Reducing measurement error with ecologically valid testing methods

‘Six solutions for more reliable infant research’ outlines important steps in improving research with the youngest of participants. Here, we suggest that a complimentary avenue for further improving reliability can be found in employing more ecologically valid tasks. Because laboratory tasks are often designed as simple, heavily controlled representations of a construct, they can be devoid of the necessary context in which the construct naturally unfolds. We posit that additional error is generated in translating laboratory task metrics back to the construct of interest. Thus, it may be necessary to supplement more traditional laboratory tasks with iterations that more closely match the research construct of interest and how it may unfold in the ‘real world’. We make suggestions for adjusting to the ‘noise’ incurred by moving research out of traditional laboratory environments, while also recognizing how these adjustments may improve the sociodemographic profiles within research studies.