DIFFERENCES IN ADULT AND CHILD FALSE MEMORIES BASED ON DIFFERENT TYPES OF ASSOCIATED WORDS

Vincent Smith Marshall University. ...................................................................................................................... Manuscript Info Abstract ......................... ........................................................................ Manuscript History Received: 05 March 2020 Final Accepted: 07 April 2020 Published: May 2020 Copy Right, IJAR, 2020,. All rights reserved.

10 seen in semantic and phonological memories. Fuzzy-trace theory proposes that there are two main different types of memory processes: verbatim and gist memory.Verbatim memory is closely linked to visuospatial memory where gist memory is more closely related to ideas and episodic memory. Previous research suggests that verbatim memory is more susceptible to making an error in phonologically and orthographically associated lists while gist memory is more susceptible to making an error in semantically associated lists (Obidziński&Nieznański 2017).
False memory studies that directly compared children to adults often times found significant differences. One study looked at why children and adults were susceptible to semantic and phonological false memories and found that phoneme awareness negatively predicted children's false recall of phonological lures (Mcgeown et al., 2014). This is surprising when considering the findings from another study demonstrated that as age increased, so did recall of the critical lure. Also, as age increased, other erroneous recalls decreased (Mcguire, London, & Wright, 2015). This was also confirmed by another study in which both younger and older children had higher numbers of spontaneous false memories than the adults, but they had lower accuracy than the adults. The DRM paradigm showed the standard results that older children and adults were more likely to produce false memories than the younger children (Otgaar et al., 2013).
Although most studies examine phonologically similar and semantically similar words, one study in particular that looked at orthographically similar words and how the creation of false memories in those words relate to dyslexia in childrenprovided evidence that impairments of memory processes were not shown for both verbatim and gist memory.
It also demonstrated that orthographically similar words were affected by dyslexia(Obidziński&Nieznański, 2017).
Previous research demonstrated that lists with all phonological associates or all semantic associates have the lowest proportion of false recall while lists with half semantic and half phonological associates have the highest proportion of false recall, and interestingly, the largest increase in false recall is from just adding a single word of the other type of associate (Ex. A list of words semantically associated to chair, such as desk and sit, improved dramatically when just one phonological associate (hair or cheer) was added (Finley et al., 2017).Although previous studies have examined semantic, phonological, and orthographical associates in some combination, there has not been a study that demonstrates the differences in these three associates in the same study. The present study is developed to test the relationship of phonologically, orthographically, and semantically associated words and false memories at two different stages of life (childhood and early adulthood).

Participants:
This study receivedIRB approval prior to collecting any data.The participants were 35young adults (18-25 years old) recruited from Marshall University and 65 underage participants(9-11 years old) recruited from local elementary schools. Participants under the age of 18 had a consent waiver signed by a parent or legal guardian and gave verbal and signed assent the day of the experiment. Participants over the age of 18 gave verbal consent the day of the experiment.Due to inconsistencies with answering (ex. answered yes to recognized the word, but marked a 1 indicating the word was not on the previous list), scores of seven of the underage participants were removed. Participants from the two local elementary schools had a mean age of 9.52 with a standard deviation of 0.599. Participants from Marshall University had a mean age of 18.89 with a standard deviation of 1.278.

Procedure:
For the all participants, DRM tasks were given (Pardilla-Delgado& Payne, 2017). The participants were presented with a list and were told this is a study list and they needed to try to remember as many words as possible from the list. Thefirst study list presented was a semantically associated DRM task with 14 words and participants were given 1 minute to study. The participants were then asked to recognize the words from a recognition list and list their certainty (1 = certain not on the list to 5 = certain it was on the list). The list contained the 14 words from before, a false recognition target, and 3 unrelated controls. The participants then repeated this DRM task for a phonologically (defined as rhyming or homophone for this study) associated list, and an orthographically associated list (defined as differing by only one letter for this study). The false targets for the three DRM tasks based on semantically, phonologically, and orthographically similar words were tree, knead, and bet. Full lists can be found in the appendix. An important part of this study is that the orthographically associated false target was not phonologically associated or semantically associated with any words on the study list, the phonologically associated false target was 11 not orthographically or semantically associated with any words on the study list, and the semantically associated false target was not phonologically or orthographically associated with any words on the list.

Results:-
A 2x2x3 repeated measures ANOVA was conducted to compare the effects of gender and age (adult or child) on when the false target was semantically, phonologically, and orthographically associated.Levene's test was violated so a significance level of .01 was used instead of .05. There was a significant difference in type of associate, F(2, 172) = 43.831, p < .001, partial η2 = .338. There were no significant interactions or other significant main effects. Post-hoc tests revealed the rating of certainty that thesemantically associated false target appeared, rating of certainty that the phonologically associated false target appeared, and rating of certainty that the orthographically associated false target appeared significantly differed from one another.

Implications:
There is a significant difference in the type of associate that was falsely remembered as expected from previous studies. The average level of certainty of orthographically related target was about "pretty certain this was on the list" while the average level of certainty of phonologically related target was near "pretty certain this was not on the list" and the average level of certainty of the semantically related target was between "pretty certain this was not on the list" and "not certain either way". This provides evidence that there should be more investigation into the differences in effects of orthographical false targets and phonological false targets and they should not be treated as the same thing. The difference between the phonologically related target and semantically related target may be statistically significant, but it is not practically significant because both are fairly close to "pretty certain this was not on the list" on average. If a recognition list were given, this could suggest that orthographically similar words may have some impact in past studies when they published the interaction as a phonologically associated word.
Another interesting result is that since fuzzy-trace theory suggests that verbatim memory is more susceptible to making an error in phonologically and orthographically associated lists while gist memory is more susceptible to making an error in semantically associated lists (Obidziński&Nieznański 2017), we may expect that the certainty of the orthographically associated and phonologically associated false targets would be similar. However, that is not the case. The phonologically associated false target had a low certainty level of less than 2, suggesting that most individuals indicated they were pretty sure this word was not on the list while the orthographically had a high certainty level of almost 4, suggesting that most individuals indicated that they were pretty sure the word was on the 12 list. This suggests that the divide between verbatim and gist memory when doing a DRM task may not be so straightforward and should be investigated further.
The lack of a significant difference in the adult and child group is not very surprising since the children included in this study were ages 9-11. A previous study (Otgaar et al., 2013) broke children into two groups and found developmental differences in younger (6-8 years old) and older children (10-12 years old) but found no difference in older children and adults.

Limitations/Future Studies:
There are a few limitations to this study that should be acknowledged. To obtain words that are phonologically similar, but not orthographically similar and vice-versa, the following definitions were used: phonologically similar words must rhyme or be homonyms (Johnson, n.d.) and orthographically similar words must differ by one letter (Love, 2012). Although these definitions are often different in different studies, this allowed the DRM tasks to test phonologically and orthographically similar words separately which very few studies have done in the past. Further studies could focus only on these differences and develop a slightly modified definition of phonologically and orthographically similar words.
This study looks at children ages 9-11 and young adults ages 18-25, but it does not look at the rest of the lifespan. There is very little research on false memories in middle adulthood (about age 35-60). This study also did not include anyyounger children (ages 6-8) or older adults; however, these populations are studied much more frequently.
Another limitation is, since the sample size was limited, there may be a counterbalancing issue in the order that the different types of associates were presented. With every group, the word lists were presented in order of semantic, phonologic, and then orthographic associated lists.
There should be a comparison done with a researcher reading the study lists aloud and the participants doing a free recall. The fact that participants visually studied the words could be a reason for the orthographically associated target having such a higher mean of certainty that it appeared. By this logic, there could possibly be similar results in favor of the phonologically associated target in such an experiment.