Introduction

The present study explores children’s narratives of an autobiographical event and the affect of misleading suggestions from a trusted authority figure (i.e. parent) regarding that event. The purpose of the present research was to enhance applied childhood eyewitness protocol and gather additional knowledge of the underlying memory and cognitive processes from a developmental perspective. Furthermore, the goal the present study was to investigate both "system variables" (e.g., the way questions are phrased) and "estimator variables" (e.g., age of child).

The Current Study

The current study was based off an initial pilot study (Poole & Lindsay, 1995); however, the present study expanded age range and sample size in an attempt to see if past results could be replicated. Also, a body touch experience was added to the procedure.

The current study addresses nine main questions, which are as follows:

1. When interviewed with open-ended questions, what is the frequency of reporting false information due to parental misinformation during the period of development between 3-8 years of age?
2. When interviewed with open-ended questions, does continued prompting promote or hinder accuracy of information being reported?
3. When interviewed with direct yes/no questions, what is the frequency of reporting false information due to parental misinformation during the period of development between 3-8 years of age?
4. When prompting follows direct yes/no questions, does the "true" answer rely in the first yes/no response or the narrative that is the result of the prompting?
5. What is the developmental trend in source monitoring performance between 3-8 years of age?
6. Fictitious touching events vs. fictitious non-touching events…are responses to one more accurate than the other?
7. What is the relative stability of true and false reports across an interval with no further intentional memory contamination?
8. What individual differences correlate with suggestibility? Do descriptive variables (e.g., parental education) and personality characteristics (e.g., talkativeness) differentiate children who have high vs. low rates of false reports?
9. Do acquiescence {passive agreement; submission w/o protest; tendency to answer yes to everything no matter what it is ("Are you an alien?" "Yes")} recall, and memory source monitoring mediate the decline in suggestibility that is often observed b/w the early preschool and elementary school years?

Participants

They tested 3 through 7 year olds.

Procedure

The present experiment was divided into three sessions.

Session 1: (Baseline condition without influence from social cues.) Each child played in a room with Mr. Science (4 activities demonstrated) for 16 minutes and then a female interviewer interviewed the child. Five open-ended questions were asked. (Tell, More, Looked, Heard, and Think

Storybooks were mailed to families 3 months after session 1. Storybooks were based around the child’s experience with Mr. Science and the books had real events and fictitious events. Parents were to read the book to their child 3 times, preferable 3 consecutive days.

Conditions: experienced-only (in the room); experienced-heard (room & book); heard-only (book); control (neither room or book); touch-heard (book); touch-control (neither room or book).

Session 2: (aprox. 3.5 months later) Interviews took place in child’s home.

Direct questions yes/no response: 20 questions~10 pairs, one of each science demonstration (8 pairs) and two pairs of questions regarding body touch. (See pg. 30).

Source monitoring: (shift focus) First, the interviewer asked the child what they remembered about the story. Then, the interviewer asked direct yes/no questions about any events that the child had not mentioned. Following (shift focus back to the room with Mr. Science), the child was told about how some events in the book really didn’t happen and some did. The child was asked to identify sources of different events through yes/no responses to 20 possible questions (10 "story" questions <did it happen in the story?> and 10 "event" questions <did that really happen?>). The interviewer retrieved the storybook at session 2.

Session 3: (one month later) Interviews at sessions 3 were identical to session 2. All interviews were recorded via tape recorder. Different female interviewers collected data in sessions 2 and 3.

Results Results Results Tables Tables Tables

Coded for four categories: 1) accurate information; 2) detail error (i.e., wrong color for an object); 3) spontaneous intrusions (i.e., off-topic talk yet somewhat related); 4) suggested information (i.e., nonexperienced events described in the story).

Session One

Free recall was measured in 2 ways: as the proportion of events that the children mentioned during free recall from the various event conditions, and as the amount of accurate and inaccurate information that they reported. As expected, the proportion of events that the children mentioned in Session one increased steadily with age, F (1,108) = 92.30, MSE = .07, p < .001.

Open-ended Prompts

Most responses were coded as "new" (91%) SUs because each age group barely mention one event out of the total of four events. New SUs were then just referred to as SUs. Findings (See Table 1, pg.32): 1) children report increased dramatically with age; 2) DETAIL ERRORS DID NOT DECREASE WITH AGE; 3) spontaneous intrusion error did decrease significantly with age. It was found there was no significant impact of continued prompting, which means that the numerous questions used here did not bias or influencing later responses. (This is good! It proves that this is not a risky practice.) There was a large (20%) spontaneous intrusion rate for the younger children (3’s & 4’s) regarding the last think prompt. This could be due to the unclear nature and non-reiteration of the topic at hand.

Sessions Two & Three: Answers to the Nine Questions

Q1: 21% of kids in session 2 described events never experienced but introduced in the book. 10% in session 3. False report rate was reduced over time. Significant age trends for experienced-only and experience-heard demonstrations appeared. (See Table 2, pg. 33 and Table 3, pg. 34). ***No tendency for reports of nonexperienced events to decline w/ age.***

$$$ Data counter to previous claims ~ Older children less suggestible than younger? = HA! Children’s free recall highly accurate? = Double HA HA! $$$

Q2: Accuracy (See Table 4 and Table 5 pg. 35). Mean number of SUs increased dramatically across age groups in all three sessions and there was a significant session effect. Due to increased practice with interview experience and story reinforced memories of actual experience. Additionally, the amount of suggested information showed significant age trends and session effects. Even after a 3 to 4 month delay, there were few detail errors about experienced events with rare spontaneous intrusions. If kids erred, it was in reporting events that only appeared in the story.

Impact of Continued Prompting: (same as in session 1) Good for eliciting more information and details, yet it is not a risky behavior.

Q3: Responses to yes/no questions (See Table 6, pg. 36 and Table 7, pg. 37). Accuracy deteriorated. The interviewer tried to elicit a description regardless of answer. Increased age was associated with an increased number of "yes" responses about experienced events and a decreased number of "yes" responses to direct questions about nonexperienced events. Younger children are not just responding "yes" but are prone to it. Exposure to misinformation increased inaccurate "yes" responding.

Q4: Which is better…initial response of yes/no or elaborate narratives? (See Tables 8 & 9, pg. 39). SUs were divided into 6 categories: accurate information; wrong event; detail error; incorrect rejection; spontaneous intrusion; or suggested information. No significant differences were found when comparing narrative length to being able to distinguish between true and false events. Yes/no responses were more accurate than narratives. (Consider kids expanding on something that they already rejected by saying "no".)

Q5: Source-Monitoring (See Table 10, pg. 40; Tables 11 &12, pg. 41; Table 13, pg. 42 and FIGURE 1, pg. 43). Source-monitoring procedure failed to be of any benefit to 3&4 year-olds. Also, source monitoring decreased false report in older children without hindering accurate recall of real events. Lastly, a substantial minority continued to report false events even after several attempts by the interviewer to suppress such actions (i.e., telling the child that it was ok to respond "no").

Q6: Touching (See Tables 6 & 7 and Tables 2 & 3). Touch influencing information increases in younger children and decreases in older children over time. Touching events are not distinguished from other non-presented events. This may be due to the socially acceptable way the questions were formed and the physical acts involved (not traumatic). Additionally, the touch-heard condition refers to an experience in the story, which may not be as memorable as an actual perceptual experience.

Q7: Stability of True and False Reports (Roughly, accurate reports = stable; Inaccurate reports = unstable). Allegations that drop out of children’s reports…be suspicious, likely to be false. Hence, it would be informative to conduct a second interview and look for consistency of the reports over a short period of time (minus what is naturally forgotten). BE CAREFUL: Some false reports are stable and some true reports are unstable.

Q8: Individual Differences: There was no consistent relationship b/w demographic characteristics and suggestibility. None of the individual difference variables predicted final source-monitoring performance. The only factor that was of any consequence was age (with age related trends as mentioned above).

Q9: Acquiescence, Recall, and source monitoring (See Table 14, pg. 45, FIGURE 2, pg. 46). Acquiescence=children who responded "yes" to control events had higher rates of false positives. Recall=recall improved while suggestibility increased. ***Better memory of real events could lead to better memory of false events, hence a greater probability for meshing them together.*** However, recall was also associated with better source monitoring which reduced suggestibility.

Discussion

~ The 8 year-olds were as likely as the younger kids to mention fictitious events in free recall. Why? Superior verbal skills and a desire to show the interviewer knowledge.

~Continued prompting did not increase reports of false information.

~Several reasons to ban yes/no questioning (feel pressure to answer, often immediate response wrong, change answer based on social cues); however, little evidence to supports these fears. In fact, the other studies have found that children respond quite accurately to yes/no questions, even very young children (Saywitz, Goodman, Nicholas, & Moan, 1991).

~The present study found that children’s responses to yes/no questions were alarming inaccurate, except for touch.

~Children will offer incoherent narratives just to keep consistency with a yes/no response.

~Source monitoring dramatically improves after age 4. (Younger children rarely reject the validity of what they said earlier.)

~Age remains the best overall predictor of accuracy.

~Older children were generally less suggestible than younger children and better able to identify the source of information

~False reports of unpleasant touch events were as other false reports regarding the science experiments.

~The authors feel that their recall measure yielded counterintuitive relationships with suggestibility and acquiescence and source monitoring, which all contribute to false report. (See Figure 2).