The ability to recognise emotion in faces in childhood and adolescence is a highly useful one and one that can gain us important insight to the childhood development. To understand the concept more fully many studies have been carried out, but with so many being flawed and contradictory it is hard to know which ones we can accept. This article attempts to critique one such study – 'Development of Emotional Facial Recognition in Late childhood and Adolescence' – and to gain further insight into it’s findings as a result.

This study investigates the development of emotional facial recognition in late childhood and adolescents. The abstract gives an effective overview of the experiment and along with the introduction, makes the aims and the concept clear. It states that much research in this area has been conducted on infants, who show a rapid development. It argues that due to most of the focus being on younger infants, this study is needed to asses the development rates of older children (aged 7-13 years) and adolescents (aged 14-18 years), and compare them to and adults (aged 25-57 years). The first paragraph of the introduction is effective in providing relevant background information on the topic area, justifying why this research would be beneficial and how findings can be applied practically to help people. It states that correctly interpreting the emotions shown through people’s faces is important because it is an indication of how that person feels internally. This means that people can interact with the environment more successfully as they will know when somebody should be avoided, for example when they are angry, or when to approach somebody; when they are smiling. The concept of testing the recognition of emotions in facial expressions is important, as it is crucial to social interactions. The study links into developmental psychology by investigating this how emotion recognition develops in normal humans. It aims to add to the understanding of this type of development. Once normal development has been studied, the findings can be used to identify affective disorders such as autism from an early age, it can also be used to help formulate a treatment for such disorders.

The reason for studying older children, adolescents and comparing them to adults, as mentioned above, was then justified by stating biological evidence that the amygdala and PFC continue to develop through adolescence and may not reach full maturity until adulthood. It identifies the amygdala as being important in recognising fear and anger, the emotions chosen to be studied in this experiment. The age ranges chosen in this study are appropriate as they show the targeted part of the development stage, through to complete maturation. What is unclear, however, is that one of the aims of the research was to be able to help detect abnormal development at an early age, so studying people over the age of seven will not help with this aim. The studies showing normal development that were mentioned on infants would be more appropriate for this as the studies were conducted on the appropriate age range.

Procedures ensured only healthy, and therefore normal participants were studied. This was achieved through a thirty two question medical and mental health questionnaire, an estimate of intelligence from the Wechsler Intelligence Scale for Children (or the adult scale for adults), and ruling out DSM-IV Axis I mental disorders. This increases construct validity and reduces the likelihood of confound due to participants not meeting the required mental standard. The adults were the parents of the children, which is good as it shows a more accurate portrayal of how the children may develop because they are genetically linked. For a study conducted in a short amount of time this is better than using unrelated adults, but if there are no time constraints a longitudinal study would be better as it would track the development of each participant over the required amount of time. This would allow the whole developmental process to be observed and show if people develop at different rates.

The study lacks external validity as all participants are from the same ethnicity and socioeconomic status. This makes generalising hard if the development of a normal brain recognising emotions in faces is influenced by learning from the environment a person grows up in. If emotion recognition is purely maturation then the results may be able to be applied. There are disproportionate amounts of males and females in each condition, which could confound the results as males and females brains may not develop at the same rate.

An experimental design was used, meaning the variables could be controlled easily. This assumes a positive ontology where results can be measured and applied to other people. There were three identification tasks; each had six photographs of a man’s face changing from one expression to another. The faces morphed from neutral to anger, neutral to fear, and fear to anger. The participants had to say which emotion they thought each picture, in each condition showed. The faces were cropped to omit hair, ears and neckline, this is a good way to increase internal validity and ensure only the expression is affecting the answer. Russel and Fernandez (1997) criticise this method as they believe that posed faces seem forced and unnatural, which may lesson the impact they have. The pictures also lack the gestures, and auditory indicators of emotions such as sighs. To eliminate order effects counterbalancing was applied.

Each of the three conditions were tested for a significant difference in which emotion the children, adolescents and adults reported seeing in each picture, using an ANOVA. As all three conditions showed a significant difference, a Bonferroni-corrected post-hoc test was conducted for each condition to distinguish who was most and least sensitive to changes in facial expression across the pictures. The results were presented fairly and clearly, and differences were shown using line graphs. The neutral-to-fear morphs and fear-to-anger morphs show similar results. Adults are significantly more sensitive to the facial expression change than children, who are the least sensitive. Adolescents are in between and do not differ significantly from either of the other groups. The neutral to anger morphs showed that adults were more sensitive to the change and that children and adolescents show similar results, and are less sensitive to change in facial expressions. Adults and adolescents had the fastest reaction times and did not differ from each other, and children were significantly the slowest group. All the results are strong as they are all within the 5% error allowance, with most significant to p less than .0001. This means that the conclusions drawn are valid. They were also supported by slope analysis.

The conclusions that emotional face recognition continues to develop through late childhood, adolescence and into adulthood is a valid conclusion as the data identifies this trend as significant. The discussion goes on to justify this research to be important as it opposes a previous belief that the development of facial processing is complete by the end of childhood. Fear and anger were chosen to be tested as Thomas et al’s theory suggests the amygdala and PFC continue to develop through adolescence. The results support this due to the adults being most sensitive to change and the children the least. Additionally fusiform gyrus which is sensitive to fear may not be fully developed by the end of childhood, and neural networks may not be mature. Anger did not follow the same pattern as the other conditions, due to adolescents performing to the same standard as children. This is attributed by Thomas et al to cultural learning. Fear is instinctive, but society requires anger to be masked which has to be learnt, implying that adults through experience understand anger better than adolescents and children. This is a weak link to the results as just because children are worse at masking anger it does not follow that they are unable to recognise it, especially as it states that they are more likely to portray anger. Limitations were discussed, such as the difficulty to generalise to other emotions, so perhaps positive emotions should be studied in this age range. It is also suggested that faces of a similar age should be used for each group as this may make them more responsive.

Russel and Fernandez (1997) support Thomson et al’s findings, stating that special neural systems are pre wired to recognise facial expressions. When children are born they do not understand language, so facial expressions convey to them how people are feeling. They state that most systems do not mature until adolescence, explaining why adolescents and adults were better at the task. In young children the facial cues may not be processed properly due to an underdeveloped fovea and lateral geniculate nucleus. They also believe experience is crucial in emotion detection as exposure to faces increases sensitivity to recognise expressions. Tottenham et al (1996) investigated facial recognition, by using either a face or a house to mask another picture. The participants were then asked what the mask was. They found a positive correlation that as the participants age increased, so did their accuracy in identifying the faces. This supports Thomas et al’s theory that when recognising facial expressions, the brain continues to mature through childhood and adolescence into adulthood.

This is a good study that corroborates other theories, and is supported by biological evidence. The methods used are appropriate and much effort is taken to keep the study valid. The study discusses limitations and suggests ideas to further the understanding of late childhood development.

Mack LeMouse

Copyrighted material; do not reprint without permission.

 

View all articles by Mack LeMouse