Feedback specificity, information processing, and transfer of training

Abstract

This study examines the effects of feedback specificity on transfer of training and the mechanisms through which feedback can enhance or inhibit transfer. We used concurrent verbal protocol methodology to elicit and operationalize the explicit information processing activities used by 48 trainees performing the Furniture Factory computer simulation. We hypothesized and found support for a moderated mediation model. Increasing feedback specificity influenced the exposure trainees had to different task conditions and negatively affected their levels of explicit information processing. In turn, explicit information processes and levels of exposure to different task conditions interacted to impact transfer of training. Those who received less specific feedback relied more heavily on explicit information processing and had more exposure to the challenging aspects of the task than those who received more specific feedback, which differentially affected what they learned about the task. We discuss how feedback specificity and exposure to different task conditions may prime different learning processes.

Introduction

Kluger and DeNisi’s (1996) meta-analytic review of 100 years of feedback research provided a watershed in that it challenged accepted wisdom regarding the effects of feedback and identified limitations in the research conducted until their review. Contrary to the widely accepted view that feedback is both necessary and beneficial, they reported that the observed effects of feedback were about equally likely to have positive, negative, or no effect on performance. They also pointed to the lack of studies examining the causal mechanisms through which feedback influences behavior and performance. Despite its extensive research history, feedback research also has been criticized for its lack of valid, cumulative findings (Annett, 1969, Ilgen et al., 1979, Kluger and DeNisi, 1996) and its many invalid textbook prescriptions (Goodman, Wood, & Hendrickx, 2004).

In response to these criticisms and to increase the field’s understanding of feedback effects, Goodman and Wood (Goodman, 1998, Goodman and Wood, 2004, Goodman and Wood, 2009, Goodman et al., 2004) have built a program of research around the impacts of different feedback properties on training performance and transfer of training, as well as around the mechanisms through which feedback can enhance or inhibit learning and subsequent transfer. Their work is based on motor and cognitive learning literatures that emphasize the guidance function of feedback and suggest that frequent, immediate feedback can provide too much guidance during training (Christina & Bjork, 1991). Consistent with the guidance hypothesis, the results of Goodman and Wood’s studies show that more specific feedback can benefit training performance yet undermine the learning required for transfer of training, particularly learning that requires performers to manage more challenging aspects of a task. As feedback specificity increases, more detailed information is provided on performers’ actions and the locations of their errors (Annett, 1969, Baron, 1988, Goldstein et al., 1968, Payne and Hauty, 1955, Wentling, 1973). Specific feedback may also provide detailed information on corrective actions. For example, less specific feedback may inform performers only of their performance levels or simply that they made errors, whereas high specificity feedback may also identify which actions were correct and incorrect, which actions caused specific errors, and what the correct responses are.

While Goodman and Wood focused on guidance as provided by feedback, guidance can be delivered through many different modes in training and work settings. A parallel body of work, begun by Frese and his colleagues (Dormann and Frese, 1994, Frese and Altmann, 1989) has examined the impacts of the levels of guidance provided by different types of error training (e.g., Gardner and Wood, 2009, Gully et al., 2002, Nordstrom et al., 1998). While the manipulations differ across error training studies, error management training is designed to provide a low level of guidance, and error avoidance training is typically designed to provide a high level of guidance. Similar results have been found in error training and feedback research, despite differences in how guidance is operationalized. Compared to low guidance, providing high guidance during training has been found to have positive or no effects on training performance and negative effects on transfer of training for more challenging tasks that require adaptive transfer of knowledge (Keith & Frese, 2008).

The seemingly contradictory effects of high guidance specific feedback on training performance and transfer of training are partially explained by the impacts of feedback specificity on the extent to which performers spontaneously engage in exploratory behavior (Goodman et al., 2004) and are exposed to different task conditions (Goodman and Wood, 2004, Goodman and Wood, 2009). Exploration and exposure to different task conditions are two distinct but related causal pathways that influence trainees’ exposure to information during training. However, these pathways do not address what trainees actually do with the information to which they are exposed. This is the subject of the current study.

The current study focuses on the learning processes that involve explicit information processing activities such as planning and evaluating actions (Berardi-Coletta et al., 1995, Keith and Frese, 2005), causal reasoning (Isenberg, 1986), and hypothesis testing (Burns and Vollmeyer, 2002, Van der Liden et al., 2003, Van der Linden et al., 2001). We build on previous research by examining explicit information processing as a causal mechanism that helps explain what performers do with the information to which they are exposed during training, and we provide the first direct test of the popular guidance hypothesis claim that high-guidance feedback undermines the information processing needed for transfer (Christina & Bjork, 1991). Extant research has yet to test the assertion that feedback decreases information processing, which is the most proximal determinant of learning. In addition, the theoretical rationale for the guidance hypothesis assertion has not been sufficiently developed. We use the Dual Space Theory of problem solving (Dunbar, 1995, Klahr and Dunbar, 1988, Simon and Lea, 1974) to develop arguments to explain how feedback specificity impacts explicit information processing activities and how explicit information processes contribute to problem solving, learning, and transfer.

We also further explore Goodman and Wood, 2004, Goodman and Wood, 2009 ideas about the effects of differences in the guidance provided by low and high specificity feedback on the extent to which trainees experience different task conditions. These authors (Goodman and Wood, 2004, Goodman and Wood, 2009) found that trainees’ feedback-induced exploratory activities resulted in different levels of exposure to unfavorable conditions, in which trainees experienced the more challenging aspects of the task, and favorable conditions, in which trainees experienced the more straightforward aspects of the task. The rules for correct responses varied across the unfavorable and favorable task conditions. The rules took the form, “If this situation, then this action is appropriate.” Those who received highly specific feedback tended to follow the guidance offered by the feedback and learned how to manage under favorable task conditions, but developed less mastery of the rules for managing unfavorable task conditions. Alternatively, those who received low specificity feedback developed greater mastery of the rules for managing unfavorable task conditions, but acquired less skill for managing favorable task conditions. In the current study we explore how the exposure to different task conditions impacts and interacts with explicit information processes to affect transfer of training. We argue that explicit information processing activities are more vital for transfer the greater the exposure to unfavorable task conditions during training.

The contributions of the current study can be summarized as follows. First, we elaborate the theoretical reasoning for and conduct the first direct test of the guidance hypothesis assertion that high-guidance feedback decreases the information processing needed for transfer of training (Christina & Bjork, 1991). Second, we extend the cumulative body of research on how feedback affects performance and learning (Goodman and Wood, 2004, Goodman and Wood, 2009, Goodman et al., 2004) to include the effects of feedback specificity on the explicit processing of the information that trainees acquire through exploratory behavior and exposure to different task conditions during training. Finally, and more generally, our study responds to the call for the study of training conditions and activities that facilitate adaptive performance (Ford & Weissbein, 1997) by addressing how the content of feedback interventions affects trainees’ ability to perform independently and adapt to changing conditions (Salas & Cannon-Bowers, 2001).