Exemplar: Determining the Best Practices for EMA Studies

Overview

Ecological momentary assessment (EMA) is a measurement methodology that utilizes the repeated collection of real-time data on participants' behavior and experience in their natural environment. While EMA allows researchers to gain valuable insights into dynamic behavioral processes, the need for frequent self-report can be burdensome and disruptive to participants. Compliance with EMA protocols is important for accurate, unbiased sampling, yet there is no "gold standard" for EMA study design. In the present study, the investigators aim to use a factorial design to identify the optimal components, or combinations of components, for achieving the highest compliance rates for smartphone-based ecological momentary assessments (EMAs). Participants will be randomized into 1 of 32 conditions in a 2x2x2x2x2 design (i.e., number of questions per EMA survey \[15 vs. 25\] x number of EMAs per day \[2 vs. 4\] x EMA prompting schedule \[random vs. fixed times\]) x payment type \[per EMA vs. based on percentage of EMAs completed\] x order of EMA item question types \[i.e., slider type first 2 weeks vs. Likert-type questions first 2 weeks\]). All participants will complete EMAs for 4 weeks. The effect of each factor on EMA compliance will be examined, as well as the effects of interactions on EMA compliance. The investigators will also examine within and between-subjects factors that influence response rates, and 2) contextual features such as environment and mood that may be associated with better compliance and participant engagement. The results of this study will have broad applications for developing best practices guidelines for future studies utilizing EMA methodologies.

Ecological momentary assessment (EMA) is a measurement methodology that utilizes the repeated collection of real-time data on participants' behavior and experience in their natural environment. EMA has been used in behavioral science for years, including research examining the environmental and psychological antecedents of cigarette smoking, substance use disorders, anxiety, eating, and sleep. EMA sampling strategies are typically time-based or event-based. Time-based sampling is solicited based on a schedule; for example, a daily diary prompted every day at the same time, or at random intervals (e.g., every 3-4 hours) each day. In contrast, event- based sampling is centered around a pre-defined event and is typically initiated by the participant. For example, a participant may be asked to initiate an assessment every time the participant smokes a cigarette. Several studies have examined the impact of EMA study design features on participant compliance. In a recent meta-analysis of compliance with EMA protocols among substance users, there was no evidence that compliance rates were associated with prompt frequency, length of assessment period, or reimbursement. Another study using a pooled dataset of 10 EMA studies found that compliance declined across days, and varied significantly depending on the time of day. In a meta-analysis of EMA studies in mental health research, a higher number of EMAs per day were associated with lower compliance, however the number of days was not. While these studies provide preliminary guidance regarding EMA study design, their results are unfortunately inconsistent, and the varied reporting across studies with respect to the definition of compliance, study design elements, and study populations make it difficult to clearly derive best practice guidelines for EMA design. In the present study, the investigators use a factorial design to identify the optimal components, or combinations of components, for achieving the highest compliance rates for smartphone-based ecological momentary assessments (EMAs). In addition, the investigators will explore the association between EMA design features and participant compliance/response rates. A factorial design is ideal for exploring these research questions because it is statistically more efficient, as this type of design needs fewer participants to answer questions about each experimental factor of interest.6 In this study, participants will be randomized into 1 of 32 conditions (i.e., number of questions per EMA survey x number of EMAs per day x EMA prompting schedule x EMA item type x payment type) to explore: 1. both within and between-subjects factors that influence response rates, and 2. contextual features such as environment and mood that may be associated with better compliance and participant engagement. The results of this study will have broad applications for developing best practice guidelines for future studies utilizing EMA methodologies.