PUB210 Article Critique Framework Online Tutoring

Article Critique Framework

Introduction

The aim of the critique for this article is its careful and argumentative analysis of the research carried out in this study for the evaluation of the study and its results. The study being critiqued in this report is on the gambling habits of young adults and how they might be associated with mental health and substance abuse problems. The research question of the study was based on the hypothesis that the involvement of an individual in gambling might be associated with a greater incidence of smoking, drinking alcohol and drug abuse. Moreover, the sociodemographic considerations were also kept in view in this study.

Study design

The study design constituted of male and female participants from the Mater-University of Queensland Study of Pregnancy (MUSP) utilizing different variables’ measurements and analysis which included: gambling and problem gambling; sociodemographic factors; substance use; and psycho-behavioral factors. The advantage of this study design lies in finding out the prevalence of gambling and its correlation with substance use in the young population of the university. However, the disadvantage is that the study has left the older population out and this is unable to present the findings with respect to a wide range of age groups. If this study had included that, it would have been more appropriate to generalize the population as a whole. The main ethical considerations of the study were informed consent, the confidentiality of participants, and adherence to data protection policies which were taken care of during both the in-person interviews as well as for those who participated via mailed questionnaires only via obtaining informed consents.

reported results (150)

The measures of association of this study included; the Canadian Problem Gambling Index (CPGI) consisting of three main sections i.e., involvement in gambling, assessment and correlates of problem gambling as previous studies have indicated this measure to have good validity and reliability; quantitative measurements of smoking per day; qualitative measures of alcohol use with three divisionary groups i.e., no alcohol use, one drink per day and more than one drinks per day; and methods of substance use included questions about frequency of illicit drugs use and then its division into three categories which are frequent users; occasional users and those who have never used. The results of this study indicated that gambling, which is an emerging problem in public health, is associated with substance use having early-onset during early adulthood life. Also, adults showing gambling problems had also more likelihood of displaying aggressive behaviors. Moreover, a bias could be attributed to only 40.8% of respondents of the study.

Error

First of all, the nonresponse error occurred from the reduction in the size of the samples collected during the 21-year follow-up period of the survey. The size reduction was such that only 48.6% of the individuals responded to this follow-up out of 7223 participants originally included in the study cohort. This might also be attributed to the selection of people for this study. This included the recruitment of pregnant women having their first clinic visit in the study out of which 8458 agreed for this research study participation. However, this number was reduced to a cohort of 7223 women who had given birth to singletons. Moreover, the sample size was further reduced when only 3512 of the total number of adults participated in the study of which 1648 were males and 1864 females. Another error can be due to an unequal number of male and female participants in the study. According to results, 44.2% of males and 36.6% of females have been reported to be involved in gambling. However, no normalization has been done with respect to the number of male and female participants. In this way, 36.6% could be an overrepresentation of the results due to a slightly larger number of female participants.

There are also chances of specification errors between the face-to-face data collection versus the data collection via emailed questionnaires. It is quite possible that an interviewer who explains the questions according to one’s mental capabilities is not transferred the same way to participants filling online forms. Also, the physical assessment by the interview taker is also missed for the later i.e., emailed questionnaire cohort thus raising another possible bias in the study.

The possible places where the measurement errors could have occurred are the socioeconomic as well as the sociodemographic differences between the participants. These included: the educational level i.e., no high school education, completed high school education, college and university level education; mode of education as being the part-time or full time; marital status as being single, married, divorced, or separated; and division of income into three levels i.e., low income, middle income, and high income. Although the division of participants according to these variables resulted in better classification of the individuals with respect to these factors and a better analysis of these factors in contributing towards the risk of gambling. However, one pitfall remained in the further classification of the individuals with respect to genders to see how these factors might have their effects with respect to the genders. Also, in case of any differences being reported, that could have given rise to further investigative questions as to the underlying reasons behind them. Moreover, errors in the comprehension of the questions might also be there due to different intellectual levels of all the groups in which the participants are divided, their ability to retrieve information and specific memories, the differences in their abilities to make judgments, and the variety of response formulation towards a question.

Confounding (250)

There have been various potential sources of confounding in the present study which included: gambling practices; gambling expenditures; the classification of gambling into subtypes; sociodemographic characteristics of participants; cross-sectional associations between gambling practice, mental health, and factors of substance use; and associations between various factors and gambling behaviors. The approaches taken for these by the authors of the research study were: frequency measurements and cross-tabulations for gambling practices, expenditures, and its subtypes; Chi-square analyses between gambling and young adults sociodemographic characteristics as well as the association between gambling, mental health, and substance abuse; and logistic regression tests and models for analyzing the association between various other factors included in the study and gambling behavior.

Other sources of residual confounding which could have been considered are the reduction in the number of samples over the 12-year survey design along with the difference in the number of male and female participants. Another confounding factor was the difference in the participants who contributed via direct interviews versus those who participated via online surveys via emails. Talking about these via one on one basis, firstly the reduction in the number of samples could have been dealt with via bootstrapping the data, and then regression modeling could have been done to see how well the model represents the original data. The same might have been applied to the gender difference in the data to get a better outlook of the results. Lastly, the difference arising with the use of online surveys could have been lessened with the help of telephonic or video conferencing.

Chance

The interpretation of the main results of this research study in light of the role that chance occurrence plays indicated that the gambling or problem gambling individuals were more likely to be involved in cigarette smoking, substance or illicit drug use. The reported confidence intervals defined for all the gambling, gambling expenditure as well as problem gambling according to the severity index are in the range of 95%. Rendering this criterion, a direct relationship was found between the money spent on gambling and the frequency of substance use. This was due to the values for these variables come in the range of the 95% confidence interval i.e., 2.3-3.6 for per week measures. Similarly, a significant association was also found between higher amounts of expenditure on gambling (more than $35) and substance abuse in the 95% confidence interval with a range of values between 3.7-7.5. therefore, the confidence interval is pretty much narrow for the above-described results and thus increasing the probability of the precision of the results more. In this study, the p-values for all the statistical analyses done were less than 0.001 which means that there is less than a 0.1% percent chance that the researchers are wrong in rejecting the null hypothesis. Since the likelihood of type I error is very low, the likelihood of type II error increases which can make the results false negative.

Conclusions

The internal validity assessment indicates that although the authors of this study have justified their prediction of association between young adult’s gambling, mental health problems, and substance abuse, a further enhancement might have been done by considering the internal biases being resulting from the difference between the sample size initially planned as well as between the genders. Also, bootstrapping the data to account for the loss present in the data might have helped in the better modeling for regression analyses. However, after these manipulations too, the results might have looked similar to what the authors have already found because, for example, the bootstrap method would only function to generate more data based on the existing data instead of adding new data to the experiment which could be a source of potential variation in the results. Moreover, the conclusions after the internal validity assessment would be similar to the authors’ conclusions in view of the confidence intervals and statistical errors.