Does test-enhanced learning transfer for triple associates?

Mem Cogn (2016) 44:24–36 DOI 10.3758/s13421-015-0547-x Does test-enhanced learning transfer for triple associates? Steven C. Pan 1 & Carol M. Wong 1 & Zachary E. Potter 1 & Jonathan Mejia 1 & Timothy C. Rickard 1 Published online: 1 September 2015 # Psychonomic Society, Inc. 2015 Abstract Test-enhanced learning and transfer for tripleassociate word stimuli was assessed in three experiments. In each experiment, training and final-test trials involved the presentation of two words per triple associate (triplet), with the third word having to be retrieved. In agreement with the prior literature on different stimuli, training through testing with feedback yielded markedly better final-test performance than did restudy. However, in contrast to the positive transfer reported for paired associate stimuli, minimal or no positive transfer was observed, relative to a restudy control, from a trained cue combination (e.g., A, B, ?) to other cue combinations from the same triplet that required a different response (e.g., B, C, ?). That result also held when two unique cue combinations per triplet were tested during training, and for triplets with low and high average associative strengths. Supplementary analyses provided insight into the overall transfer effect: An incorrect response during training appears to yield positive transfer relative to restudy, whereas a correct response appears to yield no, or even negative, transfer. Crossexperiment analyses indicated that test-enhanced learning is not diminished when two or three cue combinations are presented during training. Thus, even though learning through testing is highly specific, testing on all possible stimulus–response combinations remains the most efficient strategy for the learning of triple associates. Keywords Cued recall . Retrieval practice . Test-enhanced learning . Testing effect . Transfer * Timothy C. Rickard 1 Department of Psychology, University of California, San Diego, La Jolla, CA 92093-0109, USA The act of recalling information during a test, otherwise known as retrieval practice, enhances memory for that information, above and beyond an equivalent period of time spent restudying the same materials. This testing effect, or retrieval practice effect, has been replicated numerous times with stimuli ranging from paired associates to prose passages (Roediger & Karpicke, 2006a) and is widely regarded as one of the most robust phenomena in learning science (Butler, 2010; Carpenter, 2012). However, a limitation of the testing-effect literature to date is that the vast majority of published studies have used identical materials during the initial and final tests (for discussion, see McDaniel, Thomas, Agarwal, McDermott, & Roediger, 2013). Comparatively less is known about testing’s effectiveness for transfer to novel contexts. Alongside retention, transfer has been described as the ultimate objective of learning (e.g., Carpenter, 2012; Rohrer, Taylor, & Sholar, 2010). Instructors hope that the information that they impart to students will be flexibly accessible and will generalize to different contexts. Therefore, it is crucial to determine the extent to which test-enhanced learning transfers and whether it has limitations (Anderson & Biddle, 1975; Hinze & Wiley, 2011). In a recent review of transfer and the testing-effect literature, Carpenter (2012) concluded that testenhanced learning transfers well in three categories: (1) across temporal contexts (e.g., different retention intervals), (2) across test formats (e.g., question format alterations such as short-answer to multiple-choice), and (3) across knowledge domains (e.g., from biology to aeronautics). This classification scheme adheres to recommendations that research on transfer requires definitional clarity (Barnett & Ceci, 2002). Of the categories delineated by Carpenter (2012), the second and third include studies in which the stimulus materials used on a final test are different from those used on prior tests; among these, transfer has been demonstrated for word lists (Carpenter & DeLosh, 2006), prose passages (Butler, 2010; Mem Cogn (2016) 44:24–36 Chan, 2009, 2010; Chan, McDermott, & Roediger, 2006; Karpicke & Blunt, 2011), and science facts (Hinze, Wiley, & Pellegrino, 2013; McDaniel, Anderson, Derbish, & Morrisette, 2007; McDaniel et al., 2013; Rawson, Dunlosky, & Sciartelli, 2013). Test-enhanced learning also exhibits transfer for less commonly examined materials, such as math functions (Kang, McDaniel, & Pashler, 2011), spatial learning (Carpenter & Kelly, 2012), map learning (Rohrer et al., 2010), and even medical diagnoses (Larsen, Butler, Lawson, & Roediger, 2013). Moreover, transfer has been found when the difference between the originally tested and transfer test materials is relatively small (near transfer, involving questions on the same subject; e.g., detail questions, as in Hinze et al., 2013), as well as when it is quite large (far transfer, involving application or inference questions on separate topics; e.g., Chan et al., 2006; Johnson & Mayer, 2009; McDaniel, Howard, & Einstein, 2009). Conversely, there have been some failures to find transfer of test-enhanced learning (e.g., Agarwal, 2011; Hinze & Wiley, 2011; Tran, Rohrer, & Pashler, 2015). Such studies have employed materials similar to those in the studies described above (e.g., textbook passages, science facts, logical premises), and the transfer measured has ranged from near to far. As of this writing, however, failures to find transfer of testenhanced learning remain the exceptions in the literature. Positive transfer for paired associates Carpenter’s (2012) second category of transfer also includes studies in which the same materials are presented on the initial and final tests, but in which the stimulus and response elements have been rearranged on the final test. Strong positive transfer of that type had been observed for paired-associate words. Carpenter, Pashler, and Vul (2006) administered cued-recall tests for paired-associate English word lists in one cue-to-target direction (e.g., beach → blanket) and observed substantial transfer to the reversed, previously untested direction (e.g., blanket → beach) on a delayed final test. Similarly, Vaughn and Rawson (2014) adapted the Carpenter et al. paradigm using English–English word pairs to assess criterion-level testing (cf. Vaughn & Rawson, 2011). They replicated the finding of positive transfer for the case of subjects trained to a criterion level of one correct trial for each tested item, although the extent of transfer decreased at higher levels of criterion learning. In Carpenter et al. (2006) and Vaughn and Rawson (2014), transfer was assessed on the final test by the reversal of two elements (e.g., cue → target reversed to target → cue). This raises an important yet unaddressed question in the testing literature: Does test-enhanced learning transfer for stimuli with more than two elements, any of which could co (...truncated)