Cognitive bias in workplace investigation: Problems, perspectives and proposed solutions

A machine learning-based predictive model of causality in orthopaedic medical malpractice cases in China

PURPOSE

To explore the feasibility and validity of machine learning models in determining causality in medical malpractice cases and to try to increase the scientificity and reliability of identification opinions.

METHODS

We collected 13,245 written judgments from PKULAW.COM, a public database. 963 cases were included after the initial screening. 21 medical and ten patient factors were selected as characteristic variables by summarising previous literature and cases. Random Forest, eXtreme Gradient Boosting (XGBoost) and Light Gradient Boosting Machine (LightGBM) were used to establish prediction models of causality for the two data sets, respectively. Finally, the optimal model is obtained by hyperparameter tuning of the six models.

RESULTS

We built three real data set models and three virtual data set models by three algorithms, and their confusion matrices differed. XGBoost performed best in the real data set, with a model accuracy of 66%. In the virtual data set, the performance of XGBoost and LightGBM was basically the same, and the model accuracy rate was 80%. The overall accuracy of external verification was 72.7%.

CONCLUSIONS

The optimal model of this study is expected to predict the causality accurately.

Exploring bias in incident investigations: An empirical examination using construction case studies

INTRODUCTION

Incident investigation is a foundational tool of safety management. Determining the causal factors of any incident underpins organizational learning and subsequent positive change to processes and practices. Research of incident investigation has largely focused on what information to collect, how to analyze it, and how to optimize resultant conclusions and organizational learning. However, much less attention has been paid to the process of information collection, and specifically that of subjective information obtained through interviews. Yet, as all humans are biased and can't help being so, the information collection process is inevitably vulnerable to bias.

METHOD

Simulated investigation interviews with 34 experienced investigators were conducted within the construction industry.

RESULTS

Common biases were revealed including confirmation bias, anchoring bias, and fundamental attribution error. Analysis was also able to unpack when and how these biases most often emerged in the interview process, and the potential consequences for organizational learning.

CONCLUSIONS

Being biased to a certain degree will remain inevitable for any individual, and therefore, efforts to mitigate the effects of biases is necessary.

PRACTICAL APPLICATIONS

Increased awareness and insights can support the development of processes and training for investigators to mitigate its effects and thus enhance learning from incidents in the field prevent reoccurrence.

Reducing the impact of cognitive bias in decision making: Practical actions for forensic science practitioners

Previously published methods for reducing the impact of cognitive bias in forensic decision making have focused primarily on actions at the laboratory or organizational levels. This paper presents generalized and specific actions that forensic science practitioners can take to reduce the impact of cognitive bias in their work. Practical examples illustrating ways that practitioners can implement many of the specific actions are also provided, along with some suggestions for handling court testimony about cognitive bias. The actions presented in this paper provide a means through which individual practitioners can take ownership for minimizing cognitive bias in their work. Such actions can provide supporting evidence to stakeholders that forensic practitioners acknowledge the existence of cognitive bias and its potential influence on their work, and they can also stimulate implementation of methods that focus on solutions at the laboratory and organizational levels.