Behavioral Bias And Cognitive Mechanism Of Correlation Judgment

"Correlation judgment" is one of the core contents of economic and management decision-making.Accurate judgment of the correlation between two variables is particularly important in commercial environments such as economic management.However,a large amount of empirical evidence shows that decision-makers often exhibit cognitive biases in their correlation judgments between different events,and these biases often have a negative impact on decision quality.To explain the cognitive biases in correlation judgment,management scholars have proposed the "information sampling" hypothesis.Specifically,the information sampling hypothesis holds that decision-makers make judgments about the correlation between events based on historical factual data that they know.However,due to limited cognitive resources and data availability biases in the environment,decision-makers often exhibit systematic biases in information sampling.The information sampling hypothesis suggests that such information sampling biases are one of the important factors that contribute to inaccurate correlation judgments.The information sampling hypothesis has strong explanatory power for the bias in correlation judgment in economic and management decision-making,but there is no direct empirical evidence to support this hypothesis.This study introduces eye-tracking technology,combined with classical Bayesian models,to explain the mechanism of the generation of bias in correlation judgment in the scatter plot paradigm from the perspective of visual information sampling,and to provide direct evidence for the information sampling hypothesis.There are three studies in this paper.Study 1 reveals the behavioral performance of decision-makers in correlation judgment bias and the effect of presentation mode on bias.It is found that people tend to underestimate and the underestimation is greater under the landscape condition than under the portrait condition.Study 2 uses the decision-makers’ information sampling process in eyetracking data to mathematically model information sampling data,quantify potential biases in eye-tracking information,explore the relationship between behavioral biases and eye-tracking trajectories,and verify the mechanism of bias in correlation judgment.The study found that the visual input information sampling bias largely explains the bias in behavioral results.Study 3 is based on real-life scenarios,extending laboratory research to empirical research.By analyzing the behavioral results of MBA students with social experience and work experience in correlation judgment,and comparing the results with laboratory research,the theory is more closely linked to reality.The results show that the MBA group still underestimates,but there is no significant difference in bias between the two aspect ratios.In summary,this study combines eye-tracking experiments and behavioral experiments to reveal the behavioral performance of correlation judgment bias in scatter plots and the key role played by visual information search.This paper constructs a Bayesian learning model that has both explanatory and predictive power,explains the mechanism of correlation judgment bias from a cognitive perspective,and provides direct evidence for the information sampling hypothesis in economic and management decision-making.The research results can provide reference value and theoretical basis for improving correlation judgment in graphic visualization,as well as new methods for promoting the development of research on correlation judgment in management.