Search results

Confounding is of central importance in epidemiologic studies. Its definition has been under wide debate over the past decades. The classical definition is straightforward and easy-to-implement. Nevertheless, it is data-driven and has drawbacks. The more recent counterfactual definition captures the essential roles a confounder plays in causal inference. It would be beneficial for researchers to grasp substantive knowledge in causal structure and broadly adopt the latter definition. There are various methods of handling confounding issues. The choice of one option over another depends on various factors, including the nature of the study, sample size and rarity of events.

In this chapter, we explicate two related techniques that help quantify the sensitivity of a given causal inference to potential omitted variables and/or other sources of unexplained heterogeneity. In particular, we describe the Impact Threshold of a Confounding Variable (ITCV) and the Robustness of Inference to Replacement (RIR). The ITCV describes the minimum correlation necessary between an omitted variable and the focal parameters of a study to have created a spurious or invalid statistical inference. The RIR is a technique that quantifies the percentage of observations with nonzero effects in a sample that would need to be replaced with zero effects in order to overturn a given causal inference at any desired threshold. The RIR also measures the percentage of a given parameter estimate that would need to be biased in order to overturn an inference. Each of these procedures is critical to help establish causal inference, perhaps especially for research urgently studying the COVID-19 pandemic when scholars are not afforded the luxury of extended time periods to determine precise magnitudes of relationships between variables. Over the course of this chapter, we define each technique, illustrate how they are applied in the context of seminal strategic management research, offer guidelines for interpreting corresponding results, and delineate further considerations.

This study aims to examine the effect of organizational justice (OJ) delivery on organizational commitment (OC) while controlling variables that potentially confound this effect. These confound variable includes gender, education and tenure.

A causal quantitative field research and a cross-sectional research approach are used. A structured questionnaire is used to collect data from 463 randomly sampled employees of 13 commercial banks in Koforidua, Ghana. Structural equation modelling is used to analyze the data.

The study establishes that a statistically significant effect is made by OJ delivery on OC (p < 0.0001), even after controlling for the effects of gender, education and tenure. The model used also fits the data collected considerably (i.e. χ² = 0.258; p-value = 0.611). Thus, the relationship between OJ and OC is not confounded by how long the employee has served on the job or by the highest educational level of employees. However, this relationship was significantly confounded by gender on the basis of sharing significant covariance with them.

Even though studies exist generally on OJ and OC, the focus on Ghana and banks in particular have been skeletal and have ended in testing just the relationship without highlighting the role of confounding variables as done in this study. Apart from adding to extant literature, findings will also inform decision-making on strategies and policies to improve OJ and ultimately the employee commitment that comes with.

Postneonatal mortality (PNM), which differs from infant and perinatal mortality, has been observed in the past 25 years with respect to the health outcomes of children. While infant and perinatal mortality have been well-evaluated regarding racial differentials, there are no substantial data on PNM in this perspective. The purpose of this study was to assess whether or not social determinants of health adversely affect racial/ethnic PNM differentials in the USA.

A cross-sectional, nonexperimental epidemiologic study design was used to assess race as an exposure function of PNM using Cohort Linked Birth/Infant Death Data (2013). The outcome variable assessed PNM, while the main independent variables were race, social demographic variables (i.e. sex and age) and social determinants of health (i.e. marital status and maternal education). The chi-square statistic was used to assess the independence of variables by race, while the logistic regression model was used to assess the odds of PNM by race and other confounding variables.

During 2013, there were 4,451 children with PNM experience. The cumulative incidence of PNM was 23.6% (n = 2,795) among white infants, 24.3% (n = 1,298) among Black/African-Americans (AA) and 39.5% (n = 88) were American-Indian infants (AI), while 21.3% (n = 270) were multiracial, χ² (3) = 35.7, p < 0.001. Racial differentials in PNM were observed. Relative to White infants, PNM was two times as likely among AI, odds ratio (OR) 2.11 (95% confidence interval [CI] 1.61, 2.78). After controlling for the confounding variables, the burden of PNM persisted among AI, although slightly marginalized, adjusted odds ratio (aOR) 1.70, (99% CI 1.10, 2.65).

In a representative sample of US children, there were racial disparities in PNM infants who are AI compared to their white counterparts, illustrating excess mortality. These findings suggest the need to allocate social and health resources in transforming health equity in this direction.

This study aims to study different native languages as proxy for cultural differences in travel behaviour in multilingual Switzerland. Differences in travel behaviour between tourists from different nations (implying different cultural backgrounds) have been investigated in several studies. However, most studies do not reveal the potential confounding effects associated with cross-nation studies, such as differences in the economic, legal/regulatory or subcultural contexts of their residents. This exploratory study tries to overcome this gap by investigating differences in travel behaviour between language groups within one country: Switzerland. By holding potential confounding effects constant, our results reveal differences in travel behaviour associated with language, including travel motivation, choice of destination, choice of type of accommodation, duration of trip and travel expenses.

The study is based on data from one of the most comprehensive outbound travel behaviour survey: Travel Market Switzerland.

By holding potential confounding effects constant, our results reveal differences in travel behaviour associated with language, including choice of destination, choice of type of accommodation, duration of trip and travel expenses.

Differences in travel behaviour between tourists from different nations (implying different cultural backgrounds) have been investigated in several studies. However, most studies do not reveal the potential confounding effects associated with cross-nation studies, such as differences in the economic, legal/regulatory or subcultural contexts of their residents. This exploratory study tries to overcome this gap.

The purpose of this paper is to review the conceptual and methodological challenges of a J-shaped association between alcohol consumption (AC), coronary heart disease (CHD) and all-cause mortality. In associated papers in this journal, Skovenborg et al., 2021 reviews the evidence for the J-shaped curve, and Ellison et al., 2021 examines the advantages and drawbacks of Mendelian randomization studies of the J-shaped curve.

A number of methodological problems are common in observational research in general, and some of the methodological problems suggested for the J-shaped alcohol-CHD-associations are discussed. The extent of the methodological problems in studies of the J-shaped curve is reviewed, and the possibility that the J-shaped curve is an artifact created by reverse causality and residual confounding is discussed. Further, the issue of interaction with drinking pattern and type of alcohol is discussed.

Imprecise categorization of alcohol intake information seems to have had little effect on the J-shaped alcohol-CHD-associations, nor has it affected the ability of these studies to show increasing mortality from a range of causes with increasing AC. The problem of “sick quitters” has been resolved by large studies using lifelong abstainers or infrequent drinkers as reference group. Many studies lack information on drinking patterns with regard to regular, moderate consumption versus binge drinking. Stratified analyses by important risk factors for CHD have not significantly changed the J-shaped association observed in most epidemiologic studies.

Potential biases and residual confounding probably do not overcome the J-shaped alcohol-CDH-association observed in most epidemiologic studies; however, the existence of a J-shaped curve is challenged by some degree of uncertainty. The actual review together with the associated papers by Skovenborg et al., 2021 and Ellison et al., 2021 offers a possibility to “update your priors” and achieve greater certainty when giving your patients information on the pros and cons of alcohol intake.

The need for random assignment in sexual offending programme evaluation is clear. Decades of high dependence on weak-inference methodology, that of observational studies, has inhibited professional agreement regarding the effects of programmes. Observational studies have a place in evaluation research when more rigorous scientific designs precede them, as occurs in neighbouring fields of drug development and health. If, however, observational studies remain the only method used to evaluate sexual offending programmes, the field will continue to endure uncertainty with confident causal inferences regarding their effects remaining elusive. The paper aims to discuss these issues.

The paper takes the form of a literature review and discussion.

The case for random assignment is made alongside a rebuttal of arguments against their use.

This is an original look at the need for random assignment in sexual offending programme evaluation taking into account existing studies and discussion topics.

This study aims to examine the effect of health workers’ emotional intelligence (EI) on job performance (JP), with potential confounding variables controlled for. The confounding variables introduced are gender, education, tenure and level of access to in-service training.

A cross-sectional quantitative research design was used in this study. A self-reported questionnaire was used to collect data from 1,163 health professionals, who were selected using the simple random sampling method. Confirmatory factor analysis (CFA) was used to test a framework of hypotheses.

The resulting CFA model is of a good fit at 5 per cent significance level [chi-square (χ²) = 1.492; p = 0.222]. Moreover, the study finds that EI significantly predicts JP among health workers after controlling for the lurking variables.

Though several studies have confirmed that EI makes a significant positive effect on health workers’ JP, none of them controlled for potential confounding variables. For this reason, the effect detected in previous studies could include the influence of lurking variables and is consequently spurious. Apart from contributing to extant literature, this study controls for these lurking variables in an attempt to enhance the value of empirical evidence that supports the relevance of EI to health-care performance.