Young children’s difficulty with deception in a conflict situation

Development of Deception

Children’s deceptions and lie-telling behavior have been studied a lot in recent decades. One frequently used method is the resistance to temptation. In this paradigm, children are instructed not to peek at a toy after the experimenter has left the room. Lewis, Stanger, and Sullivan (1989) found that 38% of 3-year-olds who peeked at a forbidden toy denied peeking. Recent studies have replicated these results, showing that about one-third of 3-year-olds denied peeking, and the majority of children between 4 and 7 years of age lied (e.g. Polak & Harris, 1999; Talwar & Lee, 2002).

Another widely used method involves hiding games. In these games, children are required to use more advanced strategies that intentionally manipulate their opponents’ beliefs. For example, Sodian, Taylor, Harris, and Perner (1991) presented situations in which a competitor or a cooperator was looking for treasure. They found that 3-year-olds wiped out tracks to the treasure for both the competitor and the cooperator, whereas 4- and 5-year-olds selectively deceived only the competitor by wiping out the tracks and reinforcing the tracks for the cooperator. Furthermore, Sodian (1991) and Seno (2008) showed that 4- and 5-year-olds could tell a lie to a competitor but provided accurate information about the location of treasure to a cooperator. Kikuno (2010) also found that 5-year-olds gave false information to an enemy puppet but provided accurate information to a friend puppet. Peskin (1992) reported that 87% of 5-year-olds lied about the location of a prize to a bad puppet in order to receive the prize themselves, whereas only 29% of 3-year-olds lied.

These findings suggest that children’s deception or telling of lies starts to emerge in the preschool years. Some children as young as the age of 3 years tell lies, but these involve deception by denying, simply to avoid potential punishment (Talwar & Crossman, 2012). By the age of 5 years, children can deceive by falsifying, that is, consistently misleading an opponent by producing a false belief. This reflects their development of a theory of mind, which is defined as the ability to attribute mental states to oneself and others and to recognize that others’ mental states may differ from one’s own (e.g. Wimmer & Perner, 1983). Many studies have shown a relationship between deception and theory of mind (e.g. Newton, Reddy, & Bull, 2000; Talwar & Lee, 2008). Young children pass the false-belief task at the age of 4 or 5 years (e.g. Wellman, Cross, & Watson, 2001); therefore, deception by falsifying also emerges at this age (e.g. Peskin, 1992; Sodian, 1991). By contrast, in order to deceive by denying, it is not necessary to produce a false belief in the opponent. Therefore, deception by denying is likely to be seen earlier than deception by falsifying.

Deception in a Conflict Situation

The general method of hiding games is used to investigate whether children can deceive when they are directly asked by an opponent (a competitor or an enemy) to whom they should not give information (e.g. Peskin, 1992; Seno, 2008; Sodian, 1991). This is a typical situation in which people deceive others. However, people in daily life practice deception in another situation. We sometimes face a conflict situation in which we want to tell the truth to one person but not to another, for example, when a friend to whom we wish to give information and an opponent to whom we should not give information are present at the same time. If the friend asks us for the information, we have to control our behavior by withholding it (for a while) or by providing the information in a whisper to prevent the opponent from hearing it. These controlled behaviors in the conflict situation fulfill the important purpose of deception, which is to prevent the opponent from getting information (e.g. Chandler, Fritz, & Hala, 1989). In this example, we indirectly deceive the opponent, rather than directly deceive by falsifying or denying.

The purpose of the present study is to investigate whether children exhibit these controlled behaviors in a conflict situation. Children’s deception is thought to be less flexible or resourceful than that of adults (e.g. Newton et al., 2000). But few details about this inflexibility are available, because most previous studies that have used hiding games have mainly focused on the age at which children begin to selectively deceive. In contrast, investigating children’s deception in the conflict situation should make clear under what conditions, and to what extent, their deception is inflexible. Thus, the present investigation is significant because it reveals the essence and features of children’s deception.

Withholding information and providing information in a whisper are behaviors that can be regarded as acts of deception by concealment (e.g. Hauser, 1997). Concealment means omitting information that is important or relevant in the given context. Whiten and Byrne (1988) classified concealment as one of the five major tactical deceptions. For example, acoustic concealment was defined as occurring when an agent acts quietly, such that the target’s attention is not attracted (Whiten & Byrne, 1988). Providing information in a whisper is an example of this type of concealment. Recent studies have revealed that 36-month-olds can conceal auditory information from others during a prohibited action (e.g. Melis, Call, & Tomasello, 2010). Furthermore, 3-year-olds know that the physical presence of an opponent is an inhibiting factor when they deceive (Peskin, 1992). Indeed, Peskin (1992) demonstrated that 87% of 3-year-olds were able to physically exclude a competitor from a game and obtain their favorite things. Thus, 3-year-olds already know that it is difficult to deceive when the opponent is nearby and that they have to control their behavior by means such as concealing auditory information.

These findings suggest that young children already have the abilities to control their behavior that are required in the conflict situation. Therefore, if having these abilities is sufficient to resolve the conflict situation, young children should hesitate to give information immediately to the friend when an opponent is present and should control their behaviors to deceive by concealment (i.e. withholding information or providing information in a whisper). Thus, one prediction of the present study is that young children around the age of 5 years who pass the false-belief task should succeed in deception whether or not they are in the conflict situation, because previous studies have demonstrated that children at this age can clearly deceive (e.g. Peskin, 1992; Seno, 2008; Sodian, 1991).

However, a contrasting prediction is also possible. The situation in which only the opponent is present and the situation in which the opponent and the friend are both present differ because additional knowledge is needed in the latter situation. Successful deception by concealing in the conflict situation depends on the understanding that, if the child gives information to the friend, then this information is also shared with the opponent. Therefore, even if young children have the basic abilities required in the conflict situation, if they do not have this understanding, they may fail to prevent the opponent from getting the information.

This difference between possessing basic abilities and putting them into practice connects with the results of research on children’s understanding of secrecy, which is another act of information concealment (e.g. Bottoms, Goodman, Schwartz-Kenney, & Thomas, 2002). Although young children have the basic ability to keep a secret (e.g. Peskin & Ardino, 2003), they do not necessarily distinguish secrets from non-secrets (e.g. Anagnostaki, Wright, & Bourchier-Sutton, 2010; Kim, Harris, & Warneken, 2014). This leads to the possibility that young children do not have sufficient knowledge about how to keep information secret in the conflict situation, in which people to whom they should and should not give information are both present. In fact, young children tend to tell others about rule transgressions that they have witnessed (tattle) even in the presence of the transgressor (e.g. Ingram & Bering, 2010; Ross & den Bak-Lammers, 1998). Therefore, if young children cannot understand that giving information to the friend in the conflict situation leads to this information being shared with the opponent, they should not control their behaviors and should not hesitate to give information immediately to the friend, in spite of the presence of an opponent.

In the present study, we adapted the hiding game method to include a conflict condition in which a rabbit who is hiding has an enemy (a wolf) and a friend (a giraffe) who are present at the same time (the conflict situation), and the friend (giraffe) asks the child where the rabbit is. As in the competitive and the cooperative situations of previous studies (e.g. Seno, 2008; Sodian, 1991), we also prepared a no-conflict condition in which the enemy (wolf) and the friend (giraffe) are present at different times, and each asks the child where the rabbit is (the enemy situation and the friendship situation, respectively).

Thus, we investigated how difficult it is for young children to deceive in the conflict situation, by comparing their performance in the no-conflict and the conflict conditions. If young children can comprehend the conflict situation, they should control their behavior and deceive by withholding information about the location of the rabbit from the giraffe friend for a while or by providing information to the giraffe friend in a whisper. In contrast, if they cannot comprehend the conflict situation, they should immediately provide information to the giraffe friend, which would allow the wolf enemy to capture the rabbit.

Experiment 1

Method

Participants

Eighty-three young children participated in this experiment. Data from three children were excluded because of failure to pay attention. The remaining sample consisted of 25 4-year-olds (12 boys and 13 girls; mean age 4 years 7 months; range 4 years 1 months to 4 years 11 months), 32 5-year-olds (16 boys and 16 girls; mean age 5 years 6 months; range 5 years 0 months to 5 years 11 months), and 23 6-year-olds (10 boys and 13 girls; mean age 6 years 4 months; range 6 years 0 months to 6 years 9 months). The mean age of all of the participants was 5.5 years (SD = 0.7). All were Japanese and attended kindergarten in a mid-sized city in Japan. We did not include 3-year-olds because Japanese children successfully perform the false-belief task and the hiding game about one year later than western children (e.g. Naito & Koyama, 2006; Seno, 2008; Wellman et al., 2001). Informed consent was obtained from the principal of the kindergarten. Only children who provided verbal assent that they would take part in the research participated.

Procedure

The task was presented to each child individually as a puppet show, with three large hand-puppets: an attractive rabbit, a fierce-looking wolf, and a giraffe. The three puppets were acted out by three different experimenters. The set consisted of two identical houses which differed in color (red or blue) and one tree, all of which were large enough for the puppets to hide behind. The child sat before a table with a curtain at its far end (3.5 m from the participant) that could be lowered to occlude the puppet show. These procedures were modified from Kikuno (2010). The puppet show started with a warm-up trial, followed by the four experimental trials.

Warm-up trial

The purpose of the warm-up trial was to inform the children of the relationships among the characters. First, the children were shown that the rabbit puppet played with the giraffe puppet. The rabbit said: “I like you, giraffe.” The giraffe said: “I like you, too, rabbit. I have to go now, but we can play together later,” and the giraffe left. Next, the wolf puppet appeared and said: “I’d like to grab that rabbit.” The rabbit said: “Oh my! I don’t like that wolf. I’d better get out of here,” and the rabbit ran. The wolf chased the rabbit. The rabbit left. The wolf said: “Oh well. That rabbit is quick.” Before the wolf went offstage, the giraffe reappeared from the other side of the stage and said: “Oh no! There’s that wolf. I don’t like him, and I don’t want to see him around here. I think I’ll go home.” The wolf went offstage without noticing the giraffe. The giraffe also went offstage.

Then, the rabbit reappeared and asked two questions: “Did I like the giraffe, or did I not like the giraffe?” and “Did I like the wolf, or did I not like the wolf?” The giraffe reappeared and asked one question: “Did I like the wolf, or did I not like the wolf?” The correct answers for three questions were “like,” “not like,” and “not like,” respectively. All but three children gave correct answers to all questions. These three children were shown the warm-up puppet show again, and all then gave correct answers.

Four experimental trials

All participants were presented with four experimental trials: two in the no-conflict condition and two in the conflict condition. The order of the conditions and the house that the rabbit hid behind were counterbalanced. Figure 1 shows the story line for both conditions.

OPEN IN VIEWER

Figure 1.

Story line in Experiments 1 and 2 (from the child’s perspective).

The sequence of events in the no-conflict condition was as follows:

Phase 1: At the start of the trial, the rabbit puppet entered from the center of the back of the stage. The rabbit said to the child: “Right now I’m running away from the wolf, but I want to see my friend the giraffe later. So I’ll hide in this red (blue) house. Do you promise to save me from the wolf?” After the child responded “yes,” the rabbit hid in the red (blue) house.

Phase 2-NC was the enemy situation. The wolf puppet entered from the center of the back of the stage and asked the child: “Where did the rabbit go? Can you tell me where the rabbit is?” If the child refused to give a response, the wolf paused for about six to eight seconds, and then repeated the same request up to two more times while looking at the child. The purpose of the repetitions was to check whether the child’s refusal to respond was a confirmed use of a denying strategy for deception. If the child told or pointed to the house where the rabbit was hiding, the wolf said: “Oh! The red (blue) house!” and moved to the house and captured the rabbit. The rabbit said “Oh! Help me!” and the trial ended. In contrast, if the child told or pointed to the house where the rabbit was not hiding, the wolf moved to the house and looked for the rabbit for a while, saying: “Where is the rabbit?” Then the wolf went offstage, saying: “Oh well! I’ll go home and take a nap.” If the child refused to give a response after the third question, the wolf went offstage, saying: “Oh well! I’ll go home and take a nap.”

Phase 3-NC, the friendship situation, was presented only if the child gave false information or refused to give a response to the wolf. The giraffe entered from the center of the back of the stage and told the child: “I’m looking for the rabbit. Can you tell me where the rabbit is?” If the child told or pointed to the house where the rabbit was hiding, the giraffe moved to the house and met the rabbit, and both puppets said: “I’m glad to see you!” In contrast, if the child told or pointed to the house where the rabbit was not hiding, the giraffe moved to the house and looked for the rabbit for a while, saying: “Are you here, rabbit?” Then the giraffe went offstage, saying: “I guess I’ll go home now.” The trial then ended.

The sequence of events in the conflict condition was as follows:

Phase 1 was the same as in the no-conflict condition.

In Phase 2-C, the wolf puppet entered from the center of the back of the stage. The wolf turned to the left and said: “Hey! Here comes the giraffe.” Then the wolf turned to the right and said “There’s a tree over there. I’ll hide behind the tree and watch the giraffe for a while.” The wolf moved to hide behind the tree (from the giraffe’s perspective). However, the wolf was completely visible from the child’s perspective (see Figure 1).

In Phase 3-C, the conflict situation, the giraffe puppet entered from the center of the back of the stage and, pretending not to notice the presence of the wolf behind the tree, asked the child: “I’m looking for the rabbit. Can you tell me where the rabbit is?” If the child did not respond, the giraffe repeated the same request after a 15-second pause up to two more times while looking at the child (15 seconds and 30 seconds after the first question). In this situation, the experimenter with the rabbit puppet behind the red or blue house surreptitiously signaled to the experimenter with the giraffe puppet every 15 seconds, using a stopwatch. If the child gave no response after 45 seconds, the wolf went offstage, saying: “Oh well! I’ll go home and take a nap.” Then the giraffe turned in the direction the wolf had gone and said: “Oh dear! The wolf was here! That was scary.” Turning to the child, the giraffe said: “Hey, can you tell me where the rabbit is?” If the child told or pointed to the house where the rabbit was hiding, the giraffe went to the house and met the rabbit, and both puppets said: “I’m glad to see you!” The trial then ended. If the child told the giraffe or pointed to the house where the rabbit was hiding within 45 seconds, the giraffe began to move towards the house. At the same time, the wolf said: “Oh! The red (blue) house!” and moved to the house more quickly than the giraffe did and captured the rabbit. The rabbit said “Oh! Help me!” and the trial ended. On the other hand, if the child told or pointed to the house where the rabbit was not hiding, the giraffe moved to the house and looked for the rabbit for a while, saying: “Are you here, rabbit?” Then the giraffe went offstage, saying: “I guess I’ll go home now.” The wolf then also went offstage, saying: “I’ll go home too, and take a nap.” Finally, the rabbit appeared from the other house, and said: “The giraffe left.” and the trial ended.

Coding of responses

Children’s behaviors were classified as appropriate or inappropriate responses based on whether they prevented the wolf from getting information about the location of the rabbit. Hence, the content of the appropriate and the inappropriate responses differed between conditions.

In the no-conflict condition, children’s responses in Phase 2-NC were coded as appropriate or inappropriate. In previous studies (Peskin, 1992; Seno, 2008), correct responses were defined as those in which children provided misinformation or refused to inform. Based on these criteria, appropriate responses, which were indicative of deception, were coded into the following two categories: (1) falsifying, when the child told the wolf or pointed to the house where the rabbit was not hiding; or (2) denying, when the child kept refusing to give a response until the wolf went offstage.

In order to confirm that these categories represented appropriate responses, a pilot study was conducted with Japanese adults (N = 41). They were shown video clips of the puppet show of the no-conflict condition from Phase 1 to the scene where the wolf asked “Can you tell me where the rabbit is?” in Phase 2-NC. Then they were required to report what they would do. All adults provided a response in one of the two categories, such as “I will say [the house where the rabbit was not hiding]” or “I will say ‘No!’” The results indicated that these categories represented appropriate responses.

Some children initially refused to give a response, but then after one or two repeated requests told or pointed to the house where the rabbit was not hiding. Consistent with the criterion used in previous studies, such as Seno (2008), which classified the ways that accurate information is not given as denying or falsifying, this case was classified as falsifying. An inappropriate response was coded when the child told the wolf or pointed to the house where the rabbit was hiding. In cases in which children initially refused to give a response, but after one or two repeated requests told or pointed to the house where the rabbit was hiding, the response was classified as an inappropriate response (cf. Seno, 2008). In the present study, all but one of the children who gave appropriate responses in Phase 2-NC immediately told the giraffe or pointed to the house where the rabbit was hiding in Phase 3-NC (the one child who did not was coded as giving an inappropriate response).

In the conflict condition, the child’s responses in Phase 3-C were coded as appropriate or inappropriate. The appropriate responses were coded into one of the following four categories: (1) complete/no-response; (2) incomplete/no-response; (3) whispering; and (4) falsifying. The first two categories were used when the children’s response met two criteria, which were important in evaluating if children were hesitating about whether or not to give information to the giraffe: first, the child did not respond immediately when asked by the giraffe; and second, the child looked at least once at the wolf, which was hiding by the tree, after being asked by the giraffe. Complete/no-response was coded in the case that children did not respond for 45 seconds and the wolf went offstage. Incomplete/no-response was coded in the case that children eventually (within 45 seconds) told or pointed to the house where the rabbit was hiding. All children who were classified as complete/no-response gave accurate information to the giraffe (told or pointed to the house where the rabbit was hiding) after 45 seconds passed and the wolf had gone offstage. Whispering was coded when children controlled their behavior and gave accurate information in a hushed voice in order not to be overheard by the wolf. One boy gave accurate information by silently moving his lips and by a small movement of a bent finger-tip in both trials of the conflict condition; his response was coded as whispering, because he clearly controlled his behavior. Falsifying was coded when children told the giraffe or pointed to the house where the rabbit was not hiding. This was defined as an appropriate response because children were acting to prevent the rabbit from being captured by the wolf.

All behaviors in these four categories were behaviors that were controlled to prevent the wolf from getting accurate information and thus were indicative of deception (e.g. Chandler et al., 1989). In order to confirm whether these categories represented appropriate responses, another pilot study was conducted with the same Japanese adults (N = 41). They were shown video clips of the puppet show of the conflict condition from Phase 1 to the scene where the giraffe asked: “Can you tell me where the rabbit is?” in Phase 3-C. Then they were required to report what they would do. All but one adult provided a response in one of the four categories, such as “I won’t say anything for a while” or “I will say it in a whisper.” The results indicated that these categories represented appropriate responses.

All responses except those in the above four categories were cases in which children told the giraffe or pointed to the house where the rabbit was hiding immediately after being asked by the giraffe. These were coded as inappropriate responses. The inter-rater reliability for the coding of all responses, Cohen’s kappa, was k = 0.94. Disagreements were resolved by discussion.

Results and Discussion

Preliminary analysis

The key issue in the conflict condition is whether the children paid attention to the wolf hiding behind the tree and watching the giraffe in Phases 2-C and 3-C. To evaluate this, we examined the children’s gazes on the videotape. All children gazed at the wolf from the time that the wolf appeared to the time that the wolf hid behind the tree. When the giraffe appeared, all children moved their gaze to the giraffe. These findings indicate that all children paid attention as the wolf was hiding behind the tree and watching the giraffe. Furthermore, preliminary analysis indicated that there was no significant difference between the presentation orders of the two conditions.

Main analysis

Table 1 shows the contingency of the number of the appropriate responses in the two trials in each condition between the no-conflict and the conflict conditions. The average percentages of appropriate responses in the no-conflict condition were 36.0% for 4-year-olds, 87.5% for 5-year-olds, and 71.7% for 6-year-olds. The average percentages of appropriate responses in the conflict condition were 14.0% for 4-year-olds, 23.4% for 5-year-olds, and 30.4% for 6-year-olds. Thus, the average percentage of appropriate responses for 6-year-olds in the conflict condition was lower than that for 4-year-olds in the no-conflict condition.

OPEN IN VIEWER

Table 1.

Contingency of the number of the appropriate responses between the no-conflict and the conflict conditions in Experiment 1.

		Conflict
		2 appropriate	1 appropriate	0 appropriate
4-year-olds (N = 25)
No-conflict	2 appropriate	1	1	6
	1 appropriate	0	0	2
	0 appropriate	1	2	12
5-year-olds (N = 32)
No-conflict	2 appropriate	5	5	17
	1 appropriate	0	0	2
	0 appropriate	0	0	3
6-year-olds (N = 23)
No-conflict	2 appropriate	6	1	8
	1 appropriate	0	0	3
	0 appropriate	0	1	4

With reference to the analyses of previous studies (e.g. Sodian, 1991), the data of Table 1 were collapsed into a 2 × 2 contingency table (1 or 2 appropriate responses vs. 0 appropriate response in the no-conflict condition and 1 or 2 appropriate responses vs. 0 appropriate response in the conflict condition) in each age group.

In the no-conflict condition, 10 of 25 (40.0%) 4-year-olds, 29 of 32 (90.6%) 5-year-olds, and 18 of 23 (78.3%) 6-year-olds gave at least one appropriate response in the two trials. A chi-square test revealed an overall effect of age (χ ² (2) = 18.33, p < .001). This occurred because significantly fewer of the 4-year-olds gave appropriate responses than did the 5- and 6-year-olds (χ ² (1) = 16.65, p < .001; χ ² (1) = 7.21, p = .007). In the conflict condition, five of 25 (20.0%) 4-year-olds, 10 of 32 (31.3%) 5-year-olds, and eight of 23 (34.8%) 6-year-olds gave at least one appropriate response in the two trials. A chi-square test revealed no effect of age (χ ² (2) = 1.44, p = .487).

Next, we examined whether there were differences between conditions in each age group. For 4-year-olds, eight gave at least one appropriate response in the no-conflict condition but gave no appropriate response in the conflict condition, while three showed the reverse pattern. McNemar’s test was not significant (χ ² (1) = 2.27, p = .132). In contrast, for 5-year-olds, 19 gave at least one appropriate response in the no-conflict condition but gave no appropriate response in the conflict condition, while no one showed the reverse pattern. For 6-year-olds, 11 gave at least one appropriate response in the no-conflict condition but gave no appropriate response in the conflict condition, while only one showed the reverse pattern. McNemar’s tests were significant for 5-year-olds (χ ² (1) = 19.00, p < .001) and 6-year-olds (χ ² (1) = 8.33, p = .004).

These results indicate that young children from about the age of 5 years were clearly able to deceive in the no-conflict situation, supporting many previous studies (e.g. Seno, 2008; Sodian, 1991). However, it was difficult for them to deceive in the conflict situation.

Response patterns

In the no-conflict condition, children who gave at least one appropriate response in the two trials were classified into three categories: falsifying, denying, or a combination of both (see Table 2). Here, falsifying includes cases in which a child gave appropriate responses by falsifying on both of the trials and in which a child gave one appropriate response by falsifying on one trial and gave one inappropriate response on the other trial. Denying was classified in a similar way. Combination meant giving one appropriate response by falsifying and the other appropriate response by denying. A chi-square test revealed a significant difference among age groups for this response pattern (χ ² (4) = 10.45, p = .034). A post hoc Z test comparing the standardized residual value of frequencies with the critical Z score of ± 1.96 (for an alpha of .05) found that this result was due to greater denying among 4-year-olds and greater falsifying and less denying among 6-year-olds. These results indicate that deception by denying was greatest in younger children, and deception by falsifying increased with age. This is consistent with previous research findings.

OPEN IN VIEWER

Table 2.

Number of participants with each appropriate response pattern for the no-conflict condition in Experiment 1.

	Falsifying	Denying	Combination
4-year-olds (N = 10)	5 (3/2)	5 (5/0)	0
5-year-olds (N = 29)	18 (17/1)	9 (8/1)	2
6-year-olds (N = 18)	16 (13/3)	0 (0/0)	2

Note. The numbers within parentheses indicate the numbers of participants who gave appropriate responses for both of the two trials (left) and for one of the two trials (right).

In the conflict condition, appropriate responses for at least one trial were classified into the five categories: complete/no-response; incomplete/no-response; whispering; falsifying; and combination (see Table 3). A chi-square test revealed no difference among age groups (χ ² (8) = 10.54, p = .229). For the combination, one 4-year-old showed complete/no-response and incomplete/no-response. One of the two 6-year-olds showed incomplete/no-response and whispering. Another 6-year-old showed incomplete/no-response and falsifying. Children who were classified as incomplete/no-response waited for an average of 14.7 seconds (SD = 11.2). Overall, these results showed no differences associated with age, although incomplete/no-response was somewhat more prevalent among 5- and 6-year-olds.

OPEN IN VIEWER

Table 3.

Number of participants with each appropriate response pattern for the conflict condition in Experiment 1.

	Complete/no-response	Incomplete/no-response	Whispering	Falsifying	Combination
4-year-olds (N = 5)	1 (0/1)	1 (1/0)	0 (0/0)	2 (0/2)	1
5-year-olds (N = 10)	1 (1/0)	7 (3/4)	2 (1/1)	0 (0/0)	0
6-year-olds (N = 8)	0 (0/0)	4 (3/1)	1 (0/1)	1 (1/0)	2

Note. The numbers within parentheses indicate the numbers of participants who gave appropriate responses for both of the two trials (left) and for one of the two trials (right).

Closer examination of the responses revealed that some children carefully considered the conflict situation and paid attention to the wolf’s presence, for example, making no response for 45 seconds, saying “The wolf is around, so be careful. I’m going to whisper. Red house,” or giving accurate information by silently moving the lips and making small finger movements. These sophisticated responses indicate that some young children already appropriately grasped the conflict situation and controlled their behavior. However, such responses were infrequent, and most responses were inappropriate, with the children immediately telling or pointing to the house where the rabbit was hiding. Although appropriate responses increased with age, there were no differences among age groups, and the average percentage for 6-year-olds in the conflict condition was lower than that for 4-year-olds in the no-conflict condition. These results demonstrate that controlling behavior and deceiving in the conflict situation were difficult for young children.

We also examined children’s gazes during the giraffe’s request for information about the rabbit (Phase 3-C). The total number of trials in the conflict condition was 160, because each of the 80 participants completed two trials. There were 36 and 124 appropriate and inappropriate responses, respectively. Of the 36 appropriate responses, 33 (91.7%) involved the child at least once looking at the wolf hiding behind the tree during the giraffe’s request. In contrast, of the 124 inappropriate responses, 66 (53.2%) involved the children looking at the wolf at least once. About half of the children did not look at the wolf when the giraffe asked them about the rabbit.

Experiment 2

In Experiment 2, new conditions were prepared. In order to make the conflict situation more obvious, the wolf stood next to the giraffe instead of hiding behind the tree. However, it would be unnatural if the wolf simply stood near the giraffe when the giraffe asked the child where the rabbit is, because the giraffe also does not like the wolf. Therefore, the wolf pretended to be a bear by putting on a gentle-looking bear disguise before the giraffe appeared. Thus, because of this disguise, the giraffe did not know that the bear was really the wolf.

In Experiment 2, we compare this new conflict-disguised condition with a conflict-exposed condition, which is the same as the conflict condition in Experiment 1. Both the conflict-exposed and the conflict-disguised conditions present the same conflict situation, in which the enemy and the friend (wolf and giraffe) are together at the same time. Therefore, if children’s lower performance in the conflict condition in Experiment 1 was due to their difficulty in grasping the situation in which the wolf was hiding by the tree and the giraffe did not notice the wolf, then children’s performance in Experiment 2 should be higher in the conflict-disguised condition than in the conflict-exposed condition. In contrast, if their lower performance was due to difficulty in understanding that it is necessary to prevent the enemy from listening in a conflict situation, then performance in the conflict-disguised and conflict-exposed conditions should be equivalent.

The conflict-disguised condition raises a new question: do young children understand that the puppet which appears to be a bear is really the wolf? Flavell (1999) has demonstrated that children have substantial knowledge of the appearance–reality distinction by the age of 4 years. The participants in Experiment 2 were 5- to 6-year-olds. Therefore, they should have no problem understanding this appearance–reality distinction. In order to confirm this, we also prepared a new no-conflict-disguised condition, in which the wolf put on a bear disguise as the child watched, along with a no-conflict-exposed condition, which was the same as the no-conflict condition in Experiment 1. If children have substantial knowledge of the appearance–reality distinction, then their performance in the no-conflict-disguised and no-conflict-exposed conditions should be at the same high level.

Results and Discussion

Preliminary analysis

In the conflict-exposed condition, similar to Experiment 1, the videotape review indicated that all children gazed at the wolf from the time that it appeared to the time that it hid behind the tree. When the giraffe appeared, all children moved their gaze to the giraffe. These findings indicate that all children paid attention to the situation in which the wolf was hiding behind the tree and watching the giraffe. Furthermore, preliminary analysis found no significant differences in the presentation orders of the four conditions.

Main analysis

Table 4 shows the frequencies and the percentages of appropriate responses for the four conditions. Cochran’s Q revealed a significant difference in the percentages of appropriate responses among the four conditions, Q (3, N = 24) = 33.42, p < .001. McNemar’s pairwise comparisons indicated that appropriate responses were significantly less frequent in the two conflict conditions than in the two no-conflict conditions. There were no significant differences in frequency of appropriate responses between the two no-conflict conditions or between the two conflict conditions. The results for the two no-conflict conditions indicate that the children were able to deceive regardless of the appearance of the wolf. This means that they understood that the bear was really the wolf, consistent with previous findings that young children have knowledge of the appearance–reality distinction (Flavell, 1999). Nevertheless, in the conflict-disguised condition, most children immediately gave information to the giraffe when the nearby wolf was disguised as a gentle-looking bear, just as they did in the conflict-exposed condition when the wolf was hiding by the tree. These results suggest that children’s lower performance in the conflict condition in Experiment 1 was not due to difficulty in grasping the task, but rather to difficulty in understanding that it was necessary to prevent the enemy from listening in the conflict situation. These findings reinforce the conclusion that young children can deceive in a no-conflict situation, but it is difficult for them to deceive in a conflict situation.

OPEN IN VIEWER

Table 4.

Frequencies (percentages) of appropriate responses and number of participants with each appropriate response pattern in Experiment 2 (N = 24).

Condition	Total		Response patterns
			Falsifying		Denying
No-conflict-exposed	19	(79.2%)	4		15
No-conflict-disguised	21	(87.5%)	9		12
			Complete/no-response	Incomplete/no-response	Whispering
Conflict-exposed	8	(33.3%)	3	2	3
Conflict-disguised	7	(29.2%)	3	1	3

Response patterns

Table 4 also shows the number of participants within each appropriate response pattern. In the two no-conflict conditions, falsifying was less frequent than denying in Experiment 2, a pattern opposite to that in found in Experiment 1. This difference may have been caused by the change in method between the two experiments. The wolf repeated the same request only once in Experiment 2 but twice in Experiment 1. Falsifying included the case in which children at first refused to give a response but later told or pointed to the house where the rabbit was not hiding. Therefore, denying may have been more frequent in Experiment 2.

In the two conflict conditions in Experiment 2, responses classified as falsifying were not seen. Overall, there were no differences in the distribution of responses between the two conflict conditions. Complete/no-response was more frequent in Experiment 2. This may be due to the decrease from 45 seconds to 30 seconds before the wolf went offstage if the child gave no response.

General Discussion

The present study, by comparing the no-conflict and the conflict conditions, investigated how difficult it is for young children to deceive in a conflict situation. In the no-conflict condition, appropriate deception responses increased between the ages of 4 and 5 years. Most 5- and 6-year-olds selectively gave information in an appropriate manner: they did not give accurate information to the enemy in the enemy situation, but gave accurate information to the friend in the friendship situation. These results are consistent with previous studies of competitive and cooperative situations (e.g. Peskin, 1992; Seno, 2008; Sodian, 1991). Strategies for deception did vary with age (e.g. Talwar & Crossman, 2012): denying was prevalent among 4-year-olds, whereas falsifying was most frequently found in 6-year-olds. This trend is likely to reflect the development of theory of mind, which may be critical in misleading an opponent. Deception by denying does not necessarily require theory of mind, and the ability of children under the age of 4 years to deceive by such means has been previously demonstrated (e.g. Lewis et al., 1989; Melis et al., 2010; Polak & Harris, 1999; Talwar & Lee, 2002). By contrast, successful deception by falsifying requires the creation of a false belief in the mind of the opponent. This requires a theory of mind (e.g. Newton et al., 2000; Talwar & Lee, 2008), which develops at the age of 4 to 5 years (e.g. Wellman et al., 2001).

In the conflict conditions, most of the children in all age groups gave accurate information immediately after the giraffe (friend) asked, despite the fact that the wolf (enemy) was nearby. Although the conflict-disguised condition of Experiment 2, with the disguised wolf standing next to the giraffe, was intended to make the conflict situation easier for children to understand, most 5- and 6-year-olds immediately provided the requested information (potentially harmful for the rabbit) to the giraffe. These results indicate that comprehending the conflict situation was difficult for them.

Why is deception in the conflict situation so difficult for young children? This difficulty is probably not related to the development of first-order theory of mind, because many 5- and 6-year-olds were able to deceive by falsifying in the no-conflict condition. In addition, this difficulty probably does not arise from trouble with concealment, such as withholding information or providing information in a whisper (e.g. Whiten & Byrne, 1988); even 36-month-olds can conceal auditory information from others (e.g. Melis et al., 2010). Furthermore, 3-year-olds know that the physical presence of an opponent is an inhibiting factor when they deceive, and they are able to physically exclude a competitor in order to obtain their favorite things (Peskin, 1992). Thus, 3-year-olds already know that they have to control their behavior when the opponent is nearby, suggesting that young children have some of the abilities required for the conflict situation. Nevertheless, most 4- to 6-year-olds did not hesitate and gave information immediately to the friend in the conflict conditions.

There are several other possible explanations for this difficulty with deception in the conflict situation. First, attention limitations may have affected children’s performance. In general, it is difficult for young children to pay attention to two things at the same time (e.g. Higgins & Turnure, 1984). All the children seemed to understand the situation in which the friend and the enemy were present at the same time, but when the friend made the request, their attention may have been focused only on the asker, and they may no longer have paid attention to the rest of the scenario. In fact, the gaze analyses of Experiment 1 revealed that nearly half of the children who provided inappropriate responses did not look at the wolf when the giraffe asked them about the location of the rabbit.

Second, and more important, even if children had been able to pay attention to the friend and the enemy at the same time, it may have been difficult for them to understand that if they informed the giraffe about the location of the rabbit, then this information would also be shared with the wolf (who may then capture the rabbit). This understanding is crucial for the success of deception in the conflict situation. In Experiment 1, over half of the children who gave inappropriate responses looked at the wolf hiding behind the tree at least once when the giraffe asked them. Therefore, these children understood that the enemy and the friend were present at the same time, but they may not have understood the meaning of the conflict situation, namely, that telling the giraffe (friend) meant telling the wolf (enemy) at the same time.

This second explanation does not conflict with recent findings about children’s keeping a secret, which is another act of concealment of information (e.g. Bottoms et al., 2002). Young children have the basic ability to keep a secret (e.g. Peskin & Ardino, 2003), but do not necessarily distinguish secrets from non-secrets (e.g. Anagnostaki et al., 2010; Kim et al., 2014). Furthermore, they tend to tell others about rule transgressions that they have witnessed (tattle) even in the presence of the transgressor (e.g. Ingram & Bering, 2010; Ross & den Bak-Lammers, 1998). Thus, young children do not have sufficient knowledge to keep information secret under specified conditions. These findings support the conclusion that young children do not understand the meaning of the conflict situation: telling the friend means telling the enemy at the same time. Having the basic abilities required for the conflict situation and understanding the meaning of the situation appear to be different matters.

Although young children’s difficulty in understanding the meaning of the conflict situation is probably not related to first-order theory of mind, it may be related to an immature second-order theory of mind, which refers to the understanding of embedded mental states. Previous studies have suggested that second-order theory of mind plays a key role in many aspects of social communication beyond the preschool years (e.g. Miller, 2012). For example, the development of second-order theory of mind is related to the sophistication of deception (e.g. Talwar, Gordon, & Lee, 2007), the understanding of white lies (e.g. Broomfield, Robinson, & Robinson, 2002) and display rules (e.g. Naito & Seki, 2009), the ability to distinguish lies from jokes or irony (e.g. Sullivan, Winner, & Hopfield, 1995; Winner & Leekam, 1991), and so on. Furthermore, Peskin (1992) has suggested that concealing may be required for children to represent second-order mental states. To succeed in the conflict condition, it is not enough to understand only first-order mental states (e.g. The giraffe does not know that the wolf is hiding). It is necessary to understand second-order mental states (e.g. The giraffe does not know that the wolf intends to get the information). If children have this advanced understanding, they should notice that giving the friend information in the conflict situation leads to this information being shared with the opponent; in other words, they should hesitate to give information immediately to the friend when an opponent is present and should control their behaviors to deceive by concealment (i.e. withholding information or providing information in a whisper). Thus, the development of second-order theory of mind is important in grasping that telling the friend means telling the enemy at the same time. Therefore, performance in the conflict condition would be expected to improve from the ages of 7 to 8 years, when children begin to understand second-order mental states (e.g. Perner & Wimmer, 1985). This is consistent with previous studies that have shown that the development of second-order theory of mind is related to various aspects of social communication in the elementary school years (e.g. Broomfield et al., 2002; Miller, 2012; Naito & Seki, 2009; Talwar et al., 2007).

One limitation of the present study was that it did not investigate cognitive processes underlying the children’s responses, including executive function, which is a complex construct encompassing inhibitory control, working memory, and flexibly shifting behavior (e.g. Hughes, 2002). Some evidence suggests that inhibitory control is related to children’s deception or lie-telling behaviors (Carlson & Moses, 2001; Hala & Russell, 2001; Seno, 2008; Talwar & Lee, 2008), although the tasks used in these studies corresponded to the present no-conflict condition. The conflict situation may also require inhibitory control. For example, the appropriate response of whispering to avoid being overheard by the enemy requires inhibitory control; a “whisper task” is one example of inhibitory control tasks (e.g. Carlson, 2005; Kochanska, Murray, Jacques, Koenig, & Vandegeest, 1996). Use of such tasks may reveal that executive function is related to the ability to respond appropriately in the conflict situation.

Finally, the influence of social demand should be examined. In making responses categorized as incomplete/no-response in the conflict condition, children did not respond for a while but eventually gave accurate information before the wolf left. This may have been due to social demand. In fact, previous research has reported that even when they are asked not to tell others, young children fail to withhold the information (e.g. Bottoms et al., 2002).

In conclusion, the present study has contributed new evidence about the development of young children’s ability to deceive. Children around the age of 5 years were able to deceive in the no-conflict situation by falsifying. However, comprehending the conflict situation and deceiving were difficult for young children. Some young children were able to appropriately grasp the conflict situation and respond to it in a sophisticated manner, but such responses were infrequent. Although previous studies have demonstrated that the children of these ages are old enough to deceive by withholding information or by providing information in a whisper, most immediately provided information to the friend despite the fact that the enemy was nearby. Therefore, it seems to be difficult for most young children to understand the meaning of the conflict situation, namely, that telling the friend means telling the enemy at the same time.