“I wonder, what if, let’s try”: Sesame Street’s playful learning curriculum impacts children’s problem solving

ABSTRACT

To investigate whether educational television can enhance problem-solving process skills among preschool children, 116 three- to five-year-olds watched 12 episodes of Sesame Street – either engineering-based “playful problem solving” episodes, or episodes about social-emotional learning. Assessments were conducted on three levels: comprehension and delayed recall of educational content, parental observations of learning, and transfer of learning. Results indicated that 48–65% of children recalled problem-solving content after a 1–2 week delay, and 98% of parents reported examples of children’s real-life learning. Repeat viewing of playful problem solving episodes significantly predicted pretest-posttest gains in hands-on transfer tasks, with increases in both the variety of problem-solving heuristics used and the efficiency of children’s process. There were no significant effects on children’s solutions, suggesting children modeled process skills and did not simply learn right answers. Thus, data support Sesame Street’s educational impact on process skills, and demonstrate benefits of playful learning that can be obtained by viewing televised models.

KEYWORDS:

• Sesame street
• playful learning
• critical thinking
• problem solving
• process skills
• educational television
• preschool

Play serves a central role in children’s lives, not only in physical and social development, but in cognitive development and learning as well. Optimal learning through play occurs when activities are joyful, meaningful, actively engaging, iterative, and socially interactive. Through this sort of playful learning, children learn what Hirsh-Pasek and her colleagues have summarized as “the 6 Cs:” Collaboration, Communication skills, subject-matter Content, Critical thinking, Creative innovation, and Confidence (e.g., Golinkoff & Hirsch-Pasek, 2016; Singer et al., 2006; Zosh et al., 2017; see Fletcher et al., this issue).

Drawing upon this approach, Seasons 51 and 52 of Sesame Street adopted a curriculum focus on “playful problem solving,” with many episodes featuring characters using engineering practices of building, fixing, and improving design to solve problems or achieve goals. Throughout, characters use the refrain “I wonder, what if, let’s try!” to summarize their approach to problem solving and to prompt persistence after failed attempts (Truglio & Seibert Nast, this issue). The present research investigates whether exposure to televised characters who model playful problem solving can contribute to preschool children’s understanding of problem solving and to viewers’ own processes when solving a problem – effects that would be consistent with those found for children’s engaging in hands-on playful learning (e.g., Evans et al., 2021).

Decades of research have demonstrated the ability of Sesame Street and its international co-productions to elicit a range of educational outcomes for young children (see, e.g., reviews by Cole and Lee, 2016; Fisch & Truglio, 2001). Further, research has shown that viewing educational television programs can enhance problem solving process skills among older, school-age children (e.g., Fisch et al., 2014; Hall et al., 1990). However, it is less clear whether parallel effects on process skills can be found among preschool children, particularly regarding transfer of learning – that is, applying content or skills that were learned in one context to novel problems that the learner has not seen before. Even in classroom learning, or among adults, research consistently shows that transfer is challenging and often fails to occur (e.g., Barnett & Ceci, 2002; Detterman, 1993; cf. Lobato, 2012), making it all the more challenging for preschoolers. In the present study, these challenges were compounded by COVID-19 concerns, so that observations of children’s problem solving had to be conducted remotely via Zoom, rather than in person (see Fisch et al., this issue).

With these challenges in mind, this experimental/control, pretest/posttest study assesses children’s learning from home viewing of playful problem solving episodes of Sesame Street. It addresses the following four research questions:

RQ1:

To what degree do children exhibit comprehension and long-term recall of playful problem-solving content from the episodes (i.e., one to two weeks after viewing)?

RQ1a:

Do children understand and recall the problems shown in the episodes, and how the characters solved them?

RQ1b:

Do they recall the recurring refrain “I wonder, what if, let’s try!” that characters use to prompt problem solving?

RQ2:

Do children exhibit any indications of learning in their daily lives (as reported by their parents)?

RQ3:

To what degree does watching these episodes lead to posttest gains in children’s own hands-on problem solving (i.e., near transfer of learning from the content seen in Sesame Street to new, hands-on problems)?

RQ3a:

Does viewing playful problem-solving episodes increase the number and/or variety of problem-solving heuristics (e.g., trial and error, test hypothesis) that children use while working on new, engineering-based problems?

RQ3b:

Does viewing playful problem-solving episodes impact the efficiency of their problem-solving processes (i.e., how soon they employ a heuristic that is key to solving the problem)?

RQ3c:

Do viewers of the playful problem-solving episodes show greater gains in the sophistication of their solutions, as compared to children in the control group?

RQ3d:

Are the effects greater for children who choose to watch playful problem-solving episodes repeatedly, as compared to those who watch each episode once?

RQ4:

Do any of the above results differ as a function of children’s age, gender, ethnicity, socioeconomic status, or their parents’ level of education?

Method

Sample

Following a power analysis assuming a moderate effect size (which prescribed a sample size of N = 102), the sample consisted of 116 children (57 girls and 59 boys) and their parents, who helped facilitate children’s viewing and the Zoom-based data collection. There were 23 three-year-olds, 37 four-year-olds, and 56 five-year-olds.

The children lived in 30 different states across the United States, with the largest numbers from Texas (11 children), Florida (10), New York (10), Georgia (8), California (8), and Minnesota (7). Children identified as multiracial (28%), White (28%), Black/African-American (21%), or Hispanic/Latinx (21%). One family identified as Asian, one as Native American, and one as Albanian. All of the children and their families were either native English speakers or bilingual, so that there would be no language barriers to their understanding the English-language Sesame Street episodes. For the few children and parents who were more comfortable speaking Spanish, pretest and posttest sessions were conducted in Spanish by bilingual researchers, to allow them to express themselves more easily.

Socioeconomic status ranged widely, including annual household incomes of less than $25,000 (7%), $25,000–$50,000 (21%), $50,000-$75,000 (25%), $75,000-$100,000 (24%), and $100,000-$120,000 (23%). Most of the parents either graduated college (39%) or attended some college (30%); the rest either had post-graduate degrees (17%), some post-graduate education (3%), high school diplomas (10%), or attended some high school (1%).

Research design

The study employed a pretest-posttest experimental design with random assignment, in which all children watched 12 episodes of Sesame Street over four weeks (three episodes per week). Children were assigned to one of three groups:

• Ramp group: Watched 12 playful problem solving episodes from Seasons 51 and 52, with a focus on engineering. This included the episode “Ramp Racers,” whose educational content was related to one pair of hands-on tasks (Slides/Chutes tasks).

• Tower group: Watched the same 12 episodes as the Ramp group, except that “Ramp Racers” was replaced with “Tallest Block Tower Ever,” which was related to a different pair of hands-on tasks (Tower/Giraffe tasks).

• Control group: Children in the Control group watched 12 episodes from Season 48, with a focus on social-emotional learning and celebrating diversity.

The reason for varying one episode between the Ramp and Tower groups is that each of these episodes related to the content of one of the pairs of pretest-posttest tasks. If children learned specific content from individual episodes and applied it in similar contexts (near transfer), then (for example) the Ramp group would be expected to show growth in tasks related to the “Ramp Racers” episode, but the Tower group would not. The opposite would be true for tasks related to “Tallest Block Tower Ever,” which only the Tower group saw. On the other hand, if children applied strategies and heuristics that they acquired to many different problems (far transfer), then both groups would show growth in all of the tasks.

Children were free to re-watch episodes as many times as they wanted. Parents reported which episodes their children watched in weekly viewing logs. The logs allowed us to track which and how many episodes children viewed, and to ensure the fidelity of the treatment.

Measures

Task-based interviews were conducted remotely with children via Zoom, before and after the viewing period, by researchers who were unaware of the experimental group to which each child had been assigned. Researchers learned each child’s assignment only at the end of the posttest session, so that children in the Ramp and Tower groups could be asked the last few questions in the posttest interview, which assessed comprehension and recall of two of the episodes they had seen. Those questions could not be asked of children in the Control group, who had not seen those episodes.

During each pretest and posttest session, children were asked to complete several hands-on problem-solving tasks that involved aspects of engineering. Because the sessions were remote, we sent each family the materials, and parents facilitated the sessions by managing the Zoom call, helping to set up the materials, assisting if something was physically difficult for their child (e.g., tearing a piece of tape), and helping the researcher manage any behavioral issues (e.g., if a child started to wander away). All tasks were completed on a mat provided by the researchers, to ensure children’s actions would be visible on camera and codable. Parents were instructed not to give children any hints or help in solving the tasks. They were reminded if they forgot and started to give a hint. See Fisch et al. (this issue) for a detailed discussion of issues in remote data collection for hands-on tasks, and techniques employed to yield rich, reliable data.

Hands-on tasks

RQ3 (including its sub-questions) was addressed via three pairs of hands-on tasks, one of the pair administered in the pretest and the other in the posttest, with order counterbalanced: Slides/Chutes (about slope and friction), Tower/Giraffe (building with nonstandard materials), and Roof/Goldfish (permeable vs. impermeable materials). For full descriptions of the tasks, see Fisch et al. (this issue). The tasks in each pair were identical in their engineering content, but were made to seem different by framing their contexts and goals differently. Throughout the tasks, children had access to a “toolkit” that included materials such as tape, string, a ruler, cotton balls, markers, etc. Children were free to use any of the toolkit materials while working on the tasks, but were not required to do so.

Coding and analysis

To code children’s performance in the hands-on tasks, researchers reviewed the video recording of each session, several weeks after data collection. Each child’s performance in a task was coded in two ways, one reflecting the process they used while working, and the other representing the sophistication of the child’s solution. For process scores, a detailed coding scheme was devised to identify each specific strategy or heuristic the child used while working on the task (Table 1).

Table 1. Process score coding scheme.

Each child’s attempt to solve a given task was coded for use of the following problem-solving strategies and heuristics:
Recall information
Gather information: Ask question
Gather information: Explore materials
Manipulate: Use objects
Manipulate: Change objects
Test hypothesis
Trial & error: Conceptual (i.e., try something different while maintaining the same overall approach)
Trial & error: Motor (i.e., persistence in the face of limited physical coordination, e.g., rebuilding a structure after accidentally knocking it down)
Reapproach problem (i.e., change one’s approach to solving the task)
Write/draw
Measure: Ruler
Measure: Count
Measure: Nonstandard (e.g., measure with fingers or other object)
Estimate/approximate
Reasonableness (i.e., explicitly consider why something might or might not be a good idea before trying it out)
Related problems

Summary process scores were computed in two ways: a total score that tallied the total number of behaviors that the child employed in a task (counting repeated use of the same heuristic), and a unique score that tallied the number of different heuristics that the child used in the task. We also considered the order in which children used the heuristics via an efficiency score that reflected how soon the child thought to use a particular heuristic that we had identified as key to solving the task (i.e., how many other things the child tried before attempting the key heuristic for the first time).

Coding schemes for solution scores were individual to each pair of tasks. Each solution score coding scheme identified several levels of sophistication, based on the key engineering concepts underlying the task (e.g., properties of materials or friction) and the sophistication of the child’s solution. For example, the solution score coding scheme for the Slides task differentiated among several levels of sophistication and correctness (Table 2).

Table 2. Solution score coding scheme: Slides task.

Solution	Number of points
Successfully makes designated ramp faster than other by changing slope of one or both ramps (e.g., increasing slope of designated ramp or reducing slope of other ramp)	2
Successfully makes designated ramp faster than other by changing friction/texture of one or both ramps	2
Successfully makes designated ramp faster than other without changing slope or friction of either ramp (e.g., by changing object that slides down ramp or blocking end of ramp)	1
Changes slope or friction of either ramp but fails to make designated ramp faster (e.g., because changes slope of both ramps equally, or because raises ramp but also changes other aspect that slows object down)	1
Successfully changes slope or friction to change speed, but either thinks wrong slide is faster, changes both slides same way (so both still same speed), or makes only one slide for both friends	1
Does not change slope or friction of either ramp (even if child makes other changes, such as coloring ramp) and fails to make designated ramp faster	0
Makes no change, or gives no answer	0

If the child chooses to attempt a second solution, it is coded via a similar scheme that differs primarily in awarding an extra point for a correct solution that is different than the first solution (rather than repeating the same approach).

The process and solution score coding schemes were based on similar approaches that we had used successfully in past studies of the impact of educational television on problem solving among older, school-age children (e.g., Fisch et al., 2014; Hall et al., 1990), adapted to be age-appropriate for preschool children. They were also consistent with approaches used in research outside the context of educational television (e.g., Evans et al., 2021).

Of the 14 researchers who collected data, four researchers also participated in coding data. Each coder coded the data that he or she collected (to make it easier to resolve any issues if a particular behavior or utterance was unclear in the recording), as well as data collected by several other researchers. In this way, coders were completely unaware of children’s experimental group assignments for approximately three-quarters of the data (i.e., the data that the coder had not collected himself/herself), and were largely unaware for the remaining one-quarter. (As noted earlier, researchers learned the assignment of the children they interviewed at the end of the posttest, so that they could ask the appropriate comprehension questions. However, it would have been quite difficult to remember every child’s assignment several weeks later, by the time the data were coded.)

To assess reliability, approximately 10% of the pretest data was coded by all four coders independently. Agreement among the four coders was high (Cronbach’s alpha = .76 for process scores and .84 for solution scores, averaged across the tasks).

Posttest child and parent interviews

To investigate RQs 1 and 2, we interviewed children and parents at the end of the posttest. The child interview centered on issues of comprehension, focusing on two episodes that they had seen, as well as children’s recall of the refrain “I wonder, what if, let’s try!” which was used throughout the episodes to summarize the characters’ approach to problem solving (RQ1). The parent interview asked parents for their observations and impressions of whether (and what) their children learned from Sesame Street, and their own perceptions of the series (RQ2).

Analytic approach

Data were analyzed through a combination of descriptive and inferential statistics. Pretest-posttest data from hands-on problem-solving tasks were analyzed primarily in two ways: via General Linear Modeling (GLM) for repeated measure comparisons between the viewing groups (to assess the effect of assignment to the various experimental groups), and through correlational analyses to assess the effect of repeat viewing of a given episode. Interview data that could not be compared between viewing groups, such as comprehension and recall of content from Sesame Street episodes or the refrain “I wonder, what if, let’s try!” (which could not be asked of children in the Control group, who had not seen those episodes) were analyzed via qualitative analysis, supplemented by quantitative analysis (e.g., GLM or nonparametric statistics) to test for any impact of demographic moderators such as age, gender, ethnicity, SES, or parental education.

Results

Comprehension and long-term recall

Characters’ problems and solutions. Comprehension questions focused on children’s recall of the problem and solution in “Rainy Day Play” and “Fort Rudy,” two episodes about building roofs. In “Rainy Day Play,” Zoe and Rosita have to build a sturdy, waterproof roof to protect their chalk drawing from the rain, eventually covering a table frame with a smock. In “Fort Rudy,” Elmo and Rudy need a sturdy roof for their pillow fort, ultimately anchoring a blanket with books so that the blanket stays up. We chose these two episodes because (a) children in both the Ramp and Tower groups (N = 77) watched them and (b) their engineering content was related to the Roof/Goldfish hands-on tasks in the pretest and posttest.

Regarding RQ1a, even 1–2 weeks after viewing, substantial numbers of children were able to recount the problem and solution presented in these episodes. For “Rainy Day Play,” approximately two-thirds of the children recalled the problem either completely (it was raining, which would wash away the characters’ sidewalk chalk drawing; 18%) or partially (recalling the rain without mentioning keeping the drawing dry; 47%). Almost one-half of the children also recalled the characters’ solution of covering the drawing with a “roof” made from a table frame and a smock; 10% recalled the solution completely, and 38% showed partial recall (covering the drawing with something other than the smock).

Similarly, for “Fort Rudy,” approximately one-half of the children recalled the problem either completely (needing to make a sturdy roof for a pillow fort; 15%) or partially (e.g., needing to make a roof without mentioning sturdiness; 33%). Almost one-half of the children also recalled the characters’ solution either completely or partially; 13% recalled that they anchored a blanket with books since the blanket was not sturdy enough to stay up on its own, while 28% recalled that they used a blanket or made a roof without mentioning the books.

“I wonder … .” Throughout the 12 playful problem solving episodes from Seasons 51–52, characters repeatedly use the phrase “I wonder, what if, let’s try!” to summarize their approach to problem solving (see Truglio & Seibert Nast, this issue), and RQ1b concerned children’s recall of that key phrase. Twenty-two percent of the parents in the Tower and Ramp groups spontaneously reported that their child had begun saying the phrase at home. When we asked children in the Ramp and Tower groups directly to recall the phrase, 65% correctly recalled the entire phrase, and another 4% recalled part of it. Prompting the children by starting the phrase for them raised the numbers to 87% recalling part or all of the phrase.

Real-life learning

RQ2 asked whether children would exhibit signs of learning in their daily lives. In the parent interview, virtually all parents (98%) reported that their children learned from Sesame Street. In the Ramp and Tower groups (N = 77), 38% of the parents said their children learned about problem solving. For example, one parent of a five-year old girl noted:

It allows her to understand that there’s not just one way to do something, there’s a bunch of different ways to learn something or to figure out or solve a problem. It makes her more comfortable with getting stuff wrong; it’s the path it takes to get to the problem. It might not be immediate, but that’s okay.

Twenty-one percent talked about their children building things; for example, a parent of another five-year-old girl said:

After the episode of building a tower…both my kids went to the bathroom and I had just bought a bunch of toilet paper. They started stacking, there were probably like 17 rolls of toilet paper. And they made one big tall tower, which I thought was very funny and I knew it was from watching the show.

Seventeen percent said their children started using the phrase “I wonder, what if, let’s try;” for example, one parent noted, “Instead of getting frustrated, he’ll go, ‘I wonder, what if, let’s try.’ He used to get frustrated, but now he’ll actually stop himself and try to solve the problem himself.” Ten percent mentioned measurement, with comments such as “He measured his sister and brother using napkins and blocks. He laid them all out on the floor and measured them.” Finally, 6% said their children demonstrated increased persistence. Parents from the Control group spoke about aspects of either social-emotional learning (31%) or, more specifically, diversity (13%).

Transfer of learning

Our first pass on data analysis consisted of a set of GLM analyses to compare pretest-posttest change on each task, between the Ramp, Tower, and Control groups. Separate analyses were conducted for each type of score: total score, unique score, efficiency score, and solution score. Although all of the scores produced similar trends – with children in the Ramp and Tower groups showing greater pretest-posttest improvement than the Control group – few of these trends were statistically significant (RQ3c), so these results will not be presented here.

However, since the very first summative studies of children’s learning from Sesame Street, it has been shown repeatedly, not only that children who view Sesame Street perform significantly better than nonviewers in areas such as literacy or mathematics, but also that the more children watch, the greater their gains (e.g., Anderson et al., 2001; Ball & Bogatz, 1970; Bogatz & Ball, 1971; Borzekowski et al., 2019; Wright et al., 2001). Thus, we conducted a correlational analysis that compared pretest-posttest change in each task to the number of times children watched the episode that was most relevant to that task: “Ramp Racers” with the Slides/Chutes tasks, “Tallest Block Tower Ever” with the Tower/Giraffe tasks, and two episodes about building roofs (“Rainy Day Play” and “Fort Rudy”) with the Roof/Goldfish tasks.

Problem-solving process – variety of heuristics used

To address RQs 3a and 3d, Figure 1 presents correlations between viewing and pretest-posttest change for children’s unique scores, which reflected the variety of different problem-solving heuristics that each child used. There was a significant, moderate to large correlation¹ between the number of times that children watched the “Ramp Racers” episode of Sesame Street and their pretest-posttest gains in the Slides/Chutes tasks (which also concerned ramps; r₁₁₃ = .29, p = .002). Similarly, though only marginally significant, pretest-posttest gain in the Tower/Giraffe tasks was positively correlated with the number of times children watched “Tallest Block Tower Ever” (r₁₁₄ = .17, p = .071). There were significant, small to moderate correlations between the amount of viewing each episode about building roofs and children’s gains in the Tower/Giraffe tasks (r₁₁₁ = .19, p = .049 for “Rainy Day Play” and r₁₀₆ = .22, p = .025 for “Fort Rudy”). In each case, the more children watched the relevant episode of Sesame Street, the greater the increase in the variety of problem-solving heuristics that they applied when working on tasks involving similar principles. (In contrast, children’s total scores, reflecting their total number of actions, did not produce significant effects and are not reported here.)

Figure 1. Correlations: Number of times viewed x Change in variety of heuristics used.

Problem solving process – efficiency

RQ3b considered the sequence in which children used these heuristics – that is, how soon they came to realize that they should use a particular heuristic that was key to solving the task. Although each task could be solved in multiple ways, the heuristic Manipulate: Change Materials was key to solving all three tasks: Changing the ramps in the Slides/Chutes tasks (e.g., changing slope or friction), building a tower in the Tower/Giraffe tasks, and putting a roof on the house in the Roof/Goldfish tasks were all coded as Manipulate: Change Materials.

We created an efficiency score that reflected how soon children used Manipulate: Change Materials in a task. The efficiency score was computed by dividing the ordinal position of the first instance of Manipulate: Change Materials by the total number of behaviors that the child did in the task. For example, if a child performed seven behaviors during a task, and Manipulate: Change Materials was the fourth one used, the child’s efficiency score would be 4/7 = .57. Note that, because of the way these scores were computed, a lower score reflects greater efficiency (i.e., using Manipulate: Change Materials sooner). Thus, negative correlations with viewing mean that more viewing is associated with becoming more efficient.

Correlations between viewing and pretest-posttest change in children’s efficiency scores (Figure 2) revealed similar patterns to those found for children’s variety of heuristics, producing positive data regarding RQs 3b and 3d. Significant, small to moderate correlations were found between the number of times children viewed “Ramp Racers” and pre-post change in their efficiency in the Slides/Chutes tasks (r₁₁₃ = −.20, p = .034), viewing “Tallest Block Tower Ever” and pre-post change in the Tower/Giraffe tasks (r₁₁₃ = −.24, p = .009), and viewing each roof episode and pre-post change in efficiency in the Tower/Giraffe tasks (r₁₁₁ = −.24, p = .010 for “Rainy Day Play” and r₁₀₆ = −.23, p = .015 for “Fort Rudy”).

Figure 2. Correlations: Number of times viewed x Change in efficiency of problem solving.

Sophistication of solutions

The strongest correlations regarding change in the sophistication of children’s solutions were the same ones as in the two process score analyses, but the correlations with solution score were not statistically significant (RQs 3c and 3d). Only one correlation approached significance (Table 3).

Table 3. Correlations: Number of times viewed x Change in sophistication of solution.

	Change scores: Solution score
Number of times watched:	Slides/Chutes tasks	Tower/Giraffe tasks	Roof/Goldfish tasks
Ramp episode (Ramp Racers)	r = .17, p = .067	r = −.02, p = .838	r = .00, p = .996
Tower episode (Tallest Block Tower Ever)	r = −.01, p = .913	r = .14, p = .152	r = .01, p = .886
Roof episode (Rainy Day Play)	r = .03, p = .724	r = .08, p = .396	r = −.02, p = .843
Roof episode (Fort Rudy)	r = .00, p = .982	r = .15, p = .116	r = .07, p = .480

However, in almost all of the pretest and posttest tasks, the sophistication of children’s solutions was significantly correlated with their efficiency (Table 4). It also tended to correlate with the variety of heuristics they used, with most of these correlations either significant or approaching significance (Table 5).

Table 4. Correlations: Efficiency x Sophistication of solution.

	Solution score
Efficiency score	Pretest	Posttest
Slides/Chutes tasks	r = −.19, p = .044	r = −.31, p = .001
Tower/Giraffe tasks	r = −.31, p = .001	r = −.25, p = .007
Roof/Goldfish tasks	r = −.20, p = .032	r = −.10, p = .276

Table 5. Correlations: Variety of heuristics used x Sophistication of solution.

	Solution score
Unique score	Pretest	Posttest
Slides/Chutes tasks	r = .24, p = .010	r = .32, p = .000
Tower/Giraffe tasks	r = .27, p = .003	r = .17, p = .069
Roof/Goldfish tasks	r = .16, p = .090	r = .07, p = .483

How many viewings does it take?

Further pursuant to RQ3d, as one might expect, Table 6 shows that the number of viewings required to produce a significant effect varied across different episodes, tasks, and measures (with variation likely attributable to effectiveness of individual episodes and cognitive demands of the tasks). One viewing of “Ramp Racers” was sufficient to produce a significant effect on the variety of heuristics children used in the Slides/Chutes tasks (F_1,114 = 5.72, p = .018, Cohen’s d = .46), but three viewings were needed for a significant effect on the children’s efficiency scores (F_1,113 = 7.37, p = .008, Cohen’s d = .97). For “Tallest Block Tower Ever,” four viewings produced effects on both variety of heuristics (F_1,113 = 6.43, p = .013, Cohen’s d = 1.77) and efficiency (F_1,113 = 5.06, p = .026, Cohen’s d = 1.58). Two viewings of “Fort Rudy” produced effects on both variety of heuristics (F_1,106 = 4.13, p = .045, Cohen’s d = .64) and efficiency (F_1,106 = 4.15, p = .044, Cohen’s d = .64) in the Tower/Giraffe tasks. For “Rainy Day Play,” three viewings impacted on the variety of heuristics children used in the Tower/Giraffe tasks (F_1,111 = 4.53, p = .036, Cohen’s d = .19) and five viewings impacted on efficiency (F_1,111 = 4.12, p = .045, Cohen’s d = 1.17). Almost all of these values of Cohen’s d reflect large effects.

Table 6. Minimum number of viewings (of the relevant episode) necessary to produce significant effects.

	Smallest number of viewings for significant effect
Episode – Task	Variety of heuristics	Efficiency
“Ramp Racers” – Slides/Chutes tasks	1 viewing	3 viewings
“Tallest Block Tower Ever” – Tower/Giraffe tasks	4 viewings	4 viewings
“Fort Rudy” – Tower/Giraffe tasks	2 viewings	2 viewings
“Rainy Day Play” – Tower/Giraffe tasks	3 viewings	5 viewings

Finally, to address RQ4, we looked for potential moderating effects of child and family demographics. Findings regarding the impact of Sesame Street were consistent across all demographic groups, with no significant effects of child age, gender, ethnicity, socioeconomic status, or parent education level.

Discussion

Taken together, these data demonstrate the educational impact of Sesame Street’s playful problem solving curriculum on several levels. Most children demonstrated comprehension and recall of material from the episodes, even one to two weeks after viewing (RQ1). This was further supported by nearly all of the treatment parents’ reporting that their children had learned about subjects such as problem solving, building, and measurement (RQ2).

Yet, the most striking indicators of impact were found in children’s performance in the hands-on tasks that required children’s learning to “jump the screen” and be applied to problem solving skills in related tasks (RQ3). Although the research literature attests to the challenges of eliciting transfer of learning (e.g., Barnett & Ceci, 2002; Detterman, 1993), and despite the young age of the children in this study, children who watched relevant episodes more often showed greater gains in the variety of problem-solving heuristics they used to solve new problems, and in the efficiency of their process of problem-solving.

Impact on children’s use of problem-solving heuristics can occur through several potential mechanisms: (1) by increasing motivation (leading children to try more things in general), (2) by adding new problem-solving heuristics to their mental repertoires, and/or (3) by coming to recognize that heuristics in their repertoires are applicable to new problems. Motivation may have been a factor in this study, but cannot serve as a complete explanation because we did not find significant effects for children’s total process scores (i.e., the total number of behaviors they employed, including repeated use of the same heuristic).

It seems more likely that children learned new heuristics from the Sesame Street episodes, and/or recognized that these heuristics were applicable to the new hands-on tasks. The latter explanation is further supported by the finding that, in several cases, effects were content-specific: Repeat viewing of an episode about ramps produced effects in tasks that concerned ramps, and viewing an episode about building a tower predicted gains in tasks that entailed building a tower. Seeing ramps or a tower in the hands-on tasks may have led children to remember approaches they saw in the relevant episode of Sesame Street and realize that these approaches could be applicable in the hands-on tasks too (what the research literature refers to as near transfer – i.e., transfer of learning to problems that share similarities with the original context in which the content was learned). Note that watching an episode about ramps did not lead children to do better on tasks that related to dissimilar topics such as building towers, nor did repeat viewing of an episode about towers lead to effects on tasks involving ramps (either of which would be far transfer).

It is not clear why viewing the two episodes about roofs (“Rainy Day Play” and “Fort Rudy”) were not significantly related to change in children’s performance in the Roof/Goldfish tasks. However, two potential explanations unpack why these episodes might predict change in the Tower/Giraffe tasks: (1) Since the episodes and task both concerned building, albeit in different contexts (building a roof vs. a tower), the shared content may have led children to apply what they learned – another instance of near transfer. (2) An alternate (or, perhaps, complementary) explanation might stem from the fact that viewing of the “Tallest Block Tower Ever” episode was highly correlated with viewing “Rainy Day Play” (r₁₁₀ = .69, p = .000) and “Fort Rudy” (r₁₀₅ = .70, p = .000); more frequent viewers of “Tallest Block Tower Ever” also tended to watch “Rainy Day Play” and/or “Fort Rudy” more often. Thus, the correlations between the two roof episodes and the Tower/Giraffe tasks may simply be attributable to the children’s also watching the tower episode more often.²

Given these significant effects on children’s process of problem solving, why did we not also find significant effects for children’s solutions (although significant relationships between process and solution scores suggest that additional viewing might produce significant correlations)? One explanation might be that children are simply more exploratory in cognitive tasks than adults, even though adults may achieve greater payoffs (Liquin & Gopnik, 2022), reflected in expanded process but not necessarily a more sophisticated solution. Another explanation might be found in the long research tradition, stretching back to Inhelder and Piaget (1964), that young children do not approach problem solving as systematically as older children or adults (although we recognize that some later researchers argue that Piaget overstated this lack of systematicity; e.g., Klahr, 2021). If preschoolers approach problem solving less systematically than older children, then using a wider variety of heuristics – or even employing key heuristics sooner – would not guarantee reaching a correct solution.

Regardless, the fact that significant effects were found for scores reflecting children’s process but not their solutions suggests that the children weren’t simply memorizing correct answers from the episodes. Instead, they were modeling the processes of problem solving that they had viewed. Indeed, a few parents noticed these changes themselves. For example, one parent of a five-year-old observed:

Even when we did the similar exercises the first time [in the pretest], he did not do very well with any of the problem solving… And just watching him just now, be able to figure things out a lot quicker, he was like, “Oh, I got this. I know how to do this. I know the reason.” I felt like it was okay, maybe he did figure it out.

Another parent noted that her three-year-old daughter “was definitely more patient with her problem solving this time around. She got frustrated last time.”

Thus, viewing playful problem solving episodes of Sesame Street was associated with significant gains in children’s own processes of problem solving. There is widespread agreement among educators – including agencies such as the U.S. Department of Education and the National Research Council – that children’s development of these sorts of “21st Century skills” for critical thinking and problem solving are crucial to their future success (e.g., Alismail & McGuire, 2015; Koenig, 2011; Silver et al., 2022), attesting to the importance of finding impact on such skills, even if impact on correct answers to the immediate hands-on tasks were not also strong enough to reach significance. Often, efforts to teach process skills for critical thinking or problem solving focus on formal education in elementary or high school classrooms. The present data indicate that emergent skills can also be fostered among preschool children, through principles of playful learning and through informal education, such as educational television.

Considering the challenges of eliciting transfer in formal and informal education, this is big news for those wondering if educational television can reach beyond the learning of content to near transfer of cognitive process among young children. In this way, the present data are consistent with a substantial body of prior research demonstrating the educational benefits of Sesame Street for preschool children, extending its documented effects to process skills for STEM-based problem solving. More broadly, the results also demonstrate that benefits of playful learning can be obtained, not only through children’s own hands-on play, but also through viewing characters’ modeling playful learning on television.

Acknowledgments

We are grateful to: the Sesame Workshop staff who provided materials and input at various stages of this project, to Hammad Sheikh for statistical advice, and to Naomi Lowenthal for logistical support. Most of all, we thank the participating families, without whom the research would have been impossible.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Shalom M. Fisch

Shalom Fisch is President of MediaKidz Research & Consulting. For over 35 years, he has applied educational practice and empirical research to help create engaging, impactful educational media for children. Prior to founding MediaKidz in 2001, he served as Vice President of Program Research at Sesame Workshop.

Katelyn Fletcher

Katelyn Fletcher is a former Postdoctoral Research Fellow at Temple University’s Infant and Child Lab and consultant for the Brookings Institution Center for Universal Education. Her research interests focus on playful learning and the design, implementation, and evaluation of educational programs. Katelyn currently works in the field of education philanthropy.

Gavkhar Abdurokhmonova

Gavkhar Abdurokhmonova is a doctoral student in the Human Development and Quantitative Methodology program (concentration in Developmental Science, concentration in Neuroscience and Cognitive Science) at the University of Maryland, working with Dr. Rachel Romeo. Her research moves beyond documenting deleterious outcomes of socioeconomic disparities, and instead explores the neural mechanisms by which environments shape development, with the goal of highlighting the importance of individual differences in children’s language experiences.

Lacy Davis

Lacy Davis, Nachum Fisch, Susan Fisch, Ivelisse Seguí, Jennifer Shulman, and Carolyn Volpe are Researchers at MediaKidz Research & Consulting.

Melissa Jurist

Melissa Jurist has a decades-long career in diversity, equity, inclusion and access, currently working at Duke University. Her research interests include identifying best practice for all children in all of the STEM fields.

Randi Kestin

Randi Kestin is a freelance research and evaluation specialist whose research interests include program evaluation for children’s educational media, education reform, Israel education, public health, and market research. She finds this work most meaningful when the results help improve the lives of others.

Annelise Pesch

Annelise Pesch is a postdoctoral research fellow working at Temple University. Her research investigates social cognitive development in the preschool years including social learning, trust, play, and the impact of technology on learning and development. She leverages her research to inform the design of high-quality informal learning spaces through the Playful Learning Landscapes initiative.

Jennifer Shulman

Jennifer Tomforde has worked in the field of educational research and consulting for the past 13 years and currently coordinates a literacy outreach program.

Nava Silton

Nava R. Silton is a Full Professor at Marymount Manhattan College and Director of the Center for Health, Human Development and Creativity at the College. She also works on the mental health teams of two New York City elementary schools, and has served as a disability and developmental consultant for Disney, Nickelodeon, Apple, Netflix, and Sesame Workshop. She has created the Realabilities comic book series, The Addy & Uno off-Broadway musical, and developmentally rich tabletop games for children.

Charlotte Anne Wright

Charlotte Anne Wright is a Learning Designer and Research Specialist at Begin Learning. She led the LEGO Foundation’s Playful Learning and Joyful Parenting project as a research fellow at Temple Infant and Child Lab.

Kathy Hirsh-Pasek

Kathy Hirsh-Pasek, a Professor of Psychology at Temple University and a senior fellow at the Brookings Institute, was declared a “scientific entrepreneur” from the American Association of Psychology. Writing 17 books and 250+ publications, she is a leading professor in the science of learning who is known for translating research into actionable impact in schools (Activeplayfullearning.com), digital and screen media, and community spaces (Playfullearninglandscapes.com).

Notes

1. Traditionally, correlations of .10 are considered small effects, .30 are moderate, and .50 are large (Cohen, 1988). More recently, however, a meta-analysis by Gignac and Szodorai (2016) suggested the standards should be small = .10, “typical” (i.e., moderate) = .20, and large = .30. We have merged the two sets of standards here. For example, we refer to a correlation of .29 as a “moderate to large” effect.

2. Viewing the “Ramp Racers” episode also correlated significantly with watching “Rainy Day Play” (r₁₁₀ = .25, p = .007) and “Fort Rudy” (r₁₁₀ = .23, p = .015). But Fisher r to z tests confirmed that viewing the roof episodes was significantly more strongly correlated with viewing “Tallest Block Tower” than “Ramp Racers” (Z = 4.32, p = .000 for “Rainy Day Play” and Z = 4.58, p = 000 for “Fort Rudy”). Thus, viewing the tower episode was more likely to go hand in hand with the roof episodes than the ramp episode was.