While many studies have asked and attempted to answer the question of what causes human beings to remember certain things better than others, and certain variables have been isolated as having an effect on memory, the relationships between these variables is sometimes overlooked. Thus, this research examined the effects of concreteness, word ambiguity, and type of task on cued recall performance and the possible interactions between these variables. This research is based on the PIER model (Processing Implicit and Explicit Representations) which examines two processing methods, that of explicit intentional processing and that of implicit automatic activation of relevant information (Nelson, Schreiber, &McEvoy, 1992). PIER attempts to explain the effects of variables that relate to the underlying associative structure of words. It theorizes that everything associated to a word is activated, then irrelevant items are inhibited by context. PIER was developed based on non-ambiguous stimuli, but has been applied to ambiguous stimuli as well (Gee, 1997). Thus, one can predict that words with only one meaning (non-ambiguous) will be recalled better than words with more than one meaning (ambiguous). Also, PIER predicts and previous research has shown that, in general, targets presented in an intralist task are recalled better than those presented in the extralist task (Nelson, Schreiber, &McEvoy, 1992).
The Dual Code Theory (DCT) was developed to explain concreteness effects (Paivio, 1971, 1983). DCT postulates that when we attempt to retrieve and/or store information we use two types of processing and the result is two types of codes. Information is either encoded in a verbal (linguistic) or nonverbal (imaginal) code. DCT theorizes that concrete words are recalled more readily because they are encoded in both the verbal and the imaginal code, whereas abstract words are only encoded in the verbal code. In other words, at retrieval the concrete words have an advantage because there is more than one way to retrieve the word. Thus, we had reason to hypothesize that there would be a main effect for concreteness, such that concrete word would be recalled better than abstract words. We also expected effects based on word ambiguity, and type of task, as described above. Of particular interest was the concreteness effect for ambiguous and non-ambiguous words. The issue under investigation is whether the concreteness effect is the same for ambiguous words as it is for non-ambiguous words. Given that ambiguous words have more than one meaning, it is possible that the concreteness effect might be different depending upon which meaning was accessed.
The experiment was designed as a 2x2x2 mixed model factorial. Type of task (intralist or extralist) was tested between subjects, while concreteness (concrete or abstract words) and ambiguity (ambiguous versus non-ambiguous) were manipulated within subjects. The dependent measure was accuracy and timing of cued recall performance. Participants were 48 State University of New York, College at Fredonia students from the Introduction to Psychology classes. They received extra credit for their participation.
Two word lists of 24 cue and target word pairs were developed from a database of single response association norms (Nelson, Schrieber, & McEvoy, 1999). This normative data allowed for the control and manipulation of various characteristics known to effect cued recall performance. The words were carefully screened so that none of the cue words were related to any other words but the intended target. Also, the 24 targets used for each list were created from equal numbers of words, which were highly concrete or highly abstract and half of each of those types of words had only one meaning and the other half had more than one meaning. The following characteristics were controlled for the target words: target set size, connectivity, concreteness and frequency. The cues and the cue-target connections were controlled for the following variables: back strength, forward strength, cue set size, connectivity, concreteness, and frequency. A SUPERLAB computer program equipped with a microphone was used to present words and word pairs and to record participants' timed voice responses, while data concerning accuracy of responses was recorded by hand by the experimenter.
Subjects were randomly assigned to one of the between subjects conditions, and were tested individually within a 15 minute time frame. The participants were told they would see a list of words (extralist task) or word pairs (intralist task) and would then be asked to recall the target words when provided with a meaningfully related cue word. After signing the consent form, participants were led through a practice session to allow them to adapt to the 3 second presentation rate and to the sensitivity level of the microphone. Then the actual experiment began. Twenty-four words or word pairs were presented, followed by their related cues, and participants were then asked to recall the target word and to say it aloud into the microphone. The experimenter then recorded the accuracy of the responses, while the computer recorded the reaction time of the voiced responses.
A one-between two-within mixed model ANOVA was performed on the data. Results demonstrated that when using an alpha level of p < .05, three main effects were significant. The effect of word concreteness demonstrated that concrete words (mean proportion of correct recall = .68, SD = .24) were recalled better than abstract words (M = .59, SD = .26), F(1,54) = 16.42. Also, nonambiguous words (M = .68, SD = .25) were recalled better than ambiguous words (M = .6, SD = .26), F(1,54) = 185.41. Finally, words in the intralist task (M = .8, SD = .19) were recalled better than word in the extralist task (M = .48, SD = .21) F(1,54) = 88.88. No other effects were significant (p > .05).
The three significant main effects found in this experiment (concreteness, word ambiguity, and type of task) replicated previous findings and were also consistent with the predictions of the PIER model and of DCT theory. The purpose of this experiment was to examine the relationship of these variables in the cued recall task. The data presented here indicates that they each have independent effects on performance. However, the possibility of an interaction between concreteness and ambiguity remains intriguing. Although, this interaction was not significant in this experiment, there was a tendency toward thin interaction. Closer examination of the data reveal that it is possible that the concreteness effect is actually larger for non-ambiguous words than it is for ambiguous words. Further data collection is currently underway to more closely examine this possibility. There are a number of possible reason for why this interaction was not reliable in this study and further research is necessary to flesh this issue out.
References
Nelson, D.L., Gee, N.R., & Schreiber, T.A. (1992). Sentence encoding and implicitly activated memories. Memory & Cognition, 20, 643-654.
Paivio, A. (1971). Imagery and Verbal Processes. New York: Holt, Rinehart and Winston.
Paivio, A. (1983). The empirical case for dual coding. In J. Yuille (Ed.), Imagery, memory, and cognition: Essays in honor of Allan Paivio (pp. 307-332). Hillsdale, NJ: Erlbaum.
Gee, N.R. (1997). Implicit memory and word ambiguity. Journal of Memory and Language, 36, 253-275.