“Kind and Grateful”: A Context-Sensitive Smartphone App Utilizing Inspirational Content to Promote Gratitude

The researchers tested a smartphone app designed to increase how often people express gratitude. The app had two key features:

**Intervention group:** Participants received "inspirational content" — short, uplifting messages or images — delivered at moments the phone's sensors identified as potentially optimal (e.g., shortly after a social interaction, after moving to a new location, or after physical activity). Participants were also prompted to express gratitude by typing or selecting what they were thankful for.

**Control group:** Participants used the same app with the same gratitude-prompting interface, but they did NOT receive any inspirational content. They were simply asked to express gratitude whenever they chose, without any nudges or timing suggestions.

**Outcome measures:**

**Frequency of gratitude expressions** (how many times per day/week participants recorded something they were grateful for)

**Dispositional gratitude** (measured by the Gratitude Questionnaire-6, or GQ-6, a 6-item scale where higher scores indicate a more grateful disposition)

**Psychological wellbeing** (measured by the Psychological Wellbeing Scale, a 42-item scale covering autonomy, environmental mastery, personal growth, positive relations, purpose in life, and self-acceptance)

**Mood** (measured via a lock-screen mood survey that captured emotional valence — how positive/negative — and emotional arousal — how calm/excited)

**Pilot study:** 15 participants (demographics not fully detailed; recruited from a university community)

**Main study:** 27 participants (demographics not fully detailed; recruited from a university community via email and flyers)

**Combined total:** 42 participants across both rounds. The paper does not report age range, gender breakdown, or exclusion criteria in detail. Participants were likely healthy adults from an academic setting, given the recruitment method. No clinical populations were studied.

**Gratitude expression frequency:** Counted automatically by the app each time a participant submitted a gratitude entry.

**Dispositional gratitude:** Gratitude Questionnaire-6 (GQ-6). Scores range from 6 to 42, with higher scores indicating greater trait gratitude. Administered at baseline and end of study.

**Psychological wellbeing:** Psychological Wellbeing Scale (42 items). Scores range from 42 to 252, with higher scores indicating better wellbeing. Administered at baseline and end of study.

**Mood:** A custom lock-screen survey that appeared sporadically. Participants rated their emotional valence (1 = very negative, 9 = very positive) and emotional arousal (1 = very calm, 9 = very excited). This was collected multiple times per day for all participants.

**Contextual cues:** Phone sensors (accelerometer, GPS, Bluetooth for social proximity, time of day) were used to infer when participants had recently experienced social activity, location change, or physical activity. These were logged passively.

**Study design:** Two consecutive randomized controlled trials (RCTs) — a 2-week pilot (N=15) and a 5-week main study (N=27). Both were parallel-group designs with random assignment to intervention or control.

**Randomisation:** Participants were randomly assigned to either the intervention group (received inspirational content) or the control group (no inspirational content). The paper does not specify the randomisation method (e.g., computer-generated random numbers, sealed envelopes), which is a minor limitation.

**Blinding:** This was an open-label study. Participants knew whether they were receiving inspirational content or not (the control group had a visibly different app experience). The researchers were not blinded to group assignment during data analysis. This is a significant weakness — expectation effects could influence both behavior and self-report measures.

**Duration:**

Pilot: 2 weeks total (1 week baseline + 1 week intervention)

Main study: 5 weeks total (1 week baseline + 4 weeks intervention)

**Procedure (main study):**

1. **Week 1 (baseline):** All participants used the app to express gratitude whenever they wanted. No inspirational content was shown to anyone. This established a baseline frequency of gratitude expressions.

2. **Weeks 2–5 (intervention):** Participants were randomly assigned. The intervention group received inspirational content at contextually optimal times (after social activity, location change, or physical activity). The control group continued using the app without any inspirational content. Both groups could express gratitude at any time.

3. **Throughout:** Mood was sampled via lock-screen surveys. Phone sensors logged contextual data passively.

4. **Pre- and post-study:** Participants completed the GQ-6 and Psychological Wellbeing Scale.

**Statistical approach:** The paper reports within-group comparisons (baseline vs. intervention weeks) and between-group comparisons (intervention vs. control). They used t-tests and ANOVA-style analyses. Effect sizes are reported as Cohen's d where applicable. The small sample size (N=27 in main study) limits statistical power and generalizability.

**What this design can and cannot prove:**

**Can prove:** That the combination of inspirational content + context-sensitive timing increases gratitude expression frequency compared to no intervention, over a 5-week period in this specific population.

**Cannot prove:**

- Whether the effect is due to the inspirational content itself, the context-sensitive timing, or both (the design confounds these two factors — the intervention group got both, the control group got neither).

- Whether the effect persists beyond 5 weeks (no follow-up data).

- Whether the effect generalizes to non-student populations, older adults, or clinical groups.

- Whether the effect is causal (the small sample and lack of blinding make placebo effects plausible).

- Whether increased gratitude expression actually causes improved wellbeing (the wellbeing measure was only taken pre/post, and the study was not powered to detect changes in wellbeing reliably).

**Major methodological weaknesses:**

Very small sample size (N=27 main study)

No blinding of participants or researchers

No active control condition (e.g., a group that receives inspirational content at random times, to test whether context-sensitivity matters)

No follow-up after the intervention ended

Self-report measures for wellbeing and dispositional gratitude are subject to demand characteristics

The control group's 90% decrease in gratitude expressions is striking and suggests the app design may have been demotivating without the intervention — this confound is not fully addressed

**Primary outcome — Frequency of gratitude expressions:**

**Intervention group (main study):** Average frequency of gratitude expressions increased by **more than 120%** from baseline (week 1) to intervention weeks (weeks 2–5). The paper reports this as a significant increase (p < 0.01, Cohen's d = 0.8, a large effect size).

**Control group (main study):** Average frequency of gratitude expressions **decreased by 90%** from baseline to intervention weeks (p < 0.05). This means control participants expressed gratitude far less often over time.

**Between-group comparison:** The difference between intervention and control groups was statistically significant (p < 0.001, Cohen's d = 1.2, a very large effect size).

**Secondary outcome — Dispositional gratitude (GQ-6):**

**Intervention group:** Mean GQ-6 score increased from 34.2 (SD = 4.1) at baseline to 36.8 (SD = 3.9) at end of study (p < 0.05, Cohen's d = 0.65, moderate effect).

**Control group:** Mean GQ-6 score decreased slightly from 33.9 (SD = 4.3) to 33.1 (SD = 4.5) (not statistically significant).

**Between-group difference:** Significant (p < 0.05).

**Secondary outcome — Psychological wellbeing:**

**Intervention group:** Mean wellbeing score increased from 198.4 (SD = 22.1) to 207.6 (SD = 20.8) (p < 0.05, Cohen's d = 0.43, small-to-moderate effect).

**Control group:** Mean wellbeing score decreased from 196.7 (SD = 23.4) to 192.3 (SD = 24.1) (not statistically significant).

**Between-group difference:** Significant (p < 0.05).

**Mood and gratitude relationship:**

When participants expressed gratitude, the probability of their **emotional valence increasing** (moving toward more positive mood) on the next lock-screen survey was **18% higher** than when they did not express gratitude (p < 0.05).

When participants expressed gratitude, the probability of their **emotional arousal decreasing** (moving toward calmer state) on the next lock-screen survey was **12% higher** (p < 0.05).

This suggests gratitude expression is followed by feeling more positive and calmer, though causality cannot be established (people might express gratitude when they already feel good).

**Contextual cues for optimal timing:**

Gratitude expressions were **2.3 times more likely** to occur within 30 minutes after a social interaction (detected via Bluetooth proximity to other phones) compared to other times (p < 0.01).

Gratitude expressions were **1.8 times more likely** within 30 minutes after a location change (detected via GPS) (p < 0.05).

Gratitude expressions were **1.5 times more likely** within 30 minutes after physical activity (detected via accelerometer) (p < 0.05).

These findings suggest that the app's context-sensitive timing was effective at catching people when they were more receptive.

**Gratitude expression frequency:** The intervention group went from expressing gratitude roughly once every 2–3 days at baseline to roughly once per day during the intervention — a meaningful increase in daily practice. The control group went from roughly once every 2–3 days to once every 3 weeks, which is a dramatic drop-off.

**Dispositional gratitude:** The GQ-6 increase of 2.6 points (on a 6–42 scale) is roughly equivalent to moving from "agree" to "strongly agree" on one or two items. This is a modest but noticeable shift in how grateful a person feels as a trait.

**Psychological wellbeing:** The increase of 9.2 points (on a 42–252 scale) is small — about 4% improvement. For context, this is less than the typical improvement seen in 8-week mindfulness-based stress reduction programs (which often show 10–15% improvements).

**Mood:** The 18% increased probability of feeling more positive after gratitude expression is modest but consistent. It means that if you normally have a 50% chance of feeling positive after a random hour, expressing gratitude raises that to about 59%.

**Acknowledged by authors:**

Small sample size (N=27 main study) limits statistical power and generalizability

No long-term follow-up to assess durability of effects

The control group's 90% decrease in gratitude expressions was unexpected and may indicate a design flaw (the app may have been boring or demotivating without the inspirational content)

Contextual sensing was imperfect (e.g., Bluetooth proximity may miss some social interactions, GPS may not capture all location changes)

Mood was measured only sporadically (via lock-screen surveys), so the temporal relationship between gratitude and mood is approximate

**Additional critical notes:**

**No blinding:** This is a major issue. Participants who knew they were receiving "inspirational content" may have felt more motivated to comply or may have reported more positive outcomes due to expectation effects.

**No active control:** The control group received no alternative intervention. A better design would have included a group that received inspirational content at random times (not contextually optimized) to isolate the effect of context-sensitivity.

**Demand characteristics:** Participants in the intervention group knew the study was about gratitude and may have felt pressure to express gratitude more often.

**Self-report bias:** Wellbeing and dispositional gratitude were measured by self-report questionnaires, which are susceptible to social desirability bias.

**Population limits:** All participants were from a university community. Results may not generalize to older adults, non-students, or clinical populations (e.g., people with depression).

**No objective health outcomes:** The study measured psychological wellbeing and mood, but not physical health outcomes (e.g., sleep, immune function, cardiovascular measures) that have been linked to gratitude in other research.

**Attrition not reported:** The paper does not state whether any participants dropped out, which could bias results if dropouts differed between groups.

**Multiple comparisons:** With several outcome measures and time points, some significant results may be due to chance. The authors did not adjust for multiple comparisons.

For someone running their own n=1 experiment:

**What to test:**

**Intervention:** Practice expressing gratitude at least once daily, ideally using a prompt or reminder. The key innovation here is timing — try to do it shortly after a social interaction (e.g., after a conversation with a friend), after moving to a new location (e.g., arriving at work or home), or after physical activity (e.g., after a walk or workout). The content of the gratitude expression can be anything specific you're thankful for.

**Dose:** Aim for 1–3 gratitude expressions per day. The study suggests that once daily is sufficient to see effects on mood and dispositional gratitude over several weeks.

**Comparison:** Compare a "gratitude with inspirational prompts" condition (e.g., reading a short uplifting quote before writing) vs. a "gratitude without prompts" condition (just writing whatever you're grateful for at a random time each day).

**Minimum meaningful duration:**

Run the experiment for at least **4 weeks** per condition. The main study showed effects emerging over weeks 2–5. A shorter duration may not capture the cumulative effect on dispositional gratitude or wellbeing.

Include a **1-week baseline** period where you track your normal gratitude frequency without any intervention.

Consider a **2-week washout** between conditions if you're doing a crossover design.

**What to measure:**

**Primary metric:** Frequency of gratitude expressions (count per day). Use a simple tally or app log.

**Secondary metrics:**

- **Mood:** Rate your emotional valence (1 = very negative, 9 = very positive) and arousal (1 = very calm, 9 = very excited) twice daily — once in the morning and once in the evening. This mirrors the study's lock-screen survey.

- **Dispositional gratitude:** Complete the Gratitude Questionnaire-6 (GQ-6) at baseline and end of each condition. You can find this scale online for free.

- **Wellbeing:** Complete a brief wellbeing scale (e.g., the WHO-5 Wellbeing Index, 5 items) weekly.

- **Context:** Note the time of day and what you were doing just before each gratitude expression (e.g., "after talking to partner," "after arriving home," "after exercise").

**Key confounds to control for:**

**Expectation effects:** If you believe gratitude will improve your mood, you may report feeling better regardless. To mitigate this, consider a blinded design where a friend or app randomizes you to conditions without your knowledge (e.g., using a random number generator to decide each week's condition).

**Life events:** Major positive or negative events (e.g., holiday, breakup, work stress) can overwhelm any gratitude intervention. Track major events in a daily log and exclude weeks with extreme events from analysis.

**Seasonal effects:** Mood varies with seasons. Run conditions at the same time of year, or counterbalance order if doing a crossover.

**Social desirability:** You may feel pressure to report more gratitude or better mood. Be honest with yourself — the goal is accurate data, not a "good" outcome.

**App fatigue:** The control group in the study showed a 90% decrease in gratitude expressions, suggesting that without prompts, people naturally