Exercise Motivation · Jamie Kim

Background

Why this study

During the pandemic, gym closures and remote work made staying physically active harder for a lot of people. Prior research showed music reliably boosts workout motivation, but almost none of it looked at motivational video or spoken content, or measured motivation itself rather than physical performance. I designed a study to fill that gap.

H1: People who watch or listen to motivational content before a workout will show significantly higher motivation than people who don't.
H2: Video (with added sensory input) will motivate people more than audio alone.

Method

Study design

Design

Within-subjects · 3 sessions

Each of 15 participants completed 3 sessions in a randomized order over one week: a motivational video day, a motivational speech (audio-only) day, and a no-intervention control day, each followed by a self-selected workout.

Measurement

BREQ-2 + IMI scales

A pre-study questionnaire (BREQ-2) assessed each participant's baseline motivation style. Motivation was then measured before and after each intervention and workout using Likert-scale items from the Intrinsic Motivation Inventory.

Analysis

Linear mixed-effects model, R

Used a linear mixed-effects regression (lme4 in R) to control for random effects at the participant level, since repeated measures from the same person can't be treated as independent.

Validity safeguards

Blinding + counterbalancing

Disguised the true focus on "motivation" until the study's end, counterbalanced session order across participants, and varied question order to reduce response bias.

Behind the research

Research Operations

Timeline

15 participants × 3 randomized sessions meant scheduling 45 individual sessions inside a single week, since the within-subjects design required each person to complete all three conditions close together to keep results comparable.

Team coordination

Split responsibilities across the team of 4: two members managed participant scheduling and session counterbalancing, while the others handled data cleaning and the R analysis, so session logistics and statistical work could proceed in parallel instead of bottlenecking on one person.

Results

No significant effect, across all 3 tests

I ran 3 mixed-effects tests: pre/post-intervention motivation, pre/post-workout motivation, and overall motivation between conditions. In every test, the confidence interval for each intervention crossed zero when compared to the control condition, meaning no reliable effect either way. The audio-only condition trended slightly negative in two of the three tests.

Box plot: difference in pre- and post-intervention motivation by condition

Test 1 results — the pattern (no separation between conditions) held across all three tests I ran.

The surprise

It didn't just fail to help, for some it backfired

The more interesting finding wasn't statistical. Several participants described the "motivational" content as a source of pressure or guilt rather than encouragement:

"I felt the videos created a sense of pressure to perform."

"[The speech] made me feel pressured to do more, and I felt guilty for not extending my workout time beyond my initial commitment."

Meanwhile, some of the strongest motivation came from something I hadn't designed for at all:

"Just the fact that the study required it was motivation enough."

Validity & limitations

What I can and can't claim

No causal claim, even with significance. The naturalistic design (participants picked their own workouts, timing, and location) means I couldn't isolate the intervention as the sole cause of any change.
Small, skewed sample. 15 participants, mostly women, despite an intentional effort to recruit outside collegiate/professional athletes.
Self-report only. Motivation is inherently subjective, and the study was unmoderated, so I relied entirely on participants' own reporting.
A flawed proxy. I selected videos by YouTube popularity as a stand-in for "motivational" — in hindsight, a weak signal for actual psychological effect.

Reflection

The interventions I designed weren't motivating, and for some participants, they may have actively backfired. That's a real finding, not a failed study — the honest null result told me more about how people actually experience "motivational" content than a forced positive result would have. If I ran this again, I'd validate stimulus selection against something more rigorous than view counts, and recruit a larger, more balanced sample before drawing conclusions either way.