Learning by watching: modelling, coping models and self-efficacy

Background: why watching someone else can help

Much of what children learn, they learn by watching. Albert Bandura's social learning theory, later called social cognitive theory, explains how. Learning from a model depends on four processes: attention (the child has to notice the model), retention (remember what they saw, as images or words), production (turn that memory into action) and motivation (have a reason to try, which can come from seeing what happens to the model) (Bandura, 1986).

Two early findings matter for screens. First, children learn from models on film, not only from people in the room. In one of the "Bobo doll" experiments, nursery-school children imitated aggressive adults shown on film about as readily as live ones; a model dressed as a cartoon cat was imitated less closely, although it still made aggression more likely (Bandura, Ross & Ross, 1963). Second, the same mechanism can work in a child's favour: watching someone approach a feared thing step by step can reduce the child's own fear, which Bandura and colleagues called vicarious extinction (Bandura et al., 1967).

Bandura later linked these ideas through self-efficacy, a person's belief that they can do what a situation demands. He proposed that it shapes whether people try, how hard and for how long, and that it draws on four sources: their own past accomplishments, vicarious experience of seeing others succeed, verbal persuasion, and their physical state, with more dependable sources producing larger changes (Bandura, 1977). So watching someone get through a hard moment might help a child in three ways: it shows what to do, shows that it turns out all right, and suggests that someone like them can manage it.

How the research has been done

The evidence comes from four research traditions, each with its own designs and weaknesses.

• Fear-reduction experiments (1960s and 1970s). Children afraid of dogs or snakes watched a live or filmed model, then were tested on how close they would come to the animal, a "behavioural approach test". A review of the field noted that much of the support came from small laboratory or school studies with children who had not been referred for treatment (Ollendick & King, 1998). A meta-analysis of treatments for children's animal phobias pooled 26 independent study groups from nine reports, with 362 children averaging about 8 years old (Méndez et al., 2005).
• Classroom experiments on self-efficacy. Primary-school children watched peers or a teacher learn a maths skill, then rated their confidence and took skill tests (Schunk & Hanson, 1985); (Schunk et al., 1987).
• Vicarious fear-learning experiments. Since 2007, children aged about 6 to 10 have been shown pictures of unfamiliar animals alongside faces expressing fear or happiness, then asked about the animals and timed as they approach a box said to contain one (Askew & Field, 2007).
• Applied studies. These include films for children facing surgery (Melamed & Siegel, 1975) and peer models for young children afraid of the water (Weiss et al., 1998).

Outcomes include approach behaviour, self-reported fear, heart rate, self-efficacy ratings and skill tests. Follow-up is usually short; a 2022 systematic review called for research on how long effects last (Krause & Askew, 2022).

What the reviews and trials found

Filmed models can reduce children's fears

In the vicarious extinction studies, young children afraid of dogs watched a model gradually approach a dog and became less fearful; when the model was shown on film rather than live, the approach worked as well, and film versions with one model and with several models were compared (Bandura et al., 1967); (Bandura & Menlove, 1968). A 1998 review against criteria for "empirically supported" treatments rated filmed and live modelling as probably efficacious for childhood phobias, and participant modelling, where the child watches and then joins in with guidance, as well established (Ollendick & King, 1998). In the 2005 meta-analysis, psychological treatments for animal phobia had a very large average effect (d+ = 2.80) and modelling was the most effective procedure (d+ = 3.02), while untreated control groups also improved moderately (d+ = 0.55). The authors cautioned that most modelling studies involved children with subclinical fears (Méndez et al., 2005). Effects this large, from small early studies, should be read with caution.

Filmed modelling also reached the hospital: 60 children facing surgery who saw a film of a peer going through hospital and surgery had less anxiety before and after the operation than children shown an unrelated film (Melamed & Siegel, 1975). And a film showing increasingly active play between children, each scene ending well and narrated to highlight what the children did, raised the social interaction of withdrawn nursery-school children to the level of their classmates (O'Connor, 1969). That result also shows why replication matters. A later study that assessed 113 children, selected 32 by the same criteria and added tighter controls found no effect of a modelling film, and raised serious questions about the methods of earlier work (Gottman, 1977).

Fear is learnt by watching, and can be unlearnt

Modern experiments show that the pathway runs both ways. Children aged 7 to 9 who saw unfamiliar animals paired with scared faces developed more fearful beliefs about them, still detectable on an indirect measure three months later, and took longer to approach a box they believed held such an animal (Askew & Field, 2007). Among 97 children aged 8 to 10, a short film of peers interacting positively with an unfamiliar animal reduced fear beliefs and avoidance, even for a second animal that was not shown; a negative film increased fear beliefs; neither changed a behavioural approach task (Broeren et al., 2011). Children aged 6 to 10 learnt and unlearnt fear equally from their mother and from a stranger (Dunne & Askew, 2013), and showed little difference between adult and child models (Dunne & Askew, 2018). For reducing a fear learnt by watching, positive modelling and positive verbal information worked equally well (Reynolds et al., 2018). A systematic review of 18 experiments in 14 articles concluded that positive modelling lowered children's fear in most experiments, whether it was used to prevent, reduce or reverse fear, but that its longevity and transfer to other situations still need research (Krause & Askew, 2022).

A 2023 study shows the other side. In two experiments with 222 children aged 7 to 11, watching a film of another child performing to a negative peer audience, rather than a neutral one, raised self-reported anxiety (measured in the second experiment) when children then performed themselves (Mills et al., 2023). Showing a peer's bad experience is not neutral.

Coping models and mastery models

A mastery model performs confidently and correctly from the start. A coping model begins with errors, worries or negative self-talk, then works through them and gradually succeeds (Schunk & Zimmerman, 1995). Experiments on how a model's characteristics affect fear reduction go back at least to the early 1970s (Meichenbaum, 1971). The clearest evidence with children comes from classroom studies. Children who had struggled with maths watched peers learn fractions; watching a single coping model, several coping models or several mastery models led to higher self-efficacy, skill and training performance than watching a single mastery model (Schunk et al., 1987). Children who watched coping models also judged themselves more similar in competence to the models than those who watched mastery models (Schunk et al., 1987). A follow-up with children in on-grade-level maths classes found the same similarity pattern, and coping models who first voiced doubts about their ability raised self-efficacy more than coping models who did not, although skill did not differ between groups (Schunk & Hanson, 1989b).

The coping advantage is not universal. Among 24 water-fearful children averaging six years old, peer coping and peer mastery models, each combined with swimming lessons, both outperformed lessons alone on skill, self-efficacy and fear; the abstract reports differences between modelling and control groups, not between the two model types, and the sample was too small for many results to reach significance (Weiss et al., 1998).

Does the model need to be like the child?

Perceived similarity is thought to matter most when observers doubt their own ability, and a review of peer modelling examined when similarity leads to change (Schunk, 1987a); (Schunk & Zimmerman, 1995). Early experiments manipulated similarity directly (Kornhaber & Schroeder, 1975). The evidence on which kinds of similarity matter is mixed:

• Peers or adults. When children learned subtraction, peer models raised self-efficacy and skill more than a teacher model, and both beat no model (Schunk & Hanson, 1985); (Schunk & Zimmerman, 1995). For fear, adult and child models had similar effects (Dunne & Askew, 2018). Children under 5 seem to copy familiar behaviour from peers, to be like them, but learn new behaviour mainly from adults, whom they see as more competent (Zmyj & Seehagen, 2013).
• Age, sex and realism. Among 25 boys afraid of snakes, the film model's age and sex made no difference, but a film showing a real snake worked better than one showing a replica (Weissbrod & Bryan, 1973).
• The child as their own model. Children who watched videos of themselves solving maths problems showed higher self-efficacy and motivation than children who were filmed but did not watch (Schunk & Hanson, 1989a). Self model theory treats this as the most basic form of observational learning (Dowrick, 2012). Yet among 67 children aged 5 to 9 learning to swim, watching themselves and watching a peer worked equally well (Minkels et al., 2025).
• Peer models for autistic children. Four autistic children improved sharply on tasks they had found hard after watching typically developing peers do them correctly, and kept the gains when the peers were removed (Egel et al., 1981).

Watching versus doing

Watching is rarely the strongest source of confidence. Bandura argued that more dependable sources produce larger changes (Bandura, 1977), and a review of school research found that mastery experience, succeeding oneself, is typically the most influential source, though this varies with context (Usher & Pajares, 2008). For childhood phobias, participant modelling, which adds guided doing to watching, had stronger support than filmed modelling alone (Ollendick & King, 1998).

For whom and under what conditions

Age. Most experiments involve children of nursery and primary-school age; in the animal-phobia meta-analysis, younger children (and boys) responded better (Méndez et al., 2005).

Severity of fear. Most of the classic evidence comes from children with everyday or subclinical fears rather than diagnosed anxiety disorders (Ollendick & King, 1998); (Méndez et al., 2005).

Children who doubt themselves. Similarity and coping models appear to matter most for children who have struggled or are unsure of their ability (Schunk & Zimmerman, 1995); the classroom coping-model studies deliberately recruited such children (Schunk et al., 1987).

Temperament. After watching a peer's negative performance, children low in effortful control (the ability to regulate attention and behaviour) showed a blunted heart-rate response, which the authors interpreted as possible disengagement (Mills et al., 2023).

How it works may vary. In one swimming trial, watching seemed to work directly, by guiding attention and providing information, rather than by raising motivation (Minkels et al., 2025).

What this means in practice

For parents, teachers and anyone making stories or videos for children's hard moments:

• Show the worry, then the small step. A model who starts unsure and works through it suits children who doubt themselves (Schunk et al., 1987); (Schunk & Hanson, 1989b). Keep the worry brief and ordinary.
• End with success and a good outcome. Positive modelling reduces fear; negative modelling increases it (Krause & Askew, 2022). Avoid showing a peer's bad experience without a resolution (Mills et al., 2023).
• Make the model relatable, without overthinking it. Peers help with confidence, but adults and peers both work for fear (Schunk & Hanson, 1985); (Dunne & Askew, 2018).
• Show more than one example. Several models did as well as a coping model (Schunk et al., 1987).
• Show the real situation, gently. Seeing the actual feared thing helped more than a replica (Weissbrod & Bryan, 1973).
• Remember that adults are models too. Children picked up fear and calm from the faces of strangers as readily as from their mothers (Dunne & Askew, 2013).
• Talk about it, then do it in small steps. Positive information reduced a learnt fear as well as positive modelling did (Reynolds et al., 2018), and guided practice builds the most confidence (Ollendick & King, 1998); (Usher & Pajares, 2008).
• Get help for severe fears. If a fear is intense, lasts for months or stops a child doing ordinary things, talk to a GP or a child psychologist.

What this means for short personalised cartoons

The theory gives reasons to expect that a short film of a child-like character getting through one hard moment could help: filmed models teach (Bandura, Ross & Ross, 1963); filmed fear-reduction worked in early experiments (Bandura & Menlove, 1968); a coping arc suits children who doubt themselves (Schunk et al., 1987); and a positive outcome lowers fear (Krause & Askew, 2022). Cala Luna's free social stories and personalised cartoons are designed with this pattern in mind: one hard moment, a small step and a calm, successful ending.

There are honest gaps. Animated characters are not people: the cartoon-costumed model was imitated less closely (Bandura, Ross & Ross, 1963), and a replica of a feared animal reduced fear less than the real thing (Weissbrod & Bryan, 1973), so stylised depictions might be weaker; animated and human video models give mixed results (see video modelling and cartoons versus real video). An animated likeness is not self-modelling, whose evidence comes from real footage of the child (Schunk & Hanson, 1989a); whether a likeness adds anything is an open question (see personalisation). Most fear-learning evidence comes from laboratory tasks with unfamiliar animals, not school gates or haircuts. And showing worry too vividly could raise anxiety in some children (Mills et al., 2023). Nobody has tested Cala Luna's cartoons: their design is research-based, but their effect is unknown.

Limits of the evidence

• Old, small, laboratory studies. Much of the classic support comes from analogue studies with children who had not been referred for treatment (Ollendick & King, 1998), and the very large pooled effects come from small, early studies (Méndez et al., 2005).
• Replication. At least one prominent film-modelling result failed to replicate under tighter controls (Gottman, 1977). Few classic studies have been repeated with modern methods.
• Beliefs change more readily than behaviour. A peer film shifted fear beliefs and avoidance preferences but not a behavioural approach task (Broeren et al., 2011).
• Short follow-up. How long vicarious fear reduction lasts, and whether it transfers to new situations, is unknown (Krause & Askew, 2022).
• Mechanisms are debated. Self-efficacy is the classic explanation (Bandura, 1977), but in a recent trial watching improved swimming without changing motivational beliefs (Minkels et al., 2025).
• Context matters. The weight of each source of self-efficacy varies by domain, gender and ability (Usher & Pajares, 2008).
• Second-hand classics. Several key 1960s and 1970s papers are known today mostly through later summaries, which can simplify what was found.

Open questions for research

1. Do coping-model stories reduce anticipatory anxiety and avoidance more than mastery stories for real events, such as a first school day or a dental visit, in randomised trials with children aged 4 to 9?
2. How much worry should a model show before showing distress starts to raise anxiety (Mills et al., 2023)?
3. Do animated depictions of a feared situation work as well as real footage, given the replica effect (Weissbrod & Bryan, 1973)?
4. Is a likeness of the child a stronger model than a peer character or an adult, for self-efficacy and for behaviour?
5. Does showing the model being praised or rewarded add anything beyond showing that the situation ends well?
6. Do effects from laboratory fear-unlearning tasks carry over to real behaviour weeks later?
7. Can the classic modelling findings be replicated in preregistered, adequately powered trials?
8. Which children benefit most: does it depend on temperament, anxiety level, autism or age?

Frequently asked questions

References

1. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Prentice-Hall.
2. Bandura, A., Ross, D., & Ross, S. A. (1963). Imitation of film-mediated aggressive models. The Journal of Abnormal and Social Psychology, 66(1), 3–11. doi.org/10.1037/h0048687 · PubMed
3. Bandura, A., Grusec, J. E., & Menlove, F. L. (1967). Vicarious extinction of avoidance behavior. Journal of Personality and Social Psychology, 5(1), 16–23. doi.org/10.1037/h0024182 · PubMed
4. Bandura, A., & Menlove, F. L. (1968). Factors determining vicarious extinction of avoidance behavior through symbolic modeling. Journal of Personality and Social Psychology, 8(2, Pt. 1), 99–108. doi.org/10.1037/h0025260 · PubMed
5. Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191–215. doi.org/10.1037/0033-295X.84.2.191 · PubMed
6. Meichenbaum, D. H. (1971). Examination of model characteristics in reducing avoidance behavior. Journal of Personality and Social Psychology, 17(3), 298–307. doi.org/10.1037/h0030593
7. Kornhaber, R. C., & Schroeder, H. E. (1975). Importance of model similarity on extinction of avoidance behavior in children. Journal of Consulting and Clinical Psychology, 43(5), 601–607. doi.org/10.1037/0022-006X.43.5.601 · PubMed
8. Schunk, D. H. (1987a). Peer models and children's behavioral change. Review of Educational Research, 57(2), 149–174. doi.org/10.3102/00346543057002149
9. Schunk, D. H., Hanson, A. R., & Cox, P. D. (1987b). Peer-model attributes and children's achievement behaviors. Journal of Educational Psychology, 79(1), 54–61. doi.org/10.1037/0022-0663.79.1.54
10. Schunk, D. H., & Hanson, A. R. (1985). Peer models: Influence on children's self-efficacy and achievement. Journal of Educational Psychology, 77(3), 313–322. doi.org/10.1037/0022-0663.77.3.313
11. Schunk, D. H., & Hanson, A. R. (1989a). Self-modeling and children's cognitive skill learning. Journal of Educational Psychology, 81(2), 155–163. doi.org/10.1037/0022-0663.81.2.155
12. Schunk, D. H., & Hanson, A. R. (1989b). Influence of peer-model attributes on children's beliefs and learning. Journal of Educational Psychology, 81(3), 431–434. doi.org/10.1037/0022-0663.81.3.431
13. Schunk, D. H., & Zimmerman, B. J. (1995). Social origins of self-regulatory competence: The role of observational learning through peer modeling [Paper presentation]. Biennial Meeting of the Society for Research in Child Development, Indianapolis, IN, United States. ERIC ED381275. files.eric.ed.gov/fulltext/ED381275.pdf
14. Askew, C., & Field, A. P. (2007). Vicarious learning and the development of fears in childhood. Behaviour Research and Therapy, 45(11), 2616–2627. doi.org/10.1016/j.brat.2007.06.008 · PubMed
15. Broeren, S., Lester, K. J., Muris, P., & Field, A. P. (2011). They are afraid of the animal, so therefore I am too: Influence of peer modeling on fear beliefs and approach–avoidance behaviors towards animals in typically developing children. Behaviour Research and Therapy, 49(1), 50–57. doi.org/10.1016/j.brat.2010.11.001 · PubMed
16. Dunne, G., & Askew, C. (2013). Vicarious learning and unlearning of fear in childhood via mother and stranger models. Emotion, 13(5), 974–980. doi.org/10.1037/a0032994 · PubMed
17. Dunne, G., & Askew, C. (2018). Vicarious learning and reduction of fear in children via adult and child models. Emotion, 18(4), 528–535. doi.org/10.1037/emo0000341 · PubMed
18. Reynolds, G., Wasely, D., Dunne, G., & Askew, C. (2018). A comparison of positive vicarious learning and verbal information for reducing vicariously learned fear. Cognition and Emotion, 32(6), 1166–1177. doi.org/10.1080/02699931.2017.1389695 · PubMed
19. Krause, L., & Askew, C. (2022). Preventing and reducing fear using positive modelling: A systematic review of experimental research with children. Behaviour Research and Therapy, 148, 103992. doi.org/10.1016/j.brat.2021.103992 · PubMed
20. Mills, C., Tenenbaum, H. R., & Askew, C. (2023). Effects of peer vicarious experience and low effortful control on children's anxiety in social performance situations. Developmental Psychology, 59(5), 813–828. doi.org/10.1037/dev0001527 · PubMed
21. Ollendick, T. H., & King, N. J. (1998). Empirically supported treatments for children with phobic and anxiety disorders: Current status. Journal of Clinical Child Psychology, 27(2), 156–167. doi.org/10.1207/s15374424jccp2702_3 · PubMed
22. Méndez, X., Rosa, A. I., & Orgilés, M. (2005). Eficacia diferencial de los tratamientos psicológicos en la fobia a los animales: Un estudio meta-analítico [Differential efficacy of psychological treatments for animal phobia: A meta-analytic study]. Psicothema, 17(2), 219–226. www.psicothema.com/pi?pii=3091
23. Melamed, B. G., & Siegel, L. J. (1975). Reduction of anxiety in children facing hospitalization and surgery by use of filmed modeling. Journal of Consulting and Clinical Psychology, 43(4), 511–521. doi.org/10.1037/h0076896 · PubMed
24. O'Connor, R. D. (1969). Modification of social withdrawal through symbolic modeling. Journal of Applied Behavior Analysis, 2(1), 15–22. doi.org/10.1901/jaba.1969.2-15 · PubMed
25. Gottman, J. (1977). The effects of a modeling film on social isolation in preschool children: A methodological investigation. Journal of Abnormal Child Psychology, 5(1), 69–78. doi.org/10.1007/BF00915762 · PubMed
26. Weissbrod, C. S., & Bryan, J. H. (1973). Filmed treatment as an effective fear-reduction technique. Journal of Abnormal Child Psychology, 1(2), 196–201. doi.org/10.1007/BF00916113
27. Weiss, M. R., McCullagh, P., Smith, A. L., & Berlant, A. R. (1998). Observational learning and the fearful child: Influence of peer models on swimming skill performance and psychological responses. Research Quarterly for Exercise and Sport, 69(4), 380–394. doi.org/10.1080/02701367.1998.10607712 · PubMed
28. Zmyj, N., & Seehagen, S. (2013). The role of a model's age for young children's imitation: A research review. Infant and Child Development, 22(6), 622–641. doi.org/10.1002/icd.1811
29. Usher, E. L., & Pajares, F. (2008). Sources of self-efficacy in school: Critical review of the literature and future directions. Review of Educational Research, 78(4), 751–796. doi.org/10.3102/0034654308321456
30. Dowrick, P. W. (2012). Self model theory: Learning from the future. WIREs Cognitive Science, 3(2), 215–230. doi.org/10.1002/wcs.1156
31. Minkels, C., van der Kamp, J., & Beek, P. J. (2025). Learning the front crawl by observation: Comparing self- and other-models. Human Movement Science, 103, 103398. doi.org/10.1016/j.humov.2025.103398 · PubMed
32. Egel, A. L., Richman, G. S., & Koegel, R. L. (1981). Normal peer models and autistic children's learning. Journal of Applied Behavior Analysis, 14(1), 3–12. doi.org/10.1901/jaba.1981.14-3 · PubMed

Spotted an error, or know a study we should add? Email hello@calalunakids.com. Every study we draw on is in the research library; research terms are explained in our methods glossary.