Word Length Effect In Mixed Lists

The word length effect is the finding that the serial recall of lists of long words is worse than the recall of lists of short words. Whether is it based the articulatory duration (the phonological loop model) or the phonological complexity (the feature model), there has been a long duration debates. The results in Hulme et al (2004)'two experiments can't be explained by the two supposes, they explained it use the SIMPLE, which attribute the better recall of the mixed lists to a third factor, the class dimension. But their results may be contaminated by Guessing, for the fixed orders in the alternative mixed lists, and their random mixed lists also can be contaminated, for it can be benefit from the alternative lists by studying and other factors. we tested their results in the first experiment, then use the single random list in the experiment 2, and use an non-block design in the experiment 3.53 undergraduate students (16 in experiment 1, 22 in experiment 2, and 15 in experiment 3) participated in the study. Using Chinese characters as material, they were matched with frequency, the short items were all two-syllable and the long items were all four-syllable. Participants were told it is forbid to use the first-letter strategy in Experiment 2 and Experiment 3. The serial position reconstruction tasks were used in the three experiments.We compared the alternative mixed lists with the pure lists in the first experiment. In experiment 2 we used the random mixed lists. In experiment 3, we randomly displayed the pure lists and the alternative mixed lists, a non–Block method, and use the random mixed lists to control the guessing strategy. The repeated ANOVA were used to analysis the three experiment results.There is no significant difference between the pure short lists and the mixed lists in experiment 1 and 2, with the long items in the mixed list were better recalled than the pure long lists, no difference for the short items whether in the pure lists or in the mixed list. But in experiment 3, the short items in the alternative lists were better recalled than the pure short lists, and a significant word-length effect were also observed in the mixed lists.Our results can't be explained by the phonological loop model or the feature model; the SIMPLE can explained the first two results, but failure to explain the third experiment results. We believe that the guessing strategy must displayed in our first experiment, but more important, there is a special factor which cause the mixed lists get better recalled. We suppose the order information were coded by two ways in the mixed lists, using the temporal information and the grouping operations, the grouping operations dose not acts in the pure lists, which improve the mixed lists recall in the special tasks. We think the grouping operations take actions on the mixed lists like the syntax for the sentence, such suppose may be useful for bridge the research between the speech production and the short term memory.