Study participants

The Human Research Committee at Hampden-Sydney College approved these experiments on 2/26/2016 for all in-person sessions (24 children), and the University of Virginia Institutional Review Board for Social & Behavioral Sciences approved these experiments on 6/21/2021 under Protocol #4486 for all online data collection sessions. Recruitment and data collection only began after IRB approval was granted at each institution. Study visits for the first 24 participants occurred in college academic buildings and public library conference rooms, whereas the remaining participants had their initial and follow up meetings via Zoom due to COVID-19 precautions. Written consent obtained was from legal guardians of participants; verbal assent obtained from child participants.

The study contained 41 participants between the ages of 6 and 8 living in central Virginia and their parent or guardian (N = 82 participants). Subjects were recruited over a period of several years in two different locations in central Virginia separated by about 75 miles. Participant children came from many different types of schools and classrooms, and were recruited through Facebook posts, direct emails to area elementary schools, flyers posted on community bulletin boards, and in person recruitment outside local public libraries, museums, and child weekend sporting events in the areas. The diverse recruitment strategies and geographic distribution of participants across multiple educational settings was designed to capture the natural variation present in real-world family reading practices. The mean age of the child participants was 6.8 + /- 0.12 years old; there were no significant age differences between groups. (See Table 1 for demographics.) We did not collect data about the type of schooling the child had (private, public, or homeschooled).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographic information of study participants.

https://doi.org/10.1371/journal.pone.0340068.t001

Empathy and creativity measures

We assessed self-reported empathy using a version of the Interpersonal Reactivity Index (IRI) previously used with adults [1] and children [27], further refined by Garton and Gringart for use in 8- and 9-year-olds and validated in 413 schoolchildren [28]. We used the two-factor Garton and Gringart 12-item scale [Table 3 in [28]], where Factor 1 has six cognitive empathy questions and Factor 2 has six emotional empathy questions, with two modifications. First, we included two additional questions from the IRI Fantasy sub-scale specifically about story characters. The original IRI designed for adults had two cognitive subscales: 1) perspective taking and 2) fantasy, and two emotional subscales: 3) empathic concern and 4) personal distress [1]. When Garton and Gringart carefully validated these measures in 8–9 year olds, analysis allowed the personal distress items to be redistributed and equally loaded onto those 2 factors, cleanly separating items into a two factor measure of cognitive and emotional empathy, but discarding the Fantasy items (which describe the likelihood that a person identifies with a fictional character) [28]. Although Davis’ experimental work took a similar direction after the development of the IRI as well [29], these story fantasy questions are highly applicable to our study design. Therefore, we included two questions from Litvak and colleagues’ children’s empathy measure [27] that specifically assessed perspective-taking with fictional characters, as this directly related to the skills our study children were practicing during story reading. These questions were: “When reading a book, I try to imagine what the people in the story are thinking” and “It is easy for me to pretend that I am the star of my favorite movie”.

Second, we simplified the paper and pen answer scale from ‘Not like me at all’ through to ‘Hardly ever like me,’ ‘Occasionally like me,’ ‘Fairly like me,’ and ‘Very like me’ to a verbal response of ‘Yes,’ ‘No,’ or ‘Sometimes’ for the 6–8-year-olds in our study. Garton and Gringart’s measurement scale (designed for 8–9 year olds) was not linear because some (but not all) points on the 5-point scale included a temporal element (‘hardly ever’ and ‘occasionally’) [28]. The simplification of our scale to a ‘Yes,’ ‘No,’ or ‘Sometimes’ attempts to eliminate this potential issue for younger children (aged 6–8 years) that cannot differentiate easily between the terms ‘occasionally like me’ and ‘fairly like me.’ In our study, ‘Yes’ was scored as 2, ‘Sometimes’ was scored as 1, and ‘No’ was scored as 0.

Our decision to simplify the empathy scale response format from the original five-point Likert scale to a three-point “Yes/No/Sometimes” format was driven by developmental considerations and empirical evidence from previous research with young children. Garton and Gringart (2005) identified significant methodological concerns when adapting the adult IRI for 8–9 year olds, noting that their five-point scale was “not linear because some (but not all) points on the 5-point scale included a temporal element (‘hardly ever’ and ‘occasionally’)” and concluded that children had difficulty differentiating between terms like “occasionally like me” and “fairly like me” (p. 24). Extensive research demonstrates that children ages 6–8 perform significantly better with simplified response formats due to developmental limitations in concrete operational thinking, executive function constraints, and limited metacognitive awareness required for complex self-report measures [30,31] Coombes et al. (2021) conducted a systematic review of 81 studies with over 13,000 participants and established that children ≤7 years “think dichotomously so need two response options” while children >8 years “can reliably use a 3-point scale” [32]. Our simplified three-point scale (scored as No = 0, Sometimes = 1, Yes = 2) eliminates measurement error from inappropriate scale complexity while maintaining the ability to capture meaningful variation in empathic responses developmentally appropriate for this age group.

The empathy measures were administered verbally in individual, one-on-one sessions with trained research assistants. This administration format was chosen because it allowed research assistants to ensure that each child understood the questions and could provide appropriate responses. Research assistants were trained to present each question clearly and to rephrase questions using age-appropriate language if a child appeared confused, while maintaining the core meaning of each item, which very rarely happened. The three-point response format (“Yes,” “No,” “Sometimes”) was presented verbally and each session was conducted in a quiet, comfortable setting where children could focus on the questions without distraction.

Creativity was assessed using the Wallach and Kogan alternative uses verbal test in children, specifically examining responses for fluency and originality in divergent thinking [33].

Initial visit experience

Participants attended individual initial appointments, where the participant was randomized to a treatment group before each initial visit based on a random number generator to prevent selection bias. The child attended the initial session with a caregiver, where informed consent was obtained from both the child and caregiver. Participants were told they were participating in a study examining the impact of reading to children. At the initial visit, a study team member presented a short PowerPoint presentation that highlighted the benefits of reading aloud to your child, backed by research showing it improves school readiness and vocabulary [34,35]. The terms empathy and creativity were not mentioned during the presentation.

The parents of the participants were randomly assigned to the “read through” or the “pause” group and given packs of books accordingly. Engaging picture books were chosen that would likely be new to the children, but they were typical storybooks (not focused on an overt theme of empathy). The books we used were as follows: A Letter for Leo by Sergio Ruzzier (isbn: 9780544223608), A New Friend for Marmalade by A.M. Reynolds (isbn: 9781481420464), Cub’s Big World by Sarah Thomson (isbn: 9780544057395), Library Lion by Michelle Knudsen (isbn: 9780763637842), Nugget & Fang by Tammi Sauer (isbn: 9780547852850), Stuck with the Blooz by Caron Levis (isbn: 9780547745602), The Tooth Fairy Wars by Kate Coombs (isbn: 9781416979159). Selected short stories had some form of social conflict, such as a small polar bear getting separated from his mother, the tooth fairy fighting with a boy who wants to keep his lost tooth, or shark and minnow best friends who realize society says they should not be friends.

The read through group was given unaltered books, while the pause group was given books containing a small sticker with two question prompts placed in each book at a single story conflict point. Questions read “What do you think _(insert main character’s name)__ is feeling right now?” and “What would you do if you were _(insert main character’s name)__?”. Names were changed in each book to match the character in the specific story conflict. Parents in the pause group were instructed to ask the questions one at a time to their child. If the child responded, the parent just listened before continuing to read the story. If the child did not respond, the parent waited 30 seconds before reading further. Parents were instructed not to ask follow-up questions or to offer commentary on child responses.

The initial visit presentation included short video models of a mother and child sitting on a bed reading the picture book Are You My Mother by P.D. Eastman aloud (isbn: 9780679890478). The video for the pause group demonstrated how to read the questions “What do you think Baby Bird is feeling right now?” and “What would you do if you were Baby Bird?” and wait for answers, whereas the read through video showed a mother reading the story to a child as normal (S1 Video). Following the presentation, a study team member administered the Garton and Gringart modified IRI test and the Wallach and Kogan alternative uses test (Q1 uses for a paperclip, Q2 things with wheels) to the child [28,33], while the caregiver completed a survey with demographic information.

Study instructions and follow up

Both groups were instructed to read each book just twice over the course of the two-week study period and to put the books away while it was not study time to prevent unequal exposure. Book readings were logged by participants in a Reading Log to ensure groups followed study instructions. After two weeks elapsed, the participants were brought back for a follow up meeting where the same tests were administered (identical modified IRI and Wallach/Kogan based on Q1 uses for a newspaper, Q2 things that are round). All participants received the seven books used for the study for free. At the conclusion of the study, three participants were drawn at random to receive $50 Barnes and Noble gift cards.

We selected a two-week intervention period to balance several competing considerations. This timeframe provided families with sufficient time to complete 14 reading sessions (seven books read twice each) while maintaining a manageable commitment that would support study completion. Pilot testing with families indicated that a two-week daily reading commitment felt achievable while still requiring meaningful engagement. Longer intervention periods would have increased participant burden and likely reduced completion rates. Since research on brief empathy interventions suggests that focused, short-term interventions can produce measurable changes in empathy-related outcomes, particularly in children who are still developing these skills, we designed our intervention to test whether even a brief modification to typical reading practices could show promise as a practical, family-friendly approach to enhancing empathy and creativity.

Data analysis

Statistical analysis

For all measures except fantasy ratings, a series of 2 x 2 mixed ANOVA models were performed using the rstatix and stats packages in R [40] to analyze the effects of the treatment groups (Condition: pause intervention, “control” reading intervention), the two occasions (initial test, follow-up test), and their interaction had on each of the six measures. Because the range of possible fantasy ratings was small (0–4), we instead performed a cumulative link mixed effect model using the ordinal package [41] to analyze the same effects of treatment, occasion, and their interaction on the fantasy measure. Due to the high number of analyses, a Bonferroni correction was performed on all p-values in the six primary models (five mixed ANOVA models plus the cumulative link mixed effects model for fantasy), assuming a familywise error rate of four comparisons for models that focused on the four empathy measures and two comparisons for the two creativity measures.

To check assumptions for the ANOVA models, we visualized the distributions of the residuals of these models (S1 Fig) and checked for outliers using the rstatix package in R [40]. No extreme (> 3*IQR) outliers were found, and the normality and homogeneity of variance assumptions appeared sufficiently met for all ANOVA models except for those using emotional empathy and originality as measures. Due to the potential concerns with these two models, we double-checked the assumptions of the models using statistical assumption testing methods. More specifically, we performed Shapiro-Wilk on the residuals of the models to check normality, and performed Levene’s tests at each occasion to check homoscedasticity across the intervention groups, using the stats and car packages in R, respectively [42]. When doing so, no models had significant Levene’s test results (meaning that our assumption of homoscedasticity was met), but the model that used emotional empathy had a significant Shapiro-Wilk result (Shapiro-Wilk, p = .027). However, when performing a Box-Cox analysis (using the car package in R) to determine possible transformations, the suggested lambda of 1.41 suggested no transformation should be performed. Thus, taken holistically with the visual assumption checks, no power transformations on the data were deemed necessary to use for any measures. Because our within-subjects measure (Occasion) only had two levels, sphericity was automatically met, and no assumption test on sphericity was necessary.

Following data collection, we conducted a sensitivity analysis to determine the minimum effect sizes our study was powered to detect. This post-hoc power analysis informed interpretation of both significant and non-significant findings by clarifying which effect sizes our design could reliably detect versus those for which we lacked adequate statistical power.

Fostering empathy

Our results demonstrate that both reading groups showed significant improvements in cognitive empathy and total empathy scores from initial to follow-up visits, suggesting that daily reading routines may enhance children’s empathic abilities. This finding is consistent with research suggesting that reading facilitates the development of empathy by allowing children to engage with diverse characters and perspectives. Books serve as a shared reference for discussing characters’ emotions and thoughts, providing unique opportunities for conversations about mental states [43,44]. In addition, scaffolding, such as guided discussions around storybooks, can enhance children’s ability to extract moral lessons and apply them to their own experiences [45]. Our data suggest that the scaffolding within the book storyline may be sufficient to produce empathy changes when simply read by the caregiver, even without the additional reflection prompts during the story.

We found that cognitive empathy showed significant improvement after daily picture book reading over a two-week period, while emotional empathy did not reach statistical significance. This is consistent with previous literature suggesting that cognitive empathy may be more responsive to short-term perspective-taking practice [17]. Regular reading helps children better understand and identify with others’ emotions through perspective-taking [46], and narratives featuring protagonists can aid children in adopting new perspectives. Using picture books to address social issues integrates critical literacy skills and fosters social awareness, promoting empathy [47], and may potentially boost ethnocultural empathy [2].

Emotional empathy has been traditionally harder to change over small periods of time—it may require more time to alter deep rooted ways of identifying, processing, and acting on emotions, as it involves different brain regions than cognitive empathy [48]. Future research should investigate if emotional empathy measures may be changed with longer intervention periods or if the observed cognitive empathy improvements persist over time as children move through development. The finding that pausing to reflect during stories did not produce statistically significant empathy gains compared to reading stories without pausing suggests that children may naturally engage in perspective-taking while listening to stories or that the immersive experience of continuous reading has its own benefits [21].

Fostering creativity

Our study also found significant improvements in both creative fluency and creative originality measures after the 2-week reading experience in both groups. Creativity is often perceived as an innate talent or quality, but our results support research showing it can also be enhanced through practice, especially employing idea generation and cognitive training approaches [49]. The improvements observed in both groups could reflect genuine practice effects from the reading intervention, increased comfort with the testing procedures, or the benefits of creative practice in a supportive environment during the initial assessment.

Notably, we found that children in the pause group showed significantly greater improvements in creative fluency compared to the read-through group. This interactive scaffolding of the creative process—where children generated ideas about possible actions for story characters—represents a perspective-taking exercise that may simultaneously engage both creativity and empathetic processing. This task asked children to imagine possible storylines for the second half of the book, tapping into both creative and empathetic abilities.

Previous research examining the Alternative Uses Task in 4–6 year olds found that scores increase over this developmental period, likely because the test procedures heavily rely on children’s verbal skills, which are also developing over time [50]. However, our data revealed that older children in our sample (ages 6–8) showed lower creative originality scores than younger children, a pattern consistent with research documenting creativity declines in school-age children. Fostering creativity can be challenging in a world with rigid values, as strict rules may lead children to prioritize social acceptance over creative expression as they grow older [51]. Seminal work on creativity during development postulates a slump in creativity evident in around fourth grade [52], though this pattern varies among individual children. In fact, some children exhibit an increase in creativity during this time, with longitudinal changes in frontal neural connectivity linked to developmental improvements in creative thinking ability using functional near-infrared spectroscopy (fNIRS) [53].

The complex literature on creativity development may be the result of a modern social and educational structure that contains mixed messaging — while our culture theoretically emphasizes the importance of creative expression, assessment systems in both schools and the workforce often devalue it. The reflective reading approach in our study provides practice with imaginative and empathetic engagement through narratives, which may counteract observed age-related decreases in creative expression. Future research should examine whether these effects persist over time, as previous research showed that creativity training can still boost divergent thinking a year later [54].

Developmental timing of interventions

We chose to investigate the development of empathy skills during a developmental period (6–8 years old) when children are first exploring reading independently, but also still undergoing intense synaptogenesis that begins before birth and extends into adolescence [55], experiencing dynamic brain changes in response to environmental exposure. This developmental stage may represent a sensitive period for interventions targeting peer social skills, as early empathy training appears vital for social and emotional growth when children navigate social conflicts and relationships [2]. Forming close friendships and adapting to social norms are key developmental tasks during this stage [56], with growth partly driven by social learning opportunities at home, including modeling and reinforcement [57]. Our results demonstrate that relatively brief interventions during this time-period can produce measurable changes in creativity and empathy.

Measurement of empathy and creativity can be nuanced and complex, but there is some evidence societal levels are dropping [4,5,58,59]. This study provides more evidence that a simple bedtime reading routine can boost these key life skills. Unfortunately, reading activity levels and reading enjoyment levels have also dramatically and consistently dropped since 2010 [60,61]. Parents are aware that reading to their children has benefits, but further research should explore the broader implications of these findings for parenting and education, particularly in an increasingly digital age. This is particularly important for child development as the rise of artificial intelligence technologies makes the skills of creativity and empathy even more valuable assets for humanity.

Study limitations

The finding that both reading groups showed similar improvements in empathy and creativity, regardless of the reflection intervention, suggests several alternative explanations for our results. Study participants watched a visual model showing a mother reading to a child on her lap, and research indicates that fostering intimacy through positive touch and close physical presence during reading sessions can strengthen emotional connections and enhance empathy, and this quality time seemed beneficial for both groups. Rather than reflection-specific benefits, the observed changes may reflect the inherent value of structured, daily reading experiences that provide repeated exposure to character emotions and social situations embedded within narrative contexts. The lack of differential effects between groups could indicate that children naturally engage in perspective-taking while listening to stories, making explicit prompting redundant during this developmental period.

Our study has several important limitations that must be acknowledged. Most critically, we lacked a no-reading control group, which limits our ability to make causal inferences about the effects of reading itself. Without a control group that did not receive books or reading instructions, we cannot definitively separate the effects attributable to reading from other factors such as natural developmental changes over the two-week period, practice effects from taking the empathy and creativity measures twice, or increased parent-child interaction time regardless of reading content.

The decision not to include a no-reading control was driven by several factors: 1) resource constraints limited us to approximately 40 families, and splitting these across three groups would have substantially reduced statistical power for detecting differences between our two reading groups, 2) recruitment and retention would have been more challenging for a no-reading group, particularly since families expressed interest specifically because they wanted their children to receive books, 3) our primary research question focused on the incremental benefit of adding reflection rather than comparing reading versus no reading, and 4) ethical considerations made it difficult to justify asking families to abstain from reading for two weeks given the well-documented benefits of parent-child reading. This design limitation means our findings may simply demonstrate that adding reflective questions to reading may provide incremental benefits beyond reading alone.

Additionally, our sample was predominantly from high-income families (92.11% earning > $50,000 annually), which substantially limits the generalizability of our findings to more diverse socioeconomic populations where access to books, reading habits, and family literacy practices may differ significantly. The relatively small child sample size (N = 38) further constrains our conclusions, as post-study power analysis [62] revealed sufficient (80%) power to detect only medium to large effects (partial R² ≥ 18%) [63], but may have been underpowered to detect smaller but potentially meaningful changes. The absence of significant treatment × occasion interactions for empathy measures could reflect either a) truly similar effects of both reading approaches on empathy development, or b) smaller effect sizes that our study lacked sufficient power to detect. Given that both groups showed improvements in empathy from baseline to follow-up, it is possible that structured bedtime reading benefits empathy development regardless of whether reflection questions are included, but that our study was underpowered to detect subtle differences in the magnitude of these improvements.

The brief two-week intervention period may not have provided sufficient time for more robust empathy and creativity changes to emerge, particularly for emotional empathy, which research suggests requires longer periods to modify. Furthermore, our simplified empathy scale (Yes/No/Sometimes) and the constrained range of possible scores (0–6 for subscales) may have created ceiling and floor effects that obscured more nuanced changes in empathetic responding. Given these methodological constraints, these findings should be interpreted as preliminary evidence suggesting that consistent reading practices may support empathy and creativity development in young children, rather than definitive proof of causal mechanisms.

Future directions

Future research should address the methodological limitations identified in this study through more rigorous experimental designs. Most importantly, studies should include a true control group that receives no reading intervention, as well as additional comparison conditions such as a group receiving equivalent attention without books (such as conversation time) to isolate the specific effects of literary exposure. Additionally, longer intervention periods (6–12 weeks) would allow researchers to examine whether empathy and creativity benefits persist beyond initial novelty effects and whether reflection-based approaches show delayed benefits that emerge with sustained practice. Future investigations should also examine if observed benefits persist when reading books without a social conflict component or if reading physical books together has any advantage over reading books on an electronic device.

Measurement improvements are also needed, including the use of validated empathy scales with broader response ranges, behavioral measures of empathetic responding (not just self-report), and multiple creativity assessments to avoid fluency contamination effects. Future studies should systematically examine individual difference factors such as socioeconomic status, cultural background, neurodevelopmental differences, and developmental readiness that may moderate intervention effectiveness. Investigation of different types of books (varying in social conflict complexity, character diversity, and narrative structure) would help identify optimal literary content for empathy development while retaining a fun, connective experience for families. Finally, longitudinal follow-up assessments would determine whether observed benefits maintain over time and translate into real-world prosocial behaviors and creative problem-solving abilities.

Examining the roles of adult interactive scaffolding, presence of physical touch, or the amount of intentional time together will help develop best practices recommendations for parents and educators, teaching them how to boost skills of empathy and creativity in young children. At a time when empathy and creativity are declining in society, this research suggests that something as simple and accessible as daily bedtime reading may offer families a powerful tool to nurture these essential human capacities in the next generation.