The orthographic distinctiveness effect on direct and indirect tests of memory: delineating the awareness and processing requirements

We examined the processing and awareness requirements that mediate superior memory for distinct items in long-term memory by studying the effects of orthographic distinctiveness. Orthographically distinct words are remembered better than common words on the direct test of free recall but not on the indirect test of perceptual identification. These results suggest that the orthographic distinctiveness effect requires test awareness. But interestingly, this distinctiveness effect has also been reported in one indirect test, that of word fragment completion (Hunt & Toth, 1990). We examined the locus of the orthographic distinctiveness effect in the indirect test of word fragment completion and the direct tests of word fragment cued recall and free recall and assessed the role of awareness (Experiment 1) and conceptual processes (Experiments 2a–2c) in mediating this effect. Our results support the proposal that the distinctiveness effect depends on direct reference to the study context and further specify that this effect is mediated, to a large extent, by comparative processes even when distinctiveness emerges from surface-level differences. 2002 Elsevier Science (USA). All rights reserved. Several studies attest to the intuitive belief that people have very good memory for novel or unusual information, often called the distinctiveness effect. The von Restorff effect (superior memory for isolated items) is cited most often to demonstrate that distinctiveness aids memory (see Hunt, 1995, for details of von Restorff’s dissertation). However, the term ‘‘distinctiveness’’ has also been applied to a variety of other phenomena, including superior memory for nude pictures (Ellis, Detterman, Runci, McCarver, & Craig, 1971), bizarre sentences (McDaniel, Dunay, Lyman, & Kerwin, 1988), distinct faces (Light, Kaya-Stuart, & Hollander, 1979), atypical category members (Schmidt, 1985), and words with unusual orthographies (Hunt & Elliott, 1980; Hunt & Mitchell, 1978, 1982; Hunt & Toth, 1990). Despite an abundance of studies demonstrating that unusual or distinct information is remembered well, it remains unclear exactly how distinctiveness aids memory (see Schmidt, 1991). Journal of Memory and Language 47 (2002) 273–291 www.academicpress.com Journal of Memory and Language This research was done in partial fulfillment of the first author’s doctoral dissertation and was supported by NIH Grant R29MH57345-01 given to the second author. * Corresponding author. Present address: Department of Psychology, Washington University, One Brookings Drive, Campus Box 1125, St. Louis, MO 63130-4899, USA. E-mail address: [email protected] (L. Geraci). 0749-596X/02/$ see front matter 2002 Elsevier Science (USA). All rights reserved. PII: S0749 -596X(02)00008 -6 The following experiments are designed to address this question by exploring the nature of the processing at encoding and retrieval that mediates the distinctiveness effect in long-term memory. Experiment 1 examines the necessary retrieval conditions for obtaining the distinctiveness effect and tests whether the effect can be obtained using an indirect test of memory. Experiments 2a–2c explore the nature of the processing at encoding that could explain the effect of distinctiveness on memory tests that differ in their processing demands. We explore these issues using one particular type of distinctiveness effect, the orthographic distinctiveness effect. This effect refers to the finding that words with unusual letter combinations, such as ‘‘subpoena,’’ are remembered better than words with more common letter combinations, such as ‘‘sailboat’’ (Hunt & Elliott, 1980; Hunt & Mitchell, 1978, 1982; Hunt & Toth, 1990; Zechmeister, 1969, 1972). We chose to study this distinctiveness effect because the reported pattern of data on orthographic distinctiveness has important implications for two influential theories of memory: the distinctiveness hypothesis advanced by Hunt and colleagues (Hunt, 1995; Hunt & McDaniel, 1993; Hunt & Smith, 1996; Smith & Hunt, 2000) and the transfer appropriate processing framework advanced by Roediger and colleagues (Roediger, 1990; Roediger, Weldon, & Challis, 1989). The findings of central interest come from research examining the effect of orthographic distinctiveness on direct (or explicit) and indirect (or implicit) tests of memory (Hunt & Toth, 1990). This research shows that orthographically distinct words, such as ‘‘subpoena,’’ are remembered better than orthographically common words, such as ‘‘sailboat,’’ using a direct test of free recall (Hunt & Toth, 1990). This work also shows that the advantage of completing fragments (e.g., s_ _l b_a t) for studied words compared to nonstudied words, a measure termed ‘‘priming’’ (Graf & Schacter, 1985; Tulving, Schacter, & Stark, 1982), is greater for orthographically distinct words than for orthographically common words on an indirect test of word fragment completion (Hunt & Toth, 1990). Interestingly, although orthographically distinct words show an advantage on the indirect test of word fragment completion, they do not show such an advantage in priming on another indirect test of perceptual identification, in which participants are required to identify words that are flashed briefly on a screen (Hunt & Toth, 1990). We first consider the implications of these findings for the transfer appropriate processing framework (Roediger et al., 1989). According to this framework, performance on both direct and indirect memory tests benefits to the extent that the conceptual or perceptual processing at test recapitulates the same conceptual or perceptual processing that occurred at study. Much research supports this view (see Roediger & McDermott, 1993 for a review). For example, perceptual manipulations involving changes in modality of presentation (visual versus auditory) at study and test influence performance on indirect tests, such as the fragment completion test, which rely on the perceptual features of the stimulus (Blaxton, 1989; Graf & Mandler, 1984; Jacoby & Dallas, 1981; Rajaram & Roediger, 1993; Srinivas & Roediger, 1990), but not on indirect tests that rely on the meaning of the stimulus (Blaxton, 1989; Challis & Sidhu, 1993; Srinivas & Roediger, 1990). Conversely, conceptual manipulations, such as levels of processing where participants study words either for their semantic or phonemic properties, influence indirect tests that rely on meaning, but not perceptually driven indirect tests such as word fragment completion (e.g., Blaxton, 1989; Hamann, 1990; Srinivas & Roediger, 1990). Performance on direct tests also demonstrates the importance of the match in processing. For example, free recall, which relies, for the most part, on generative or elaborative processes, is influenced by conceptual manipulations, such as the levels of processing (Craik & Tulving, 1975) and the generation of an item (Jacoby, 1978; Slamecka & Graf, 1978) but not by perceptual manipulations, such as changes in modality (Rajaram & Roediger, 1993). While most direct tests of memory are largely driven by conceptual processing, some direct tests, such as fragment cued recall, in which the percept is partially available at test, are influenced by both conceptual (Roediger, Weldon, Stadler, & Riegler, 1992) and perceptual processes (see Blaxton, 1989; Challis & Sidhu, 1993). The reported pattern of the orthographic distinctiveness effect across various memory tests (Hunt & Toth, 1990) is inconsistent with the transfer appropriate processing framework because it shows the effect of an ostensibly perceptual variable such as orthography on a conceptual test but not a perceptual test. For example, orthographic distinctiveness influences free recall, which is considered to be a conceptually driven test, but not perceptual identification, which is 274 L. Geraci, S. Rajaram / Journal of Memory and Language 47 (2002) 273–291