For various reasons ranging from cost to a lack of alternatives, self-report is a common data collection method. However, anyone who has used a self-report data collection method would be well aware of the limitations of this method. Limitations primarily focus on the accuracy of responses, and can include such considerations as an intentional reporting bias (e.g., the participant wants to be seen in a positive light), and an unintentional reporting bias (e.g., the participant simply forgot that they had engaged in the targeted behaviour). I myself have used self-report methodology in a number of different research projects, therefore I am always interested in studies which investigate the validity of self-report measures by comparing the findings with other methodologies.
One recent study compared self-report findings with the findings from home-based observations: Osborne, Shibl, Cameron, Kendrick, Lyons, Spinks, Sipe, and McClure report that in some instances, the self-report responses were 100% in agreement with some of the observations held in the homes of 32 families, while in general the Authors concluded that self-report methodologies can confidently be used in instances where observation may not be feasible. The Authors note that knowing that a home visit would be occurring may have encouraged participants to more accurately report items in the self-administered survey; however, interestingly the Authors also noted that over-reporting of safe practice was demonstrated in approximately half of the items, while under-reporting occurred for one-third of items, suggesting that self-report biases are a complex phenomenon indeed.
Choosing a data collection method can require consideration of multiple factors including strengths and limitations associated with each approach, and validation studies can help us understand the potential magnitude of some of these strengths and limitations.