Proximal Antecedents and Correlates of Adopted Error Approach: A Self-Regulatory Perspective

ABSTRACT

The current study aims to further investigate earlier established advantages of an error mastery approach over an error aversion approach. The two main purposes of the study relate to (1) self-regulatory traits (i.e., goal orientation and action-state orientation) that may predict which error approach (mastery or aversion) is adopted, and (2) proximal, psychological processes (i.e., self-focused attention and failure attribution) that relate to adopted error approach. In the current study participants' goal orientation and action-state orientation were assessed, after which they worked on an error-prone task. Results show that learning goal orientation related to error mastery, while state orientation related to error aversion. Under a mastery approach, error occurrence did not result in cognitive resources “wasted” on self-consciousness. Rather, attention went to internal-unstable, thus controllable, improvement oriented causes of error. Participants that had adopted an aversion approach, in contrast, experienced heightened self-consciousness and attributed failure to internal-stable or external causes. These results imply that when working on an error-prone task, people should be stimulated to take on a mastery rather than an aversion approach towards errors.

Keywords:

• attribution of failure
• error handling
• self regulation

Notes

1. For the original purpose of testing effects of error approach, we used an error approach manipulation, using the error approach items as a manipulation check. As our manipulation check, however, revealed no effects, F(2, 57) = .59, p =n.s. for error aversion items; F(2, 57)=.16, p = n.s, for error mastery items, we concluded that our error approach manipulation was not successful. Further, ANOVAs conducted on theoretically relevant variables and control variables did not reach significance. In addition, Box's test of equality of covariance matrices indicated that the covariance matrices did not differ significantly between the three conditions, Box's M = 193.39, F(132, 8716) = 1.05, p = .33. We therefore treated the dataset as a one-group cross-sectional survey, conducted in a controlled task setting.