figure 9.4. Developmental trends in false memory for distractors that are semantically related to list targets and developmental trends in false memory for distractors that are phonologically related to list targets, as reported by Holliday and Weekes (in press).
were extended downward to include children as young as Dewhurst and Robinson's (2004) youngest participants, perceptual illusions would be more seductive than semantic illusions.
This article answers an important question: are age increases in false memory for semantically related words observed for procedures other than the DRM paradigm? Of course, FTT's analysis of false memory would expect age increases in other tasks for which the ability to connect meaning across different words supports false recall/recognition and for which it is difficult to use recollection rejection to suppress false memories. In the research that Howe reports, children's recall of DRM lists was compared to their recall of another familiar type of list: categorized lists. Whereas DRM lists are generated by selecting the first few associates of critical distractors, categorized lists are constructed by selecting the most common exemplars from production norms for familiar taxonomic categories (e.g., animals, flowers, furniture; see, e.g., Battig & Montague, 1969). In Experiment 1, a total of 180 children were tested from each of three age levels (5-, 7-, and 11-year-olds). Half the children at each age level recalled eight DRM lists, and the other half recalled eight categorized lists. The important new finding, as would be expected on theoretical grounds, was that false recall increased with age (from 0.16 to 0.27 to 0.32) for both categorized lists and DRM lists. Further, false recall for the two types of lists did not differ in reliability, and their age trends were the same. Thus, age increases in false recall did not seem to depend on whether false-memory items are associatively related or taxonomically related to targets.
Howe's (2006) second experiment incorporated an additional novel design feature, gist cuing. A further sample of 180 children, evenly divided with respect to the same three age levels, studied and recalled only categorized lists. However, they did so under one of two conditions: category cuing or no cuing. In the former condition, children were told the name of the category just before each list was read (e.g., "the words that you are about to hear are all names of animals"), whereas children in the latter condition were told nothing. FTT predicts that if limitations in younger children's meaning-connection abilities contribute to age increases in false memory, false memory should be higher in the cuing condition. It was not, which does not support the prediction. However, this is a null effect, and such null results are always difficult to interpret, as we mentioned earlier in connection with studies of the suggestibility of children's memories of traumatic experiences involving their bodies. In this particular instance, other data suggest that the null effect may be a Type 2 error (i.e., a failure to reject a false null hypothesis). Evidence for this possibility may be found in Experiment 2 of Brainerd et al. (2004). The same cuing procedure was used in connection with false memory for categorized lists, and in line with FTT's prediction, children displayed more false memory following category cues than following no cues. There was also an important procedural difference: false memory was measured with recognition rather than recall tests. Further evidence that, as expected on theoretical grounds, gist cuing increases children's false memories and that this effect is not confined to recognition can be found in DRM experiments by Brainerd et al. (2004) and by Holliday, Reyna, and Brainerd (2006). In both studies, the authors cued the meanings of DRM lists just before each list was presented for half the children and provided no pre-presentation meaning cues for the other half of the children. Brainerd et al.'s children (6-, 11-, and 14-year-olds) showed elevated levels of false recall of critical words in the gist-cuing condition relative to the no-cues condition. Likewise, Holliday et al.'s children (9-, 11-, 13-, and 15-year-olds) displayed higher levels of false recall and false recognition of critical distractors in the gist-cuing condition relative to the no-cues condition.
Howe's (2006) final experiment incorporated the further design modification of pictorial presentation. As Schacter, Israel, and Racine (1999) noted, FTT predicts that, other things being equal, false memory should be reduced by any manipulation that enhances verbatim memory for individual targets, while leaving gist memory unaffected. Howe pointed out that pictorial presentation of lists is often considered to be just such a manipulation. Moreover, considering that the false-memory responses that are made for lists that repeatedly cue the same meanings, whether categorized or DRM, have been consistently found to be more fragile in children, verbatim-strengthening manipulations would be expected to have especially marked effects in children. To evaluate these ideas, categorized lists were again studied and recalled by 5-, 7-, and 11-year-old children. The lists were again read aloud to children, but now each word was illustrated by a picture of the named object as it was read. As expected on theoretical grounds, this presentation method suppressed false recall of unpresented exemplars to very low levels (below 20% at all age levels), and it completely eliminated the age increases that have been so consistently obtained with oral presentation (see also Ghetti et al., 2002).
Brainerd et al. (2006)
These investigators asked whether developmental increases in the DRM illusion were more than quantitative by studying possible qualitative differences between the false memories of children and adults. They also measured these increases in a new way by implementing an ability-based definition of "development" With respect to qualitative change, Brainerd et al. studied whether children exhibit any or all of five effects that have been obtained in adult DRM studies: (1) list strength (some DRM lists produce high rates of false recall and false recognition, whereas other lists produce low rates); (2) recall inflation (on an immediate memory test, false memory for critical distractors is higher if participants have responded to a prior recall test than if they have not) and delayed inflation (on a delayed memory test, false memory for critical distractors is higher if participants have responded to an immediate recall test than if they have not); (3) delayed stability (when memory tests are postponed for a few days, true memory levels decline, but false memory levels for critical distractors remain relatively stable); (4) thematic intrusion (false recall is dominated by intrusions that are consistent with list themes); and (5) true-false dissociation (true recall and true recognition of DRM targets correlate negatively with false recall and false recognition of critical dis-tractors). To the extent that children display all of these effects, Brainerd et al. proposed, the indicated conclusion is that children's reduced levels of false memory are not due to qualitative changes in the nature of the DRM illusion, whereas the conclusion that qualitative change is involved is forced to the extent that children do not display these effects. Two experiments were reported, in which the children were between the ages of 7 and 11 and in which both recognition and recall tests were used to measure false memory for critical distractors. None of the six effects was observed in the youngest children, and five of them were found to emerge with age, all of which is consistent with the qualitative-change view.
In the second of their two experiments, Brainerd et al. (2006) compared developmental trends in false memory for two different definitions of development: the usual chronological-age definition and an ability-based definition that was based on children's school performance. Participants came from two age levels, 7- and 11-year-olds. Within each age level, children of two ability levels were included: children whose school performance was within the normal range and children who had been classified as learning disabled (but whose measured IQs were in the normal range). Learning-disabled children provide a further test of FTT's hypothesis that younger children's meaning-connection limitations contribute to age increases in false memory because learning-disabled children's limitations in this area are even greater (e.g., Swanson, 1991). The children studied and recalled DRM lists in the usual way. Nondisabled children displayed the usual developmental increases in false memory, with intrusions of critical distractors increasing from 0.16 in younger children to 0.30 in older children. The ability measure produced the same picture of developmental increases within chronological age levels. The intrusion levels were 0.16 and 0.06, respectively, for nondisabled and disabled 7-year-olds, and the corresponding intrusion levels for nondisabled and disabled 11-year-olds were 0.30 and 0.19, respectively. Although false memory increased with chronological age in both nondisabled and disabled children, the contribution of ability to false memory is dramatically illustrated by the fact that the intrusion rate for disabled 11-year-olds (0.19) was not reliably higher than the intrusion rate for nondisabled 7-year-olds.
Brainerd and Reyna (2007)
This research was aimed at securing stronger tests of the theoretical ideas that have dominated our discussion of false memory in the narrative and word tasks. According to those ideas, predicting the direction of developmental trends in false memory turns on three considerations: (1) that participants store and retrieve dissociated verbatim and gist traces of target materials; (2) that verbatim and gist retrieval are opponent processes in false memory (i.e., verbatim retrieval suppresses false memories while gist retrieval supports it); and (3) that developmental lags exist in extracting the semantic information from target materials, especially semantic relations among events that share familiar meanings. As we have seen, this leads one to expect age increases in false memory in situations in which false-memory items tap meaning relations that are repeatedly cued by target materials (as in the sentences that compose narratives, DRM lists, or categorized lists). Although this expectation has been consistently confirmed, an even stronger prediction follows: the same unpresented items should exhibit opposite age trends, depending on whether the target material allows semantic relations to be formed for those words. Imagine that children study a single long list of words whose meanings are all familiar. For instance, suppose that the words are exemplars of everyday categories, such as colors, flowers, foods, and the like. Further, imagine that the list exemplifies some category once to half the children (e.g., the color category is exemplified by red), but the list exemplifies that same category multiple times to other children (e.g., color is exemplified by black, blue, brown, green, pink, purple, red, and white). Finally, suppose that a recognition test is presented on which the related distractors color and yellow appear. The theoretical analysis makes the surprising prediction that these distractors will display opposite age trends for the two groups of children: (a) the false-alarm rate should decrease with age if the list contained only a single color exemplar because developmental variability in false alarms to unpresented exemplars will be chiefly controlled by improvements in verbatim suppression, and (b) the false-alarm rate should increase with age if the list contained a large number of color exemplars because developmental variability in false alarms to unpresented exemplars will be chiefly controlled by improvements in the ability to connect the gist across target words.
We tested these predictions in experiments with 6-, 10-, and 14-year-olds. Children listened to two different lists and responded to a recognition test following each list. Each list contained exemplars of taxonomic categories that are very familiar to children. For three of the categories (e.g., animals, colors, men's names), the list contained eight exemplars, and for the other three categories (e.g., clothing, flowers, furniture), the list contained a single exemplar. On the test list, there were two false-memory items for each category: the category label and an unpresented exemplar (e.g., color and yellow). The prediction was that these distractors' false-alarm rates should increase with age if eight color exemplars had been presented but decrease with age if only a single color exemplar had been presented. The results are plotted in Figure 9.5, where it can be seen that, indeed, false-alarm rates increased with age when categories had been cued eight times but decreased when they had been cued once.
We also tested a control sample of children from the same age range to determine whether it was semantic relatedness in particular that was responsible for the contrasting developmental trends in Figure 9.5. These children participated in an experiment that used the same design, except that phonological relatedness rather than semantic relatedness was the
Was this article helpful?
Parenting is a challenging task. As a single parent, how can you juggle work, parenting, and possibly college studies single handedly and still manage to be an ideal parent for your child? Read the 65-page eBook Single Parenting Becoming The Best Parent For Your Child to find out how. Loaded with tips, it can inspire, empower, and instruct you to successfully face the challenges of parenthood.