I examined behavioral changes in web surveys, focusing on shifts in attentiveness over time. Using response-time Gaussian mixture modeling, respondents were classified into three types: consistently inattentive, deteriorating (initially attentive but later disengaged), and consistently attentive. For the deteriorating group, a Weibull model estimated the timing of lapses relative to cumulative time, revealing a sharp decline in attentiveness at seven–nine minutes. These findings challenge static views of inattentiveness as a fixed trait and support a dynamic framework of attentional decline during survey participation. By quantifying when and how inattention occurs, this approach provides insights into inattentive responding patterns and supports real-time monitoring and adaptive design.
Yoshihisa Kaneko (Sun,) studied this question.