Part B vs. Part A has been associated with set shifting

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

EXECUTIVE FUNCTIONING AND GAMBLING 657

performance (Arbuthnott & Frank, 2000) and with perseverative errors onthe Wisconsin Card Sorting Task (Kortte, Horner, & Windham, 2002).These findings, and indeed a substantial research literature with the TMT,suggest that poor performance on the TMT can be considered an indicator ofthe tendency to perseverate in one response mode and to have difficultyshifting from one mental set to another. These aspects of the TMT make itwell suited for the current research, as problem gambling is defined as ‘per-sistent and recurrent maladaptive gambling behavior characterized by aninability to control gambling’ (APA, 1994). Thus, the lack of executive con-trol that makes set switching on the TMT difficult for some older adults mayalso make it more difficult for them to control their perseverative (i.e., per-sistent and recurrent) gambling. If so, then a relationship should emergebetween performance on the TMT and gambling problems among olderadults who gamble.

STUDY 1

Study 1 assessed whether the TMT correlates with gambling problemsamong older adults who gamble. Participants were recruited at a gamblingestablishment, and the study itself was conducted on the premises of theestablishment. To avoid interrupting patrons of the establishment for longerthan necessary, all measures were chosen with brevity in mind. The primarymeasures were a self-report gambling problems scale and the TMT. Becauseexecutive functioning is related to impulsivity (Hinson et al., 2003; Nigg,2000), and because impulsivity is strongly related to gambling problems(Steel & Blaszczysnki, 1998), this experiment also included a measure ofimpulsivity. If the TMT predicts gambling problems beyond their mutualrelationship with impulsivity, then the TMT should account for variance ingambling problems independent of that predicted by impulsivity.

Method

Participants

Sixty-four older adults (age 60–85, M = 70.9, SD = 6.6; 45 female)were recruited from people playing poker machines in Sydney, Australia,and paid $20 AUD (∼$15 US) for their participation. The study itself wasconducted in a quiet area on the gambling premises.

Procedure

Participants first completed a demographic scale that included age,gender, and education. Because the most commonly used self-report scale forgambling problems – the South Oaks Gambling Screen (SOGS; Lesieur &Blume, 1987) – has a number of items that are inappropriate for older adults

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

658 WILLIAM VON HIPPEL ET AL.

who may have only recently developed gambling problems, 6 items werechosen from this scale and an additional five items were chosen from theGambler’s Anonymous Scale. The scoring of the items was also changedfrom yes/no to a 5-point scale with the verbal labels of Never, Rarely,Sometimes, Often, and Very Often. Scores were averaged to form a gamblingproblems scale with a possible range from 1 to 5 (M = 1.99, SD = .90,a = .94). Participants then completed a 10-item version of the Eysenckimpulsivity scale (Eysenck & Eysenck, 1991) in which items were chosenthat seemed most relevant to older adults. Responses to these items wereprovided on the same scale as the revised gambling scale (M = 2.42,SD = .53, a = .74).

Participants then completed the Trail Making Test. To minimize theimpact of processing speed on the TMT performance, a residual score wascomputed such that the variance in Part B accounted for by Part A wasremoved from the measure. This residual score was computed by regressingPart B onto Part A, and then saving the unstandardized residual from thisanalysis to serve as the measure of TMT performance.1

Results

Analyses revealed that the gambling problems scale was correlatedwith impulsivity (r = .49, p < .001) and marginally correlated with theresidual TMT (r = .23, p = .07). Impulsivity was uncorrelated with the resid-ual TMT (r = .02, ns). To assess whether the TMT predicted independentvariance in gambling problems beyond that accounted for by impulsivity,age, gender, and education, the gambling problems scale was regressed onresidual TMT, impulsivity, age, gender, and education. This analysisrevealed that impulsivity (β = .41, p = .001) and the residual TMT (β = .26,p < .05) both accounted for independent variance in gambling problems. Noother variables were significant predictors in this analysis.

Discussion

Consistent with predictions, an association emerged between the TMT andself-reported gambling problems among older adults recruited from a gamblingestablishment. Furthermore, the association between the TMT and gamblingproblems was independent of the relationship between these variables andimpulsivity. Nevertheless, several questions remain. First, it was unclearwhether the TMT accounts for variance in gambling problems independent ofthe frequency with which people gamble. In Study 1 participants reported ontheir gambling problems, but not the frequency with which they gamble, andthus it is impossible to know if the TMT is associated with gambling problems

1The results were functionally identical when raw scores were used for the TMT in both studies.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

EXECUTIVE FUNCTIONING AND GAMBLING 659

independent of the actual frequency with which people gamble. To address thisissue, participants in Study 2 were asked how much they play poker machines. Ifpoor executive functioning manifests itself in more frequent gambling, then theTMT would be unlikely to account for unique variance in gambling problemsbeyond a measure of gambling frequency. In contrast, if poor executive func-tioning leads to gambling problems independent of gambling frequency, then thetwo measures should account for independent variance in gambling problems.Additionally, to ensure that poker machine playing was not just a proxy for timespent at the gambling establishment, participants were also asked how muchthey play bingo. Poker playing was expected to predict gambling problems, butbingo is less likely than poker machines to be associated with gambling prob-lems (Productivity Commission, 1999). As a consequence, if the measuresassess actual time spent gambling, then time spent playing poker should be pre-dictive of gambling problems but time spent playing bingo should not.

Second, no evidence was provided for the validity of the gamblingproblems scale. If self-reported gambling problems are of importance amongolder adults, they are likely to be associated with financial difficulties (APA,1994), which in turn have the potential to lead to depression (Battersby,Tolchard, Scurrah, & Thomas, 2006; MacDonald, McMullan, & Perrier, 2004).Thus, Study 2 included measures of financial difficulties and depression.

Third, gambling problems have also been associated with superstition(Raylu & Oei, 2004), and superstition has been linked to poor executivefunctioning (Wain & Spinella, 2007). If executive control is a unique predic-tor of gambling problems, it should also account for variance in gamblingproblems beyond that predicted by superstition. Finally, no measure wastaken in Study 1 to ensure that participants were not suffering from incipientdementia. To address this concern, participants in Study 2 were given theMini-Mental Status Exam (MMSE; Kukull et al., 1994), thereby enabling anassessment of whether the TMT is associated with gambling problemsamong people who do and do not show signs of incipient dementia.

STUDY 2

Method

Participants

One hundred and sixty-five older adults (age 60–94, 99 female) wererecruited from people playing bingo or poker machines at gambling sites inBrisbane, Australia, and paid $20 AUD (∼$15 US) for their participation.

Procedure

Participants first completed a demographic scale as in Study 1. Theorder of the remaining measures was counterbalanced, with the exception of

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

660 WILLIAM VON HIPPEL ET AL.

the measure of financial distress and the dementia screening (which weregiven last). To assess frequency of gambling, participants indicated, on5-point scales that ranged from Never to Daily, how often they play bingoand poker machines. Participants also indicated how much time they typi-cally spend playing bingo and poker machines. The latter questions wereresponded to on a 5-point scale that ranged from None to Over 6 hours. Toassess executive functioning, participants completed the TMT as in Study 1.

To tap a wide variety of superstitious beliefs, participants completed a15-item superstition scale that was composed of items from the SuperstitiousBeliefs in Gambling Scale (Joukhador, Blaszczynski, & Maccallum, 2004),the Predictive Control subscale of the Gambling Related Cognitions Scale(Raylu & Oei, 2004), and a few items based on observations of pokermachine gamblers (e.g., ‘When I’m playing the pokies [poker machines], Isometimes feel a big win is coming up and I want to stay on my machinerather than switch or take a break.’). Responses were provided on a yes/noformat, and scores were summed to form a scale with a possible range of 0 to15 (a = .79). Participants also completed a 6-item version of the impulsivitytest from Study 1, with responses provided on a yes/no format, and scoreswere summed to form a scale with a possible range from 0 to 6 (a = .64).

To assess gambling problems, participants also completed a 10-itemversion of the SOGS, which was used in place of the blended gamblingproblems scale from Study 1. The 10 items chosen for administration onlyreferred to consequences and problems with gambling, and did not explicitlyask about frequency of gambling (e.g., ‘Have people criticized your bettingor told you that you had a problem, regardless of whether you thought it wastrue?’). Responses were provided on a yes/no format, and scores were aver-aged to form a scale with a possible range from 0 to 1 (a = .77). Participantsalso completed a 3-item version of the InCharge Financial Distress Scale(Prawitz et al., 2006). Responses were provided on 7-point scales, and scoreswere averaged to form a scale with a possible range from 1 to 7, with higherscores indicating greater financial distress (a = .83).

To assess depression, participants completed a 14-item version of theGeriatric Depression Scale (GDS; Sheikh & Yesavage, 1986). Responseswere provided on a yes/no format, and scores were summed to form adepression scale with a possible range of 0 to 14 (a = .73). The final task forparticipants was dementia screening, which relied on the Mini-Mental StatusExam (MMSE; M = 26, SD = 3.23). Participants with scores greater than 26out of 30 are considered to be of normal mental status (i.e., showing no signof dementia; Kukull et al., 1994).2

2The reliabilities reported above were computed with the full sample. Reliabilities with the sub-samplescoring higher than 26 on the MMSE were equivalent or higher than these.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

EXECUTIVE FUNCTIONING AND GAMBLING 661

Results

In the current sample, 84 participants scored above 26 on the MMSE.Because almost half the sample showed signs of possible dementia, the analyseswere conducted with the entire sample, but are also reported separately withthe subsamples that scored above 26 and that scored 26 or less wheneverthose subsamples differed from each other. Descriptive statistics for bothsubsamples are presented in Table 1. The only variables to differ signifi-cantly by MMSE classification were age and impulsivity (p < .05). Correla-tions between all of the variables were then computed separately for the twosubsamples (see Table 2).

As can be seen in Table 2, the gambling problems scale was correlatedwith frequency of playing the poker machines in both subsamples and withtime spent playing poker machines among the high MMSE subsample.Gambling problems were not correlated with frequency or time spent play-ing bingo in either subsample. Gambling problems were correlated with impul-sivity only in the low MMSE subsample and with financial distress only in thehigh MMSE subsample. In none of these cases was the difference between thecorrelations significant across the two subgroups (zs < 1.42, ps > .15).Gambling problems were not significantly correlated with the GDS in eithersubsample, but this relationship was significant when collapsed across bothsubsamples (r = .26, p = .001). Gambling problems were also correlated withsuperstition in both subsamples. Gambling problems were only significantlycorrelated with the residual TMT in the high MMSE subsample, although thedifference between the correlations did not approach significance (z = .23,p > .80). Gambling problems were not correlated with scores on the MMSEwhen collapsed across both subsamples (r = .10, p > .20).

TABLE 1. Descriptive Statistics for the Variables in Study 2, as a Function of MMSE Scores

MMSE ≤ 26 MMSE > 26

Mean SD Mean SD

Age 71.16 6.38 68.40 6.55Poker time 2.73 .94 2.73 .95Poker frequency 2.74 .99 2.89 1.13Bingo time 1.55 .97 1.82 1.09Bingo frequency 1.62 1.11 1.95 1.20Gambling problems .19 .16 .24 .21Financial distress 3.28 1.14 3.01 1.32Depression .16 .15 .17 .17Residual TMT −1.90 28.93 .81 35.60Impulsivity 2.35 1.56 3.02 1.75Superstition .22 .20 .23 .21

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

662 WILLIAM VON HIPPEL ET AL.

TA

BL

E 2

.C

orre

latio

ns B

etw

een

Var

iabl

es in

Stu

dy 2

, as

a Fu

nctio

n of

MM

SE

Sco

res

(MM

SE

> 2

6 ab

ove

diag

onal

, MM

SE

≤ 2

6 be

low

dia

gona

l)

Age

Poke

r ti

me

Poke

r fr

eque

ncy

Bin

go

tim

eB

ingo

fr

eque

ncy

Gam

blin

g pr

oble

ms

Fina

ncia

l di

stre

ssD

epre

ssio

nR

esid

ual

TM

TIm

puls

ivity

Sup

erst

ition

Age

−.19

−.18

.03

−.01

−.05

−.09

.00

.05

−.03

.12

Pok

er ti

me

.17

.44*

**−.

01−.

12.3

7***

.14

.12

.01

−.12

.18

Poke

r fr

eque

ncy

−.10

.36*

*−.

24*

−.32

**.3

5**

.10

−.03

.00

−.05

.16

Bin

go ti

me

−.11

−.23

−.31

*.7

9***

−.10

.06

.22*

.08

−.04

.08

Bin

go f

requ

ency

−.12

−.12

−.34

**.9

2***

−.10

.01

.17

.04

−.10

.09

Gam

blin

g pr

oble

ms

−.02

.18

.30*

.03

−.01

.34*

*.1

9.2

3*.0

4.3

0**

Fina

ncia

l dis

tres

s−.

03.0

1−.

16.1

4.1

4.1

1.4

4***

.04

.15

.20

Dep

ress

ion

−.03

.12

−.03

.05

.09

.23

.03

.20

.06

.12

Res

idua

l TM

T.0

6.0

0.3

4**

−.23

−.24

.19

.03

.05

−.03

.04

Impu

lsiv

ity−.

20.1

8.2

0.1

0.0

4.2

5*−.

25*

.09

.09

.04

Supe

rstit

ion

.10

.19

.20

−.11

−.14

.29*

−.01

.12

.29*

.20

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

EXECUTIVE FUNCTIONING AND GAMBLING 663

To assess whether the TMT accounted for independent variance ingambling problems (beyond that accounted for by impulsiveness, supersti-tion, frequency/time spent playing poker machines, and the MMSE), andwhether gambling problems had downstream consequences for financial dis-tress and depression, a regression-based causal model was estimated follow-ing the procedures outlined in Baron and Kenny (1986). Age, gender, andeducation were also included as control variables in this model (none ofwhich predicted unique variance in gambling problems). In the first step ofthe model, scores on the SOGS were regressed on impulsiveness, supersti-tion, frequency/time spent playing poker machines, the MMSE, the residualTMT, and age, gender, and education. Consistent with predictions, the TMTwas found to account for independent variance in gambling problems,beyond that accounted for by all other variables (see Figure 1). When thisregression equation was run separately in the high and low MMSE sub-samples, the residual TMT measure accounted for independent variance ingambling problems in the high MMSE subsample (β = .26, p < .02), but notin the low MMSE subsample (β = .04, p > .80).

FIGURE 1. Regression-based path analysis of the effects of poker machine playing, executivefunctioning, impulsivity, and superstition on gambling problems, and the subsequent mediated impactof gambling problems on depression via financial distress. Path coefficients are standardizedregression weights. The coefficient in parentheses between gambling problems and depressionrepresents the direct effect when the mediator is included in the model, the coefficient not inparentheses represents the direct effect when the mediator is not included in the model. Gender, age,education, and scores on the Mini-Mental Status Exam were included in the model as controlvariables, but were not significant predictors of gambling problems, and thus are not displayed in thefigure. Note: *p < .05; **p < .01.

Time spentPlaying Poker

FrequencyPlaying Poker

Residual TMTScores

Impulsivity

GamblingProblems

FinancialDistress

Depression

.28**

.28**

.07

.18*

.18*

.19*

(.22*)

.20*

Superstition

.21*

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

664 WILLIAM VON HIPPEL ET AL.

In the second step of the model, financial distress was regressed onimpulsiveness, superstition, frequency/time spent playing poker machines,the MMSE, the residual TMT, the SOGS, and age, gender, and education.Consistent with predictions, gambling problems predicted financial distress.In the third and fourth steps of the model, depression was regressed on all ofthese variables either with (step 3) or without (step 4) financial distress as apredictor. Consistent with predictions, these analyses revealed that financialdistress mediated the impact of gambling problems on depression (seeFigure 1). A Sobel test revealed that this mediational pathway was signifi-cant (z = 2.03, p < .05). The results of the entire model were essentiallyequivalent (although slightly stronger) when the model was estimated onlywith participants with MMSE scores greater than 26. Additionally, whencompeting models were examined by reordering the variables, no evidencefor significant mediation emerged.

Discussion

The results of Study 2 provide further evidence that the TMT is associ-ated with gambling problems among older adults. In this study, older adultswho performed poorly on this measure of executive functioning also reportedgreater gambling problems. Furthermore, these self-reported gambling prob-lems were associated with financial distress, which in turn predicted depres-sion. These findings suggest that the self-reported gambling problems ofthese older adults were significant in nature, and were having an impact onother important aspects of their lives. Although these findings await corrobo-ration with longitudinal data, they provide suggestive evidence that decline inexecutive control among older adults can lead to substantial gambling prob-lems and to subsequent financial difficulties and mental health problems.

The results of Study 2 also address the possibility that it was incipientdementia that underlies the relationship between executive functioning andgambling problems. In Study 2 the relationship between the TMT andgambling problems was non-significantly stronger among older adults whoperformed better on the cognitive screening test compared to older adultswhose scores suggested possible dementia. Thus, it seems to be the case thatexecutive functioning plays at least as much of a role in gambling problemsamong people whose cognitive functioning is otherwise intact as it doesamong people who might be suffering from dementia. Additionally, the rela-tionship between gambling problems and financial distress emerged prima-rily among participants who showed no signs of incipient dementia. Thus, itseems possible that gambling problems may be only one of many sources offinancial hardship for people who, by virtue of their cognitive deficits, arelikely to have difficulty controlling their finances.

The findings of Study 2 also suggest that the association between exec-utive functioning and gambling problems exists independently of individual

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

EXECUTIVE FUNCTIONING AND GAMBLING 665

differences in impulsivity or superstition. Although superstition was a stron-ger predictor of gambling problems than any of the other variables, execu-tive functioning nevertheless accounted for significant variance beyond thatexplained by levels of superstition. This finding suggests that the type ofexecutive control tapped by the TMT is associated with gambling problemsin a manner that is independent of the type of disordered thinking evident insuperstition.

Lastly, it should also be noted that the TMT predicted problem gam-bling independent of the frequency and duration that participants played thepoker machines. Although frequency and duration of poker playing werethemselves predictors of gambling problems, it seems to be the case thatpoor executive functioning is not associated with gambling problemsbecause people who have difficulty with the TMT play more often or forlonger durations. Rather, it appears to be the case that poor executive func-tioning is associated with worse gambling decisions. It remains a questionfor future research whether poor executive functioning is associated withmaking larger bets, persevering at gambling in the absence of wins, or otherpossible behaviours that are associated with gambling problems.

GENERAL DISCUSSION

The findings from two studies provide evidence for the role of executivefunctioning in problem gambling in older adults. In both studies, older adultswho showed poor executive functioning on the Trail Making Test were morelikely than older adults who showed good executive functioning to reportgambling problems. In Study 1 this relationship was independent of individ-ual differences in impulsiveness, suggesting that executive functioning wasrelated to gambling problems independent of the relationship between exec-utive functioning and impulsivity. Study 2 replicated this finding and alsowent beyond Study 1 in several important ways.

First, Study 2 demonstrated that the relationship between executivefunctioning and gambling problems was independent of individual differ-ences in superstition, suggesting that executive functioning was related togambling problems independent of the relationship between executivefunctioning and the type of disordered thinking evident in superstition.Second, Study 2 revealed that the relationship between executive function-ing and gambling problems was also independent of frequency and amountof gambling, suggesting that poor executive control is not related to gam-bling problems through a mutual relationship with gambling frequency oramount. Third, Study 2 went beyond Study 1 by demonstrating that theseself-reported gambling problems appear to be important, as they were asso-ciated with self-reported financial distress and depression. Fourth, Study 2demonstrated that the relationship between executive functioning and

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

666 WILLIAM VON HIPPEL ET AL.

gambling problems emerged among people whose scores on a cognitivescreening test suggest that they do not suffer from incipient dementia, andindeed this relationship appeared stronger among these individuals thanamong people whose scores suggest dementia as a possibility.

Although the current research does not indicate why executive func-tioning might be more strongly associated with gambling problems amongpeople who show fewer signs of dementia, there are some relevant hints inthe data. First, it does not seem to be the case that restriction of range amongpeople with low scores on the MMSE is the source of possible differences.As can be seen in Table 1, the means and standard deviations for the variousmeasures are highly comparable in the two subsamples. Second, an alterna-tive possibility is that people with higher MMSE scores might be more capa-ble of accurately reporting their degree of gambling problems. Consistentwith this possibility, the gambling problems scale was (nonsignificantly)more strongly correlated with time spent playing the poker machines amongpeople with MMSE scores greater than 26 than among people with MMSEscores less than or equal to 26. Inconsistent with this possibility, however,gambling problems were (nonsignificantly) more strongly correlated withimpulsiveness among people with low rather than high MMSE scores.Furthermore, gambling problems were significantly correlated with fre-quency of poker machine playing and with superstition in both groups ofparticipants. Thus, at this point it is unclear if there really is a differencebetween the two subsamples, and if so, what the cause of it might be. Futureresearch might untangle this problem by including objective corroborationof self-reported gambling problems among people with low vs high MMSEscores and/or other indicators of incipient dementia.

The findings from these studies are consistent with predictions, butthere are weaknesses in the current approach that must be noted. First andmost importantly, the results were cross-sectional and correlational, and thusit is impossible to rule out other causal patterns, or the possibility that therelationship is driven by an unmeasured third variable. For this reason,longitudinal research will be necessary to establish the causal relationshipthat is suggested by the current results. Second, the findings of these studiesrely on self-reported gambling problems, and factors such as impressionmanagement, self-deception, or memory loss might cloud the relationshipbetween actual problem gambling and self-reported gambling problems. Forthis reason, future studies should assess behavioral indicators of gamblingproblems. For example, perseverance in gambling tasks has been shown to behigher among pathological gamblers than among normal adults (Goudriaan,Oosterlaan, de Beurs, & van den Brink, 2005). If older adults who gamblerely on executive control to prevent themselves from gambling excessively,those with poor executive functioning should have more difficulty stoppinggambling than those with good executive functioning.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

EXECUTIVE FUNCTIONING AND GAMBLING 667

It should also be noted that these results do not suggest that older gamblershave more problems with executive functioning than older non-gamblers,nor do they suggest that executive functioning deficits lead all older adults toexperience gambling problems. Rather, they provide initial evidence thatdeclines in executive control might cause older adults who already gamble todevelop gambling problems. This evidence can only be taken as tentative atthis point in time, as longitudinal data are necessary to establish a causal rolefor executive decline.

Finally, it is worth mentioning that the current research relied on onlya single measure of executive functioning, which is primarily an indicatorof set switching and perseveration. There are a variety of executive func-tions, however, and a wide assortment of tasks designed to measure them.Although the Trail Making Test is well studied and easy to administer ingambling establishments, it would be highly beneficial to broaden thescope of measurement of executive dysfunction. From the current studiesit is unclear whether there is something special about the set switching andperseveration aspects of the Trail Making Test that predict gambling prob-lems among older adults, or whether other aspects of executive controlmight also predict gambling problems. For example, it seems possible thatmeasures of executive functioning that tap behavioral restraint or inhibi-tion (see Lustig, Hasher, & Zacks, 2007; Yoon, May, & Hasher, 2000)might also predict gambling problems among older adults, as gamblingproblems themselves reflect a lack of restraint. This is a question for futureresearch, as the current results clearly represent only a first step in theexamination of the relationship between age, executive functioning, andgambling problems.

CONCLUSIONS

A growing literature suggests that executive decline brought about by frontallobe atrophy leads to a variety of social and cognitive consequences (seeLustig et al., 2007; von Hippel, 2007; West, 1996). The current studiesextend this prior work by demonstrating that executive decline is associatedwith gambling problems among older adults who gamble. These results raisethe possibility that increased gambling among older adults might not alwaysbe an issue of personal choice, as some older adults might have difficultyengaging in self-control when gambling due to losses in executive functioning.Future research should attempt to corroborate these findings longitudinally.If the causal relationship suggested by the current findings is supported inlongitudinal research, such findings would have important public policyimplications, and indeed some protections for older adult gamblers might beappropriate. Such a pattern of findings would also suggest new avenues oftreatment of gambling problems based on factors known to improve

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

668 WILLIAM VON HIPPEL ET AL.

executive functioning among older adults, such as aerobic exercise (Krameret al., 1999) and circadian rhythms (Hasher et al., 1999).

ACKNOWLEDGEMENTS

Support for this research was provided by grants from the Australian Research Council.Preparation of this manuscript was facilitated by a fellowship at the Institute for AdvancedStudy, Berlin.

Original manuscript received March 31, 2008Revised manuscript accepted March 5, 2009

First published online May 29, 2009

REFERENCES

American Psychiatric Association (APA) (1994). Diagnostic and Statistical Manual ofMental Disorders (4th ed.). Washington, DC: American Psychiatric Association.

Arbuthnott, K., & Frank, J. (2000). Trail Making Test, Part B as a measure of executivecontrol: Validation using a set-switching paradigm. Journal of Clinical and ExperimentalNeuropsychology, 22, 518–528.

Army Individual Test Battery (1944). Manual of Directions and Scoring. Washington, DC:War Department, Adjutant General’s Office.

Banfield, J. F., Wyland, C. L., Macrae, C. N., Munte, T. F., & Heatherton, T. F. (2004). Thecognitive neuroscience of self-regulation. In R. F. Baumeister & K. D. Vohs (Eds.),Handbook of self-regulations. New York: Guilford.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in socialpsychological research: Conceptual, strategic, and statistical considerations. Journal ofPersonality and Social Psychology, 51, 1173–1182.

Battersby, M., Tolchard, B., Scurrah, M., & Thomas, L. (2006). Suicide ideation andbehaviour in people with pathological gambling attending a treatment service. Interna-tional Journal of Mental Health and Addiction, 4, 233–246.

Eysenck, H. J., & Eysenck, S. B. G. (1991). Manual of the Eysenck Personality Scales.London: Hodder & Stoughton.

Goudriaan, A. E., Oosterlaan, J., de Beurs, E., & van den Brink, W. (2005). Decision makingin pathological gambling: A comparison between pathological gamblers, alcohol depen-dents, persons with Tourette syndrome, and normal controls. Cognitive Brain Research,23, 137–151.

Goudriaan, A. E., Oosterlaan, J., de Beurs, E., & van den Brink, W. (2006). Neurocognitivefunctions in pathological gambling: A comparison with alcohol dependence, Tourettesyndrome and normal controls. Addiction, 101, 534–547.

Hasher, L., Zacks, R. T., & May, C. P. (1999). Inhibitory control, circadian arousal, and age.In D. Gopher & A. Koriat, (Eds.), Attention and performance XVII: Cognitive regulationof performance: Interaction of theory and application (pp. 653–675). Cambridge, MA:The MIT Press.

Henry, J. D., von Hippel, W., & Baynes, K. (in press). Social inappropriateness and executivecontrol in older adulthood. Psychology and Aging.

Hinson, J. M., Jameson, T. L., & Whitney, P. (2003). Impulsive decision making and workingmemory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29,298–306.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

EXECUTIVE FUNCTIONING AND GAMBLING 669

Joukhador, J., Blaszczynski, A., & Maccallum, F. (2004). Superstitious beliefs in gamblingamong problem and non-problem gamblers: Preliminary data. Journal of GamblingStudies, 20, 171–180.

Koechlin, E., Ody, C., & Kouneiher, F. A. (2003). The architecture of cognitive control in thehuman prefrontal cortex. Science, 302, 1181–1185.

Kortte, K. B., Horner, M. D., & Windham, W. K. (2002). The Trail Making Test, Part B:Cognitive flexibility or ability to maintain set? Applied Neuropsychology, 9, 106–109.

Kramer, A. F., Hahn, S., Cohen, N. J., Banich, M. T., McAuley, E., Harrison, C. R., Chason, J.,Vakil, E., Bardell, L., Boileau, R. A., & Colcombe, A. (1999). Aging, fitness andneurocognitive function. Nature, 400, 418–419.

Kukull, W. A., Larson, E. G., Teri, L., Bowen, J., McCormick, W., & Pfanschmidt, M. L.(1994). The Mini-Mental State Examination score and the clinical diagnosis of dementia.Journal of Clinical Epidemiology, 47, 1061–1067.

Lesieur, H. R., & Blume, S. (1987). The South Oaks Gambling Screen (SOGS): A new instru-ment for the identification of pathological gamblers. American Journal of Psychology,144, 1184–1188.

Lustig, C., Hasher, L., & Zacks, R. T. (2007). Inhibitory deficit theory: Recent developmentsin a ‘new view’. In D. S. Gorfein & C. M. MacLeod (Eds.), The place of inhibition incognition. (pp. 145–162). Washington, DC: APA.

MacDonald, M., McMullan, J. L., & Perrier, D. C. (2004). Gambling households in Canada.Journal of Gambling Studies, 20, 187–236.

Miller, E. K., & Cohen, J. D. (2001). An integrative theory of prefrontal cortex function.Annual Review of Neuroscience, 24, 167–202.

Nigg, J. T. (2000). On inhibition/disinhibition in developmental psychopathology: View fromcognitive and personality psychology and a working inhibition taxonomy. PsychologicalBulletin, 126, 220–246.

Petry, N. M. (2005). Pathological gambling: Etiology, comorbidity, and treatment. Washington,DC: APA.

Potenza, M. N., Leung, H.-C., Blumberg, H. P., Peterson, B. S., Fulbright, R. K., Lacadie, C. M.,Skudlarski, P., & Gore, J. C. (2003). An fMRI Stroop task study of ventromedialprefrontal cortical function in pathological gamblers. American Journal of Psychiatry,160, 1990–1994.

Prawitz, A. D., Garman, E. T., Sorhaindo, B., O’Neill, B., Kim, J., & Drentea, P. (2006).Incharge Financial Distress/Financial Well-Being Scale: Development, administration,and score interpretation. Financial Counseling and Planning, 17, 34–50.

Productivity Commission (1999). Australia’s Gambling Industries: Inquiry Report. Retrievedfrom http://www.pc.gov.au/inquiry/gambling/finalreport/index.html

Pushkar, D., Basevitz, P., Arbuckle, T., Nohara-LeClair, M., Lapidus, S., & Peled, M. (2000).Social behavior and off-target verbosity in elderly people. Psychology and Aging, 15, 361–374.

Raylu, N., & Oei, T. P. S. (2004). The Gambling Related Cognitions Scale (GRCS): Develop-ment, confirmatory factor validation and psychometric properties. Addiction, 99, 757–769.

Raz, N., & Rodrigue, K.M. (2006). Differential aging of the brain: Patterns, cognitive corre-lates and modifiers. Neuroscience and Biobehavioral Reviews, 30, 730–748.

Sheikh, J. I., & Yesavage J. A. (1986). Geriatric Depression Scale (GDS): Recent evidenceand development of a shorter version. In T. L. Brink (Ed.), Clinical gerontology: A guideto assessment and intervention (pp. 165–173), New York: Haworth.

Steel, Z., & Blaszczinski, A. (1998). Impulsivity, personality disorders, and pathologicalgambling. Addiction, 93, 895–905.

Stewart, B. D., von Hippel, W., & Radvansky, G. A. (2009). Age, race, and implicitprejudice: Using process dissociation to separate the underlying components. Psychologi-cal Science, 20, 164–168.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014

670 WILLIAM VON HIPPEL ET AL.

Stuss, D. T., & Knight, R. T. (Eds.). (2002). Principles of frontal lobe function. New York:Oxford.

Tang, Y., Whitman, G. T., Lopez, I., & Baloh, R. W. (2001). Brain volume changes on longi-tudinal magnetic resonance imaging in normal older people. Journal of Neuroimaging, 11,393–400.

von Hippel, W. (2007). Aging, executive functioning, and social control. Current Directionsin Psychological Science, 16, 240–244.

von Hippel, W., & Dunlop, S. M. (2005). Aging, inhibition, and social inappropriateness.Psychology and Aging, 20, 519–523.

von Hippel, W., & Gonsalkorale, K. (2005). ‘That is bloody revolting!’ Inhibitory control ofthoughts better left unsaid. Psychological Science, 16, 497–500.

von Hippel, W., Silver, L. A., & Lynch, M. E. (2000). Stereotyping against your will: Therole of inhibitory ability in stereotyping and prejudice among the elderly. Personality andSocial Psychology Bulletin, 26, 523–532.

Wain, O., & Spinella, M. (2007). Executive functions in morality, religion, and paranormalbeliefs. International Journal of Neuroscience, 117, 135–146.

West, R. L. (1996). An application of prefrontal cortex function theory to cognitive ageing.Psychological Bulletin, 120, 272–292.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

alif

orni

a Sa

nta

Cru

z] a

t 13:

05 1

0 O

ctob

er 2

014