FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W1979476747> ?p ?o ?g. }

• W1979476747 endingPage "1677" @default.
• W1979476747 startingPage "1675" @default.
• W1979476747 abstract "A paper based on one of only a handful of existing national gambling surveys is to be welcomed, especially one that engages us in debate about the definition, nature and measurement of such a controversial and intriguing construct as ‘pathological gambling’. Toce-Gerstein et al. (2003) report results from the US National Opinion Research Center study which was commissioned by the US National Gambling Impact Study Commission, carried out across the USA and reported 4 years ago (NORC 1999). Another, similarly sized national US telephone gambling survey was reported by Welte et al. (2001), but the NORC study remains the only other nation-wide survey to have been carried out in the United States, and to my knowledge the United States remains one of only five countries where national gambling surveys have been conducted (the others being Australia, Britain, New Zealand and Sweden). NORC developed a new 17-item scale with between one and three items corresponding to each of the 10 criteria for pathological gambling stated in the fourth edition of the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (APA 1994). It produces a score between zero and 10. For those who are concerned with the psychometrics of such instruments, Toce-Gerstein et al.'s detailed presentation of how each DSM-IV criterion relates to the total score (Fig. 1a–j in their paper) is of considerable interest. Some criteria, such as chasing losses, preoccupation with gambling and lying to conceal the extent of involvement with gambling, turned out to be ‘low threshold’ criteria: the majority of those with a total score of 3 or 4 (i.e. just below the score of 5 which adherents to DSM often take as the lower bound for pathological gambling) met those criteria and an asymptote was reached thereafter. By contrast, some other items, such as jeopardizing or losing a relationship, job or educational or career opportunity because of gambling, and committing illegal acts to finance gambling, turned out to be ‘high threshold’ criteria, with only a minority reaching those criteria until a total score of 8 was reached. Chasing losses appeared to be something of rogue item as even at a total score of only 1 or 2, 50% reported this behaviour, and a factor analysis suggested that chasing was a factor all of its own. I confess to being one who becomes excited about such things, not only because of the intrinsic fascination of psychometrics, but also because we carried out a somewhat similar analysis based on the results of the first British Gambling Prevalence Survey (BGPS) (Sproston et al. 2000; Orford et al. 2003a). Questions designed to assess the 10 DSM-IV criteria were asked of all participants who acknowledged engaging in gambling in the previous 12 months (n = 5550). We also found a strong first factor (accounting for 40% of total variance) with all but one of the individual criteria loading 0.4 or above (Orford et al. 2003b). Again, the exceptional item was chasing, which loaded only 0.31 on the first general factor. In the light of Gerstein et al.'s (2003) paper I have looked in more detail at how individual criteria related to total DSM scores in our data. The results look similar. Comparing those whose total score was 1–2, 3–4 and 5 or above (numbers are too small to break the latter group down further) the percentage meeting the chasing losses criterion varies little (51%, 54%, 60%), whereas the gradient for committing crime to finance gambling is steep, particularly at the higher end (<1%, 13%, 50%). The poor performance of the chasing losses item in both our sets of data is interesting because some have argued that the phenomenon of chasing lies at the very heart of gambling addiction (e.g. Dickerson 1990; Lesieur 1994). The explanation may lie in the fact that chasing losses is comparatively common and therefore with low ‘positive predictive value’ for serious problem gambling. Alternatively, the DSM-IV criterion, and the questionnaire items based upon it, may be missing the point about chasing. O’Connor's (2000) detailed study of chasing suggests that there might be more to it than ‘going back the next day to recoup losses’. It may also include in-session chasing, ‘chasing’ (if that is the right word) wins as well as losses, careless betting in the face of losses, and important cognitive elements such as the anticipation of being able to escape debts and other problems. However, Toce-Gerstein et al. (2003) go beyond an examination of the performance of individual items, arguing strongly that their data provide evidence for a hierarchy of separable gambling disorders that are ‘distinguished qualitatively as well as quantitatively’, and which are related to each other according to a ‘stepwise progression of severity’. That is surely to go too far. It is not clear to me how the cross-sectional set of data with which they are working could be used to differentiate between a hierarchical stepwise model of distinct disorders and a continuum model. In fact, Toce-Gerstein et al. set up no such criteria before analysing their data, but rather used their data post hoc to suggest support for the first of those models. Even then their use of the results is a little loose in places. For example, inter-item correlations are extracted selectively from the complete correlation matrix in order to support the argument that the three items reported most commonly (preoccupation, escape and lying) form a coherent set which characterize ‘problem gambling’ which is distinct from ‘low-severity pathological gambling’ and ‘high-severity pathological gambling’ (e.g. the correlation of 0.30 between lying and bailout is not mentioned). Small, and presumably non-significant, differences between correlation coefficients are treated as if they are of significance. The addiction and allied fields are replete with attempts to erect hierarchies and other typologies which have not stood the test of time. A particularly instructive recent example is that of bulimia. The DSM-IV criteria for bulimia left out of account a large number of people who were recognized to have an eating disorder. That recognition led to a proposal for a distinct category termed ‘binge eating disorder’ (e.g. Fairburn & Wilson 1993; Hay & Fairburn 1998). Others preferred ‘eating disorders not otherwise specified (EDNOS)’. Yet others concluded that bulimic eating disorders were best thought of as lying on a continuum ranging from bulimia of purging type through bulimia of a non-purging kind, to binge-eating disorder, the least severe. In fact it seems to be a general rule that DSM definitions of addictive disorders, such as alcohol or drug dependence, bulimia nervosa or pathological gambling, turn out not to be sufficiently comprehensive, requiring those who treat or study addictive behaviours to invent new category labels for those who are omitted, such as alcohol or drug abuse, binge eating disorder or problem gambling (Orford 2001). The use of different labels inevitably suggests qualitative differences for which there may be no evidence. An alternative model, for which we thought we had found some evidence in data from the British Gambling Prevalence Survey (Orford et al. 2003b), is a multi-dimensional one. A strong first unrotated factor, as found by Toce-Gerstein et al. (2003), is certainly not inconsistent with the existence of two or more interpretable factors. Some would take their second factor eigenvalue of 1.1 as an indication that two factors should be rotated. When we performed that with our data we found two easily interpretable factors. One, accounting for 29% of total variance had high loadings for illegal acts, risked relationship and bailout. The second, accounting for 25% of total variance, had high loadings for preoccupation, tolerance, withdrawal, escape, and chasing. We interpreted the first as ‘gambling-related problems’ and the second as ‘feeling dependant’. Like Toce-Gerstein et al., we concluded that the defining and measuring of gambling problems is beset with difficulties, but our preference was for a solution in terms of two or more facets or dimensions rather than a number of distinguishable disorders. It would be useful if a psychometrician could advise us on whether there are data that could support one of those positions against the other. Meanwhile, I recall the arguments between psychologists and psychiatrists at the Institute of Psychiatry in London, going back to the 1960s, regarding the best model of depression. Has that argument been settled yet to everyone's satisfaction? Finally, there were some controversial features of the NORC study which should be mentioned. One was the complicated way in which the authors of that study attempted to correct for the probable under-estimate of problem gambling prevalence obtained from a random population survey, by drawing a supplementary survey of ‘patrons’ of gambling venues and using a complex method of weighting the two resulting estimates. The combined prevalence estimate turned out to be out of line with the rather higher prevalence estimate obtained in US state surveys. The basic survey was carried out by telephone with just under 2500 adults, which might be thought a rather small number for estimating a problem affecting a small percentage of the population. Another unusual feature of the NORC study was the decision not to use the standard screening instrument that has been used in most national surveys, namely the SOGS (the South Oaks Gambling Screen: Lesieur & Blume 1987). SOGS was used alongside a scale based on DSM-IV criteria in the US national survey reported by Welte et al. (2001) and in the British Gambling Prevalence Survey (Sproston et al. 2000; Orford et al. 2003a). NORC (1999) argued that SOGS was based on the outdated revised third edition of the DSM criteria. One of the principal advantages of SOGS is that it is a standard set of questions, unlike DSM-IV, which is a set of criteria for which there is no corresponding standard set of questions. For example, the 10 DSM-IV criteria were translated into 10 separate questions for the BGPS and 17 different questions for the NORC scale. The illegal acts criterion, to take one example, was translated for the BGPS as, ‘Have you committed a crime in order to finance gambling or to pay gambling debts?’, and for the NORC scale as, ‘Have you ever written a bad check or taken money that didn’t belong to you from family members or anyone else in order to pay for your gambling?’. A further problem that has bedevilled the standard measurement of problem gambling is whether it is more appropriate to ask for symptoms relating to ‘life-time’ or ‘last 12 months’ (or any other interval). When it was assumed that ‘once a pathological gambler, always a pathological gambler’, then the life-time method was preferred. More recently, with recognition that such an assumption is unsafe, the last 12 months method has been more common. It appears that NORC used both, although strangely it is not clear until quite late on in the Toce-Gerstein et al. (2003) paper that their analysis was based on the life-time method. That seems a rather important detail." @default.
• W1979476747 created "2016-06-24" @default.
• W1979476747 creator A5072784822 @default.
• W1979476747 date "2003-12-01" @default.
• W1979476747 modified "2023-09-24" @default.
• W1979476747 title "The fascination of psychometrics: commentary on Gerstein et al . (2003)" @default.
• W1979476747 cites W1974551483 @default.
• W1979476747 cites W2055983016 @default.
• W1979476747 cites W2062988353 @default.
• W1979476747 cites W2069647770 @default.
• W1979476747 cites W2073714611 @default.
• W1979476747 cites W2157599268 @default.
• W1979476747 cites W2397479638 @default.
• W1979476747 cites W2476715233 @default.
• W1979476747 doi "https://doi.org/10.1111/j.1360-0443.2003.00578.x" @default.
• W1979476747 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/14651498" @default.
• W1979476747 hasPublicationYear "2003" @default.
• W1979476747 type Work @default.
• W1979476747 sameAs 1979476747 @default.
• W1979476747 citedByCount "5" @default.
• W1979476747 countsByYear W19794767472014 @default.
• W1979476747 countsByYear W19794767472019 @default.
• W1979476747 crossrefType "journal-article" @default.
• W1979476747 hasAuthorship W1979476747A5072784822 @default.
• W1979476747 hasConcept C118552586 @default.
• W1979476747 hasConcept C15744967 @default.
• W1979476747 hasConcept C171606756 @default.
• W1979476747 hasConcept C70410870 @default.
• W1979476747 hasConceptScore W1979476747C118552586 @default.
• W1979476747 hasConceptScore W1979476747C15744967 @default.
• W1979476747 hasConceptScore W1979476747C171606756 @default.
• W1979476747 hasConceptScore W1979476747C70410870 @default.
• W1979476747 hasIssue "12" @default.
• W1979476747 hasLocation W19794767471 @default.
• W1979476747 hasLocation W19794767472 @default.
• W1979476747 hasOpenAccess W1979476747 @default.
• W1979476747 hasPrimaryLocation W19794767471 @default.
• W1979476747 hasRelatedWork W1991298437 @default.
• W1979476747 hasRelatedWork W2054723510 @default.
• W1979476747 hasRelatedWork W2147740381 @default.
• W1979476747 hasRelatedWork W2155138768 @default.
• W1979476747 hasRelatedWork W2416758480 @default.
• W1979476747 hasRelatedWork W2515615111 @default.
• W1979476747 hasRelatedWork W4231073527 @default.
• W1979476747 hasRelatedWork W4243766169 @default.
• W1979476747 hasRelatedWork W4254295689 @default.
• W1979476747 hasRelatedWork W2994342817 @default.
• W1979476747 hasVolume "98" @default.
• W1979476747 isParatext "false" @default.
• W1979476747 isRetracted "false" @default.
• W1979476747 magId "1979476747" @default.
• W1979476747 workType "article" @default.