Self-confidence and performance goal orientation interactively predict performance in a reasoning test with accuracy feedback Nadin Beckmann a, ⁎, Jens F. Beckmann b , Julian G. Elliott c a Australian School of Business, Accelerated Learning Laboratory, University of New South Wales, Sydney, NSW 2052, Australia b School of Psychology, University of Sydney, Brennan MacCallum Building (A18), Sydney, NSW 2006, Australia c School of Education, Durham University, Leazes Road, Durham DH1 1TA, United Kingdom abstract article info Article history: Received 30 January 2008 Received in revised form 5 September 2008 Accepted 13 September 2008 Keywords: Reasoning test Accuracy feedback Goal orientation This study takes an individual differences' perspective on performance feedback effects in psychometric testing. A total of 105 students in a mainstream secondary school in North East England undertook a cognitive ability test on two occasions. In one condition, students received item-specific accuracy feedback while in the other (standard condition) no feedback was provided. While accuracy feedback had, on average, no effect on test performance, differential effects were observed. Performance goal orientation and self- confidence moderated feedback effects on performance. The provision of accuracy feedback seemed to have performance optimising effects upon test takers with high performance goal orientation in combination with low self-confidence. The implications of these findings in relation to the validity of findings from psychological testing are discussed. © 2008 Elsevier Inc. All rights reserved. 1. Introduction Performance feedback is widely considered to be a prominent agent of learning processes or at least of improved performance. However, research has shown that, counter-intuitively, providing feedback does not always improve task performance. Meta-analytic studies have provided evidence that feedback effects on performance are often quite variable; on some occasions, feedback improves performance, in others, no effect can be found, and yet in other instances, feedback can negatively influence performance (Bangert- Drowns, Kulik, Kulik, & Morgan, 1991; Kluger & DeNisi, 1996). Relatively little is known about feedback effects in psychometric assessment settings, although the few studies published reveal a similarly mixed picture. Item feedback has been found to enhance performance in cognitive ability and achievement tests (e.g., Bethge, Carlson, & Wiedl, 1982; Betz, 1977; Birenbaum & Tatsouka, 1987; Carlson & Wiedl, 1979; Dash & Rath, 1984; Dillon, 1981), have no effect on test performance (e.g., Delgado & Prieto, 2003; Roos, Wise, & Plake, 1997; Stankov & Crawford, 1997; Wise, Plake, Pozehl, Boettcher- Barnes, & Lukin, 1989), and, in other studies, negatively affect test performance (e.g., Delgado & Prieto, 2003; Strang & Rust, 1973; Wise, Plake, Eastman, & Boetcher & Lukin, 1986). The provision of feedback is the central feature of dynamic testing (for an overview see Lidz & Elliott, 2000). Dynamic tests typically consist of items similar to those usually utilised in intelligence tests (e.g., reasoning problems) ordered in increasing complexity. However, dynamic tests differ procedurally from traditional measures in that the examinee is provided with feedback about their responses to items. Because this offers an opportunity to learn during the test session, it is argued that performance on such measures offers more valid indication of cognitive abilities (e.g., learning potential) than test scores obtained from traditional feedback-free intelligence tests (Grigorenko & Stern- berg, 1998; Guthke & Beckmann, 2003; Guthke & Stein, 1996). The simplest and, hence, most common form of feedback provided in dynamic tests and other cognitive ability tests concerns whether a response to a test item is accurate or not (in dynamic tests, e.g., Guthke & Beckmann, 2000; other intelligence tests, e.g., Stankov & Crawford, 1997; and adaptive achievement tests, e.g., Tonidandal, Quinones, & Adams, 2002). However, the mechanisms by which this very simple form of feedback affects test performance are far from understood. On a theoretical level Carlson and Wiedl (1992, 2000) discuss three functions of feedback in cognitive ability testing: modification, compensation and inhibition. With regard to simple accuracy feed- back it could be argued that it can modify knowledge related to problems by informing test takers about the appropriateness of the problem solving strategies they have employed; it might compensate for factors hindering performance by relocating the test taker's focus of attention to the task, and it might inhibit performance-reducing behaviours by motivating test takers to work harder. We argue however, accuracy feedback could also undermine the confidence of test takers and thus, negatively affect their test performance. Heterogeneous result patterns as observed in feedback research would seem to suggest the involvement of moderating variables. Those can lie in the situation (e.g., level of elaborateness of feedback) Learning and Individual Differences 19 (2009) 277–282 ⁎ Corresponding author. Tel.: +61 2 9931 9187; fax: +61 2 9931 9199. E-mail addresses: nadinb@agsm.edu.au (N. Beckmann), jensb@psych.usyd.edu.au (J.F. Beckmann), joe.elliott@durham.ac.uk (J.G. Elliott). 1041-6080/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.lindif.2008.09.008 Contents lists available at ScienceDirect Learning and Individual Differences journal homepage: www.elsevier.com/locate/lindif