Applying Retrieval-Induced Forgetting to Children’s Testimony THOMAS L. PHENIX 1 * and HEATHER L. PRICE 2 1 Campion College at the University of Regina, Regina, Canada 2 University of Regina, Regina, Canada Summary: When a crime has been committed, investigators must obtain as much and as accurate information as possible from witnesses. Initial incomplete interviews may pose a potential cost to unretrieved information. We explored retrieval-induced forgetting (RIF) in children’s autobiographical memory for related episodic events. Children experienced four related play sessions across 4 days, followed by word stem retrieval practice of half of the items, and a cued recall task that took place either 15 minutes or 2 hours later. Categorized details were experienced either within a single day or across multiple days. The emergence of RIF was dependent upon the length of delay between practice and test phases and the temporal distribution of the episodic events. Our observations demonstrate that RIF can be a subtle phenomenon whose occurrence depends upon whether the retrieval context supports an environment of retrieval interference. Copyright © 2012 John Wiley & Sons, Ltd. Imagine that a young child has been repeatedly abused by her father. During the subsequent investigation, the young child is questioned about her memory for the abusive incidents. The investigative interviewers understandably focus on the recall of details they deem most relevant to charging the father with a criminal offense. Because memory retrieval is focused on particular details of the abuse and is inevitably incomplete, the child is not asked to report details currently deemed irrelevant to the investigation or about details of which the investigator is unaware. However, a well-established phenomenon, called retrieval-induced forgetting (RIF), may compromise this strategy of retrieving only relevant details. Anderson, Bjork, and Bjork (1994) demonstrated that targeting specific memories for retrieval may result in greater difficulty when subsequently attempting to retrieve previously non-retrieved memories. Thus, the investigators’ attempts to access specific targeted details may make it difficult to recall other related details, details that could become a subsequent focus of the defense attorney. The child witness, by successfully retrieving a targeted subset of details and through no fault of her own, now may face a serious disadvantage that can be capitalized on by the defense. Retrieval of a targeted memory frequently occurs in a competitive retrieval environment. Related non-targeted memories may be highly activated alongside highly activated targeted memories. In order to efficiently retrieve these targeted memories, Anderson et al. (1994) argued that the targeted memory inhibits non-targeted memories, thereby enhancing the distinction between targeted and non-targeted memories. This distinctiveness facilitates easy retrieval of the targeted memory but at the cost of forgetting non-targeted memories on subsequent retrieval attempts. Anderson et al. (1994) developed a systematic method to study RIF. Their method involves four separate phases: study, practice, distraction, and test. The initial study phase involves partic- ipants studying category–exemplar pairs (e.g., FRUIT–apple, FRUIT–banana, DRINKS–scotch). Participants are then asked in the practice phase to retrieve and say aloud appropriate exemplars (e.g., say ‘apple’) given the word stems of category–exemplar pairs (e.g., FRUIT–ap____). This set of stimuli was referred to as the RP+ condition. In the practice phase, the stimuli used are half of the exemplars from half of the study phase categories. After a short distrac- tion task, participants enter the final test phase in which they are asked to recall all the exemplars previously presented during the study phase. As expected, recall is greatest for exemplars that were retrieved during the practice phase (e.g., FRUIT–apple; RP+ condition). More intriguing is that exemplars from practiced categories not retrieved during the practice phase (e.g., FRUIT–banana; referred to as the RPÀ condition) are less likely to be successfully recalled when compared with the baseline (i.e., unpracticed category and unpracticed exemplars; e.g., DRINKS–scotch). This baseline was referred to as the Nrp condition. This difference in recall rates between non-retrieved exemplars from practiced categories (i.e., RPÀ) and the baseline is the RIF effect (i.e., Nrp). Retrieval-induced forgetting is a robust phenomenon. Researchers have reported observing the signature marks of RIF in episodic (Anderson et al., 1994; Ciranni & Shimamura, 1999), semantic (Campbell & Phenix, 2009; Phenix & Campbell, 2004; Starns & Hicks, 2004), visuo-spatial (Ciranni & Shimamura, 1999), as well as in eyewitness (MacLeod, 2002) and autobiographical memory (Harris, Sharman, Barnier, & Moulds, 2010), and after short (Baran, Wilson, & Spencer, 2010; Chan, 2009; Saunders & MacLeod, 2002) and long delays (Ford, Keating, & Patel, 2004; Migueles & Garcia-Bajos, 2007) between practice and test phases. Thus, RIF has been reliably demonstrated in a variety of contexts. Already well-established in the laboratory, Shaw, Bjork, and Handel (1995) were the first researchers to extend the relationship between the RIF phenomenon to eyewitness testimony. Since their pioneering work, other researchers have examined RIF in the context of eyewitness mem- ory (Camp, Wesstein, & De Bruin, 2012; Garcia-Bajos, Migueles, & Anderson, 2009; MacLeod, 2002; Macrae & MacLeod, 1999; Migueles & Garcia-Bajos, 2007; Saunders & MacLeod, 2002). The implications of RIF for eyewitness memory are substantial. As argued elsewhere (Macrae & *Correspondence to: Thomas L. Phenix, Psychology, Campion College at the University of Regina, Regina, Saskatchewan, CANADA, S4S 0A2. E-mail: tom.phenix@uregina.ca Copyright © 2012 John Wiley & Sons, Ltd. Applied Cognitive Psychology, Appl. Cognit. Psychol. 26: 796–801 (2012) Published online 5 August 2012 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/acp.2861