Pretending with realistic and fantastical stories facilitates executive function in 3-year-old children

Abstract

Fictional stories can affect many aspects of children’s behavior and cognition, yet little is known about how they might help or hinder children’s executive function skills. The current study investigated the role of story content (fantasy or reality) and mode of engagement with the story (pretense or a non-pretense control) on children’s inhibitory control, an important component of early executive function. Sixty 3-year-olds were randomly assigned to hear a fantastical or realistic story and were either encouraged to engage in pretense or a non-pretense activity related to the story. They then completed Less is More, a test of inhibitory control. Story content had no impact on children’s inhibitory control; children performed equally well after hearing a fantastical or realistic story. However, children who engaged in story-related pretend play showed greater inhibitory control than those who engaged in a non-pretense activity. We found no interaction between story content and play engagement type. These results held controlling for baseline inhibitory control, receptive vocabulary, age, gender, affect, and propensity toward pretense. Therefore, mode of play engagement with a story was more important in promoting children’s inhibitory control skills than the degree of realism in the story.

Fiction takes many forms for young children, including stories and pretending. In practice, these different types of fiction are often combined, as when children incorporate known narratives in their play and teachers use stories as inspiration for play-based classroom activities (Welsch, 2008). The stories can be realistic in nature or, as is frequently the case, they might include elements of fantasy. Nearly 80% of children’s books contain fantastical themes; when we include television and movies, the rate of fantasy content jumps to 91% (Goldstein & Alperson, 2019). Yet, evidence is sparse and often contradictory as to how pretending and fantastical story content influence children’s cognition.

One set of early abilities that might be influenced by fiction is executive function skills (EF), defined as the neurocognitive processes that contribute to the conscious self-regulation of thought and action (Carlson, Zelazo, & Faja, 2013; Jones, Bailey, Barnes, & Partee, 2016; Zelazo, Carter, Reznick, & Frye, 1997). Inhibitory control—the capacity to inhibit automatic, impulsive thoughts or actions that stand in the way of a goal—is the focus of the current study and a key component of executive function (Carlson & Moses, 2001). Complex inhibition tasks require a child to keep a rule in mind, respond in light of the rule, and inhibit a response that does not align with that rule (Garon, Bryson, & Smith, 2008). For example, in the Less is More task (Carlson, Davis, & Unger, 2005), children are given the option of a large or a small pile of treats, but the rules indicate that the pile the child chooses will be given away; they keep the unchosen pile of treats. The ability to inhibit such responses develops rapidly around age 3, and most children succeed on simple versions of these measures by age 4 (Carlson, 2005). Importantly, executive functions in the early childhood years have been shown to predict a host of important outcomes such as school readiness (Allan, Hume, Allan, Farrington, & Lonigan, 2014; Blair & Raver, 2015; Blair & Razza, 2007; Duckworth & Carlson, 2013; McClelland, Acock, Piccinin, Rhea, & Stallings, 2013), SAT scores (Shoda, Mischel, & Peake, 1990), and even criminal outcomes in adulthood (Moffitt et al., 2011). Although early EF development is tightly linked to the development of the prefrontal cortex in the first few years of life (Carlson et al., 2013), it is also reliant on environmental inputs (Center on the Developing Child, 2011). Thus, research into the various factors that might build early EF skills is important.

The current study investigates the potential impact of fiction on EF. At the most basic level, EF is tied to fiction by a shared developmental timeline: in the preschool years, children are exposed to a wealth of fantasy themes in books, television, and games (Goldstein & Alperson, 2019) and pretend play is in full swing (Singer & Singer, 1990). More direct evidence comes from experimental and training studies. For example, Thibodeau and colleagues (2016) conducted a 5-week randomized control intervention to assess the effects of dramatic pretend play on EF. Children were assigned to engage in pretense (e.g., pretend to take a trip to the moon), non-pretense play (e.g., ball games, coloring), or a no treatment control group. EF improved only for those in the pretense group. Role play as fictional characters can also improve EF and related abilities. Compared to controls, children who pretended to be characters such as Batman, Superman, or Rapunzel exhibited greater cognitive flexibility (White & Carlson, 2016), perseverance (White et al., 2017), delay of gratification (Karniol et al, 2011) and emotion regulation (Grenell et al., 2018). This growing body of evidence suggests that fictional contexts can be powerful facilitators of EF. However, the question remains as to which aspects of these fictional activities are responsible for EF gains. In each case, fiction was fantastical and presented within the context of pretend play. Thus, the goal of this study was to assess whether EF is facilitated by story content (fantasy versus reality), mode of engagement (pretending versus other forms of play), or both.

Fantasy in stories

Children can learn a variety of lessons from fictional stories. Infants and toddlers can imitate actions of characters (Simcock & DeLoache, 2006) and learn words from picture books (Chiong & DeLoache, 2012; Genea, Pickard, & DeLoache, 2008). By the time they reach the preschool years, children can learn solutions to difficult problems (Holyoak, Junn, & Billman, 1984) and gather facts on topics such as cultural traditions (Bonus & Mares, 2015; Mares & Sivakumar, 2014).

It is also around 3 years of age that children begin to grasp the distinction between fantasy and reality (see Weisberg, 2013 for a review). Fantastical elements of stories are those that do not or could not occur in the real world. They might involve, for example, imagined objects or creatures (e.g., flying carpets, monsters), make-believe places (e.g., a fairy land), or impossible events (e.g., walking through walls) (Hopkins & Weisberg, 2017). Once children understand that such elements are not real, how does it affect their ability to learn from fantastical fiction? Results are mixed.

Fantasy can be helpful for learning in some contexts, but harmful in others (for a review, see Hopkins & Weisberg, 2017). On one hand, fantastical events might better orient children’s attention to story content because they are unusual (Weisberg, Hirsh-Pasek, Golinkoff, & McCandliss, 2014). Weisberg and colleagues (2015) found that preschoolers learn new words better when they are introduced within a fantastical story than a more realistic one. Other studies show limited benefits of fantasy. For example, 3- to 6-year-old children transfer problem-solving solutions best when stories have moderate levels of fantasy (Richert & Schlesinger, 2016). Children also transfer knowledge when magic occurs before a solution is presented in a story, but not if the fantasy elements happen after the solution (Hopkins & Lillard, 2016 as cited in Hopkins & Weisberg, 2017). On the other hand, children might quarantine information from fantasy contexts, believing that it is not relevant to the real world (Leslie, 1987; Richert & Smith, 2011). In one study, 3- to 5-year-olds heard fantastical or realistic stories that included facts about “popple” flowers (e.g., smelling popples will give a person the hiccups). Of children who heard the realistic version, 72% transferred information about these flowers to questions about the real world; only 24% of children who heard the fantastical story did so (Walker, Gopnik, & Ganea, 2015). Children also generalize fewer lessons from fantasy characters than realistic ones (Richert, Shawber, Hoffman, & Taylor, 2009). Thus, it appears that learning from fantasy is possible, but success might depend on several important factors including amount of fantasy and the type of information conveyed (Hopkins & Weisberg, 2017).

Little attention has been paid to the connection between stories and EF; even less consideration has been given to the relative role of fantastical versus realistic content. Evidence from a related area of research confirms the link between fantasy and inhibition: Children with a high fantasy orientation (i.e., the tendency to engage in fantastical worlds; Sharon & Woolley, 2004; Singer & Singer, 1990; Taylor, 1999) display better inhibition than children with lower levels of fantasy orientation (Pierrucci, O’Brien, McInnis, Gilpin, & Barber, 2013). However, given the correlational nature of the research, it is unclear whether fantasy, EF, or both are driving this relation. Long-standing theories shed some light on why one might expect fantasy to facilitate EF skills. First, Vygotsky (1967) noted the importance of imaginative activities in the development of children’s ability to think about objects and events outside of their current experience (i.e., internal systems of representation) and thus to free themselves from the control of external stimuli. Through practice with considering alternatives to reality in fictional narratives, children might strengthen their ability to think about alternatives to their real-world choices. The theory of psychological distancing (Sigel, 1970) further suggests that considering scenarios more distant from reality should provide a bigger boost to children’s EF than closer ones. In line with this idea, Carlson et al. (2005) found that children’s inhibitory control improved with symbolic distance from task stimuli. Three- to 4-year-olds’ Less is More scores improved significantly over no treatment controls when the experimenter substituted mouse symbols for small rewards and elephant symbols for large rewards. Likewise, mentally stepping away from one’s current situation (e.g., temptation, distractions) could provide an opportunity for reflection on behavior, making inhibition of impulses more likely (Zelazo, 2004). Indeed, several studies show that psychologically distancing oneself from the here and now by taking an outsider’s perspective can improve children’s EF (Mahy et al., 2020; Prencipe & Zelazo, 2005; White & Carlson, 2016) and emotion regulation (Kross, Duckworth, Ayduk, Tsukayama, & Mischel, 2011).

Only a handful of experiments have investigated how fiction, whether realistic or fantastical, affects EF abilities. Kesek, Cunningham, Packer, and Zelazo (2011) investigated the role of implicit priming, versus explicit instructions, on 4-year-old children’s ability to delay gratification. Children were given choices between small immediate rewards and larger delayed rewards (e.g., 1 sticker now or 4 stickers later). But first, children in the implicit priming conditions heard a realistic story about a boy named Jamie who was going on a trip to the fair with his grandfather. Jamie either liked to maximize rewards or he liked to get things right away. Children’s decisions to wait, or not wait, for larger rewards mirrored Jamie’s decisions in the story. Explicit instruction to “get the most rewards” or “get rewards as soon as you can” did not affect their choices. This study illustrates the power of fiction to facilitate EF. However, it does not provide evidence as to whether fantastical stories might be better or worse than realistic stories.

Another set of studies specifically examined the impact of fantasy on children’s EF, but focused mainly on television shows (Lillard, Drell, Richey, Bogusweski, & Smith, 2015; Lillard & Peterson, 2011). Four- to 6-year-old children watched one of a set of fantastical or realistic cartoons, engaged in play, or read a book with the experimenter, then completed a battery of EF tasks. Compared to realistic shows, playing, or reading, children’s EF was significantly worse after watching fantastical TV shows (Lillard et al., 2015). Lillard and colleagues suggested that fantasy could overload children’s ability to process information because frequent violations of real-world expectations might tax cognitive resources. However, it is important to note that, children did not experience the same decrements to EF performance after watching Arthur, a low-fantasy cartoon show, or after reading a fantastical book. Thus, further research is needed to investigate whether relative differences in performance post realistic or fantastical fiction holds for storybooks as well.

Pretend play

Another common source of fiction in the early years is pretend play (Thompson & Goldstein, 2019). Although the term pretend play is often used interchangeably with dramatic play, make-believe play, or imaginative play, it need not be fantastical; pretend play can also have realistic themes. What is crucial is that in pretense, players act and think “as-if…” (Garvey, 1991; Reynolds, 1976). In other words, regardless of the reality of a situation, they act in accordance with internally-generated alternatives (Lillard, 1993). Pretending might involve, for example, using a banana to place a play telephone call (i.e., object substitution), throwing an imaginary birthday party with dolls, piloting a make-believe spaceship to the moon, or creating imaginary companions (Fein, 1981; Garvey, 1991; Leslie, 1987; Taylor, 1999). Pretend behaviors emerge early in life and development follows a fairly consistent timeline throughout the preschool years (for review, see Fein, 1981; Garvey, 1991; Grenell & Carlson, 2016; White, Carlson, & Zelazo, 2020). In the second year of life, children start to attribute pretend characteristics to objects. Gradually, play becomes more complex over the course of the next few years (Howes, Unger, & Seidner, 1989; Woolley & Tullos, 2008) so that pretense can involve abstract concepts, like a world in which miniature elephants serve tiny ice cream cones. By the preschool years, pretend play is in “high season” (Singer & Singer, 1990).

Both theory and research support a link between pretense and EF. Vygotsky famously claimed that “a child’s greatest self-control occurs in play (p. 99, 1978)” and believed that pretending affords opportunities through which children can build EF. Theoretical arguments as to why pretense could be related to EF are similar to those for fantasy. Successful engagement in pretend play requires a child to suppress their current view of reality, in order to see and act upon objects or ideas in a manner relevant to the pretend scenario (i.e., decoupling pretense from reality). The fact that a teacup is empty in real life does not prohibit a child from treating it as if it were truly full of liquid in a pretend tea party (Harris, 2000). In this way, pretense could provide children with practice in inhibition and flexible thought, such as the representation of counterfactual ideas (Gopnik & Walker, 2013). Moreover, pretending can create psychological distance from one’s current situation in several ways by mentally transporting the player to a different time, location, or role. It is also necessarily hypothetical, another dimension of psychological distancing (Sigel, 1970). This mental separation could boost EF by decreasing the salience of immediate stimuli, allowing for reflective, rather than impulsive, responses (for a review, see Carlson & White, 2013). Correlational data also provide some support for a connection between pretending and EF. Carlson, White, and Davis-Unger (2014) found that 3- to 5-year-olds who were good at representing pretense ideas and actions also had superior EF abilities. However, Pierucci and colleagues found that, whereas fantastical play and cognition were related to higher EF, realistic play was not (Pierucci, O’Brien, McInnis, Gilpin, & Barber, 2014). We must turn to experimental evidence for further causal support for the effects of pretense on EF.

In a now classic set of experiments, Dias and Harris (1988, 1990; see also Richards & Sanderson, 1999) provided evidence that pretense can improve children’s inhibition and flexible thinking. Children were presented with syllogisms such as “Tot is a fish. All fish live in trees. Where does Tot live?” When problems were presented in a non-pretense manner (e.g., only verbally), preschoolers showed great difficulty in overcoming their bias toward reality; they insisted that Tot would live in water, as is the case for real fish. However, if experimenters presented the problem in a pretend mode—for example, by using toys to act out the problem or using make-believe intonation—children were able to come to the correct conclusion that Tot lived in a tree. Successful responses to these syllogisms required children to inhibit known realities in order to reason about the premises as if they were true (Dias & Harris, 1988). As such, although syllogisms are not a typical test of EF, this research provides some evidence to suggest that pretense can improve EF abilities.

Two studies have taken this idea a step further by comparing the effects of pretense and non-pretense interventions. Goldstein and Lerner (2017) conducted a training study with low-income preschoolers in which the children were randomly assigned to book reading, block building, or dramatic play games. Following 24 sessions, children in the dramatic play condition outperformed their peers on measures of emotion regulation. An important caveat to note here is that the dramatic play games, although mostly realistic (e.g., pretending to be a chef), did contain some fantasy themes (e.g., pretending to be a superhero), so we cannot be certain that these effects were due to pretense alone. In another study, Saltz, Dixon, and Johnson (1977) trained preschoolers for 8 months on thematic fantasy play (i.e., play about fairy tales), realistic play, fantasy story discussion, or non-fantasy play activities (e.g., art, workbooks). In the first year of the study, children in the play conditions exhibited better inhibition than controls; they were able to guard a toy without touching it longer than children in the non-play conditions. There were no differences between fantasy and non-fantasy groups. However, two years later, the effect of pretense failed to replicate using a different inhibitory control task (i.e., take a pretend trip to the moon, but do not get out of the spaceship). Thus, although theory and research point to the possibility that pretense could facilitate EF performance, further experimental evidence is needed to confirm this effect independent of fantasy content.

Current Study

The goal of the current study was to explore how two forms of fiction, fantasy and pretense, affect 3-year-olds’ inhibitory control. Specifically, we used a 2×2 factorial design to examine the individual effects of fantastical story content and children’s mode of engagement with stories (i.e., pretense or non-pretense activities). To address fantasy content, children heard either a story that contained fantastical themes or one that was realistic. To address engagement, half of the children in each story group were encouraged pretend while acting out the scenes of the book. The other half were encouraged to engage in a non-pretense activity related to the book. Inhibitory control was measured immediately afterward using the Less is More task (Carlson et al., 2005). We also measured several potential confounds including baseline EF, receptive vocabulary, and propensity to engage in pretend play. Importantly, because positive affect is common in play (Krasnor & Pepler, 1980; Russ & Dillion, 2011) and it has been shown to facilitate EF performance (Yang & Yang, 2014), we also assessed children’s affect during the intervention. Following previous research and Vygotsky’s (1978) theory that pretending promotes representational flexibility, we hypothesized that a pretense activity would facilitate inhibitory control to a greater extent than a non-pretense activity. We also hypothesized that fantasy and reality stories would affect inhibitory control differently; however, given the mixed evidence on the relation between EF and fantasy content in storybooks, we did not have an a priori prediction for the direction of this effect.

Method

Participants

Participants were 60 typically developing 3-year-olds (M = 42.6 months, SD = 1.82, range = 39 to 46; 33 girls), recruited from a participant pool at a Midwestern US university. Four additional children were tested but excluded because they did not complete the study or because of equipment failures. According to parent report of ethnicity, 48 children in the final sample were white, 3 were white/Asian, 2 were white/African American, 1 was Pacific Islander, and 1 was African American/Native American; 5 parents did not answer this question. All children were English speakers and seven children also spoke a second language within the home. Median family income was $85,000-$100,000 per year (with 13 families choosing not to report income). All families received $10 and a child-sized T-shirt for participating in the study.

Procedure

Children were tested in a laboratory setting by one of two experimenters. Participants were randomly assigned to one of four conditions (n = 15 each) in a 2(story content: fantasy, reality) x 2(engagement: pretense, non-pretense) design (see Table 1). All experimental manipulations were administered immediately prior to administration of the Less is More EF task. Children also completed receptive vocabulary (PPVT, Dunn & Dunn, 2007) and a baseline measure of EF (peg tapping, Diamond & Taylor, 1996) before the story task manipulation and a pretense engagement task (phone task; Tahiroglu, Mannering, & Taylor, 2011; Taylor, Sachet, Maring, & Mannering, 2013) following Less is More. Sessions were videotaped and lasted 45–60 min. IRB approval was granted by the University of Minnesota.

Table 1.

Experimental Design

		Engagement condition
		Pretense	Non-pretense activity
Story content condition	Reality	Heard a realistic story Encouraged to pretend while placing cutouts on a storyboard	Heard a realistic story Encouraged to place cutouts on a storyboard
	Fantasy	Heard a fantastical story Encouraged to pretend while placing cutouts on a storyboard	Heard a fantastical story Encouraged to place cutouts on a storyboard

Story content manipulations.

Children were randomly assigned to hear one of two stories written for the current study: a fantastical story about Planet Opposite, which depicted scenes that do not or could not occur in the child’s life, or a corresponding realistic story about Funtown, which depicted no violations of reality (see Figure 1). Planet Opposite used the remote setting of another planet to introduce the fantastical nature of the story (Dias & Harris, 1998). It described a world in which everything is topsy-turvy (e.g., “fish live in the sky and butterflies swim underwater”, and “a mouse is as big as a house, but an elephant is teeny tiny”). Funtown mirrored the storyline of Planet Opposite using realistic events (e.g., “butterflies live in the sky and fish swim underwater”, and “an elephant is as big as a house, but mice are teeny tiny”). Both stories were nine pages long and portrayed the same set of characters. Most, but not all, of the events depicted in these two stories were direct opposites of one another. See Appendix for full story scripts.

Figure 1.

Example Pages From the a) Realistic and b) Fantastical Stories

Story engagement manipulations.

As the experimenter read the stories, they sat with children on the floor next to a poster-sized storyboard. The felt storyboard depicted a house next to a small lake, with a background of green fields and a blue sky. For each page of the story, the experimenter provided children with small felt cutouts that matched the characters and/or objects mentioned on each page (e.g., fish and butterflies for the page depicted in Figure 1).

Children were randomly assigned to engage in pretense or a non-pretense control activity related to the story they heard. Eliciting active participation from children when reading has been shown to improve learning outcomes (Ard & Beverly, 2004; Casey, Erkut, Ceder, & Young, 2008; Han, Moore, Vukelich, & Buell, 2010; Strouse, Doherty, & Troseth, 2013). Therefore, we designed an active control condition in an attempt to avoid confounds related to participation level.

Pretense.

Children were instructed that they were going to “pretend” that everything that happened in the storybook happened on the storyboard as well. By setting the stage for pretend play, we aimed to elicit a pretense mindset in which children might begin to think more flexibly (Vygotsky, 1978). We also encouraged children to express their thoughts verbally or behaviorally. The experimenter proceeded to read the first two pages of the story (i.e., an introduction to the location and a scene about getting dressed in the morning) aloud. For the second page of the story, the experimenter demonstrated pretense using the cutout characters by making them move and talk before placing the cutouts on the board. The experimenter instructed the child “Now it’s your turn to pretend” before reading the rest of the story.

Non-pretense control.

Children in the non-pretense conditions were instructed to simply place the felt cutouts on the board to match the scenes described in the story. They were not given encouragement to pretend. The experimenter read the first two pages of the story aloud and placed the corresponding pieces on the storyboard without engaging in pretense actions or voices, then told the child, “Now it’s your turn to put the pieces on the board.”

In all conditions, the experimenter read each of seven subsequent pages for the children and then placed the appropriate cutouts in front of them. If children placed the cutouts on the board incorrectly (e.g., a child who heard about Planet Opposite placed fish in the water or a child who heard about Funtown put them in the sky), the experimenter provided feedback one time by rereading that page of the story and gave the child another chance to place the cutout.

Measures

Target EF Task.

Immediately following the experimental manipulation, children completed the Less is More EF task (Carlson et al., 2005) at a small, child-sized table. In this task, children are presented with two piles of treats, one large (five treats) and one small (two treats), and must point to the smaller pile in order to keep the larger pile for themselves. Children were allowed to try two treat choices and their favorite treat was used in the task to ensure they were motivated to obtain the reward. The experimenter then introduced a naughty monkey puppet named Chris and instructed children that, “The box you point to goes into Chris’s cup and he keeps those treats, and you get to keep the treats in the other box and they’ll go into your cup.” Following a rule-check and a practice trial, participants completed 16 test trials with a rule reminder after the eighth trial. Performance was measured by the number of times children pointed to the smaller pile of treats (ranging from 0–16).

Baseline EF.

Peg tapping (Diamond & Taylor, 1996) served as a baseline measure of EF. Children were provided with a small wooden dowel and instructed to tap once if the experimenter tapped twice, and twice if the experimenter tapped once. Following practice trials, the children were presented with 16 counterbalanced test trials. Performance was measured by the number of correct responses (ranging from 0–16).

Receptive vocabulary.

The Peabody Picture Vocabulary Test 4th edition (PPVT-IV; Dunn & Dunn, 2007) is a standardized measure of receptive vocabulary, which has been shown to correlate with both pretense and EF (Taylor & Carlson, 1997). Following standard task instructions, children were presented with a series of cards, each showing four pictures. The experimenter would say a word depicted by a picture on the card and ask the child to point to the correct picture. Raw scores were used for analyses.

Propensity for pretense.

Children’s propensity toward engagement in pretense was measured using the toy phone task (Tahiroglu et al., 2011; Taylor et al., 2013). Children were asked if they would like to pretend to call someone to play a game. The experimenter presented children with a toy phone and proceeded to observe children’s pretense actions with the phone, recording six potential behaviors: picking up the receiver, dialing, holding the phone to their ear, talking to the friend, pausing as if taking turns while the friend is talking, and reporting what the friend said. Each action received one point so that total scores could range from 0–6.

Affect coding

Because pretense is characterized as pleasurable (Krasnor & Pepler, 1980), it is possible that improved EF performance could be attributed to increases in affect (Yang & Yang, 2014) as a side effect of pretending, rather than the act of pretending itself. To address this issue, we examined children’s affect during the story activity. Two coders who were blind to the hypotheses of the study rated the affect of 54 participants during the storyboard task using video recordings. (Due to technical issues, six recordings were of insufficient quality to accurately observe children’s affect and were not coded.) Based on existing scales used to code positive, neutral, and negative affect in young children (LaFreniere & Sroufe, 1985; Lozoff et al., 2007; Winsler & Wallace, 2002), we developed a detailed five point scale to assess children’s affect as they completed the storyboard task (see table 2). Interrater reliability was good (ICC = .89). Discrepancies in affect scores were settled by a third coder.

Table 2.

Affect Coding Rubric

Score	Description
5	Highly positive affect: Child is energetic, frequently laughs, giggles, smiles and/or verbally confirms enjoyment of task
4	Somewhat positive affect: Child occasionally smiles or laughs, is engaged in the task, may briefly mention enjoyment of task
3	Neutral affect: Child engages in task but shows no overt signs of enjoyment or displeasure
2	Somewhat negative affect: Child shows some signs of displeasure or frustration, occasionally frowns or fusses, may briefly mention disinterest in task
1	Highly negative affect: Child frequently frowns, shows frustration or anger, fusses, attempts to leave the room, and/or verbally confirms dislike for task

Results

Preliminary Analyses

Preliminary analyses suggested that randomization was successful: Conditions did not differ by age, sex, receptive vocabulary (M = 72.60, SD = 13.38), baseline EF (M = 7.25, SD = 4.76), or propensity for pretense (M = 3.72, SD = 1.71). Baseline EF (peg tapping) was the only measure that significantly predicted scores on Less is More, r(59) = .33, p = .01 (see Table 3 for full correlation matrix). Less is More scores did not differ by gender, p = .27.

Table 3.

Bivariate Correlations

	Age	Receptive Vocabulary	Baseline EF	Propensity for Pretense
Less is More	.03	.21	.33**	.05
Age (in months)		.16	.17	−.14
Receptive Vocabulary			.30*	.02
Baseline EF				−.08

Note.

^*p < .05

^**p ≤ .01

Surprisingly, children showed less positive affect in pretense conditions (M = 3.32, SD = 0.90) than non-pretense conditions (M = 3.88, SD = 0.77), F(1, 52) = 6.05, p = .02, η_p² = .10. Affect did not differ by story content.

Condition Differences

Using Less is More scores (total correct out of 16) as the dependent variable, a 2(story content: fantasy, reality) x 2(engagement: pretense, non-pretense control) ANOVA yielded a significant main effect of engagement type, F(1, 56) = 13.82, p < .001, η_p² = .20. As shown in Figure 2, children who were encouraged to engage in pretend play prior to the Less is More task had higher inhibitory control scores than children who engaged in the non-pretense activity (M_pretense = 10.90, SD = 3.39, 95CI [9.57, 12.23]; M_non-pretense = 7.40, SD = 3.78, 95CI [6.07, 8.73]). Story content had no effect on subsequent inhibitory control (M_fantasy = 9.33, SD = 3.69, 95CI [8.00, 10.67]; M_reality = 8.97, SD = 4.30, 95CI [7.63, 10.30]), F(1, 56) = 0.15, p = .70, η_p² =.00. We also found no interaction between engagement and story content, F(1,56) = 0.10, p = .75, η_p² = .00. These results held when controlling for age, gender, affect, baseline EF, propensity for pretense, and receptive vocabulary (both when covariates were entered individually and simultaneously in the model).

Figure 2.

Mean EF Performance by Condition (Story Content x Mode of Engagement)

Note. Error bars represent 95% CI.

Pretense and Learning Across Trials

It was possible that the effects of pretense would taper off quite quickly when children returned to real-world activities. To explore the pattern of performance across trials, we divided Less is More scores into two blocks of eight trials, which were separated by a rule reminder. Given the previous results showing no effect of story content, we collapsed the two types of stories for this analysis. A 2(engagement: pretense, non-pretense) x 2 (block: first 8 trials, last 8 trials) repeated measures ANOVA, confirmed a significant main effect of pretense, F(1, 58) = 13.94, p < .001, η_p²= .19. It also yielded a significant main effect of block, F(1, 58) = 17.49, p <.001 η_p² = .23; children tended to show better inhibition as the task progressed. However, we found no interaction between distancing condition and block, F(1, 58) = 1.61, p = .21, η_p² = .03, suggesting that learning occurred at a similar pace for children who were in the pretense and non-pretense conditions. Overall, it is important to note that although all children showed improvements in LIM scores across trials (replicating Carlson et al., 2005), the effect of pretense held even after accounting for block and interaction effects. Those in the pretense conditions immediately showed superior inhibitory control and were able to sustain it across trials (Figure 3).

Figure 3.

Mean EF Performance for Pretense and Non-Pretense Conditions by Task Block

Note. Error bars represent 95% CI.

Discussion

In the current study, we set out to examine the individual contributions of two, often overlapping, types of fiction on children’s EF: fantastical stories and pretend play. Story content had no impact on EF. Reading stories with realistic or fantastical themes resulted in nearly identical inhibitory control scores. However, mode of engagement with the stories did influence EF performance. Regardless of the story they acted out, children who were encouraged to engage in story-based pretend play significantly outperformed their peers who engaged in a non-pretense activity. The effect size was large (η_p²= .20) and pretense-related improvements endured across task trials. Overall, this brief (approximately 10 min) episode of pretense improved 3-year-old children’s scores by nearly 50% over the non-pretense play conditions, even when they were faced with real, tempting rewards during the task, as opposed to substituting the rewards with symbols as in Carlson et al. (2005).

In line with our hypothesis, encouraging children to engage in pretend play facilitated EF. The effects of pretense were robust, holding over and above a comprehensive set of control variables including baseline EF and propensity to engage in pretend play. Of note was the fact that they also held when controlling for affect, thus suggesting that the benefits of pretending are due to more than simply increased positive mood (Yang & Yang, 2014). Comparisons to previous studies of the Less is More task confirm that scores in the non-pretense condition were closely aligned with prior results. Carlson et al. (2005) found that 3-year-olds had a .49 probability of making a correct response in the standard condition; similarly, children averaged just under 50% correct in our non-pretense conditions. Therefore play condition differences in this study are likely due to pretense boosting EF scores, not non-pretense activities hurting them. This research corroborates previous results which showed that training pretense (e.g., Thibodeau et al., 2016) and short-term role play manipulations (e.g., Karniol et al., 2011; White & Carlson, 2016; White et al., 2017) also can improve EF. The current study further extends prior evidence to suggest that pretense might be driving the results of such interventions.

The question as to why pretense improved EF in this study remains unanswered. The pretend play manipulations did not provide explicit messaging about inhibitory control strategies. As noted above, affect also cannot account for the difference in Less is More scores between pretense and non-pretense conditions. Nor does it appear that the pretense manipulation relied on overt engagement in pretend play. We observed few instances, in either of the engagement conditions, of verbal or behavioral elaboration during the placement of storyboard pieces like making a character talk or providing additional unscripted movements. Children in the pretense conditions exhibited such behaviors on an average of only 0.86 (out of 7 possible) pages of the story, whereas children in the non-pretense control group did so for an average of 0.63 pages (p = .50). Rather, children in both groups tended to simply place the pieces on the storyboard as depicted in their respective story. Even despite direct encouragement in the pretense conditions, neither group was likely to engage in much observable pretend play. Thus, we must consider the possibility that the mechanism is representational: the flexible mindset that children enter into during pretense could have primed greater flexibility on Less is More. This idea aligns well with Vygotsky’s (1978) theory that pretense contributes to internal systems of thought through which children can break free from stimulus control, as well as psychological distancing theories (Sigel, 1970) that similarly promote the role of separation from one’s immediate context. Consistent results for psychological distancing through role-play have been reported for older preschoolers (4- to 6-year-olds) using a variety of self-regulation measures including card-sorting (White & Carlson, 2016), persistence (White et al., 2017), and emotion regulation (Grenell et al., 2019).

One distinct feature of the present study is that there was a demarcation between pretense and the EF task rather than asking children to pretend while performing the task. This suggests the benefits of pretense can carry over into more “serious” activities. However, if the “as-if” stance of pretense (Garvey, 1991; Reynolds, 1976) is thought to occur at a representational level that is distinct from reality (Friedman & Leslie, 2005; Harris, 1994, 2000; Lillard, 1993, 2001; Olson, 1991), how does pretense cross this boundary to impact EF in real life? Although we made an effort to signal that the pretense was finished (i.e., moving to a new location, away from the storyboard and play activity), one plausible explanation is that children might have seen the Less is More task as a continuation of pretense. The intervention also could have operated at a more implicit level. Children might have been primed by the pretense episode to take on a more distanced and flexible mindset that was then applied to the EF task. This could be especially effective for children because priming taps into automatic processing systems and thus bypasses top-down control (Kesek et al., 2011). If the flexible thought processes that characterize pretense carry over into the EF task, inhibiting a first response could be less effortful. It remains unknown how long children would benefit from such priming; however, it is important to note that children in the pretense condition performed better than those in the non-pretense condition across both test blocks, suggesting the effect was not fleeting. In the long-term, practice with psychological distancing via pretense could allow children with underdeveloped EF skills to experience what it is like to perform at a more advanced level (“a head taller,” Vygotsky, 1967). In turn, through continued use, the desired or “correct” way of responding could become more facile and eventually form a new habit (e.g., Hebb, 1946). Further research will be necessary to explore this potential explanation.

Contrary to our hypothesis, we found no evidence that fantasy content affected inhibitory control, for better or for worse. Although theory (e.g., Vygotsky, 1978; Sigel, 1970) would have predicted a boost from fantasy content, previous empirical evidence has shown mixed results. Some studies have revealed benefits of fantasy, others benefits of reality; our findings lie firmly in the middle. In a review of the literature on children’s understanding of fantasy and reality in fiction, Hopkins and Weisberg (2017) note that inconsistencies might be explained by differing levels of fantasy content across interventions. Indeed, this was the case for Lillard, Drell, and colleagues’ (2015) research on EF and television programming. They found that high levels of fantasy in cartoons made children’s EF plummet compared to no-fantasy play, but low-fantasy cartoons did not have the same effect. The fantasy content in Planet Opposite could be described as moderate. The story took place in another world and all plot lines could not or do not typically happen in real life. However, some aspects of the story were certainly more fantastical (e.g., fish flying, house-sized mice) than others (e.g., the custom of wearing pants on your head). Highly fantastical elements like magic or anthropomorphized animals did not occur. It is possible that a moderate level of fantasy was not sufficient to affect EF. Relatedly, children might not represent our fantastical and realistic stories as members of two separate categories of fiction; given the between subjects design, with each child hearing only one story, this was not a question we could answer in our study. Future research should probe children’s ability to distinguish between the stories used in a study and examine whether explicit categorization is necessary for stories to affect behavior. More work is needed to investigate different levels of fantasy on EF across media contexts and types of play.

The results of this study highlight the powerful impact of realistic and fantastical pretense on EF. Together with previous experiments (Grenell et al., 2018; Karniol et al., 2011; White & Carlson, 2016; White et al., 2017), interventions (Elias & Berk, 2002; Saltz, Dixon, & Johnson, 1977; Thibodeau et al., 2016), and longitudinal studies (White, Thibodeau, Palermo, & Mikulski, 2021), we are beginning to paint a detailed picture of how and in what contexts pretense helps children achieve better self-regulation. These studies have important implications for pretense-based curricula in early education settings and interventions to strengthen early EF. Of course, pretense is only one of many contributors to EF development (see Carlson, Zelazo, & Faja, 2013 for a review). Thus a comprehensive approach to developing EF skills in young children should consider an array of factors such as caregiving (Bernier, Carlson, & Whipple, 2010; Hughes & Ensor, 2005), SES (Blair & Razza, 2007; Noble, Norman, & Farah, 2005; Obradović, 2010), schooling (Lillard & Else-Quest, 2013), training programs (Diamond & Ling, 2016), and real-life experiences (Lillard & Taggart, 2019), as well as pretend play.

Limitations and future directions

The results of the current study must be interpreted in light of its limitations. First, the independent variable consisted of only one EF task score. Given that one-task variables can be statistically noisy, future work should consider the use of more comprehensive EF batteries (Willoughby, Wirth, & Blair, 2011). Because the duration of the effect of pretense on EF is unknown, researchers will need to pay particularly close attention to order effects. Second, one might argue that a fictional story that hinges on opposites could teach to the EF test (Lillard, Dore, Hopkins, & Smith, 2015), so to speak. Less is More requires children to choose an action that is the opposite of typical behavior (i.e., point to the treats you want to give away, rather than the ones you want to keep). However, we found no evidence that stories containing opposites affected children’s EF performance. Third, the sample size was relatively small. This could have contributed to our inability to find an effect of story content or an interaction between content and pretense. However, the confidence intervals and effect sizes obtained here (near 0 for story content and the interaction, but large for mode of engagement) suggest that the current results are likely to accurately reflect population means. Nonetheless, research with larger sample sizes is needed to corroborate the results of the current study and to explore the effects of individual differences (e.g., baseline EF or pretense abilities) on the efficacy of pretense interventions for EF. Lastly, our pretense and non-pretense conditions were designed to minimize action differences across groups while maximizing mindset differences and pretending was, to some extent, constrained by the narrative of a book. This form of pretending is not necessarily representative of children’s naturalistic pretense. As such, one must be cautious in generalizing the current findings to all types of pretend play. Future work would benefit from further exploration of children’s actions and responses during pretense episodes in order to link specific behaviors to EF outcomes.

A crucial area for future study is to understand how age could impact the effects of pretense and fantasy on EF. Each of these abilities change rapidly across the first few years of life. At 3 years of age, children are typically gaining inhibitory control, but at age 2 they were tackling simpler working memory skills, and by age 4 they are already working to develop more advanced set shifting abilities (Garon et al., 2008). At the same time, pretend play becomes more complex and decontextualized (see White, Carlson, & Zelazo, 2020 for a review) while children gain clarity on how fantasy applies—or does not apply—to the real world (Flavell, Flavell, & Green, 1989; Richert & Schlesinger; 2016; Shtulman, 2009; Shtulman & Carey, 2007). With so many moving parts, we cannot simply assume that the effects found here will hold at different ages and for other facets of EF.

Finally, future research should continue to investigate how pretend play content affects children’s cognition. Beyond fantasy themes, children can learn from a variety of lessons embedded in pretense. For example, after engaging in realistic pretend play with relevant content, children show improved understanding of object functions (Hopkins, Dore, & Lillard, 2015), emotion regulation (Goldstein & Lerner, 2017), and causality (Buchsbaum et al., 2012). What other lessons could children learn through pretense? To what extent would children learn these lessons if they were presented within a fantastical, versus realistic, pretense scenario? We must also keep in mind the possibility that some pretense content might hinder cognition, including EF. More research is needed to investigate negative character traits (Karniol et al., 2011), violent themes (Berk, Mann, & Ogan, 2006; Dunn & Hughes, 2001), and highly fantastical events (Lillard, Drell, Richey, Bogusweski, & Smith, 2015; Richert & Schlesinger, 2016). This area of research is only beginning to take shape and there is still much to learn about how pretending contributes to burgeoning EF abilities in early childhood.

Conclusion

In summary, the current study found that encouraging pretense based on realistic and fantastical stories improved children’s inhibitory control. As such, it lends support to the long-held view that pretending can facilitate the development of self-regulatory abilities in the early years of life. It also provides important new evidence to suggest that, at least when it comes to inhibitory control skills, the degree of fantasy in pretend play might matter less than engaging the pretense mode itself. Overall, this research underscores the importance of further experimentation to determine how and in what contexts various aspects of fiction influence children’s cognition.

Highlights.

1 3-year-old children heard realistic or fantastical stories

2 They were encouraged to engage in story-based pretend play or a control activity

2 Children in the pretense group exhibited increased inhibitory control

3 Story content (realistic or fantastical) did not affect inhibitory control

Appendix

Realistic Story	Fantastical Story
Have you heard about Funtown? It’s a great place where kids have lots of fun!	Have you heard about Planet Opposite? It’s a silly place where everything is backwards!
Time to wake up and get dressed. Put on your favorite pants!	Time to wake up and get dressed. Put on your pants on your head!
Let’s go to the park! In Funtown, everybody walks with a friend.	Let’s go to the park! On Planet Opposite, everybody walks on their hands.
Look at the trees! They grow really big!	Look at the trees! They grow upside down!
Let’s race to the zoo! In Funtown, when someone says “Go!” we run really fast!	Let’s race to the zoo! On Planet Opposite, stop means go. So, when I say stop, run really fast!
In Funtown, many animals are at the zoo! Kitties say “Meow” and doggies say “Woof! Woof!”	On Planet Opposite, many animals are at the zoo! Kitties say “Woof! Woof!” and doggies say “Meow!”
Butterflies live in the sky ... and fish swim underwater.	Fish live in the sky ... and butterflies swim underwater.
An elephant is as big as a house, and mice are teeny tiny!	A mouse is as big as a house, and elephants are teeny tiny!
And at the end of the day, when it’s time to sleep, the moon comes up to say goodnight! What a great day in Funtown!	And at the end of the day, when it’s time to sleep, the sun comes up to say goodnight! What a great day on Planet Opposite!