01.10.2013 Views

Gambling Among Young People, 837 kB

Gambling Among Young People, 837 kB

Gambling Among Young People, 837 kB

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

that first presented the tool, the two MAGS scales had an internal consistency with<br />

Cronbach alpha values of .83 and .87 (41). In a later study that only included the first<br />

scale the internal consistency was much lower at .65 (compared to .83) (48). The stability<br />

has not been tested. Finally the reliability of Lie/Bet has been proved to be good. Kappa<br />

values of .81 have been measured in the test-retest (43).<br />

general comments on the validity of the instruments<br />

The most commonly used questionnaires, SOGS-RA and DSM-IV-J, have been criticised<br />

by the Canadian researcher Ladouceur, who says that their estimates of gambling problems<br />

in young people are too high. One explanation for this might be that children do<br />

not understand the meaning of the questions. This has been tested by letting children<br />

and young people appreciate their gambling problems with and without an explanation<br />

of the meaning of the questions. With an explanation of the meaning of the questions,<br />

children, in particular those under the age of 12, reported fewer negative gambling consequences<br />

than they did before the questions had been explained to them. The decrease<br />

in the prevalence of pathological gambling was 50–73 per cent for SOGS-RA and<br />

20–29 per cent for DSM-IV-J (50, 51). However, according to Derevensky and Gupta<br />

from the International Centre of Youth Problem <strong>Gambling</strong> and High-Risk Behaviors in<br />

Canada, the tools are suitable for children over the age of 12 and the criticism is hence<br />

unjustified (53).<br />

Further criticism put forward by Ladouceur is that DSM-IV-J has been used<br />

erroneously in several studies. Instead of requiring four criteria out of nine to be fulfilled<br />

in order to classify a person as a probable pathological gambler, in several cases four<br />

affirmative answers out of a possible 12 have been used (52). According to Derevensky<br />

and Gupta this erroneous classification has not, however, had an impact on the results,<br />

according to new calculations based on the correct system (30, 53).<br />

In a comparative study of SOGS-RA, DSM-IV-J and GA, the prevalence of problem<br />

gambling varied depending on the tool. DSM-IV-J identified 3.4 per cent, SOGS-<br />

RA 5.3 per cent and GA 6 per cent problem gamblers or pathological gamblers in the<br />

same group of young people. The variation between the tools was greater for girls,<br />

1 per cent (DSM-IV-J), 1.5 per cent (SOGS-RA) and 3.5 per cent (GA) (46). The fact<br />

that girls reported fewer gambling problems than girls indicates that all these tools have<br />

a problem with conceptual validity since none of them presume a gender difference.<br />

Moreover, it is impossible to state which prevalence figure lies closest to the truth as<br />

there is no external “golden standard”. In another study where SOGS-RA was compared<br />

to MAGS-7, the tools did not really correspond in their classification of risk and<br />

problem gamblers respectively. They did, on the other hand, correspond fairly well<br />

when it came to problem free gamblers. Of the in total 269 individuals who according<br />

to one or other of the tools were assessed as risk gamblers, only 39 of them were classified<br />

by both tools, which gives a concordance for risk gamblers of 14.5 per cent. For<br />

problem gamblers the concordance was 20.5 per cent while for problem free gamblers<br />

the concordance was 82.2 per cent (48).<br />

The self-assessment tools are based on young people reporting negative consequences<br />

of gambling. One study asked young people to instead say whether they felt they had a<br />

gambling problem or not, which 1 per cent said they did compared with 3–6 per cent<br />

problem gamblers based on the self-reported negative consequences (54). This can<br />

perhaps be explained by the fact that young people themselves do not always know<br />

whether they have a gambling problem or not, partly because the problem is not wellknown<br />

and partly because the consequences of the problem may seem more vague to<br />

young people as compared to adults who for example are unable to pay their bills or

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!