OU blog

Personal Blogs

Alfred Anate Mayaki

Scoping Review: Bayesian Inference in R

Visible to anyone in the world
Edited by Alfred Anate Mayaki, Friday, 15 Mar 2024, 16:16

This one is focused on Bayesian inference. Bayes’ Theorem in combination with dynamic stochastic general equilibrium theory is really popular as a tool for monetary policy (where I first encountered its use) but also has applications in other areas such as biology and, in this case, people analytics.

Now, there are three R packages of interest. The first is a package oriented around the Markov Chain - MCMCpack (Martin, 2006), the second is deBInfer (Boersch-Supen, 2016), and most recently, there is INLA (Gomez-Rubio, 2020).

More on these later.

References

Boersch-Supen, P. H., Ryan, S. J. and Johnson, Leah, R. (2016) “deBInfer: Bayesian Inference of dynamical models of biological systems in R”, Special Feature: Technological Advances at the Interface between Ecology and Statistics, 8(4), pp. 511-518, Available at: https://doi.org/10.1111/2041-210X.12679 (Accessed on 15 March 2024)

Gomez-Rubio, V. (2020) ‘Bayesian inference with INLA’, 1st Edition, New York: Chapman and Hall, Available at https://doi.org/10.1201/9781315175584 (Accessed on 15 March 2024)

Martin, A. D. and Quinn, Kevin. M. (2006) “Applied Bayesian Inference in R using MCMCpack”, R News, 6(1), pp. 2-7 – Available at: https://deepblue.lib.umich.edu/bitstream/handle/2027.42/116223/rnews06.pdf?sequence=1 (Accessed on 15 March 2024)

-------

This post was written by Alfred Anate Mayaki, a student on the MSc in HRM. It was inspired by an Economic Issues article by Sarah Brown and John Sessions entitled “Absenteeism, Presenteeism, and Shirking”.


Permalink Add your comment
Share post
Alfred Anate Mayaki

Started from the Bottom: Bayesian SPNE and Probability in HRM

Visible to anyone in the world
Edited by Alfred Anate Mayaki, Wednesday, 13 Mar 2024, 11:08

Bayes’ application to HRM is limited to event probability but is a topic that is mentioned in passing in a paper on shirking and presenteeism by S. Brown (2004) recommended by Dr. Andrew Bryce (Sheffield), which was written over 20 years ago this year.

Brown (2004) reads as follows:

“Such ‘shirking’ is potentially costly to firms and may incite them to undertake monitoring. BST** envisage a monitoring technology in which there is some probability, α < 1, of each absentee’s true state of health being revealed to the firm.”

After deciding to initiate a brief scoping review for the B812 literature topic of choice (‘Wellbeing’). I thought I would check in with the blog and provide some justification and background for this choice of theme.

This spurious love affair with Bayes’ theorem has loomed over my educational learnings but only in its form as sub-game perfect in non-cooperative game theory. Big thanks to Melvyn Coles, Pierre Regibeau, and Franco Squintani for their lectures and classes from our days in Colchester on Economics. 

I started the HRM course in Nov 2023 and while I am still somewhat aware of some concepts surrounding Bayes, things have changed. Nowadays, Bayes’ theorem (10+ years on) is being used in combination with what we call supervised learning and algorithmic techniques such as neural networks.

So, how do we proceed? Perhaps, it is wise to proceed with caution. A brief scoping review will get me up to speed and updated with new research as much as is feasibly possible.

References

Brown, S. and Sessions, John (2004) “Absenteeism, Presenteeism and Shirking”, Economic Issues, 9(1), pp. 15-22 – Available at: https://econpapers.repec.org/article/eisarticl/104brown.htm (Accessed on 13 March 2024)

**Barmby, T. A., Sessions, J. G. and Treble, J. G. (1994) “Absenteeism, Efficiency Wages and Shirking”, Scandinavian Journal of Economics, 94(4), pp. 561-566 – Available at: https://www.jstor.org/stable/3440797 (Accessed on 13 March 2024)

-------

This post was written by Alfred Anate Mayaki, a student on the MSc in HRM, and was inspired by the author's previous learnings and experiences.


Permalink Add your comment
Share post
Attachments:
application/pdfLCCC Organogram.pdf
Alfred Anate Mayaki

Predicting Dysfunctional Presenteeism: The Value of Bayesian Inference and Cognitive Load Theory

Visible to anyone in the world

Fox, Park and Lang (2007) famously wrote about secondary task reaction times (STRTs) in the context of psychology and communications - a sector I'm very familiar with. In the intervening period, further studies have been conducted on the workplace learning environment as a contextualised example of STRT research (Sewell, Santhosh, and O’Sullivan, 2020). Now, according to Kinman and Grant (2021), there is an estimated cost of approximately £4,000 per UK employee lost due to time spent at work whilst being clinically unwell (Patel, et. al., 2023). This is also known as what we refer to as presenteeism. This blog post hypothesizes that a highly significant correlated relationship exists between low STRTs and a type of presenteeism known as therapeutic presence. In my blog, I build on this theme of research by suggesting that it may be possible to alleviate the 'sunk' cost of low productivity in work which is lost due to the inefficiencies of working while being unproductive. I assume we have access to quality data, which isn't always the case. The blog considers the use of Bayesian inference to predict therapeutic presence (or what many refer to as ‘dysfunctional presenteeism’). 

What is ‘Dysfunctional’ Presenteeism?

Dysfunctional presenteeism can be defined as a type of presenteeism which can occur when an employee remains actively at work despite being clinically unwell (Bryan, Bryce, and Roberts, 2022; Henderson, and Smith, 2022). What we know about presenteeism is very little, particularly in instances of remote work (Schmitz, Bauer, and Niehaus, 2023). Dysfunctional presenteeism can vary in motive and context, but a recent survey estimated that UK employees spend roughly 2 weeks per annum actively working whilst being clinically unwell, which is quite a significant amount of time. Notably, what I have learned today is that presenteeism is frequently framed through the lens of ‘surface acting’ (Correia Leal, et. al., 2023; Patel, et. al., 2023:842).

STRTs and Cognitive Load Theory

How does STRT fit in to this? Well, my thinking is that low secondary responses are a clear indicator of the risk of presenteeism at work. Now, although the predominant focus of much of the literature on cognitive load emerges as research that is based on secondary task responses (Lang, et. al., 2006, Sewell, Santhosh, and O’Sullivan, 2020:1133), the significance of STRTs to cognitive load is insurmountable. As with Lang et. al., 2006:370), we know that when participants perform primary and secondary tasks within what we refer to as the STRT ‘paradigm’, the primary and secondary tasks are clearly defined[1].

Applying Bayes’ Theorem on Inference

Woolridge (2013) and as demonstrated recently in a paper by Saramago and Claxton (2020). Currently working on this in an essay.

Predictive Analytics in the Workplace

Now, Analytics is not necessarily ethical, especially when applied based on Ford’s principles. It is seemingly often pitted against the self-governance of employees by their peers[2], Analytics through surveillance is thus touted as the means of managerial control least appropriate for the digital age. In one instance, the UK's Low Carbon Contracts Company and the Electricity Settlement Company, a publicly owned renewable energy intermediary owned by the Secretary of State for Business, demonstrate through its management organizational chart, a distinctively unique feature (See Appendix Figure 1). Data Analytics is housed separately from People and Strategy. What does this tell us?

Conclusion

With quality data on both the threshold of secondary tasks and length of productive work for individual workers, this blog suggests that statistical inference techniques such as those based on Bayesian estimation, can predict the reduced work productivity of an individual employee (who may or may not be working remotely) instances of increased levels of stress, cognitive overload and eventual presenteeism?

[1] In this context, primary tasks involve the simultaneous observation and recollection of a media source such as television or film, whilst the secondary task involves a recordable activity. The idea is (the hypothesis, so to speak) that response times for the second task reduce as the primary task increases in difficulty (Lang, et.al., 2006).

[2] Including approaches such as self-evaluation of work performance.

References

1.     Byran, M. L., Bryce, A. M. and Roberts, J. (2022) “Dysfunctional presenteeism: Effects of physical and mental health on work performance”, The Manchester School, 90(4), pp. 409-438 – Available at: https://doi.org/10.1111/manc.12402 (Accessed on 23 November 2023)

2.     Correia Leal, C., Ferreria, A. I. and Carvalho, H. (2023) “Hide your sickness and put on a happy face: The effects of supervision distrust, surface acting, and sickness surface acting on hotel employees’ emotional exhaustion”, Journal of Organisational Behaviour, 44(1), pp. 871-887 – Available at https://doi.org/10.1002/job.2676 (Accessed on 27 February 2023)

3.     Fox, J. R., Park, B., and Lang, A. (2007) “When Available Resources Become Negative Resources: The Effects of Cognitive Overload on Memory Sensitivity and Criterion Bias”, Communication Research, 34(3), pp. 277-296. Available at: https://doi.org/10.1177/0093650207300429 (Accessed on 23 February 2024)

4.     Henderson, A.A. and Smith, C.E. (2022) “When does presenteeism harm productivity the most? Employee motives as a key moderator of the presenteeism–productivity relationship,” Journal of Managerial Psychology, 37(6), pp. 513–526. Available at: https://doi.org/10.1108/JMP-08-2020-0446 (Accessed on 26 February 2024)

5.     Kinman, G. and Clements, A. J. (2023) “Presenteeism: the case or action”, Occupational Medicine, 73(4), pp. 181-182 – Available at: https://doi.org/10.1093/occmed/kqad033 (Accessed on 26 February 2024)

6.     Kinman, G. and Grant, C. (2021) “Presenteeism during the COVID-19 pandemic: Risk factors and solutions for employers”, Society of Occupational Medicine, Available at: https://www.som.org.uk/Presenteeism_during_the_COVID-19_pandemic_May_2021.pdf (Accessed on 26 February 2024)

7.     Lang, A., Bradley, S. D., Park, B., Shin, M. and Chung, Y. (2006) “Parsing the Resource Pie: Using STRTs to Measure Attention to Mediated Messages”, Media Psychology, 8, pp. 369-394 – Available at: https://doi-org.libezproxy.open.ac.uk/10.1207/s1532785xmep0804_3 (Accessed on 26 February 2024)

8.     Patel, C., Biron, M., Sir Cooper, C. and Budhwar, P. S. (2023) “Sick and Working: Current challenges and emerging directions for future presenteeism research”, Journal of Organizational Behaviour, 44(1), pp. 839-852 – Available at https://doi.org/10.1002/job.2727 (Accessed on 27 February)

9.     Saramago, P., Claxton, K., Welton, N. J. and Soares, M. (2020) “Bayesian econometric modelling of observational data for cost‐effectiveness analysis: establishing the value of negative pressure wound therapy in the healing of open surgical wounds,” Journal of the Royal Statistical Society. Series A, Statistics in society, 183(4), pp. 1575–1593. Available at: https://doi.org/10.1111/rssa.12596 (Accessed on 27 February)

10.   Schmitz, H., Bauer, J. F. and Niehaus, M. (2023). Working Anytime and Anywhere - Even When Feeling Ill? A Cross-sectional Study on Presenteeism in Remote Work. Safety and Health at Work14(4), 375–383 – Available at: https://doi.org/10.1016/j.shaw.2023.11.001 (Accessed on 26 February 2024)

11.   Sewell, J.L., Santhosh, L. and O’Sullivan, P.S. (2020) “How do attending physicians describe cognitive overload among their workplace learners?” Medical education, 54(12), pp. 1129-. Available at: https://doi.org/10.1111/medu.14289 (Accessed on 26 February 2024)

------

This post was written by Alfred Anate Mayaki, a student on the MSc in HRM, and was inspired by the author's previous learnings and experiences.


Permalink
Share post

This blog might contain posts that are only visible to logged-in users, or where only logged-in users can comment. If you have an account on the system, please log in for full access.

Total visits to this blog: 74685