Content uploaded by Anna Bussu
Author content
All content in this area was uploaded by Anna Bussu
Content may be subject to copyright.
THE EFFECT OF SOCIO-ECONOMIC AND EMOTIONAL
FACTORS ON GAMBLING BEHAVIOUR
Anna Bussu
Claudio Detotto
WORKING PAPERS
2013/05
CONTRIBUTI DI RICERCA CRENOS
FISCALITÀ LOCALE E TURISMO
LA PERCEZIONE DELL’IMPOSTA DI SOGGIORNO E DELLA
TUTELA AMBIENTALE A VILLASIMIUS
Carlo Perelli
Giovanni Sistu
Andrea Zara
QUADERNI DI LAVORO
2011/01
TEMI ECONOMICI DELLA SARDEGNA
!"#!$
CENTRO RICERCHE ECONOMICHE NORD SUD
(CRENOS)
UNIVERSITÀ DI CAGLIARI
UNIVERSITÀ DI SASSARI
CRENOS was set up in 1993 with the purpose of organising the joint research
effort of economists from the two Sardinian universities (Cagliari and Sassari)
investigating dualism at the international and regional level. CRENoS’ primary
aim is to improve knowledge on the economic gap between areas and to provide
useful information for policy intervention. Particular attention is paid to the
role of institutions, technological progress and diffusion of innovation in the
process of convergence or divergence between economic areas. To carry out its
research, CRENoS collaborates with research centres and universities at both
national and international level. The centre is also active in the field of
scientific dissemination, organizing conferences and workshops along with other
activities such as seminars and summer schools.
CRENoS creates and manages several databases of various socio-economic
variables on Italy and Sardinia. At the local level, CRENoS promotes and
participates to projects impacting on the most relevant issues in the Sardinian
economy, such as tourism, environment, transports and macroeconomic
forecasts.
www.crenos.it
info@crenos.it
CRENOS – C AGLIARI
VIA SAN GIORGIO 12, I-09100 CAGLIARI, ITALIA
TEL. +39-070-6756406; FAX +39-070- 6756402
CRENOS - S ASSARI
VIA TORRE TONDA 34, I-07100 SASSARI, ITALIA
TEL. +39-079-2017301; FAX +39-079-2017312
Title: THE EFFECT OF SOCIO-ECONOMIC AND EMOTIONAL FACTORS ON GAMBLING BEHAVIOUR
ISBN: 978 88 84 67 817 1
First Edition: March 2013
© CUEC 2013
V i a I s M i r r i o n i s , 1
09123 C a g l i a r i
T e l . / F a x 070 291201
w w w . c u e c . i t
1"
"
The effect of socio-economic and emotional factors
on gambling behaviour
Anna Bussua c, Claudio Detottob
a Department of Political Science, Communication Sciences and Information
Engineering, University of Sassari,
b Department of Economics and Business (DiSEA) and CRENoS,
University of Sassari
Abstract
Gambling represents a channel through which some relevant aspects of our social
life, such as audacity, competition and risk, manifest themselves. Gambling is both a
pleasing diversion and a way of socialisation, where gratification and problematic
issues alternate. Most gamblers are social players who participate in games without
any relevant implications on their life, regardless of how frequently they engage in
the activity. Unfortunately, in some cases gaming activities can have a dramatic
impact on the player to the point that he/she has little control over them. In such
cases, the approach to gaming can be defined as critical or even pathological.
Pathological gambling is a serious form of addiction that causes gamblers to suffer
from social and financial problems as they constantly look for ways to increase their
“dose”.
This study proposes a bivariate ordered probit approach aimed at examining the
emotional factors of gambling expenditures and problematic behaviour or addiction
while also controlling for socio-economic determinants. It is based on a survey
among 1,315 gamblers in Sardinia (Italy) in the time span from June 2004 to March
2005. To measure gambling-related problems and gaming addiction we use survey
responses on the existence of problems caused by game participation (in terms of
psychological, relational, economic, labour difficulties directly linked to gambling)
and on the need for help and/or the intention to stop the gambling experience.
The findings show that women bet less than men and that income and gambling
frequency are positively correlated with the amount of money allocated to gambling.
Furthermore, having a sense of omnipotence and being willing to replay in case of a
win increase the propensity to bet more money. Notably, women have a higher
probability to be problematic gamblers after controlling for all other characteristics.
Income is negatively associated with problematic gamblers while those who
experience guilt or frustration after a loss and bet a higher amount of money have a
higher probability of exhibiting gambling-related problems.
Those who have other players in their family (wife/husband, children,
brother/sister, parents and grandparents), do not play alone and gamble for many
hours a day have a higher probability to become pathological gamblers. In addition,
income positively affects the probability to have pathological consequences while
education is negatively correlated to it. Finally, experiencing satisfaction in case of a
2"
"
win, disappointment in case of loss and excitement in the middle of the game is
negatively associated with pathological players.
c Corresponding author: email: abussu@uniss.it - Department of Political
Science, Communication Sciences and Information Engineering,
University of Sassari, Piazza Università 11, Sassari (I-07100
Jel Codes: C35; D01; D81; D87
Psycinfo Codes: 3233
Keywords: problem gambling; risk factors; emotional factors; gambling
behaviour
3"
"
1. Introduction
Gambling refers to any activity involving a bet and whose final
outcome is essentially influenced by aleatory elements and, in some
cases, by a gambler’s skill. The interest in carrying out a research on
gambling activities arises from the fact that it involves large sections of
the adult population in many social aspects such as competition,
boldness, risk-propensity and consumption choices (Zuckerman, 1983;"
Zuckerman and Khulman, 2000).
The gambling industry is very large all around the world and it
accounts for several billions of dollars. In the United States one-third of
all adults regularly participate in games (Narayanan and Manchanda,
2012) and, according to “H2 Gambling Capital” (2012), a consultancy
based in London, the country exhibits the highest gross win in the world
($80.45 billions) while Italy, with an average per capita loss of $517 and
gaming revenues of $19.05 billion, is the leading country in Europe and
the 4th in the world. It is not a surprise that some of the European
countries that suffer the most from the economic crisis are at the top of
this ranking, namely Spain, Greece, Italy and Ireland. As a matter of fact,
economic literature have highlighted the role of business cycles on
gambling and how during economic recessions lotteries are seen as a way
to increase disposable income, especially among people on a low-wage1.
Such phenomenon is particularly evident in urban areas due to the higher
supply of games and the higher presence of potential clients (Imbucci
1997, 1999; Sarti and Triventi, 2012). In fact, the main incentive for
participation in a lottery is the disproportion between the very low cost
of the ticket and the potentially very high winning prizes, which makes it
accessible also to risk-adverse and low-income consumers. However,
since lotteries and games are an easy way to collect resources, many
states tend to incentivize such activities during recessions in order to
finance public needs. "
It is worth noticing that gaming activity has not just an
economic motivation but it is also a pleasant and compensatory
diversion and an interesting chance for people to come together,
socialise and share special moments (Conlisk, 1993; Chantal, 2001).
"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
1 For many years gambling has been considered a recession-proof business but it
does not seem to be the case in last economic crisis, probably due to industry
saturation in many developed countries (Tripoli, 2009)."
4"
"
Unfortunately, an inappropriate relationship with it, characterized by a
lack of self-control on the player’s part, can result in serious
consequences relating to psychological, relational and economic aspects
of the player’s life (Zuckerman, 1999; Zuckerman and Khulman, 2000).
The problems associated with gambling activity occur when it unleashes
unstoppable and uncontrollable impulses that, in the long run, may
dramatically affect the gambler's life. In such cases, we speak of
dependency or, more accurately, of addiction. While the first term refers
to medical issues and is more closely associated with the physiological
consequences of substance taking, the concept of “addiction” refers to
the psychological dimensions that drive the individual towards the
desired object, producing behavioural effects such as compulsion and
loss of control (Patrizi and Bussu, 2008). In this sense, pathological
gambling can be defined as an “addiction without a substance” because
of the uncontainable impulse that leads the individual to engage in an
activity that is in itself rewarding but progressively affects both the
subjective capacity to manage the behaviour of game and other spheres
of activity (Langer, 1975; Langer and Roth, 1975).
As in all other types of addiction, the gambler may increase
his/her "dose of gambling", betting even more in order to achieve the
same level of excitement (tolerance) and/or recover his/her losses.
Notably, a strong state of anxiety is experienced when he/she is unable
to play or has decided to stop playing (abstinence) (Custer, 1982). In
fact, when gambling activity becomes predominant, gamblers could, with
a lot of time and effort, conceive, a (kind of) parallel life that is often
deliberately hidden to others.
According to Hulen and Burns (1998), gamblers can be
classified in several ways, including by psychological needs, motivation,
addiction degree, and typology of game. For example, players classified
by psychological needs are of two types: action gamblers or escape gamblers.
The former prefer active games in which the skill and strategy of the
player are crucial, such as cards, sports betting etc.; the latter, especially
when they are women, participate just to escape from stressful situations
or events and select games where the luck component prevails (Hulen
and Burns, 1998).
The majority of gamblers fall under the category of social
gamblers, who can play regularly or occasionally for fun or socialising
without losing control over their own actions, thus avoiding any negative
consequences. In contrast problematic gamblers suffer from psychological,
relational and affective disorders along with economic problems due to
5"
"
their little control over the game (Dickerson, 1984). When the player
approaches the stage of despair and does not have any self-control over
gambling activity, he/she can be considered a pathological gambler (Custer,
1982; Dickerson, 1984). According to Steel and Blaszczynski (1996), on
the one hand, pathological gamblers affected by impulsivity and
antisocial personality are more at risk of experiencing negative
consequences as a result of their behaviour; on the other hand, those
who develop significant problems of gambling addiction could become
more impulsive and antisocial in response to attempts to deal with their
plight.
Notably, people invest money in the game not only to increase
their expected income but also to obtain intangible benefits, such as
entertainment, or other intangible goods, such as excitement and
enjoyment (Blanco et al. 1996; Loba et al., 2001). Undeniably, among the
various needs that gambling is able to satisfy, the search for pleasure is
perhaps the most obvious and at the same time the most complex to
analyse (for a detailed literature review see Johansson et al., 2009). Many
players gamble in order to get the instant gratification associated with the
risk of the bet (Zuckerman, 1983). For such reason, many people will
prefer to participate in games that give a higher level of excitement, such
as poker, horse racing, sports betting, etc.
The emotional elements are not the sole factors influencing
gambling activities. In fact, socio-economic and demographic factors
affect individual preferences and risk-aversion and, consequently, a
player’s attitude to gambling. For example, Mikesell (1991) and Eaton
(2000) show that income and gambling expenditures are positively
correlated, although the share of per capita spending on gambling
decreases as income becomes higher. According to Sawkins and Dickie
(2002) and Worthington et al. (2007), age positively affects the
propensity to gamble among American and Australian players,
respectively, while a negative correlation between age and gambling is
found by Scott and Garen (1994), Niffenegger and Muuka (2001) and
Welte et al (2004). Interestingly, Mikesell (1991) shows that betting
expenditure increases with age, although there is a turning point at
averagely 44 years old.
Other empirical studies find that different ethnical groups are
associated with different attitudes to gambling activities (Clotfelter and
Cook, 1987; Livernois 1987; Scott and Garen, 1994; Stranahan and Borg
1998a, 1998b; Liu, 2006; Welte et al., 2004; Tan et al., 2010). Education
has a significant and negative impact on game consumption, which
6"
"
means that more educated people are more risk adverse (Scott and
Garen, 1994; Stranahan and Borg, 1998a, 1998b). Niffenegger and
Muuka (2001) empirically observe that people with tertiary education, on
average, spend more in lotteries than other educated groups.
As highlighted by the abovementioned literature, game
expenditures and gambler typology are affected by individual
characteristics. Unfortunately, many empirical analyses are limited by the
presence of latent variables that could lead to biased estimates and
misleading inference. The aim of this study is to overcome such
empirical limits by examining the determinants of the spending
behaviour of a sample of gamblers and their status of
dependence/problems via a bivariate ordered probit model. Precisely,
this econometric model is a system of equations, which can overcome
the latent variables problem by controlling for potential endogenous
variables and correlations between the residuals of each equation.
The study uses a sample of 1,315 gamblers collected through
“face to face” questionnaires in Sardinia (Italy) between June 2004 and
March 2005 in various typical gambling venues, such as game rooms,
bingo rooms and sport betting shops, and it takes into account several
types of game, such as lotteries, video poker and casino games. The
questionnaire draws inspiration from the South Oaks Gambling Screen
(SOGS) (Lesieur and Blume, 1987). Italy is an important case study not
only because of its very high per capita expenditure in gaming activity
compared to other developed countries, such as the US, the UK,
Germany and Japan, but also due to a recent state intervention that will
expand the gaming supply dramatically (Hooper, 2012) and might
further increase Italians’ propensity to gamble. Moreover, the analysis is
focused on Sardinia – an island of 1.6 millions inhabitants – which
reduces any problems of heterogeneity that may arise from cultural
differences among Italian regions.
An important issue concerns the measuring of gambling
addiction and gambling-related problems. Since it is impossible to
measure such phenomena objectively, we use the subjective responses of
the interviewed. Precisely, in our study the problematic gambler is
identified by a set of items investigating the presence of psychological,
economic, relational, labour, emotional and sexual problems directly
related to game activities, while the pathological gambler is associated
not only with gaming problems but also with those respondents who
7"
"
state their need for help and/or their intention to stop the gambling
experience. In order to limit the well-known problems2 of
underestimation – due to the gamblers’ reluctance to manifest their
condition – and overestimation – owing to some people trying to get
attention – different items have been compared in order to verify the
robustness of the results.
Notably, the need for help in the case of pathological gambling
behaviour is not always related to the gamblers’ awareness of being
suffering from an addiction but rather to the impact that critical and
stressful situations and emotional components have in their lives.
Unfortunately, the most important barriers preventing the gamblers
from stopping their betting activity and asking for counselling are
shame, denial and social factors and not a lack of information or trust in
local support agencies (Evans and Delfabbro, 2005).
The individual factors under study refer to a broad range of
characteristics: age, income, education, family status, presence of other
gamblers in the respondent’s family, attitude to playing alone and so on.
Furthermore, a set of emotional indicators are considered in order to
estimate the effects of the emotions felt during the game and after a
win/loss on gambling expenditures and the probability of the player
being a social, problematic or pathological gambler.
The paper is organised as follows. Section 2 describes the
econometric approach and the dataset in detail. The results of the paper
are presented in Section 3. Finally, Section 4 concludes the paper.
2. Empirical approach
Following the empirical literature on gambling behaviour
(Delfabbro and Thrupp, 2003; Worthington et al., 2007; Tan et al.,
2010), this study proposes the bivariate ordered probit model illustrated
below to explore the impact of socio-economic and emotional factors
on gambling expenditures and gambling-related problems or addiction.
To be precise, by using a survey of 1,315 gamblers in Sardinia from June
2004 to March 2005, the following econometric model is estimated:
"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
2 See Johnson et al. (1998) for a detailed review of the literature."
8"
"
𝑦!!=𝑋!!𝛽!+𝑒!!!!(1𝑎)
𝑦!!=𝑦!!𝛾+𝑋!!𝛽!+𝑒!!!!(1𝑏)
for i = 1, 2, …, n. Model (1a)-(1b) constitutes a system of
equations (Sajaia, 2012), where y1i, BET, is associated to four stated
gambling expenditure classes (1 for bets of less than 10 euros, 2 for bets
between 11 and 50 euros, 3 for bets between 50 and 300 euros, 4 for
bets higher than 300 euros), while y2i represents two binary variables: 1)
the respondents’ subjective social representations of problems caused by
gambling activities (PROBLEMS); 2) the need for help and/or the
intention to stop the gambling experience (PATHOLOGY). In this
sense, two different systems are regressed by using one indicator at a
time. Notably, 21.2% of the sample claimed to be engaging in gambling
activities due to relationship problems with family and friends,
economic problems, sexual difficulties and psychological stress, while
12.5% of respondents declared to also have asked for help and/or to
have had the intention to stop the gambling experience. BET indicates
the daily amount of money spent on gambling: 56.3% of the respondents
bet less than 10 euros, 27.44% between 11 and 50 euros, 14.09%
between 50 and 300 euros, 2.2% more than 300 euros.
With this model it is possible to have the expected joint
dependence of both dependent variables. To be precise, the endogenous
y1i is simultaneously determined with y2i. Hence, X1 and X2 are matrices
of observables, β1 and β2 are vectors of parameters, γ is a scalar
representing the effect of y1i on y2i. Finally, e1 and e2 are two error terms,
assumed to be jointly normal with correlation coefficient ρ and
uncorrelated with the explanatory variables, i.e. E(Xi,e1i) = 0 and E(Xi,e2i)
= 0. The parameters in the system of equations (1a)-(1b) are identified
only if X1 ≠ X2, i.e. at least one column of X1 should not be present in
X2 (Sajaia, 2012). Notably, the simultaneous bivariate ordered probit
model expressed in equations (1a)-(1b) is simplified in a seemingly
unrelated specification when γ = 0, as shown in the following system:
𝑦!!=𝑋!!𝛽!+𝑒!!!!(2𝑎)
𝑦!!=𝑋!!𝛽!+𝑒!!!!(2𝑏)
In the above specification, y1i has no effect on y2i but the system
takes into account the correlation between the two error terms, e1 and e2,
increasing the efficiency of the estimates β1 and β2 (Greene, 2003).
9"
"
According to the statistical test on the parameter γ, specification (1a)-
(1b) or (2a)-(2b) is considered.
The matrices Xi includes the following socio-economic and
behavioural factors. AGEi represents the age of the i-th respondent,
while FEMALEi is a dummy variable that has a value of one if the
player is female. SINGLE, DIVORCED, WIDOWER and MARRIED
are dummies that indicate the family status of the gamblers.
DISTANCE measures the distance in kilometres between the
respondents’ residences and their habitual gambling places. The expected
sign is not obvious. On the one hand, since the higher are the distances
the higher is the cost of transfers, we might expect long distances to be
associated with low bets due to budget constraints. On the other hand, a
positive relationship between distances and bets could be also expected
because players can reduce the number of transfers and increase the
amount of money they play as the distance increases.
EDUCATION and INCOME indicate individual education and
income level, respectively. According to Winters et al. (1993) and
Ladouceur et al. (1999), a negative relationship exists between education
and the risk of pathological gambling. A positive correlation between
INCOME and gambling expenditure is expected since higher income
level can be associated with the allocation of higher amount of money to
gambling consumption.
TIME_EXPERIENCE measures the number of years passed
since the first gambling experience. As gambling expenditures exhibit a
positive trend, a positive sign is expected (Bolen and Boyd, 1968).
ALONE is a dummy that has a value of one if the gambler usually plays
alone.
GENETIC and FAMILY are two dummies that have a value of
one if the parents and grandparents were gamblers and if other members
of the family (wife/husband and children) gamble regularly. This way, we
can control for genetic and/or emulative behaviour of gamblers. In both
cases a positive relationship is expected (Gupta and Derevensky, 1997;
Bergh et al., 1997).
WEEK_TIMES and N_HOURS indicate the number of times
per week and the number of hours per day in which respondents gamble.
Since pathological gamblers tend to increase their “dose” over time, a
positive sign is expected for both variables.
The last set of variables includes emotional dummies that can
affect gamblers’ behaviour. To be precise, the feelings and psychological
processes of gamblers during their gambling activities are collected
10"
"
through survey questionnaires. The information is divided in three
groups according to whether such feelings were experienced during the
game, in case of win or in case of loss. Each respondent can indicate up
to three choices per group.
The first set of dummies represents the feelings in case of win:
WIN_EUPHORIA, WIN_PLEASURE, WIN_SATISFACTION,
WIN_REPLAY (it has a value of one when the gambler feels the desire
to replay immediately), WIN_OMNIPOTENCE (it equals one when the
gambler feels a sense of omnipotence) and WIN_OTHERS. The second
group of covariates represents the feelings in case of loss:
LOSS_GUILT, LOSS_FRUSTRATION, LOSS_DISAPPOINTMENT,
LOSS_EXCITEMENT, LOSS_ANGER, LOSS_REDEEM,
LOSS_LOW_SELF_ESTEEM and LOSS_HELPLESSNESS. Finally,
the third set accounts for the gamblers’ emotions during the game:
DURING_EUPHORIA, DURING_PLEASURE,
DURING_SATISFACTION, DURING_ANXIETY,
DURING_EXCITEMENT, DURING_FRUSTRATION,
DURING_ANGER and DURING_OTHERS. Table 1 shows a short
description of all these variables.
All the abovementioned variables are included in both matrices
X1 and X2, except for DISTANCE, which is included in X1 and not in
X2 as it only affects gambling consumption. Table 2 summarises the
descriptive statistics of all the variables in use.
[TABLE 1 HERE]
[TABLE 2 HERE]
3. Results and discussion
In a first stage, equations (1a)-(1b) are estimated by using a two-
stage approach. The simultaneous specification is tested through the
statistical test on the γ parameter. If it is statistically different from zero,
the simultaneous specification of (1a)-(1b) is displayed, otherwise the
seemingly unrelated specification of (2a)-(2b) is regressed and presented.
Table 3 shows the results of the bivariate order probit using
both gambling expenditures and the respondents’ subjective social
representations of gambling-related problems as dependent variables.
Looking at the bottom of the columns (1) and (2) of Model (1), the
significance of the ρ statistic (=-0.624; p-value < 0.10) and the
11"
"
Likelihood Ratio (LR) test of independent equations (=12.63; p-value <
0.01) provide evidence that the residuals in the two equations are
correlated, thus justifying the use of a bivariate model. Furthermore, the
γ statistic (=0.824; p-value < 0.01) is highly significant, which indicates
an endogenous relationship between the dependent variables. As one can
expect, γ is positive, i.e. the higher is the average bet the more likely it is
that problems will arise. In other words, by using a bivariate ordered
probit model it is possible to control for the effect of the amount of
money played by gamblers on their subjective statement through the
parameter ρ, the presence of latent variables not included into the
analysis and through the parameter γ, which could affect both dependent
variables. Furthermore, all the cut-offs (/CUT11, /CUT12, /CUT13 and
/CUT21) statistical tests and the Wald test on the joint significance of
the coefficients are quite beyond the critical value (at 99% level
confidence). They can be considered as an index of goodness of fitting
of the model.
Notably, age and family status have no effects on the amount of
money played by the players. On average, women bet less money than
men while income is positively related with bet values. Such a result is
quite intuitive since the higher the disposable income, as in the case of
male and wealthy individuals, the higher the amount of money allocated
for gambles. Furthermore, Wärneryd (1996), Sawkins and Dickie (2002)
and Welte et al. (2004) show that women are more risk averse than men,
which leads to a reduction in gambling participation. The non-linearity
of the income effect is also tested by including a quadratic
transformation of INCOME into the model (see column (1) of Model
(2)) but its coefficient is not statistically different from zero.
Interestingly, DISTANCE, WEEK_TIMES and N_HOURS
increase the probability of gamblers betting higher values, which
indicates that bets rise as the distance travelled and the weekly and daily
frequency increase. Such findings empirically confirm the effects of
problematic gambling showing that gamblers constantly look for ways to
increase their “dose” both in terms of time and money allocated for
gambling activities. As indicated in Model (2), two interaction variables,
namely “WEEK_TIMES×DISTANCE” and “N_HOURS×
DISTANCE”, are inserted among the regressors in order to check for
the presence of a trade-off between the distance travelled and the
frequency of gambling activities. The rationale is that one could reduce
the time frequency of gambling activity as the distance to reach the
place increases in order to minimise travel costs. According to the
12"
"
statistical test on the significance of the coefficients in column (1) of
Model (2), no trade-offs seem to appear.
In relation to the emotional factors that might have an effect
during or at the end of the play, two indicators, namely WIN_REPLAY
and WIN_OMNIPOTENCE, are statistically significant. To be precise,
those gamblers who exhibit the willing to replay or a sense of
omnipotence after a win tend to have a higher probability to bet more
than other players. In such a model, no effect seems to derive from the
emotions felt during the gambling activities or after a loss. In this sense,
emotions associated with positive events (win) seem to play a relevant
role in explaining the amount of money allocated for gambling.
Column (3) and (4) of Model (1) show the effects of regressors
on the respondents’ subjective assessment of having health, wealth,
affective or relational problems directly stemming from their game
participation. As before, the variables FEMALE and INCOME are
statistically significant. Interestingly, women and low-income people are
more likely to state their problems. A possible explanation of these last
findings is that individuals with lower income might be more easily
affected by economic problems due to more binding budget constraints.
Those players who experience guilt (LOSS_GUILT) and
frustration (LOSS_FRUSTATION) when they lose are more likely to be
problematic gamblers. Hence, one can say that negative feelings arising
from a loss can be a positive factor as players who experience them are
more likely to recognise their condition and problems. In other words,
the sense of guilt and frustration could be used as an indicator allowing
specialists to identify gamblers at risk of problematic condition.
Table (4) shows the results of the seemingly unrelated bivariate
ordered probit regression3 using the four bet classes (BET) and the
subjective statement of need for help and/or the intention to stop the
gambling experience along with the problematic condition
(PATHOLOGY). The variable PATHOLOGY is more binding than the
one used before (PROBLEMS) because players are aware not only of the
problems arising from gambling activity but also of the advantages they
could obtain by stopping it. In this sense, admitting their need for help
"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
3 In a preliminary analysis, the presence of the endogenous term γ has been
tested and the null hypothesis of absence of endogeneity cannot be rejected.
Hence, Model (2a)-(2b) is regressed and presented."
13"
"
or the intention to stop gambling means that the benefits, both in
emotional and monetary terms, associated with such games have become
less relevant than the costs they entail.
In the bottom of Column (1)-(2)-(3)-(4) of Model (1) diagnostic
tests are presented. All the cut-offs (/CUT11, /CUT12, /CUT13 and
/CUT21) statistical tests and the Wald test on the joint significance of
the coefficients are quite beyond the critical value (at 99% level
confidence), which indicates a more than satisfactory goodness of fitting
of the model. The ρ parameter statistical test (=0.371; p-value < 0.01)
and the LR test of independent equations (=13.69; p-value < 0.01) are
also quite beyond the critical values.
Columns (1)-(2) of Model (1) represent the estimates of the
factors that can affect bet choices. As before, FEMALE, INCOME,
DISTANCE, WEEK_TIMES, N_HOURS, WIN_REPLAY and
WIN_OMNIPOTENCE are significant while the quadratic term of
INCOME and the interaction variables “WEEK_TIMES×
DISTANCE” and “N_HOURS×DISTANCE” in Column (1)-(2) of
Model (2) are not statistically different from zero.
In Columns (3)-(4) of Model (1), the results of the second
equation of the system are shown. EDUCATION, ORIGIN, FAMILY
and N_HOURS are positively related with the probability of becoming a
pathological gambler while ALONE and INCOME are negatively
correlated to it. Notably, a high gaming frequency and the presence of
other players in the family (wife/husband, children, brother/sister,
parents and grandparents) increase the willingness to stop gambling
activity. Furthermore, those gamblers who bet alone are less likely to
admit that they need help or that they have the intention to stop
gambling.
The feelings in case of win/loss and during gambling affect the
probability of respondents being suffering from malaise. Precisely, the
self-reported emotions of satisfaction in case of win
(WIN_SATISFACTION), disappointment
(LOSS_DISAPPOINTMENT) in case of loss and excitement during the
game (DURING_EXCITEMENT) decrease the probability of
respondents being suffering from a pathologic condition. Hence, such
emotions lead to a reduction in the benefits arising from the game and
make gamblers desire to stop playing or ask for help in this matter.
14"
"
4. Conclusions
The relationship between socio-economic variables and
gambling behaviour has been widely analysed by scholars, while, to our
knowledge, the impact of psychological and emotional factors have not
been fully explored, despite their importance in theoretical literature. In
fact, the emotions felt at the end, both in case of win or loss, and during
the game affect individual utility, since they increase costs and benefits
associated with game output, and, consequently, the probability to switch
from a social gambler typology to a problematic or a pathological one.
Gambling can be explained with the need of sensation seeking where the
positive reinforcement is linked to the anticipatory arousal felt during
the game (Zuckerman, 1979), depending on the player and on game
typologies (Coventry and Brown, 1993; Le Breton, 1995).
Understanding the risks of gambling and investigating how to
maintain a proper relationship with games represent the first step in
providing social policies that effectively contain gaming problems.
Gaming is not negative in itself since it reflects some relevant aspects of
our social life, such as audacity, competition and risk. Hence, promoting
prevention campaigns and providing psychological interventions is
necessary not only in the presence of gambling addiction or problems
but also in case of recreational gambling activities as it incentivizes
responsible approaches to gaming.
By employing a bivariate ordered probit approach, this paper
aims to examine both the socio-economic and the emotional
determinants of game behaviour, in terms of expenditures and
probability for a player to become a problematic or pathological
gambler, among a survey of 1,315 players in Sardinia (Italy). The self-
reported information about the existence of problems due to game
participation (in terms of economic, psychological, labour difficulties
directly linked to gambling, etc.) and also the need for help and/or the
intention to stop the gambling experience are taken as measures of
gambling-related problems and gambling addiction, respectively.
The findings show that being male, gambling with high
frequency, having a sense of omnipotence and being willing to replay in
case of a win are positively associated with a higher average gaming
consumption. Such findings are perfectly in line with recent literature
(Breiter et al., 2001; Kuhnen and Knutson, 2005; Lee and et., 2007)
indicating that positively aroused feelings may incentivize risk taking.
Female players show, ceteris paribus, a higher probability of being
problematic gamblers. Income is negatively associated with problematic
15"
"
gambling while those who experience guilt and frustration after a loss
and bet a higher amount of money have a higher probability of
exhibiting gambling-related problems.
Those who have other players in their family (wife/husband,
children, brother/sister, parents and grandparents), do not play alone
and gamble for many hours a day have a higher probability of being
pathological gamblers. The familial gambling is a relevant aspect as far
as problematic gamblers are concerned. In this regard, by employing a
survey analysis in Montreal, Gupta and Derevensky (1997) find that
"86% of children who gamble regularly reported gambling with family
members".
Income positively affects the probability of having pathological
consequences while education is negatively correlated with them. Finally,
experiencing satisfaction in case of a win, disappointment in case of a
loss and excitement in the middle of the game is negatively associated
with pathological players.
However, when interpreting these results, one should be aware
of some caveats. Firstly, although self-related data offers many
information about emotions and health of respondents, some problems
could arise from underestimation and overestimation of
problematic/pathological gamblers rates, which might reduce the
explanative power of our models. To avoid such bias, different items
have been compared in order to verify the robustness of respondents’
answers. Secondly, the results of the present study may have limited
generalizability since it focuses only on a specific area of Italy but, at the
same time, heteroscedasticity problems arising from different cultural
factors among Italian regions are more likely to be avoided. Finally, as a
further step of this research and subject to data availability, the analysis
will be extended to the estimation of the risk factors – both emotional
and socio-economics – of gambling behaviour in different typologies of
game.
16"
"
References
Blanco, C., Orensanz-Munor, L., Blanco-Jerez, C., & Sàiz-Ruiz, J.
(1996). Pathological gambling and platelet MAO activity: A
psychobiological study. American Journal of Psychiatry, 153,
119–121.
Bergh, C., Eklund, T., Sodersten, P., & Nordin, C. (1997). Altered
dopamine function in pathological gambling. Psychological
Medicine, 27, 473–475.
Bolen, D. W., & Boyd, W. H. (1968). Gambling and the gambler: a
review and preliminary findings. Archives General Psychiatry,
18, 617-630.
Breiter, H.C., Aharon, I., Kahneman, D., Dale, A., & Shizgal, P.,
(2001). Functional imaging of neural responses to
expectancy and experience of monetary gains and losses.
Neuron, 30, 619–639.
Chantal, Y., Vallerand, R. J., & Vallières, E. F. (2001). Motivation
and gambling involvement. The Journal of Social Psychology,
135, 755-763.
Clotfelter, C. T., & Cook, P. J. (1987). Implicit taxation in lottery
finance. National Tax Journal, 40, 533–546.
Conlisk, J., (1993). The utility of gambling. Journal of Risk and
Uncertainty, 6, 255-275.
Coventry, K. R., & Brown, R. I. F. (1993). Sensation seeking,
gambling and gambling addictions. Addiction, 88, 541-554.
Custer, R. L. (1982). Pathological gambling. In Whitfield A. (Ed.),
Patients whit alcoholism and other drug problems. New
York: Year Book Publisher.
Delfabbro P., & Thrupp L. (2003) “The social determinants of
youth gambling in South Australian adolescents”, Journal of
Adolescents, 26, 313-330.
Dickerson, M. G. (1984). Compulsive gamblers. New York: Longman
Group Ltd.
Eaton, D.H. (2000) The Kentucky lottery. Research Article. College of
Business and Public Affairs, Murray State University,
Murray, KY. [WWW document]. URL
17"
"
http://www.murraystate.edu/qacd/cbpa/bber/articles/eat
on.pdf (accessed on 11 March 2013).
Evans, L., Delfabbro, P. H. (2005). Motivators for Change and
Barriers to Help-Seeking in Australian Problem Gamblers,
Journal of Gambling Studies, 21, 2, 133-155.
Greene, W. H. (2003). Econometric Analysis. (5th Edition). New
Jersey: New York University.
Gupta, R., & Derevensky, J. (1997). Familial and social influences
on juvenile gambling behavior. Journal of Gambling Studies,
13, 179-192.
H2 Gambling Capital (2012). Leading Global Gambling Nations –
Asia and eGaming Continue to Out Perform. [WWW
document]. URL http://www.h2gc.com/article/leading-
global-gambling-nations--asia-and-egaming-continue-to-
out-perform (accessed on 11 March 2013).
Hooper, J. (2012). Gambling addiction takes its toll in Italy. The
Guardian. [WWW document]. URL
http://www.guardian.co.uk/world/2012/dec/25/gamblin
g-addiction-takes-toll-taly (accessed on 11 March 2013).
Hulen, D., & Burns, P. (1998). Differences in Pathological Gamblers in
Arizona. Arizona Council on Compulsive Gambling.
[WWW document]. URL
http://azccg.org/Action_or_Escape.html (accessed on 11
March 2013).
Imbucci, G. (1997). Il gioco. Lotto, totocalcio, lotterie. Storia dei
comportamenti sociali. Venezia: Marsilio.
Imbucci, G. (1999). Il gioco pubblico in Italia. Storia, cultura e mercato.
Venezia: Marsilio.
Johansson, A., Grant, J. E., Kim, S. W., Odlaug, B. L., Götestam,
K. G. (2009). Risk Factors for Problematic Gambling: A
Critical Literature Review. Journal of Gambling Studies, 25, 67-
92.
Johnson, E. E., Hamer, R. M., & Nora, R. M. (1998). The lie/bet
questionnaire for screening pathological gamblers: A
follow-up study. Psychological Reports, 83, 1219-1224.
Kuhnen, C. M., & Knutson, B. (2005). The Neural Basis of
Financial Risk Taking. Neuron, 47, 763-770.
18"
"
Ladouceur, R., Boudreault, N., Jacques, C., & Vitaro, F. (1999).
Pathological gambling and related problems among
adolescents. Journal of Child & Adolescent Substance Abuse, 8,
55–68.
Langer, E. J. (1975). The illusion of control. Journal of Personality and
Social Psychology, 32, 311-328.
Langer, E.J., & Roth, J. (1975). Heads I win, tails it's chance: The
illusion of control as a function of the sequence of
outcomes in a purely chance task. Journal of Personality and
Social Psychology, 32, 951-955.
Le Breton, D. (1995). La passione del rischio. Torino: Gruppo Abele.
Lesieur, H., & Blume, S. B. (1987). The South Oaks Gambling
Screen (The SOGS): a new instrument for the
identification of pathological gamblers. American Journal of
Psychiatry, 144, 1184-1188.
Livernois, J. R. (1987). The redistributive effect of lotteries:
evidence from Canada. Public Finance Quarterly, 15, 339–
351.
Lee, H. P., Chae, P. K., Lee, H. S., & Kim, Y. K. (2007). The five-
factor gambling motivation model. Psychiatry Research, 150,
21–32
Liu, L. (2006). Casinos winning big by betting on Asians.
Washington Post. The Associated Press.
Loba, P., Stewart, S. H., Klein, R. M., & Blackburn, J. R. (2001).
Manipulations of the features of standard Video Lottery
Terminal (VLT) games: Effects in pathological and non-
pathological gamblers. Journal of Gambling Studies, 17, 297–
320.
Mikesell, J. L. (1991). Lottery expenditure in a non-lottery state.
Journal of Gambling Studies, 7, 89–98.
Narayanan, S., & Manchanda, P. (2012). An empirical analysis of
individual level casino gambling behavior. Quantitative
Marketing and Economics, 10, 27-62.
Niffenegger, P., & Muuka, G. N. (2001). To gamble or not to
gamble? An empirical survey of the gambling preferences
of residents of Marshall County, Kentucky. Journal of
Business and Public Affairs, 28, 94–102.
19"
"
Patrizi, P., & Bussu, A. (2008). Il giocatore d’azzardo: la narrazione
tra socialità e problematicità. Rassegna di Psicologia, 3, 59-83.
Sajaia, Z. (2012) “Maximum likelihood estimation of a bivariate
ordered probit model: implementation and Monte Carlo
simulations”, The Stata Journal, forthcoming.
Sarti, S., & Triventi, M. (2012). Il gioco d'azzardo: l'iniquità di una
tassa volontaria. La relazione tra posizione socio-
economica e propensione al gioco. Stato e Mercato, 96, 503-
533.
Sawkins, J. W., & Dickie, V.A. (2002). National lottery
participation and expenditure: preliminary results using a
two stage modelling approach. Applied Economics Letters, 9,
769–773.
Scott, F., & Garen, J. (1994). Probability of purchase, amount of
purchase, and the demographic incidence of the lottery tax.
Journal of Public Economics, 54, 121–143.
Steel, Z., & Blaszczynski, A. (1996). The factorial structure of
pathological gambling. Journal of Gambling Studies, 12, 3–20.
Stranahan, H. A., & Borg, M. O. (1998a). Separating the decisions
of lottery expenditures and participation: a truncated Tobit
approach. Public Finance Review, 26, 99–117.
Stranahan, H. A., & Borg, M. O. (1998b). Horizontal equity
implications of the lottery tax. National Tax Journal, 51, 71–
82.
Tan, A. K. G., Yen, S. T., & Nayga Jr., R. M. (2010). Socio-
demographic determinants of gambling participation and
expenditures: evidence from Malaysia. International Journal of
Consumer Studies, 34, 316–325.
Tripoli, L. (2009). Wasn’t the gaming law business supposed to be
recession-proof? Gaming Law Review and Economics, 13, 23-
26.
Zuckerman, M. (1979). Sensation seeking: beyond the optimal level of
arousal. New Jersey: L. Erlbaum Associates.
Zuckerman, M. (1999). Vulnerability to Psychopathology: A biosocial
Model. (1st ed.). Washington: American Psychiatric
Association.
20"
"
Zuckerman, M., & Khulman, D. M. (2000). Personality and risk
taking: Common biosocial factors. Journal of personality, 68,
999-1028.
Zuckerman, M. (1983). La ricerca di forti sensazioni, Psicologia
Contemporanea, 59, 26-32.
Wärneryd, K. E. (1996). Risk attitudes and risky behavior. Journal of
Economic Psychology, 17, 749-770.
Welte, J. W., Barnes, G. M., Wieczorek, W. F., Tidwell, M. C. O.,
& Parker, J. C. (2004). Risk factors for pathological
gambling. Addictive Behaviors, 29, 323–335.
Winters, K. C., Stinchfield, R., & Fulkerson, J. (1993). Patterns and
characteristics of adolescent gambling. Journal of Gambling
Studies, 9, 371–386.
Worthington, A.C., Brown, K., Crawford, M., & Pickernell, D.
(2007). Socioeconomic and demographic determinants of
household gambling in Australia. Review of Economics of the
Household, 5, 209–222.
21"
"
Table 1. Description of variables
Variable
Description
BET
A categorical variable that has a value of 1 for bets of
less than 10 euros; 2 for bets between 11 and 50
euros; 3 for bets between 50 and 300 euros; 4 for bets
higher than 300 euros.
PROBLEMS
A dummy variable that has a value of 1 if the
respondent states to be suffering from gambling-
related problems, namely economic, relational, labour,
emotional and sexual problems.
PATHOLOGY
A dummy variable that has a value of 1 if the
respondent states to be in need of help and/or
wanting to stop the gambling experience.
AGE
It represents the age of the respondent.
FEMALE
A dummy variable that has the value of 1 if the
respondent is female.
DISTANCE
A categorical variable that represents the distance
travelled to go to the gaming venue: it has a value of 1
for distances of less than 5 km; 2 for distances
between 6 and 25 km; 3 if distances are between 26
and 40 km; 4 if they are between 41 and 60 km; 5 if
they are between 60 and 100 km; 6 if they are higher
than 100 km.
SINGLE
A dummy variable that has the value of 1 if the
respondent is single.
DIVORCED
A dummy variable that has the value of 1 if the
respondent is divorced.
WIDOW/ER
A dummy variable that has the value of 1 if the
respondent is a widow/er.
MARRIED
A dummy variable that has the value of 1 if the
respondent is married.
EDUCATION
A categorical variable that, accordingly to the highest
education degree, has a value of 1 if the respondent is
illiterate or unschooled; 2 if he/she has a primary
school dipolma; 3 if he/she has a middle school
diploma; 4 if he/she has a secondary school diploma;
5 if he/she has a tertiary degree; 6 if he/she has a
post-graduate degree.
INCOME
A categorical variable that has a value of: 1 for
incomes of less than 10,000 euros; 2 for incomes
between 10,000 and 15,000 euros; 3 for incomes
between 15,000 and 20,000 euros; 4 for incomes
between 20,000 and 30,000 euros; 5 for incomes
between 30,000 and 40,000 euros; 6 for incomes
higher than 40,000 euros.
TIME_EXPERIENCE
A categorical variable that indicates how old the
respondent was when he/she gambled for the first
time. It has a value of 1 if he/she was less than 15
22"
"
years old; 2 if he/she was between 15 and 18 years old;
3 if he/she was between 18 and 25 years old; 4 if
he/she was between 26 and 30 years old; 5 if he/she
was between 31 and 45 years old; 6 if he/she was
between 46 and 60 years old; 7 if he/she was older
than 60.
GENETIC
A dummy variable that has a value of 1 if the
respondent’s father, mother or grandparents gamble.
FAMILY
A dummy variable that has a value of 1 if the
respondent’s family (wife/husband or his/her
children) gambles.
WEEK_TIMES
A categorical variable that indicates how many times
the respondent gambles in a week. It has a value of 1 if
he/she never gambles; 2 if he/she does it once; 3 if
twice; 4 if three times; 5 if more than 3 times.
N_HOURS
A categorical variable that indicates for how long the
respondent gambles in a day. It has a value of 1 if
he/she never gambles; 2 if he/she gambles for less
than 30 minutes; 3 if he/she gambles for 30 to 1 hour;
4 if he she gambles for 1 to 2 hours; 5 if he/she
gambles for 2 to 4 hours; 6 if he/she gambles for
more than 4 hours.
ALONE
A dummy variable that has a value of 1 if the
respondent usually plays alone.
WIN_EUPHORIA
A dummy variable that has a value of 1 if the
respondent is euphoric in case of win.
WIN_PLEASURE
A dummy variable that has a value of 1 if the
respondent feels pleasure in case of win.
WIN_SATISFACTION
A dummy variable that has a value of 1 if the
respondent feels satisfied in case of win.
WIN_REPLAY
A dummy variable that has a value of 1 if the
respondent feels an urge to try again in case of win.
WIN_OMNIPOTENCE
A dummy variable that has a value of 1 if respondent
feels a sense of omnipotence in case of win.
WIN_OTHERS
A dummy variable that has value of 1 if the
respondent feels a different emotion from the ones
listed before in case of win.
LOSS_GUILT
A dummy variable that has a value of 1 if the
respondent feels a sense of guilt in case of loss.
LOSS_FRUSTRATION
A dummy variable that has a value of 1 if the
respondent feels frustrated in case of loss.
LOSS_DISAPPOINTMENT
A dummy variable that has a value of 1 if the
respondent is disappointed in case of loss.
LOSS_EXCITEMENT
A dummy variable that has a value of 1 if the
respondent is excited in case of loss.
LOSS_ANGER
A dummy variable that has a value of 1 if the
respondent is angered in case of loss.
LOSS_REDEEM
A dummy variable that has a value of 1 if the
23"
"
respondent feels some need of redeeming
himself/herself in case of loss.
LOSS_LOW_SELF_ESTEEM
A dummy variable that has a value of 1 if the
respondent has low self-esteem in case of loss.
LOSS_HELPLESSNESS
A dummy variable that has a value of 1 if the
respondent feels helpless in case of loss.
DURING_EUPHORIA
A dummy variable that has a value of 1 if the
respondent is euphoric during the game.
DURING_PLEASURE
A dummy variable that has a value of 1 if the
respondent feels pleasure during the game.
DURING_SATISFACTION
A dummy variable that has a value of 1 if the
respondent feels satisfied during the game.
DURING_ANXIETY
A dummy variable that has a value of 1 if the
respondent is anxious during the game.
DURING_EXCITEMENT
A dummy variable that has a value of 1 if the
respondent is excited during the game.
DURING_FRUSTRATION
A dummy variable that has a value of 1 if the
respondent feels frustrated during the game.
DURING_ANGER
A dummy variable that has a value of 1 if the
respondent is angry during the game.
DURING_OTHERS
A dummy variable that has a value of 1 if the
respondent feels a different emotion from the ones
listed during the game.
24"
"
Table 2. Descriptive statistics
Variable
Obs
Mean
Std. Dev.
Min
Max
BET
1228
1.622
0.806
1
4
PROBLEMS
1259
0.212
0.409
0
1
PATHOLOGY
1315
0.125
0.331
0
1
AGE
1254
34.719
12.446
14
79
FEMALE
1300
0.226
0.418
0
1
DISTANCE
1248
1.514
1.022
1
6
SINGLE
1297
0.570
0.495
0
1
DIVORCED
1297
0.355
0.479
0
1
WIDOW/ER
1297
0.048
0.213
0
1
MARRIED
1297
0.027
0.162
0
1
EDUCATION
1301
4.108
1.216
1
6
INCOME
1069
2.727
1.478
1
6
TIME_EXPERIENCE
941
3.040
1.300
1
7
GENETIC
1315
0.030
0.170
0
1
FAMILY
1315
0.023
0.152
0
1
WEEK_TIMES
1255
3.245
1.205
1
5
N_HOURS
1229
2.944
1.399
1
6
ALONE
1235
0.481
0.500
0
1
WIN_EUPHORIA
1315
0.344
0.475
0
1
WIN_PLEASURE
1315
0.490
0.500
0
1
WIN_SATISFACTION
1315
0.487
0.500
0
1
WIN_REPLAY
1315
0.173
0.378
0
1
WIN_OMNIPOTENCE
1315
0.036
0.186
0
1
WIN_OTHERS
1315
0.049
0.215
0
1
LOSE_GUILT
1315
0.111
0.314
0
1
LOSE_FRUSTRATION
1315
0.062
0.242
0
1
LOSE_DISAPPOINTMENT
1315
0.576
0.494
0
1
LOSE_EXCITEMENT
1315
0.056
0.230
0
1
LOSE_ANGER
1315
0.254
0.435
0
1
LOSE_REDEEM
1315
0.234
0.424
0
1
LOSE_LOW_SELF_ESTEEM
1315
0.013
0.113
0
1
LOSE_HELPLESSNESS
1315
0.031
0.174
0
1
DURING_EUPHORIA
1315
0.270
0.444
0
1
DURING_PLEASURE
1315
0.449
0.497
0
1
DURING_SATISFACTION
1315
0.251
0.434
0
1
DURING_ANXIETY
1315
0.247
0.431
0
1
DURING_EXCITEMENT
1315
0.176
0.381
0
1
DURING_FRUSTRATION
1315
0.018
0.134
0
1
DURING_ANGER
1315
0.047
0.212
0
1
DURING_OTHERS
1315
0.104
0.306
0
1
25"
"
Table 3. Results of the simultaneous bivariate ordered probit
regression
Model (1)
Model (2)
Dependent
variable: BET
Dependent
variable:
PROBLEMS
Dependent variable:
BET
Dependent
variable:
PROBLEMS
Coeff.
Std.
Er.
Coeff.
Std.Er.
Coeff.
Std.Er.
Coeff.
Std.Err.
AGE
0.005
(0.005)
0.009
(0.008)
0.005
(0.005)
0.005
(0.008)
FEMALE
-0.239*
(0.136)
0.312*
(0.168)
-0.243*
(0.136)
0.316**
(0.154)
DISTANCE
0.118**
(0.050)
0.215
(0.146)
SINGLE
-0.379
(0.254)
-0.144
(0.382)
-0.392
(0.254)
0.021
(0.354)
DIVORCED
-0.279
(0.246)
-0.265
(0.378)
-0.288
(0.246)
-0.100
(0.354)
WIDOW/ER
-0.069
(0.311)
-0.423
(0.400)
-0.071
(0.312)
-0.302
(0.383)
EDUCATION
0.030
(0.037)
0.001
(0.049)
0.028
(0.037)
-0.012
(0.044)
INCOME
0.125***
(0.033)
-
0.130***
(0.046)
-0.017
(0.136)
-0.015
(0.153)
0.023
(0.021)
-0.020
(0.023)
TIME_EXPERIENCE
-0.051
(0.042)
-0.072
(0.067)
-0.048
(0.042)
-0.041
(0.063)
GENETIC
0.116
(0.256)
0.720
(0.404)
0.095
(0.260)
-0.487
(0.446)
FAMILY
0.389
(0.270)
0.904
(0.592)
0.371
(0.270)
0.562
(0.599)
WEEK_TIMES
0.267***
(0.044)
-0.044
(0.118)
0.248***
(0.057)
-0.122
(0.095)
0.023
(0.025)
N_HOURS
0.323***
(0.040)
-0.111
(0.129)
0.395***
(0.053)
-0.193*
(0.097)
-0.044
(0.027)
ALONE
-0.032
(0.100)
0.005
(0.125)
-0.032
(0.100)
0.012
(0.113)
WIN_EUPHORIA
0.110
(0.126)
-0.075
(0.157)
0.101
(0.126)
-0.092
(0.142)
WIN_PLEASURE
0.037
(0.118)
-0.075
(0.147)
0.040
(0.118)
-0.063
(0.134)
WIN_SATISFACTION
0.006
(0.120)
-0.064
(0.150)
0.013
(0.121)
-0.055
(0.137)
WIN_REPLAY
0.367***
(0.138)
-0.009
(0.227)
0.366***
(0.138)
-0.123
(0.197)
WIN_OMNIPOTENCE
0.723***
(0.225)
-0.156
(0.379)
0.702***
(0.226)
-0.360
(0.316)
WIN_OTHERS
-0.031
(0.243)
0.029
(0.295)
-0.044
(0.245)
0.021
(0.273)
LOSE_GUILT
-0.208
(0.150)
0.368**
(0.181)
-0.216
(0.150)
0.346**
(0.177)
LOSE_FRUSTRATION
-0.018
(0.183)
0.434*
(0.251)
-0.034
(0.183)
0.342
(0.255)
LOSE_DISAPPOINTMENT
-0.148
(0.113)
0.147
(0.145)
-0.146
(0.114)
0.152
(0.130)
LOSE_EXCITEMENT
0.190
(0.186)
0.154
(0.267)
0.197
(0.187)
0.038
(0.249)
LOSE_ANGER
-0.041
(0.119)
0.037
(0.147)
-0.051
(0.119)
0.043
(0.134)
LOSE_REDEEM
0.114
(0.118)
-0.047
(0.153)
0.116
(0.118)
-0.078
(0.138)
LOSE_LOW_SELF_ESTEEM
-0.042
(0.338)
0.427
(0.414)
-0.051
(0.341)
0.355
(0.398)
LOSE_HELPLESSNESS
-0.050
(0.266)
-0.267
(0.356)
-0.078
(0.266)
-0.171
(0.326)
DURING_EUPHORIA
-0.192
(0.124)
0.109
(0.164)
-0.189
(0.124)
0.152
(0.144)
DURING_PLEASURE
-0.143
(0.118)
-0.105
(0.183)
-0.141
(0.118)
-0.034
(0.168)
DURING_SATISFACTION
-0.019
(0.126)
-0.111
(0.161)
-0.013
(0.126)
-0.074
(0.148)
DURING_ANXIETY
0.027
(0.134)
0.047
(0.167)
0.042
(0.134)
0.028
(0.152)
DURING_EXCITEMENT
0.048
(0.135)
-0.024
(0.167)
0.053
(0.135)
-0.031
(0.152)
DURING_FRUSTRATION
-0.049
(0.378)
0.315
(0.446)
-0.078
(0.380)
0.302
(0.421)
DURING_ANGER
-0.005
(0.255)
0.331
(0.324)
-0.022
(0.255)
0.253
(0.308)
DURING_OTHERS
-0.181
(0.190)
0.175
(0.241)
-0.170
(0.191)
0.189
(0.217)
RHO
-0.624*
(0.316)
-0.809*
(0.212)
GAMMA
0.824***
(0.241)
0.958***
(0.143)
26"
"
/CUT11
2.164***
(0.497)
2.132***
(0.486)
/CUT12
3.254***
(0.504)
3.221***
(0.495)
/CUT13
4.789***
(0.547)
4.760***
(0.507)
/CUT21
1.763***
(0.683)
1.488**
(0.212)
WALD TEST
Chi2(36) = 269.66***
Chi2(36) = 269.27***
LR TEST OF INDIP. EQNS.
Chi2(1) = 12.63***
Chi2(1) = 14.61
N. OBS
710
710
LOG LIKELIHOOD
-885.536
-883.857
Notes: 1) Wald-statistic is a test where all slope coefficients are jointly zero. 2) Standard
errors in parenthesis. 3) ***, ** and * indicate significance at the 1%, 5% and 10% levels,
respectively.
27"
"
Table 4. Results of the seemingly unrelated bivariate ordered
probit regression
Model (1)
Model (2)
Dependent
variable: BET
Dependent
variable:
PATHOLOGY
Dependent
variable: BET
Dependent
variable:
PATHOLOGY
Coeff.
Std.Er.
Coeff.
Std.Er.
Coeff.
Std.Er.
Coeff.
Std.Err.
AGE
0.004
(0.005)
0.013
(0.007)
0.005
(0.005)
0.012
(0.008)
FEMALE
-0.231*
(0.135)
-0.260
(0.178)
-0.232*
(0.135)
-0.245
(0.179)
DISTANCE
0.093*
(0.049)
0.152
(0.185)
SINGLE
-0.353
(0.254)
-0.420
(0.338)
-0.361
(0.229)
-0.414
(0.313)
DIVORCED
-0.258
(0.246)
-0.592
(0.387)
-0.259
(0.230)
-0.586
(0.385)
WIDOW/ER
-0.040
(0.311)
-0.313
(0.413)
-0.055
(0.313)
-0.289
(0.415)
EDUCATION
0.026
(0.037)
0.193***
(0.067)
0.025
(0.037)
0.195***
(0.066)
INCOME
0.126***
(0.033)
-0.133**
(0.067)
0.012
(0.136)
-0.018
(0.261)
INCOME^2
0.017
(0.020)
-0.024
(0.040)
TIME_EXPERIENCE
-0.050
(0.042)
0.015
(0.057)
-0.048
(0.042)
0.019
(0.057)
GENETIC
0.162
(0.253)
0.770**
(0.350)
0.165
(0.314)
0.779**
(0.354)
FAMILY
0.376
(0.269)
1.022***
(0.374)
0.373
(0.331)
1.034***
(0.322)
WEEK_TIMES
0.250***
(0.044)
0.049
(0.062)
0.298***
(0.075)
0.050
(0.062)
WEEK_TIMES * DISTANCE
-0.033
(0.042)
N_HOURS
0.335***
(0.040)
0.157**
(0.075)
0.316***
(0.068)
0.157**
(0.076)
N_HOURS * DISTANCE
0.013
(0.039)
ALONE
-0.068
(0.099)
-
0.630***
(0.214)
-0.078
(0.100)
-
0.635***
(0.243)
WIN_EUPHORIA
0.105
(0.126)
-0.199
(0.179)
0.105
(0.126)
-0.200
(0.179)
WIN_PLEASURE
0.032
(0.118)
-0.059
(0.165)
0.034
(0.118)
-0.060
(0.166)
WIN_SATISFACTION
0.006
(0.120)
-
0.616***
(0.169)
0.009
(0.121)
-
0.626***
(0.169)
WIN_REPLAY
0.359***
(0.138)
0.389
(0.186)
0.364***
(0.138)
0.398*
(0.237)
WIN_OMNIPOTENCE
0.667***
(0.222)
0.172
(0.290)
0.682***
(0.223)
0.164
(0.290)
WIN_OTHERS
-0.018
(0.242)
-0.154
(0.304)
-0.014
(0.243)
-0.145
(0.304)
LOSE_GUILT
-0.190
(0.149)
0.210
(0.189)
-0.185
(0.149)
0.222
(0.189)
LOSE_FRUSTRATION
0.076
(0.179)
0.170
(0.245)
0.066
(0.179)
0.181
(0.245)
LOSE_DISAPPOINTMENT
-0.142
(0.113)
-0.358*
(0.209)
-0.132
(0.113)
-0.367*
(0.209)
LOSE_EXCITEMENT
0.191
(0.186)
-0.275
(0.291)
0.193
(0.187)
-0.299
(0.292)
LOSE_ANGER
-0.042
(0.119)
0.084
(0.162)
-0.043
(0.119)
0.080
(0.162)
LOSE_REDEEM
0.110
(0.117)
0.319
(0.203)
0.115
(0.117)
0.311
(0.163)
LOSE_LOW_SELF_ESTEEM
-0.030
(0.336)
-0.956
(0.809)
-0.063
(0.338)
-0.899
(0.891)
LOSE_HELPLESSNESS
-0.015
(0.262)
-0.556
(0.518)
-0.020
(0.263)
-0.564
(0.551)
DURING_EUPHORIA
-0.179
(0.123)
0.359
(0.232)
-0.181
(0.124)
0.372
(0.232)
DURING_PLEASURE
-0.123
(0.118)
0.191
(0.168)
-0.121
(0.118)
0.196
(0.168)
DURING_SATISFACTION
-0.010
(0.126)
-0.149
(0.179)
-0.023
(0.126)
-0.146
(0.179)
DURING_ANXIETY
0.016
(0.134)
0.133
(0.187)
0.010
(0.135)
0.121
(0.187)
DURING_EXCITEMENT
0.059
(0.134)
0.442*
(0.244)
0.048
(0.135)
0.446*
(0.243)
DURING_FRUSTRATION
-0.141
(0.350)
0.095
(0.424)
-0.146
(0.352)
0.099
(0.425)
DURING_ANGER
-0.188
(0.251)
-0.464
(0.725)
-0.207
(0.253)
-0.438
(0.726)
DURING_OTHERS
0.030
(0.188)
0.369
(0.259)
0.014
(0.189)
0.392
(0.259)
RHO
0.371***
(0.093)
0.382***
(0.093)
/CUT11
2.113***
(0.416)
2.076***
(0.507)
28"
"
/CUT12
3.195***
(0.424)
3.158***
(0.514)
/CUT13
4.701***
(0.441)
4.675***
(0.525)
/CUT21
2.652***
(0.708)
2.835***
(0.775)
WALD TEST
Chi2(36) = 270.01***
Chi2(39) = 270.94***
LR TEST OF INDIP. EQNS.
Chi2(1) = 13.69***
Chi2(1) = 14.28***
N. OBS
714
714
LOG LIKELIHOOD
-744.194
-743.156
Notes: 1) Wald-statistic is a test where all slope coefficients are jointly zero. 2) Standard
errors in parenthesis. 3) ***, ** and * indicate significance at the 1%, 5% and 10% levels,
respectively.
Ultimi Contributi di Ricerca CRENoS
I Paper sono disponibili in: Uhttp://www.crenos.itU
13/04
Luc Bauwens, Edoardo Otranto, “Modeling the
Dependence of Conditional Correlations on Volatility”
13/03
Oliviero A. Carboni, Claudio Detotto, “The economic
consequences of crime in Italy”
13/02
Pasqualina Arca, Gianfranco Atzeni, Luca Deidda,
“Economics of bankruptcy exemption: Signaling value
of collateral, cost of credit and access to credit”
13/01
Miguel Casares, Luca Deidda, Jose E. Galdon-Sanchez,
“Business cycle and monetary policy analysis with
market rigidities and financial frictions”
12/36
Maria Chiara Di Guardo, Raffaele Paci, “M&A and
knowledge flows in the European Union’s Neighboring
Countries”
12/35
Raffaele Paci, Emauela Marrocu, “Tourism and regional
growth in Europe”
12/34
Fabio Cerina, “Endogenous Growth and Sustainable
Tourism”
12/33
Michele Battisti,Filippo Belloc, Massimo Del Gatto,
“Unbundling technology adoption and tfp at the firm
level. Do intangibles matter?”
12/32
Massimo Del Gatto, Filippo di Mauro, Joseph Gruber,
Benjamin Mandel, “The “Revealed” Competitiveness of
U.S. Exports ”
12/31
Marinella Cadoni, Roberta Melis, Alessandro Trudda,
“Financial Crisis: a new measure for risk of pension
funds assets”
12/30
Oliviero A. Carboni, Paolo Russu, “Global Indeterminacy
in a Tourism Sector Model”
12/29
Oliviero A. Carboni, Paolo Russu, “A Model of Economic
Growth with Public Finance: Dynamics and Analytic
Solution”
12/28
Maria Giovanna Brandano, Claudio Detotto, Marco Vannini,
“Comparative efficiency of producer cooperatives and
conventional firms in a sample of quasi-twin
companies”
12/27
Bianca Biagi, Maria Giovanna Brandano, Dionysia Lambiri,
“Does tourism affect house prices? Some evidence
from Italy”
12/26
Gabriele Cardullo, Maurizio Conti, Giovanni Sulis, “Sunk
Capital, Unions and the Hold-Up Problem: Theory and
Evidence from Sectoral Data”
12/25
Fabio Cerina, Fabio Manca, “Catch me if you learn:
development-specific education and economic growth”
12/24
Andrea Pozzi, Fabiano Schivardi, “Demand or
productivity: What determines firm growth?”
12/23
Dimitri Paolini, Juan de Dios Tena, “Institutional
Complexity and Managerial Efficiency: A Simple
Model”
12/22
Miguel Jara, Dimitri Paolini, Juan de Dios Tena,
“Management Efficiency in Football: An Empirical
Analysis of two Extreme Cases”
12/21
Marta Foddi, Stefano Usai, “Regional innovation
performance in Europe”
12/20
Juan de Dios Tena, Claudio Detotto, Dimitri Paolini, “Do
managerial skills matter? An analysis of the impact of
managerial features on performance for the Italian
football”
Finito di stampare nel mese di Maggio 2013
Presso Copy…Right! studio grafico & stampa digitale
Via Turritana 3/B – Tel. 079.200395 – Fax 079.4360444
07100 Sassari
www.crenos.it
