Designing a Diary or Experience Sampling Study: Sampling Schedule and Interval

Experience Sampling Methods (ESM) or diary studies involve prompting participants to report on their experiences, thoughts, emotions, or behaviors at specific times. A key decision is the sampling schedule, which defines when and how often participants are prompted (e.g., via app notifications). This balances data richness with participant burden—too frequent, and it causes dropout; too infrequent, and you miss variability.

I'll focus on your query about determining the sampling schedule and interval, drawing from established ESM research (e.g., from scholars like Mihaly Csikszentmihalyi and Daniel Kahneman). If you provide more details (e.g., study goals, duration, participant type), I can refine this.

1. Key Factors to Determine Your Sampling Schedule

Before picking an interval, consider:

• Study Objectives: What are you capturing? Short-term fluctuations (e.g., mood swings) need frequent sampling; long-term trends (e.g., daily habits) need less. For volatile phenomena like stress or pain, aim for granularity; for stable ones like life satisfaction, weekly suffices.


• Participant Burden: Aim for 20-40 prompts per day max to avoid fatigue. Retention drops if prompts exceed 30-60 minutes total daily time.


• Duration: Short studies (1-7 days) can handle denser schedules; longer ones (e.g., 4-12 weeks) need sparser ones to prevent burnout.


• Feasibility: Use technology (e.g., apps like Pacifica for ESM) for random or fixed prompts. Pilot test for compliance (target 70-80%).


• Phenomenon Timescale: Align with natural rhythms—e.g., hourly for work stress, daily for sleep patterns.


• Ethical Considerations: Include opt-out options, debrief on burden, and get IRB approval if academic.

Common pitfalls: Over-sampling leads to "prompt fatigue"; under-sampling misses events. Evidence from meta-analyses (e.g., Conner et al., 2018 in Psychological Assessment) shows optimal schedules yield high validity without attrition.

2. Recommended Sampling Intervals and Schedules

Intervals refer to the time between prompts (e.g., every 2 hours). Schedules can be fixed, random within blocks, or event-based. Here's a framework with examples:

• High-Frequency Sampling (for Momentary or Intraday Data):


• Interval: Every 1-3 hours.


• When to Use: Capturing real-time changes, like emotions during daily activities or pain in chronic illness studies.


• Example Schedule:


• Wake-to-sleep window (e.g., 8 AM-10 PM): 6-10 prompts/day.


• Random within fixed blocks (e.g., every 90 minutes during active hours, avoiding sleep).


• Pros/Cons: Rich data on fluctuations; high burden (dropout ~20-30%). Evidence: In Csikszentmihalyi's flow studies, 8 prompts/day revealed nuances in engagement.


• Tip: Limit to 1-2 weeks max.


• Moderate-Frequency Sampling (Balanced for Most Studies):


• Interval: Every 4-6 hours, or 3-5 times/day.


• When to Use: General well-being, work productivity, or social interactions—common in psychology and health research.


• Example Schedule:


• Fixed times: Morning (upon waking), midday (e.g., 12 PM), evening (e.g., 6 PM), bedtime.


• Or random: 4 prompts/day between 9 AM-9 PM, stratified (e.g., one per quadrant of the day).


• Pros/Cons: Good compliance (80%+ response rates); captures daily variation without overload. From van Berkel et al. (2018) in CHI Proceedings, this yields reliable ecological momentary assessments.


• Tip: Ideal for 1-4 week studies; pair with a daily wrap-up survey.


• Low-Frequency Sampling (for Longitudinal or Retrospective Data):


• Interval: Daily, every other day, or 2-3 times/week.


• When to Use: Broader patterns, like habit formation, journaling for therapy, or end-of-day reflections.


• Example Schedule:


• End-of-day prompts (e.g., 8 PM daily) for a "diary" feel.


• Or event-contingent: Prompt after key events (e.g., meals or stressors), supplemented with 1-2 fixed daily checks.


• Pros/Cons: Low burden, high retention (90%+); less granular (misses intra-day shifts). Studies like those in Journal of Personality and Social Psychology use this for trait-like variables.


• Tip: Best for 4+ week studies; use as a hybrid with occasional intensive bursts.

3. Implementation Tips

• Random vs. Fixed: Random intervals (e.g., via algorithms in tools like Qualtrics or custom apps) reduce anticipation bias but ensure coverage (e.g., no two prompts <30 min apart).


• Duration and Total Prompts: For a 7-day study, aim for 20-50 total prompts/participant. Track via compliance metrics.


• Adaptations: For shift workers, sync to their schedule; for sensitive topics (e.g., mental health), allow snooze/delay.


• Tools: Free/open-source: LimeSurvey for diaries; paid: LifeData or mEMA for ESM apps with geolocation/behavioral triggers.


• Validation: Run a pilot with 10-20 participants to adjust—measure response time, completion rates, and qualitative feedback.

This setup should give you a solid starting point. What's the focus of your study (e.g., topic, population)? Any preferences on frequency? I can help with questionnaire design or analysis next.