A Question to Big Data: When Do Injured Consumers Stay Satisfied?

摘要

In a first, the author leverages big data to ask when does Injurious Consumption Outcome impact Overall Customer Satisfaction? I combine parallel computing and machine learning with Latent Instrumental Variables to build a predictive model to study more than 4.6 million injuries between 2003 and 2012 reported at U.S. hospital emergency departments. First, I find that on its own Injurious Consumption Outcome's impact on Overall Customer Satisfaction is non-significant, but it is the sensitivity of the consumer product injury that hurts Overall Customer Satisfaction. Second, there are limits to the impact of high Consumer Sentiment on Overall Customer Satisfaction—in the presence of a Serious Injury a positive Consumer Sentiment is effective in a product context. Third, economic Growth Prospects, appear to have a broader affect than Consumer Sentiment. Finally, in case of Injurious Consumption Outcome, Gender does not have much of a role to play when the outcome of interest is Overall Customer Satisfaction.. As a robustness check, a propensity score matched quasi-experiment is also carried out on this big data.