METHODS OF DATA COLLECTION
Measures can also be categorized according to the format used to gather the data. There is a variety of methods available for collecting data in behavioral intervention trials. These methods include self-reports, observer or informant reports, or group assessments and may involve the use of checklists, questionnaires, standardized assessment instruments, rating scales, surveys, interviews, or focus groups. They may also involve direct observation or direct measurement of some phenomena of interest (e.g., heart rate, brain activity, weight). Many studies include a variety of assessment methods. For example, one might obtain ratings of IADL performance for an individual as well as from his or her caregiver or measure stress using the Perceived Stress Scale and levels of catecholamine in the blood. One might also gather cost data or data regarding service utilization such as emergency room visits using an existing database.
Clearly, there are strengths and weaknesses associated with each of these methods. An advantage of self-report methods is that they allow individuals to describe their own feelings, perceptions, and experiences. Use of self-reports via standardized instruments can also be a relatively quick and inexpensive way to gather data; think of the stress measurement example alluded to earlier. However, potential weaknesses include things such as response biases (e.g., social desirability bias, extreme or acquiescent responding) or an individual may not respond truthfully, may not recall what is being assessed (e.g., “Did you have any difficulty doing XYZ in the past month?”), or may not understand the questions. Informant reports are often gathered as a way to verify or supplement self-report data or because an individual is not capable of accurately responding (e.g., someone with a cognitive impairment). Informants can offer a different perspective or enrich the data reported by an individual. Similar to self-reports data, informant reports are subject to biases; it may be difficult to identify informants and inclusion of informant data may add cost to a study. There may also be issues with informed consent and institutional review boards in cases where patients are unable to respond. In clinical assessments, there is often some degree of subjectivity.
Physiological indices or biomarkers are also increasingly being integrated into behavioral intervention research studies to enhance an understanding of the impact of an intervention. For example, it is common in cognitive intervention trials to include brain imaging to help unravel why an intervention results in a change in cognitive performance. Careful consideration needs to be given to the inclusion of physiological measures. The choice of these measures must be well justified and based on an understanding of the potential conceptual/empirical link between the biological/physiological process (es) and the intervention. It is also important to have updated and calibrated equipment, and members of the research team who are skilled in protocols for data collection and analysis. Other considerations in using these measurement approaches are related to data management and storage, cost of including these types of measures, and issues related to participant safety.