This paper reviews the extensive research examining developmental patterns in normal fear. Areas of focus include age, gender, and socioeconomic status differences in fear content, prevalence, and intensity. The structure and stability/duration of normal fears are also discussed. Finally, the crosscultural research in this area is reviewed. Cross-sectional and longitudinal studies indicate that fear decreases in prevalence and intensity with age. There are also major changes in the content of normal fear over the course of development. Such changes are characterised by a transition from infant fears which are related to immediate, concrete, and prepotent stimuli, and which are largely noncognitive, to fears of late childhood and adolescence which are related to anticipatory, abstract, and more global stimuli and events. Fears of late childhood and adolescence are also more cognitive. One of the prominent themes during these years is social evaluation. It is noteworthy that fears have been demonstrated to be largely transitory in nature, decreasing significantly in number and intensity over time and with maturation. Future research into normal fear should more closely examine the validity of current assessment techniques. There is also a need for research into the developmental correlates of fear, including individual differences and affective environmental experiences. In particular, research examining the correlates of normal fear within a developmental psychopathology framework is recommended.

A large sample of children and adolescents were screened for test anxiety using the Test Attitude Inventory (Spielberger, 1980). Subjects with low and high test anxiety were then compared on self-report measures of trait anxiety, depression, and fear and then asked to report their thoughts and level of distress following an imagined test. Academic grades and performance on standardised achievement and ability tests were also obtained. High test-anxious children and adolescents reported higher levels of trait anxiety, depression, and fear as well as greater distress and cognitive interference during the imagined test. They also obtained lower grades and performed more poorly on the standardised measures. Age effects moderated these findings. Discussion focuses on the clinical and developmental implications of the findings.

In this study, we examined the prevalence and nature of nonclinical panic attacks in 649 Australian youth and explored the relations between such attacks and measures of social support, stress, anxiety, depression, and fear. Full-blown attacks (attacks involving four or more symptoms with rapid onset) were reported by 104 of the youth (16%). Attacks were reported more frequently by girls than boys; however, age was unrelated to panic-attack status. Heightened levels of anxiety and fear, as well as stress in the family and lack of family support, were related to attack status. Path-analytic procedures supported a working model consisting of stress, social support, and emotional distress as related to panic status. Limitations of the self-report data on panic attacks and the other measures used in this study are acknowledged.

This paper describes some of the current problems with the assessment of children's anxiety. The main assessment methods — structured clinical interview and self-report — often have problems with validity and reliability while less frequently used measures such as physiological recording, behavioural observation, and thought-listing also have their attendant difficulties. Possible ways of reducing some of these problems using a developmental perspective are discussed.