Age-Related Changes In Emotion Recognition Across Childhood: A Meta-Analytic Review

Riddell, C., Nikolić, M., Dusseldorp, E., & Kret, M. E. (2024). Age-related changes in emotion recognition across childhood: A meta-analytic review. Psychological Bulletin, 150(9), 1094–1117. https://doi.org/10.1037/bul0000442

Rationale

This meta-analysis aimed to comprehensively examine how children’s emotion recognition abilities develop from early childhood to early adolescence, and how various methodological factors influence this developmental trajectory.

While many individual studies have investigated children’s emotional recognition, there has been little systematic quantitative analysis of this large body of research to date.

As the authors note, “despite the wealth of research on this topic, the only meta-analysis on typically developing children’s emotion recognition competencies was performed by McClure (2000)” (p. 1106). However, the previous meta-analysis focused on gender differences rather than broader developmental patterns.

Understanding the typical developmental trajectory of emotion recognition is important because these abilities are associated with various positive psychosocial outcomes for children, including better academic performance (Izard et al., 2001), higher self-confidence (Widen, 2019), and increased social adjustment (Goodfellow & Nowicki, 2009).

Conversely, emotion recognition deficits have been linked to various childhood disorders (Collin et al., 2013). By synthesizing findings across many studies, this meta-analysis aimed to provide a more comprehensive picture of how these crucial skills develop.

Additionally, given the wide variety of methods used to study emotion recognition in children, it was unclear how much task characteristics and other methodological factors may influence results.

As the authors state, “the extent to which task- and stimulus-related features, that, along with age, may influence these developmental trajectories is currently unclear” (p. 1095).

Analyzing how these factors moderate developmental trajectories can help inform best practices for assessing emotion recognition in future research and clinical applications.

Method

The authors conducted a systematic review and meta-analysis adhering to PRISMA guidelines. They searched the databases APA PsycInfo, Web of Science Core Collection, and Linguistics and Language Behavior Abstracts in January 2021.

Search strategy and terms

The authors provide an example of their search strategy for APA PsycInfo, which included terms related to emotion recognition, child age groups, and various stimulus/task types.

They screened abstracts and full texts, with two authors independently reviewing a subset to ensure reliability.

Inclusion and exclusion criteria

Studies were included if they:

• Examined emotion recognition in 2-12 year old children
• Tested recognition of specific emotion categories using external cues
• Included basic emotions (happiness, sadness, anger, fear, disgust, surprise)
• Used only typically developing children
• Examined at least two different emotion categories
• Were written in or translated to English
• Provided necessary statistics for effect size calculation

Studies were excluded if they:

• Used difficulty adjustment procedures
• Only examined valence/arousal judgments rather than specific emotions
• Used only vignettes without external cues

Statistical measures

The authors used a multilevel meta-analytic approach to account for dependencies in the data. They calculated a corrected proportion index as their effect size measure to adjust for chance performance and differences in response options across studies.

They conducted moderation analyses using metaregression to examine how various factors influenced emotion recognition accuracy.

Results

Overall Analysis

• The pooled estimate of emotion recognition accuracy across all studies (excluding free-labeling tasks) was 88%, significantly above chance. Age was a significant predictor of accuracy, with recognition improving across childhood.
• Task type also significantly moderated accuracy, with forced-choice tasks yielding higher accuracy than match-to-sample tasks.
• There were some regional differences, with European samples showing higher accuracy than North American samples. The authors found no significant gender differences in accuracy.
• For free-labeling tasks, which were analyzed separately, the pooled accuracy estimate was lower at 60%. Age was also a significant predictor of accuracy for these tasks.

Emotion-Specific Analysis

• Happiness was the most accurately recognized emotion across ages, followed by anger, surprise, sadness, disgust, and fear.
• The order of recognition accuracy for emotions was consistent across age, with all emotions showing similar developmental trajectories of improvement.
• Some stimulus characteristics moderated recognition of specific emotions. For example, color stimuli were recognized more accurately than monochrome stimuli for fearful expressions only.
• Validated stimuli were also recognized more accurately than unvalidated stimuli for fearful expressions.

Insight

This comprehensive meta-analysis provides important insights into how children’s emotion recognition abilities develop across childhood.

The finding that recognition accuracy improves steadily for all basic emotions, but with a consistent order of accuracy (happiness highest, fear/disgust lowest), extends previous qualitative reviews.

This suggests there may be both universal and emotion-specific factors influencing recognition development.

The study also highlights how methodological choices can significantly impact measured recognition abilities in children. The large differences in accuracy between forced-choice and free-labeling tasks (88% vs. 60%) underscore the importance of task selection when assessing these skills.

This extends previous individual studies showing task-dependent results by quantifying the magnitude of these effects across a large body of research.

The lack of gender differences in accuracy contrasts with some previous findings of small female advantages. This null result from a large pooled sample suggests any gender differences may be smaller or less consistent than previously thought.

The regional differences in accuracy between European and North American samples are intriguing and warrant further investigation.

Future research

Future research could examine potential cultural, educational, or other factors that may contribute to these differences.

An important area for future research is examining children’s specific error patterns or misattributions in emotion recognition, rather than just overall accuracy.

This could provide more nuanced insights into how emotion concepts develop.

Additionally, expanding research to include a wider range of cultures and non-basic emotions would help determine how universal these developmental patterns are.

Strengths

The study had many methodological strengths including:

• Comprehensive literature search and systematic review following PRISMA guidelines
• Large pooled sample size (31,101 children across 129 studies)
• Use of multilevel meta-analytic models to account for dependencies in the data
• Correction of accuracy scores for chance performance and differing numbers of response options
• Separate analyses for overall recognition and specific emotions
• Examination of multiple potential moderating factors (age, task type, stimulus characteristics, etc.)
• Inclusion of both published and unpublished literature to mitigate publication bias

Limitations

The authors acknowledge several limitations of the study:

• Focus only on basic emotions limits generalizability to other emotion categories
• Majority of included studies were from Western, educated, industrialized, rich, and democratic (WEIRD) societies, limiting cross-cultural insights
• Analysis of only correct responses without examining misattributions or error patterns
• Inability to directly compare free-labeling and forced-choice tasks statistically due to methodological differences
• Potential overrepresentation of certain stimulus types (e.g. static photographs) in the literature

These limitations suggest caution in generalizing findings to non-basic emotions or non-Western populations. The focus on overall accuracy without error analysis also provides an incomplete picture of children’s developing emotion concepts.

Implications

This meta-analysis has significant implications for both research and applied fields related to children’s emotional development.

For researchers, it highlights the importance of carefully considering task characteristics when designing studies or interpreting results, as factors like response format can substantially impact measured recognition abilities.

The finding of consistent developmental trajectories across emotions supports theories proposing innate readiness to recognize basic emotions, though cultural learning likely shapes this development.

For clinical practice, the results provide benchmarks for typical emotion recognition development across childhood. This could aid in identifying children with atypical trajectories who may benefit from intervention.

However, the large impact of task type suggests clinicians should use multiple assessment methods for a comprehensive picture of a child’s abilities.

In educational settings, the findings support introducing more complex emotions later in curricula, as children show lower recognition accuracy for emotions like fear and disgust across ages.

The steady improvement in recognition across childhood also suggests potential benefits of emotion education throughout the school years.

The regional differences in accuracy highlight the need for culturally-appropriate norms and potentially tailored emotional education approaches in different geographical areas.

Overall, this meta-analysis underscores the complexity of emotion recognition development and the need for nuanced, multi-method approaches in both research and applied settings dealing with children’s emotional competencies.

References

Primary reference

Riddell, C., Nikolić, M., Dusseldorp, E., & Kret, M. E. (2024). Age-related changes in emotion recognition across childhood: A meta-analytic review. Psychological Bulletin, 150(9), 1094–1117. https://doi.org/10.1037/bul0000442

Other references

Collin, L., Bindra, J., Raju, M., Gillberg, C., & Minnis, H. (2013). Facial emotion recognition in child psychiatry: A systematic review. Research in Developmental Disabilities, 34(5), 1505-1520.

Goodfellow, S., & Nowicki, S., Jr. (2009). Social adjustment, academic adjustment, and the ability to identify emotion in facial expressions of 7-year-old children. The Journal of Genetic Psychology, 170(3), 234-243.

Izard, C. E., Fine, S., Schultz, D., Mostow, A., Ackerman, B., & Youngstrom, E. (2001). Emotion knowledge as a predictor of social behavior and academic competence in children at risk. Psychological Science, 12(1), 18-23.

McClure, E. B. (2000). A meta-analytic review of sex differences in facial expression processing and their development in infants, children, and adolescents. Psychological Bulletin, 126(3), 424-453.

Widen, S. C. (2019). Emotion perception and recognition. In S. Hupp & J. Jewell (Eds.), The encyclopedia of child and adolescent development (pp. 1-11). Wiley.

Keep Learning

Some Socratic questions for a college class to discuss this paper could include:

1. How might cultural differences in emotional display rules or vocabulary impact the developmental trajectory of emotion recognition? How could future research address this?
2. What are the pros and cons of using basic emotion categories versus other frameworks (e.g. dimensional models) to study emotion recognition development?
3. How might the development of emotion recognition abilities interact with other aspects of cognitive and social development in childhood?
4. What ethical considerations should researchers keep in mind when designing emotion recognition studies with children?
5. How might emerging technologies (e.g. virtual reality, machine learning) be used to advance our understanding of emotion recognition development?
6. Given the findings on task effects, what would be an ideal multi-method approach to assessing a child’s emotion recognition abilities in a clinical setting?
7. How might atypical emotion recognition development impact a child’s social and academic experiences? What interventions could potentially help?