Why childcare workers work when ill: organizational position matter

ABSTRACT

Working despite being ill is a challenging behaviour in the childcare sector. Whereas previous research on presenteeism often ignores whether a subject is a leader or an employee, we examine herein how organizational position affects presenteeism and the reasons for presenteeism among childcare workers. By analysing data from a two wave panel of 827 employees and leaders from the occupational group of childcare workers, we find that presenteeism is higher among employees than among leaders, however, not when we control for absenteeism. Employees and leaders have different reasons for going to work despite being ill. For instance, leaders have a higher probability of going to work ill because they want to avoid creating a backlog of work tasks. In contrast, employees have a higher probability of presenteeism out of consideration for colleagues. Overall, the findings provide insights into how presenteeism is affected by different organizational positions in the childcare sector.

KEYWORDS:

• Presenteeism
• childcare workers
• leader
• employee
• Bayesian negative binomial regression models

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data which support the findings of this study are available from the corresponding author upon reasonable request. The data are not publicly available, due to the anonymity of the respondents.

Notes

1. https://www.dst.dk/da/Statistik/emner/borgere/husstande-familier-og-boern/boernepasning

2. In this study, we chose a set of weakly informative priors, where the priors for the expected mean and the intercept are specified as a normal distribution with mean 0 for both parameters and a standard deviation of 1 for the intercept and 0.5 for the beta coefficients. This suggests that the effect is centred around zero, which implies enforcing little prior information in the analysis and letting the data speak in the analysis. Because the models are multilevelled and run with a random intercept for each individual, we also need a prior for sigma, where we chose a Student_t(3, 0, 2.5) prior for the negative binomial models and a Cauchy(0.1) prior for the logistic regressions. Furthermore, because negative binomial models draw on a beta distribution, it requires a prior for the shape parameter. In these models, we use gamma[0.01, 0.01] for the shape parameter.

3. Bivariate correlations of presenteeism are presented in Appendix A1 and A2.

4. Calculated using the hypothesis in the brms package. I’: 90%-CI for one-sided and 95%-CI for two-sided hypotheses.”*”: For one-sided hypotheses, the posterior probability exceeds 95%; for two-sided hypotheses, the value tested against lies outside the 95%-CI. Posterior probabilities of point hypotheses assume equal prior probabilities.

5. https://www.rdocumentation.org/packages/BayesianFirstAid/versions/0.1/topics/bayes.t.test