Results of an analysis of survey data which includes data from inappropriate respondents (respondents who do not devote an appropriate amount of attentional resources when answering questions or whose answers for two questions are contradictory) are untrustworthy. To address this problem, an instructional manipulation check or directed questions scale can be used to identify such respondents. However, survey companies are not willing to use such tools for ethical reasons. In the present study, using eleven machine learning models and six exploratory variables, a prediction model which can judge whether a respondent is inappropriate is developed. The model shows that two explanatory variables, the maximum number of consecutive items on a scale to which a respondent answered with the same response option and response time, are effective for the prediction. The model can reduce the percentage of inappropriate respondents in the analyzed data, which leads to an improvement in the trustworthiness of the analysis results.
Measurement of subjective probability may theoretically be achieved based on the decision making tasks which require participants to choose between two gambles with known and unknown outcome probabilities. However, this approach is known to suffer from the effect of a human cognitive factor known as the ambiguity aversion. Moreover, because this approach is not based on statistical model, the estimation precision of the subjective probability cannot be evaluated. In the current study, we introduce the cumulative prospect theory model to this problem, and derive its Fisher information matrix. Using this information, we propose an adaptive presentation of the decision making tasks. Simulation studies and an empirical application confirmed that the derived Fisher information corresponds well with the empirical posterior standard deviation, and that the proposed adaptive task selection method performs much better than selecting the tasks at random. Furthermore, adaptive task selection which fixes the rewards of the two gambles was found to perform worse than the unconstrained ones. We conclude that unconstrained adaptive task selection is desirable for measurement of subjective probability under ambiguity.
Customers who hunt only for sale items in shopping are commonly referred to as cherry pickers. However, the quantitative identification of cherry pickers remains controversial in the relevant literature. This study proposes a new approach for identifying cherry pickers with a latent transition model. The proposed approach is sufficiently flexible such that the associated latent class vary over time with the purchasing attitudes of customers. Also, the covariates of the model include the demographic characteristics and the purchase experience of individual customers. The investigation for the purchase experience effect is so important for establishing marketing strategies. In practice, the inference is performed with a stage estimation method. This study illustrates the proposed approach with a purchase data set on an E-commerce website. The results reveal the dynamic patterns of the purchasing attitudes including the latent states of cherry-picking behavior.
This study examined the difference between authors and the consistency in each author's writing styles, both were the basis of authorship verification. We analyzed 88 academic papers on psychology written by 22 authors and focused on the rates of “non-content words”, “bigram of parts-of-speech”, “bigram of postpositional particles”, “positioning of commas”, ”words before period”, and “Kanji, Hiragana, and Katakana” in the papers. Next, symmetric Kullback-Leibler divergence distances between the papers were calculated. To examine the author differences in writing styles, using hierarchical Bayesian modeling, we compared the distances between papers written by the same author with those by different authors. Furthermore, to examine author consistency in writing styles, we compared the distances of short durations (under five years) between papers written by the same author with the longer durations (over five years). These results supported the hypothesis that there exist author differences and consistency in writing styles.
The objective in this research is to propose an index for customer relationship management instead of utilizing RFM or share of wallet. This study examined whether the number of days from contract to initial usage can be used as the index for customer management. For this purpose, an empirical analysis was conducted by using a Bayesian model with purchase history data of a single year and three years regarding two kinds of credit cards and a customer master data. The results of the empirical analysis showed the shorter duration to the contract day, the amount of money for the one hundred eighty days from initial usage became higher. These results clarified that the days can be utilized as the index for customer management. In the conventional customer management index such as RFM, it is necessary to store purchase history data for several months in a data warehouse, however, this proposed method does not need to store transaction data since the duration data can be utilized. Hence, this index can promptly find problematic customers.