Social and non-social fear in preschoolers and prospective associations with lying about cheating

Abstract

Little is known about the development of children’s lying. The present study examined whether observed social and non-social fear in preschoolers predicts children’s consistent cheating (N = 460; M = 4.3 years of age) and consistent lying about cheating. When left alone, 155 (34%) children cheated in both games conducted. Of these consistently cheating children, 54 (35%) children lied about their cheating after both games, whereas the remaining 101 children confessed to cheating after at least one game. Children’s temperamental fear did not predict consistent cheating. However, non-social (but not social) fear did predict consistent lying. Children with lower levels of non-social fear were more likely to lie. These findings suggest that non-social fear plays a role in the development of children’s antisocial lying.

Keywords

children’s antisocial lying children’s cheating non-social fear social fear

Lying has received much attention since the beginning of modern psychological research (Burton, 1963; Hartshorne & May, 1928) and this continues to the present time (Popliger, Talwar, & Crossman, 2011; Rotenberg, Betts, Eisner, & Ribeaud, 2012; Stouthamer-Loeber, 1986; Talwar & Lee, 2011). It is important to understand how lying develops in young children for its practical relevance in settings such as pedagogical and legal ones, but also for its implications on theories about children’s moral and social development (Talwar & Lee, 2008). In the last two decades, there have been several advancements in our understanding of the normative development of children’s lying, especially in relation to sociocognitive factors such as theory-of-mind understanding and executive functioning (Lee, 2013). However, as Lee’s (2013) review points out, we still know very little about the role of other psychological factors such as temperament in explaining individual differences in children’s tendency to lie. In an attempt to address this gap, here, we examined the influence of temperament, more specifically temperamental fearfulness, on young children’s lie-telling behaviour. In the present study, we focus on antisocial lies (i.e., “lies motivated by self-interest that violate trust and rules of communication and/or are intended to harm others or to avoid just punishment,” Talwar & Crossman, 2011, p. 145), excluding the consideration of prosocial lies (i.e., “lies told for the benefit of others, perhaps at personal cost,” Talwar & Crossman, 2011, p. 150), or “white lies” (Stouthamer-Loeber, 1986). Antisocial lying is regarded as a serious problem behaviour, not only because of the immediate consequences, but also because of its impact on the child’s future (Stouthamer-Loeber, 1986).

In her review, Stouthamer-Loeber (1986) found that antisocial lying is more prevalent in referred children than in normal children, and that among referred children, lying is more prevalent in children with conduct problems than in children with neurotic problems. This is in line with a cross-sectional study by Stouthamer-Loeber and Loeber (1986) who found that across 4th-, 7th- and 10th-grade, children’s antisocial lying was related to several behavioural problems such as fighting and stealing. One of the few longitudinal studies on this topic showed that parent and teacher reports on antisocial lying in 5-year-old children predicted offenses 15 years later (Michell & Rosa, 1981, cited in Stouthamer-Loeber, 1986). Yet, this (prospective) association with antisocial behaviour problems seems to be restricted to consistent antisocial lying (i.e., frequent antisocial lying) in contrast to inconsistent antisocial lying (i.e., occasional antisocial lying), which seems to be a rather normative child behaviour and unrelated to antisocial behaviour problems (Gervais, Tremblay, Desmarais-Gervais, & Vitaro, 2000; Stouthamer-Loeber, 1986; Talwar & Crossman, 2011). For instance, Gervais et al. (2000) found that consistent antisocial lying was associated with antisocial behaviour problems. Therefore, it is important to differentiate consistent antisocial lying from inconsistent antisocial lying when studying antisocial lying.

Most previous studies of lying in children have cross-sectionally examined associations with demographic factors and children’s cognitive capability (Burton, 1963; Hartshorne & May, 1928; Lewis, Stanger, & Sullivan, 1989; Polak & Harris, 1999; Talwar & Lee, 2011). A significant amount of this previous research was based on temptation situations, in which children were instructed not to peek at a toy or to cheat in a game when left alone. This line of research examining whether children cheat and whether they lie about their cheating (which is considered an antisocial lie, e.g. Talwar & Crossman, 2011) has yielded interesting findings (Lewis et al., 1989; Polak & Harris, 1999; Talwar & Lee, 2011).

Studies based on temptation situations, in which an attractive toy is promised for winning a game, show that the prevalence of cheating ranges from 82–88% in children aged 2–6- years (Lewis et al., 1989; Talwar & Lee, 2002), but lower (i.e., about 50%) in children aged 6–11 years (Talwar, Gordon, & Lee, 2007). This suggests that cheating is more normative in younger children than in older children. Indeed, in a study of 3–8-year-old children (n = 150), the prevalence of cheating did not differ by age (Talwar & Lee, 2008), whereas in another sample (i.e., 6–11-year olds, n = 172), older children cheated less (Talwar et al., 2007). When no attractive toy was promised for winning a game (Talwar & Lee, 2002), the prevalence of cheating was somewhat lower (i.e., 76% compared to 82%), but still the majority cheated. The prevalence of cheating does not seem to differ by gender (Talwar et al., 2007; Talwar & Lee, 2008). Finally, cheating does not seem to be related to theory of mind functioning (Talwar & Lee, 2008).

Results regarding lying show that the majority of children lies about cheating, with prevalence estimates varying from 62% in 2–3-year-old children (Lewis et al., 1989) to 93% in 6–11-year-old children (Talwar et al., 2007). Across different cultures, different forms of normative antisocial lie-telling behaviour (i.e., for self-protection or for personal reward) increases during early childhood and continue to increase until adolescence (Arentt Jensen, Jensen Arnett, Feldman, & Cauffman, 2004; Talwar & Crossman, 2011). In contrast, the prevalence of consistent antisocial lying is lower and stable over time (Stouthamer-Loeber, 1986). Results about the effect of gender on lying are mixed. For instance, Lewis, Stanger and Sullivan (1989) observed more lying among girls but other studies found no gender differences (Talwar & Lee, 2002, 2008). Finally, the capability to cover up lies (i.e., “the consistency between their initial lie with subsequent statements,” p. 867) seems to be positively related to theory of mind functioning (Talwar & Lee, 2008).

A limitation of previous studies based on temptation situations is the study design of using only one temptation situation (Lewis et al., 1989; Talwar & Lee, 2002, 2011), which makes it impossible to distinguish consistent antisocial lying from inconsistent antisocial lying (Gervais et al., 2000; Stouthamer-Loeber, 1986; Talwar & Crossman, 2011). Therefore, it remains unclear whether previous findings about demographic correlates apply to consistent cheating and lying or to inconsistent cheating and lying. Indeed, there is evidence suggesting that developmental trajectories differ for consistent and inconsistent antisocial lying (Gervais et al., 2000; Stouthamer-Loeber, 1986; Talwar & Crossman, 2011). Studies using at least two temptation situations are needed to better distinguish consistent cheating/lying (i.e., in two situations) from inconsistent cheating/lying. The present study attempts to fill this gap, it is our aim to predict consistent cheating and antisocial lying by using two temptation situations.

Despite the large amount of research on broadly based measures of antisocial behaviour and conscience development, there is a lack of research examining psychological precursors of children’s (antisocial) lying in itself (Gervais et al., 2000; Popliger et al., 2011; Rotenberg et al., 2012). A psychological factor that plays an important role in conscience development (Asendorpf & Nunner-Winkler, 1992; Fowles & Kochanska, 2000; Kochanska, 1995; Kochanska & Aksan, 2006) and in the development of antisocial behaviour (Gao, Raine, Venables, Dawson, & Mednick, 2010; Pardini, 2006; Raine, Reynolds, Venables, Mednick, & Farrington, 1998) is temperamental fearfulness. As consistent antisocial lying is associated with antisocial behaviour (Gervais et al., 2000; Stouthamer-Loeber, 1986) and may be (in the reversed form) an indicator of conscience development, temperamental fearfulness may also play a role in the development of consistent antisocial lying. This hypothesis is in line with the theoretical proposition (Talwar & Crossman (2011, p. 165) that consistent antisocial lying may develop when children’s early lying is not successfully deterred (p. 165). This proposition is also in line with the fact that antisocial lying is seen as a risky behaviour (Stouthamer-Loeber, 1986).

Various theoretical models have been proposed to explain this association (Hoffman, 2003; Kochanska, Murray, & Coy,1997; Mowrer, 1960; Raine, 2002), like avoidance learning (Mowrer, 1960), fear conditionability (Gao et al., 2010; Raine, 2002) and Hoffman’s theory of internalization (2003). Hoffman hypothesized parental discipline techniques results in an optimal anxiety level in relatively fearful children, but in insufficient anxiety level in relatively fearless children to trigger attention to parental messages and to promote internalization. Based on these theoretical models, we hypothesize that children with relatively low levels of temperamental fear may be more likely to cheat and lie consistently. To the best of our knowledge, this postulated association between temperamental fear and children’s consistent cheating and lying has never been directly tested in a prospective study.

Most previous studies examining the association between temperamental fear and conscience development and antisocial behaviour (Asendorpf & Nunner-Winkler, 1992; Fowles & Kochanska, 2000; Kochanska et al., 1997) used broad measures of temperamental fear, including social as well as non-social fear. However, there are reasons to expect that non-social and social fear may have a different effect on children’s development. First, social and non-social fear are only moderately correlated (Dyson, Klein, Olino, Dougherty, & Durbin, 2011; Majdandžić & Van Den Boom, 2007). Second, social and non-social fear are differently associated with developmental outcomes (Dyson et al., 2011; Kochanska & Radke-Yarrow, 1992). For instance, Kochanska and Radke-Yarrow (1992) found that toddlers’ observed social fear predicted inhibition towards an unfamiliar peer, whereas observed non-social fear predicted patterns of play. Third, fear conditioning studies suggest that fear conditioning is based on non-social aversive stimuli, such as a loud noise (e.g. Gao et al., 2010) or a painful pressure (e.g. Birbaumer et al., 2005). Therefore, it seems important to differentiate between social and non-social fear, when examining its effect on children’s development.

In sum, the aim of the present study was to investigate whether social and non-social fear are prospectively associated with 4-year-old children’s consistent cheating and lying. Children’s gender and age, and mother’s educational level and psychopathology, were included as control variables because they may be related to temperamental fear, cheating or lying (Barker, Copeland, Maughan, Jaffee, & Uher, 2012; Dierckx et al., 2011; Dyson et al., 2011; Lewis et al., 1989; Stouthamer-Loeber, 1986; Suveg, Shaffer, Morelen, & Thomassin, 2011). Based on Hoffman’s theory of internalization (2003), avoidance-learning models (Gao et al., 2010; Mowrer, 1960; Raine, 2002) and the finding that consistent lying is (prospectively) associated with antisocial problem behaviours (e.g. Gervais et al., 2000), we expect that consistent cheating and consistent lying is predicted by less temperamental fear. We explore whether social and non-social fear independently predict consistent cheating and lying. We found no studies examining the independent effect of social fear on conscience or antisocial development, as previous studies aggregated social and non-social fear scores (Asendorpf & Nunner-Winkler, 1992; Fowles & Kochanska, 2000; Kochanska et al., 1997). In contrast, there is some evidence of the independent effect of non-social fear on antisocial development (Birbaumer et al., 2005; Gao et al., 2010; Raine, 2002).

Method

Subjects

For the present study, we used data from a subsample of the Generation R Study, an ongoing population-based prospective birth cohort from foetal life onward in Rotterdam, the Netherlands (Tiemeier et al., 2012). Children in this subsample, usually referred to as the Focuscohort, were born between February 2003 and August 2005 and form a prenatally enrolled birth cohort. At age 36 months, children visited our research centre with an accompanying parent to attend the Laboratory Temperament Assessment Battery (Lab-TAB). At age 48 months, home visits were performed to observe children’s cheating and lying. The subgroup for the current study consisted of the 603 children who attended the assessments. The current analyses are conducted in 460 (76%) children for whom data on the determinant (i.e., temperamental fear) as well as the outcomes (i.e., cheating and lying) were available. The Medical Ethics Committee of the Erasmus Medical Center, Rotterdam, has approved the study. Written informed consent was obtained for all participants.

Temperament, 36 months

Temperament was observed at age 36 months using two episodes of the Laboratory Temperament Assessment Battery Preschool Version (Goldsmith, Reilly, Lemery, Longley, & Prescott, 1999). The Lab-TAB is a widely used, standardized instrument for observational assessment of early temperament. The Stranger Approach and Jumping Spider episodes (from the Preschool Version) were selected to assess fearfulness. The Stranger Approach episode evokes distress in the presence of a stranger (social fear), while the Jumping Spider episode elicits a startle response (non-social fear). Coders received extensive training and their reliability was established on a set of 25 DVDs before data were coded. Coders were not involved in administering the Lab-TAB episodes. As suggested in the manual, episodes were ordered, such that the Stranger approach was conducted first.

Social fear (stranger approach)

In the Stranger Approach episode, the child deals with social fear when a novel, slightly threatening stranger approaches. The situation is modelled after real-life events. During the observation, the child was left alone in a room. After some time, a female stranger disguised as a male wearing a dark coat, sunglasses, and a baseball cap, entered the room and asked standard questions to the child in a neutral tone of voice while slowly approaching the child. The episode was divided into nine epochs for scoring. Intensities of fear expressions, distress vocalizations, activity decrease, approach, avoidance, gaze aversion, verbal hesitancy, and nervous fidgeting were scored in each epoch as specified in the manual. Averages were computed for each of the child’s response or parameter across epochs. The Cronbach’s alpha was .40, which is modest probably because the coding of social fear is based on different, partly unrelated, response channels (i.e., vocal, bodily and facial) (Dyson et al., 2011; Majdandžić & Van Den Boom, 2007). The Stranger Approach task relates considerably to parent-ratings of social fear (e.g. Dyson et al., 2011), indicating evidence for convergent validity. Interrater reliabilities (ICCs, single measures, n = 25) for these averages ranged between .71 and .97. In the next step, a fear composite for this episode was formed by first reversing the approach average, then converting all averages into z scores, and finally combining them into one overall composite by taking the mean of all averages (z scores) as suggested by the manual (for more details, see Kok et al., 2012; Szekely, 2014).

Non-social fear (jumping spider)

The Jumping Spider episode is designed to elicit a startle/fear reaction based on an unexpected event, which is generally a predominant source of fear among preschool children. In this episode, the child and the experimenter were seated across from each other at a child-sized table on which there was a toy spider in a cage. The experimenter introduced a toy spider as a soft, fuzzy animal that does not bite and prompted the child to pet it. When the child’s hand was approximately 5 cm from the spider, the experimenter made it “jump.” This process was repeated four times in total. After the four trials, or when the child clearly did not want to pet the spider anymore, the experimenter explained to the child that it was only a toy, and allowed the child to make the spider jump if he/she wanted to. The episode was divided into five epochs for scoring. Intensity of fear expressions, vocal distress, bodily fear, approach, withdrawal, gaze aversion, and startle response were scored in each epoch. Averages were computed for each child response or parameter across epochs. The Cronbach’s alpha was .67. Interrater reliabilities (ICCs, single measures, n = 25) for these averages ranged between .66 and .95. In the next step, a fear composite for this episode was formed by converting all averages into z scores, and combining them into one overall composite by taking the mean of all averages (z scores) (for more details, see Kok et al., 2012; Szekely, 2014).

Cheating and lying, 48 months

The procedure to assess cheating and lying followed previous studies (Asendorpf & Nunner-Winkler, 1992; Hartshorne & May, 1928; Kochanska et al., 1997; Lewis et al., 1989). Children were observed playing two games with rules, which if followed, made winning impossible. As a result, children were tempted to cheat in these temptation situations (Hartshorne & May, 1928; Talwar & Lee, 2002). The experimenter first explained the rules of the game, telling the child not to transgress any of the rules and made sure the rules were clear to the child. The “dart game” involved throwing balls at a dartboard on the floor without stepping over a line that was too far from the dartboard to have a reasonable chance of success. In the “frog game,” the child had to find three brown frogs hidden under a cloth without looking, whereas there were only green frogs under the cloth. There were only a few brown frogs on the table that were used to explain the game to the child. To make sure the rules were understood the child was given the opportunity to practice. In these practice trials the children won as in the dart game, the dartboard was very close to the child and, in the frog game, there were only brown frogs under the cloth.

The child was promised an attractive prize for winning. In both games, the child was then left alone for 3 min to play the game while being videotaped. After 2 min, the experimenter called from the adjoining room that she would come back in a minute. After 1 min, she warned the child again before opening the door. When the experimenter returned, she asked standard questions: 1) “How did you play the game?” 2) “Did you follow the rules?” 3) “Did you cross the line? (Did you peek under the cloth?)” 4) “Did you throw more than three balls? (Did you take more than three frogs?)” 5) “Did you put balls manually on the dart? (Did you take brown frogs from the table?)” 6) “Was it hard to remember the rules?” Then the experimenter “discovered” she had set the game up wrong; she adapted the game and let the child play the game again to win a prize.

Coding was similar for both games (see e.g. Asendorpf & Nunner-Winkler, 1992; Kochanska et al., 1997). Children’s behaviour was coded as cheating if, for example, they moved the line closer to the dart or lifted the cloth in case of the frog game. Cheating behaviours were coded (yes/no) for each of 60 3-second intervals. Previous studies (Gervais et al., 2000; Stouthamer-Loeber, 1986; Talwar & Crossman, 2011) suggested that it is important to differentiate inconsistent lying from consistent lying, as the former seems to be a normative child behaviour, whereas the latter seems to be problematic behaviour and related to other antisocial behaviours. Because we conducted two cheating games, we were able to make a distinction between consistently cheating children (i.e., children who cheated in both games), children who cheated on one task and those who did not cheat.

Among the consistently cheating children, we distinguished consistent lie-tellers and inconsistent lie-tellers. To this end, we scored children’s verbal responses during the short post-task interview (see also Talwar & Lee, 2002). These verbal responses included children’s spontaneous confessions and those obtained by the questions. Verbal responses were classified into three categories: 1) saying “no” or shaking the head “no”; 2) saying “yes” or nodding the head “yes”; and 3) giving no answer. The consistent cheaters were categorized as liars, when they replied “no” to all of the questions or gave no answer, and they were categorized as confessors if they responded “yes” to at least one of the questions. This criterion for distinguishing lie-tellers and confessors was based previous studies (Lewis et al., 1989; Talwar & Lee, 2002). We differentiated between consistent lie-tellers (children that lied in both games) and inconsistent lie-tellers (children that confessed in at least one of the two games).

Control variables

Information about mother’s educational level was based on questionnaires completed by mothers during pregnancy. Mother’s educational level was defined by the highest accomplished educational level and, in line with the definition of Statistics Netherlands (2004), categorized from low (≤12 years of education) to high (≥18 years of education). Maternal psychopathology was based on the Brief Symptom Inventory (BSI), a validated self-report questionnaire (Derogatis & Melisaratos, 1983), completed by mothers when children were 36 months old. For this study, we used mean scores for the depression, anxiety and hostility scales.

Statistical analyses

First, descriptive statistics about children’s cheating and lying, temperamental fear and control variables were calculated. Second, Chi-square tests were conducted to explore the association between gender and cheating and between gender and lying. Third, we used analyses of variance (ANOVAs) to assess whether cheating and lying were related to children’s temperamental fear and age, and mother’s psychopathology and educational level. Fourth, bivariate associations between independent and control variables were explored in a correlation matrix. Finally, logistic regression analyses were performed to examine whether social fear and non-social fear predicted children’s cheating and lying after accounting for control variables. No multivariate outliers were identified in the regression analyses.

The percentage of missing values in the population for analysis was 1% for mother’s educational level, 3% for social fear and 8% for the BSI-scales. We conducted multiple imputation techniques for missing values using version 20 of IBM SPSS statistic. The results from the analyses with Multiple Imputation techniques did not differ from those without imputed missing values. Therefore, we only present the analyses using multiple imputations techniques for missing values.

Results

Of the 460 children, 49 (11%) did not cheat on either game, 256 (56%) cheated on one game and 155 (34%) cheated on both games (consistent cheaters). The chi-square statistic was significant for cheating across the two tasks (Chi-square = 10.53, df = 1, p < .01). Among the consistent cheaters, we differentiated between inconsistent lie-tellers (children that confessed in at least one of the two games) (n = 101, 65%) and consistent lie-tellers (children that lied in both games) (n = 54, 35%).

Table 1 presents descriptive statistics for the dependent and independent variables included in the main analyses. The Chi-square test showed that the percentage of girls and boys who consistently cheated did not significantly differ. Likewise, ANOVAs showed that children’s age, social fear and non-social fear, and mother’s educational level, depression, anxiety and hostility did not differ between consistent cheaters and inconsistent cheaters.

Table 1.

Descriptive statistics and effect sizes (Cramer’s V and η_p ²) for analyses comparing consistent cheating and inconsistent cheating and for analyses comparing consistent lying and inconsistent lying.

	Total (N = 460)		Inconsistent cheating (n = 305, 66%)	Consistent cheating (n = 155, 34%)	χ²or F	Effect size	p value
Child characteristics
Boys, n (%)	223 (48.5)		142 (63.7)	81 (36.3)	1.34	.05	.25
Girls, n (%)	237 (51.5)		163 (68.8)	74 (31.2)
Age (months), mean (sd)	51.49 (1.36)		51.51 (1.33)	51.49 (1.43)	.03	.00	.87
Social fear, mean (sd)	.42 (.08)		.41 (.08)	.42 (.08)	.01	.00	.94
Nonsocial fear, mean (sd)	.35 (.15)		.34 (.15)	.31 (.15)	.08	.00	.78
Maternal characteristics
Educational level, mean (sd)	3.96 (.95)		4.08 (.91)	3.96 (1.02)	1.50	.00	.22
Depression, mean (sd)	.09 (.22)		.11 (.23)	.09 (.20)	.48	.00	.49
Anxiety, mean (sd)	.15 (.23)		.15 (.24)	.15 (.22)	.02	.00	.90
Hostility, mean (sd)	.16 (.22)		.16 (.23)	.16 (.19)	.03	.00	.86
			Inconsistent lying (n = 101, 65%)	Consistent lying (n = 54, 35%)	χ²or F	Effect size	p value
Child characteristics
Boys, n (%)	–	–	58 (71.6)	23 (28.4)	3.10	.10	.08
Girls, n (%)	–	–	43 (58.1)	31 (41.9)
Age (months), mean (sd)	–	–	51.16 (1.20)	52.11 (1.63)	17.27	.10	< .001
Social fear, mean (sd)	–	–	.42 (.08)	.42 (.08)	.01	.00	.92
Nonsocial fear, mean (sd)	–	–	.36 (.15)	.31 (.15)	4.31	.03	< .05
Maternal characteristics
Educational level, mean (sd)	–	–	3.84 (1.01)	4.19 (1.01)	4.08	.03	< .05
Depression, mean (sd)	–	–	.12 (.24)	.03 (.09)	5.90	.04	<. 05
Anxiety, mean (sd)	–	–	.16 (.23)	.13 (.20)	.30	.00	.59
Hostility, mean (sd)	–	–	.17 (.19)	.15 (.20)	.52	.00	.47

Note. P values from chi-square test or ANOVA, Variables child age (range 47–60 months) continuously measured; child social fear (observed score range 0.22–0.69), child non-social fear (observed score range 0–0.76), maternal depression (observed score range 0–2.0), maternal anxiety (observed score range (0–2.5) and maternal hostility (observed score range (0–2) are continuous variables, for all variables higher scores indicate more symptoms, possible scores ranges were 0–1 for child social and non-social fear; 0–4 for maternal psychiatric symptoms; education is an ordinal variable (range 0–5; from only primary education to higher academic education).

Of the 155 consistent cheaters, 54 (35%) consistently lied, whereas 101 (65%) children confessed after at least one game. As hypothesized, observed non-social fear was lower among consistent lie-tellers than among inconsistent lie-tellers. Scores for observed social fear did not differ between consistent lie-tellers and inconsistent lie-tellers. The percentage of consistent lie-tellers did not differ significantly between boys and girls. Older children and children with more highly-educated mothers or less depressive mothers were more likely to lie consistently. Table 2 presents the correlations between the independent and control variables. Notably, social and non-social fear were not significantly correlated. Higher levels of maternal hostility were associated with lower levels of non-social fear in their children. Higher levels of maternal education were associated with lower levels of maternal depression. The three subscales of mother’s psychopathology (i.e., depression, anxiety and hostility) were positively correlated.

Table 2.

Correlations between temperamental fear and control variables.

	Child Gender	Child Age	Maternal Educational level	Maternal Depression	Maternal Anxiety	Maternal Hostility	Child Nonsocial fear	Child Social fear
Child demographics
Child Gender (boy = 0, girl = 1)	1
Child Age	.03	1
Maternal characteristics
Educational level	.01	.02	1
Depression	.05	.02	-.14**	1
Anxiety	-.01	.02	-.06	.65**	1
Hostility	-.01	-.04	-.04	.47**	.50**	1
Child characteristics
Nonsocial fear	.00	.06	-.06	-.08	-.05	-.10*	1
Social fear	-.03	.06	.02	.00	.02	-.04	.02	1

Note. Values are Pearson correlation coefficients; *p<.05, **p<.01. n=460 for child age and gender, n=456 for maternal education, n=425 for maternal anxiety and maternal depression, n=423 for maternal hostility, n=460 for non-social fear and n=446 for social fear.

Predicting children’s consistent cheating and lying

To examine whether temperamental fear predicts children’s consistent cheating and lying after adjusting for children’s gender and age, maternal educational level and maternal depression, anxiety and hostility scores, two multivariate logistic regression models were used (Table 3). First, we ran a logistic regression on consistent cheating, with inconsistent cheating as the reference category. In line with the bivariate analyses, temperamental fear (i.e., social and non-social fear) did not predict consistent cheating.

Table 3.

Predicting children’s consistent cheating and lying.

	Logistic Regression (N = 460)			Logistic Regression (N = 155)
	Consistent cheating versus Inconsistent cheating			Consistent lying versus Inconsistent lying
	OR		95% CI	OR		95% CI
Child demographics
Gender (boy=0, girl=1)	.81		.55, 1.20	2.31*		1.05, 5.07
Age	1.00		.86, 1.15	1.84***		1.34, 2.52
Maternal characteristics
Educational level	.87		.70, 1.06	1.12		.74, 1.70
Depression	.50		.14, 1.78	.00**		.00, .21
Anxiety	1.22		.38, 3.91	3.64		.33, 40.78
Hostility	1.25		.44, 3.53	1.13		.11, 11.89
Child characteristics
Social fear	1.00		.82, 1.22	.1.12		.73, 1.70
Nonsocial fear	.96		.79, 1.17	.61*		.40, .92
Constant	1.53		.00, 2699.23	.00***		.00, .00
Chi-square		4.05			41.91***
Degrees of freedom		8			8
Nagelkerke R²		.03			.33

Note. *p<.05, **p<.01, ***p < .001.

Second, we conducted a logistic regression analysis on consistent lying. The inconsistent lie-tellers were the reference group. In line with the bivariate analyses, relatively low scores of observed non-social fear predicted consistent lie-telling. Social fear did not predict children’s consistent lying, independent of whether or not we adjusted for non-social fear (OR = 1.16, p = .50). Following the approach of Julious (2004), we tested the difference in effect between social fear and non-social fear. Since the 84% confidence intervals did not overlap (OR = .62*; 84% CI = .45, .85 for non-social fear and OR = 1.30; 84% CI = .94, 1.79 for social fear) (Julious, 2004), we concluded that the effect of non-social fear on children’s consistent lying was different from the effect of social fear.

Discussion

The present study of preschool children examined whether observed social and non-social fear at 3 years predicted children’s consistent cheating and lying at 4 years, after adjustment for demographic factors such as children’s gender and age. A lower level of observed non-social fear predicted children’s consistent lying, whereas temperamental fear did not predict children’s consistent cheating.

In contrast to previous study designs using one temptation situation to define cheating and lying (e.g. Lewis et al, 1989; Polak & Harris, 1999), we used two temptation situations, in order to better distinguish consistent cheating and lying (i.e., in two situations) from inconsistent cheating and lying. Previous studies (Gervais et al., 2000; Stouthamer-Loeber, 1986; Talwar & Crossman, 2011) suggested that it is important to differentiate consistent lying from inconsistent lying, as the former seems to be associated with antisocial problem behaviours whereas the latter seems more normative child behaviour and not associated with problem behaviours. This idea is supported by the present results, as 66% of the children did not cheat or inconsistently cheated, whereas only 34% cheated consistently. Likewise, among the consistent cheaters 65% confessed on at least one game, whereas only 35% lied on both games.

In line with our hypothesis, low scores for observed non-social fear predicted consistent lying. However, social fear did not predict consistent lying. This difference in effect of non-social and social fear on consistent lying appeared to be significant, implying that the effect of non-social fear on children’s lying is different from the effect of social fear. These findings are in line with studies suggesting that social and non-social fear is differently associated with children’s developmental outcomes (Dyson et al., 2011; Kochanska & Radke-Yarrow, 1992). Our results are consistent with avoidance learning (Mowrer, 1960) and fear conditioning (Gao et al., 2010, Raine et al., 1998), which seem to be based on non-social aversive stimuli such as a loud noise (e.g. Gao et al., 2010) or a painful pressure (e.g. Birbaumer et al., 2005) or, in this case, a toy spider. Talwar and Lee (2011) hypothesized that consistent antisocial lying may develop when children’s early lying is not successfully deterred. Possibly, lying is less likely to be deterred in children with relatively low levels of non-social fear because of a poor fear conditioning (Birbaumer et al., 2005; Gao et al., 2010; Raine et al., 1998) or poor avoidance learning (Mowrer, 1960). An alternative explanation might be that those children who were better able to suppress their psychological (and possibly physiological) fear-reactions in a non-social, truly fear-inducing situation had better inhibitory control skills. Inhibitory control (the ability to inhibit automatic or prepotent responses to stimulation for the purposes of goal-oriented activity) is an important component of executive functioning (Miyake, Friedman, Emerson, Witzki, & Howerter, 2000) and has consistently been related to preschoolers’ ability to successfully deceive others (Carlson, Moses, & Hix, 1998; Evans, Xu, & Lee, 2011; Talwar & Lee, 2008). According to the activation-decision-construct model (Walczyk, Roper, Seemann, & Humphrey, 2003), a cognitive-based theory of deception, the decision to tell the truth or a lie requires inhibition to suppress critical details of the truth when a lie is made.

In contrast with our hypothesis, consistent cheating was not predicted by temperamental fear. It is possible that temperamental dimensions like self-regulation may be more important in predicting cheating than temperamental fear. It should be noted that in the present study (and in previous studies), cheating takes place in a rather non-social context as the child is alone when playing the game, whereas lying takes place in a social context. It is an interesting question for future studies whether temperamental fear does predict cheating when it takes place in a social context.

When interpreting these results, some limitations must be kept in mind. First, social fear and non-social fear were each based on only one Lab-TAB episode. Observed temperament has been found to vary across situations within dimensions (Majdandžić and Van Den Boom, 2007). Moreover, the internal consistency for the Stranger Approach task was modest. This could lead to some misclassification and an underestimation of the true associations. Therefore, the fact that we did not find an effect of social fear on cheating and lying should be interpreted with caution. However, the low internal consistency for the Stranger Approach possibly suggests that social fear is expressed by different, and partly unrelated, responses channels (i.e., vocal, bodily and facial) (Dyson et al., 2011; Majdandžić & Van Den Boom, 2007). Moreover, the interrater reliabilities in the current study for the Jumping Spider as well as the Stranger Approach were acceptable.

Second, in the Jumping Spider episode (used to measure non-social fear) an experimenter was present. In theory, the fear response displayed in this situation could (at least) partly reflect a social fear response. However, the children were familiar with the experimenter in the Jumping Spider episode in contrast to “the stranger” in the Stranger Approach. Moreover, the spider was introduced when the children were already at ease with the experimenter. Therefore, it seems very unlikely that the fear response as measured in the Jumping Spider seems to be triggered by the experimenter instead of the spider. Indeed, previous studies have shown that the Stranger Approach and the Jumping Spider are rather valid tasks to measure social and non-social fear, respectively (e.g. Dyson et al., 2011; Majdandžić & Van Den Boom, 2007). Future studies should make clear whether these results could be replicated using other tasks than the Stranger Approach and Jumping Spider to measure social and non-social fear, respectively.

Despite these limitations, this study is the first to show that temperamental (non-social) fear plays a role in the development of children’s consistent antisocial lying.

Footnotes

Funding

The Generation R Study is made possible by financial support of the Netherlands Organization for Health Research and Development (ZonMw Geestkracht 10.000.1003). The work of Eszter Szekely and Catherine Herba was supported by a grant of the Sophia Children’s Hospital Foundation for Scientific Research (SSWO-2007-514). The work of Henning Tiemeier was supported by a grant from the Netherlands Organization for Health Research and Development (ZonMw Vidi grant 017.106. 370).

References

Arnett Jensen

Jensen Arnett

Feldman

S. S.

Cauffman

(2004). The right to do wrong: Lying to parents among adolescents and emerging adults. Journal of Youth and Adolescence, 33 (2), 101–112.

Asendorpf

J. B.

Nunner-Winkler

(1992). Children’s moral motive strength and temperamental inhibition reduce their immoral behavior in real moral conflicts. Child Development, 63 (5), 1223–1235.

Barker

E. D.

Copeland

Maughan

Jaffee

S. R.

Uher

(2012). Relative impact of maternal depression and associated risk factors on offspring psychopathology. British Journal of Psychiatry, 200 (2), 124–129. doi:10.1192/bjp.bp.111.092346

Birbaumer

Veit

Lotze

Erb

Hermann

Grodd

Flor

(2005). Deficient fear conditioning in psychopathy: A functional magnetic resonance imaging study. Archives of General Psychiatry, 62 (7), 799–805.

Burton

R.V.

(1963). Generality of honesty reconsidered. Psychological Review, 70 (6), 481–499.

Carlson

S. M.

Moses

L. J.

Hix

H. R.

(1998). The role of inhibitory processes in young children’s difficulties with deception and false belief. Child Development, 69 (3), 672–691.

Derogatis

L. R.

Melisaratos

(1983). The brief symptom inventory: An introductory report. Psychological Medicine, 13 (3), 595–605.

Dierckx

Tulen

J. H. M.

Tharner

Jaddoe

V. W.

Hofman

Verhulst

F. C.

Tiemeier

(2011). Low autonomic arousal as vulnerability to externalising behaviour in infants with hostile mothers. Psychiatry Research, 185(1–2), 171–175. doi:10.1016/j.psychres.2009.09.014

Dyson

M. W.

Klein

D. N.

Olino

T. M.

Dougherty

L. R.

Durbin

C. E.

(2011). Social and non-social behavioral inhibition in preschool-age children: Differential associations with parent-reports of temperament and anxiety. Child Psychiatry & Human Development, 42 (4), 390–405. doi:10.1007/s10578-011-0225-6

10.

Evans

A. D.

Lee

(2011). When all signs point to you: Lies told in the face of evidence. Developmental Psychology, 47 (1), 39–49.

11.

Fowles

D. C.

Kochanska

(2000). Temperament as a moderator of pathways to conscience in children: The contribution of electrodermal activity. Psychophysiology, 37, 788–795.

12.

Gao

Raine

Venables

P. H.

Dawson

M. E.

Mednick

S. A

. (2010). Association of poor childhood fear conditioning and adult crime. American Journal of Psychiatry, 167(1), 56–60. doi:10.1176/appi.ajp.2009.09040499

13.

Gervais

Tremblay

R. E.

Desmarais-Gervais

Vitaro

. (2000). Children’s persistent lying, gender differences, and disruptive behaviours: A longitudinal perspective. International Journal of Behavioral Development, 24 (2), 213–221.

14.

Goldsmith

H. H.

Reilly

Lemery

K. S.

Longley

Prescott

(1999). Laboratory Temperament Assessment Battery: Preschool Version. Madison: University of Wisconsin Press.

15.

Hartshorne

May

M.A.

(1928). Studies in the nature of character. Vol. 1. Studies in deceit. New York, NY: Macmillan.

16.

Hoffman

M. L.

(2003). Empathy and moral development. Implications for caring and justice. Cambridge, UK: Cambridge University Press.

17.

Julious

(2004). Using confidence intervals around individual means to assess statistical significance between two means. Pharmaceutical Statistics, 3 (3), 217–222. doi:10.1002/pst.126

18.

Kochanska

(1995). Children’s temperament, mothers discipline, and security of attachment – Multiple pathways to emerging internalization. Child Development, 66 (3), 597–615.

19.

Kochanska

Murray

Coy

K. C.

(1997). Inhibitory control as a contributor to conscience in childhood: From toddler to early school age. Child Development, 68 (2), 263–277. doi:10.2307/1131849

20.

Kochanska

Radke-Yarrow

(1992). Inhibition in toddlerhood and the dynamics of the child’s interaction with an unfamiliar peer at age five. Child Development, 63 (2), 325–335. doi:10.1111/j.1467-8624.1992.tb01630.x

21.

Kochanska

Aksan

(2006). Children’s conscience and self-regulation. Journal of Personality, 74 (6), 1587–1617.

22.

Kok

van IJzendoorn

M. H.

Linting

Bakermans-Kranenburg

M. J.

Tharner

Luijk

M. P. C. M.

Tiemeier

. (2012). Attachment insecurity predicts child active resistance to parental requests in a compliance task. Child: Care, Health and Development, 39 (2), 277–287.

23.

Lee

(2013). Little liars: Development of verbal deception in children. Child Development Perspectives, 7 (2), 91–96.

24.

Lewis

Stanger

Sullivan

M. W.

(1989). Deception in 3-year-olds. Developmental Psychology, 25 (3), 439–443.

25.

Majdandžić

Van Den Boom

D. C.

(2007). Multimethod longitudinal assessment of temperament in early childhood. Journal of Personality, 75 (1), 121–168. doi:10.1111/j.1467-6494.2006.00435.x

26.

Miyake

Friedman

N. P.

Emerson

M. J.

Witzki

A. H.

Howerter

(2000). The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: A latent variable analysis. Cognitive Psychology, 41 (1), 49–100.

27.

Mowrer

O. H.

(1960). Learning and behavior. New York, NY: Wiley.

28.

Pardini

D. A.

(2006). The callousness pathway to severe violent delinquency. Aggressive behavior, 32, 590–598.

29.

Polak

Harris

P. L.

(1999). Deception by young children following noncompliance. Developmental Psychology, 35 (2), 561–568.

30.

Popliger

Talwar

Crossman

(2011). Predictors of children’s prosocial lie-telling: Motivation, socialization variables, and moral understanding. Journal of Experimental Child Psychology, 1 (10), 373–392.

31.

Raine

Reynolds

C. A.

Venables

P. H.

Mednick

S. A.

Farrington

D. P.

(1998). Fearlessness, stimulation seeking, and large body size at age 3 years as early predispositions to childhood aggression at age 11 years. Archives of General Psychiatry, 55 (8), 745–751.

32.

Raine

(2002). Biosocial studies of antisocial and violent behavior in children and adults: A review. Journal of Abnormal Child Psychology, 30, 311–326.

33.

Rotenberg

K. J.

Betts

L. R.

Eisner

Ribeaud

(2012). Social antecedents of children’s trustworthiness. Infant and Child Development, 21, 310–322.

34.

Székely

Tiemeier

Jansen

PW.

Jaddoe

VW.

Hofman

Verhulst

FC.

Herba

CM.

(2014). Maternal depressive symptoms are associated with low fearfulness in preschoolers. Journal of Clinical Child & Adolescent Psychology, 43 (5), 791–798.

35.

Statistics Netherlands. (2004). Standaard onderwijsindeling 2003 [Standard classification of education 2003]. Voorburg/Heerlen: Statistics Netherlands.

36.

Stouthamer-Loeber

(1986). Lying as a problem behavior in children: A review. Clinical Psychology Review, 6 (4), 267–289. doi:10.1016/0272-7358(86)90002-4

37.

Stouthamer-Loeber

Loeber

(1986). Boys who lie. Journal of Abnormal Child Psychology, 14 (4), 551–564.

38.

Suveg

Shaffer

Morelen

Thomassin

(2011). Links between maternal and child psychopathology symptoms: Mediation through child emotion regulation and moderation through maternal behavior. Child Psychiatry & Human Development, 42 (5), 507–520. doi:10.1007/s10578-011-0223-8

39.

Talwar

Crossman

. (2011). From little white lies to filthy liars: the evolution of honesty and deception in young children. Advances in Child Development and Behavior, 40, 139–179

40.

Talwar

Gordon

H. M.

Lee

(2007). Lying in the elementary school years: Verbal deception and its relation to second-order belief understanding. Developmental Psychology, 43(3), 804–8 10. doi:10.1037/0012-1649.43.3.804

41.

Talwar

Lee

(2002). Development of lying to conceal a transgression: Children’s control of expressive behavior during verbal deception. International Journal of Behavioral Development, 26 (5), 436–444. doi:10.1080/01650250143000373

42.

Talwar

Lee

(2008). Social and cognitive correlates of children’s lying behavior. Child Development, 79 (4), 866–881.

43.

Talwar

Lee

(2011). A punitive environment fosters children’s dishonesty: A natural experiment. Child Development, 82 (6), 1751–1758. doi:10.1111/j.1467-8624.2011.01663.x

44.

Tiemeier

Velders

F. P.

Szekely

Roza

S. J.

Dieleman

Jaddoe

V. W. V.

Verhulst

F. C

. (2012). The Generation R Study: A review of design, findings to date, and a study of the 5-Httlpr by environmental interaction from fetal life onward. Journal of the American Academy of Child and Adolescent Psychiatry, 51 (11), 1119–1135.

45.

Walczyk

J. J.

Roper

K. S.

Seemann

Humphrey

A. M.

(2003). Cognitive mechanisms underlying lying to questions: Response time as a cue to deception. Applied Cognitive Psychology, 17 (7), 755–774.