Clinical and developmental research must place fundamental priority on careful measurement, and it is good to get regular reminders of that in different settings and cultures; several papers issued forth this month serve this purpose. It is fair to say that the priority placed on measurement in a research setting may not always get applied in practical and applied settings. So, for example, questions are rarely asked about the reliability and validity of measures used to assess parent-child relationship quality at clinic intake; however, all research studies using parent-child relationship assessments would be closely scrutinized. There is a great deal of effort in translating treatments from the laboratory setting to the clinic and community setting; there is perhaps too little emphasis on translating evidence-based assessment methods to clinical and applied settings. As a result, we should all be on the look-out, in each issue, for instances of measures that could be equally valuable in large-scale research studies and in applied settings.

One of the inventive papers in this issue searches and explores Google to gather an inexpensive index of neighborhood quality. Including neighborhood data in psychological research has a solid history, particularly in research focused on delinquency. Postal code, zip code and other geographic and economic indicators have been used to index the larger community setting of the individual, although the distant nature of these metrics from the actual neighborhood quieted enthusiasm. Alternative measures in which researchers interview neighborhood residents directly may be more telling, but requires an intensity that is often not practical. Odgers and colleagues1 report a Google-based method for assessing some key features of neighborhoods. They report that a Google database may have enough coverage and detail to index broad measures of neighborhood decay and dangerousness, as implied by a healthy convergence of raters’ views of street images and other metrics of neighborhood disadvantage. The effects may not be especially large, and there remain uncertainties about the role of the neighborhood and behavioral outcomes (e.g. what are the psychological processes linking perception of a depriving neighborhood and engaging in antisocial behavior?). And, individuals may be embedded within different “neighborhood units” or traverse unit boundaries, and the climate of some neighborhoods may be especially likely to decline with daylight. Nonetheless, incorporating readily available data sets from the internet into research may offer researchers a useful tool. Furthermore, the article highlights the potential benefits of exploiting existing databases that may make the expensive process of data collection more efficient and accessible.

The need for high quality measurement is obvious in the case of psychiatric epidemiology. Accurate counts of mental disorder should inform health care system decisions and the educational, juvenile justice, and other systems in society; they are foundational to understanding and improving public health. Vincente et al. offer important new data on the epidemiology of psychiatric disorders in Chilean children based on DSM-IV measures (DISC-IV) and definitions. The authors’ conclusions echo the conclusions of many other authors of different epidemiological reports: the rates of impairing psychiatric disorders in children are high (about one-quarter overall), the rate of service use of affected individuals in the past 12 months is low and variable, and the rates of disorder are disproportionately concentrated in families according to low socio-economic status. A different sort of epidemiological approach is offered by Yoshimasu and colleagues in their Rochester (MN) follow-up study of ADHD cases and matched controls. They find that ADHD is associated with a wide range of disorders by late adolescence. Like Vincente et al., their calibration for disorder is DSM-IV, but the measurement approach was different, depending heavily on medical record data. The similarities (using DSM-IV definitions) and differences (applying different methods of collecting symptom data) in measurement approach between the papers may be of interest but were not themselves the focus of scientific scrutiny. That is not the case in the study by Bevaart and colleagues. In a study based in the Netherlands, Bevaart et al. show that rates of disorder vary by ethnicity, with children from immigrant minority families showing marginally to substantially higher rates of problem behavior than Dutch children according to both parent and teacher report. Perhaps more important was the finding that problem perception varied nearly as much across subpopulations, and did so in different ways from the rates of problems. That is, assessing symptoms and impairment can offer only a modest glimpse into the dynamics leading to referral and engagement in treatment. The report of Bevaart et al. is not the first to draw distinctions among (a) the existence of a clinical problem according to symptom expression, (b) the perception of that behavioral disturbance as a problem by parents, and (c) parental perception that there is a need for care/intervention. But, it is a happy coincidence that the paper appears alongside other large-scale studies of symptoms cited above because it highlights a useful tension in the research literature about what is measured and for which purpose. It has been well-documented that the percentage of children who are referred for child mental health services is much lower than the percentage of children with clinically significant symptoms. That may be the key gap that needs addressing, but it is clear that, no matter what the clinical measures say, perceptions need also to be part of the measurement. O’Driscoll and colleagues offer yet another reminder of the difficulty in assessing perceptions of clinical symptoms, and the possible impact of stigma. Their use of the implicit attitudes test showed that children and adolescents may hold implicit (rather than or in addition to explicit) biases about age-mates with ADHD or depression – although the pattern of findings is a bit complicated and depends on disorder and age and sex of respondent. These data are an important extension to Bevaart et al. in showing that (self-) perception measures need also to be challenged in research.

Perhaps only a minority of the numerous measures that are used in child psychology and psychiatry research are efficient, user-friendly, clinically useful, and interpretable across diverse populations. That suggests that more work is needed on how best to capture and translate the concepts that suit our measurements.