[Proposal] A Meditation Study: Investigating the Mechanisms and Covariates of Mindfulness Meditation for Improved Well-Being

*This study has undergone piloting and is currently being run with undergraduate university students in an online format. Replicating this study on Prolific with a representative sample would have valuable implications for our understanding of the mechanisms in which meditation improves well-being, as well as the efficacy and equitable applications of mindfulness interventions.
Below is an abbreviated proposal, but please see the full pre-registered protocol for additional details.

Importance and Motivation

The field of mindfulness and meditation research has grown exponentially in the past two decades. One of the most supported findings is that meditation reduces stress and anxiety; however, little work has focused on the underlying mechanisms, and there are major sources of bias in the field that have yet to be addressed.
Components of Meditation
There are likely multiple components of [mindful] meditation that may causally contribute to reductions in stress and anxiety and improved well-being, such as slow breathing, meta-awareness, and a detachment from thoughts. Therefore, comparing the effects of each component is most feasible in the context of a single session of meditation. This will not only elucidate the mechanisms by which meditation improves well-being, but also quantify observable improvements in a short-term context.
Reducing Bias
Results of meditation interventions may be inflated due to a number of common practices in the field that introduce bias (see full protocol for more details). Therefore, the current study aims to address these issues by measuring and manipulating expectations of benefits (i.e., testing the placebo effect), implementing active control groups, using blind recruitment and random assignment of a representative sample, avoiding mention of the words “meditation” or “mindfulness” (unless specifically manipulated), and addressing exploratory questions to better understand the average effectiveness and generalizability of mindfulness meditation for different people.

The goals of the current study are to test the immediate effects of meditation on state well-being, to discern whether these effects result from the physical practice of slow breathing or the mental practice of mindfulness, and to assess how any observed benefits are influenced by a person’s expectations of improvement, changes in the valence of their thoughts, or personal characteristics.


This is a 3-part study, where each part follows a basic pre/post intervention design, collects the same measures, and has planned Primary, Secondary, and Tertiary analyses (which correspond to Primary, Secondary, and Tertiary variables). Each part differs only by the interventions and information given to participants, but all interventions are presented as a “20-minute relaxation exercise.”

Part 1: Effects of a Single Session of Meditation (Meditation vs. Control)
Part 2: Mechanisms of Meditation (Meditation vs. Control vs. Breath vs. Mind)
Part 3: Meditation Expectations (Label Bias & The Placebo Effect)

Design Overview

Part 1 will compare a mindful Meditation with an informational Control condition, and assess the role of internal experiences (expectations and thoughts) on any observed benefits. Then, Part 2 will compare Meditation and Control conditions with two additional interventions that contrast specific components of meditation: Breath (a slow breathing exercise with no mindfulness or thought-related instructions) and Mind (a mind-wandering exercise with no breath or mindfulness instructions). Again, the role of internal experiences will be determined and compared to results of Part 1. Finally, Part 3 will assess the causal role of expectations by the inclusion or exclusion of a placebo statement or the word “meditation” in the exercise description, all of which will be mindful meditation. Conclusively, this study aims to (Part 1) quantify observable changes in well-being due to a single session of meditation, while considering the role of internal experiences on the strength of the effect; (Part 2) clarify which components of meditation are necessary or sufficient to improved well-being within a single session, again looking at the role of internal experience; and (Part 3) understand the role of bias and expectation in meditation, and how much improvement in well-being can be contributed to a placebo effect.


Each intervention is a 20-minute, guided relaxation exercise. Participants in each intervention are told that they will participate in a “relaxation exercise”, avoiding indications of the words, “mindfulness” and “meditation” (except when this is manipulated in Part 3). Each intervention follows the same structure, with nearly perfectly matched timing, word count, and duration of instructions (with the exception of the control condition, which only matches the duration of instruction). All interventions are pre-recorded and voiced by a professional meditation and mindfulness workshop instructor, Nadia Horvath. Full scripts, with timing indicators and preliminary descriptions, are available in the Appendix of the full pre-registered protocol.


The following measures are collected in all 3 Parts of the study and correspond to the Primary, Secondary, and Tertiary analyses.

Primary Variables
Primary Analyses will look at improvements in well-being as a result of interventions. In this study, well-being is defined as and measured by state anxiety, state stress, and general mood. Normalized difference scores pre- to post-intervention will be analyzed to determine improvement in well-being.

Secondary Variables
Secondary Analyses will look at the role of internal experiences during the intervention on improvements in well-being. In this study, two aspects of internal experiences will be measured: expectations (how much participants believe the “relaxation exercise” will make them feel better or worse) and thought valence (how positive or negative the participants thoughts were during the exercise). The participants’ expectations of improvements, as well as the valence of their thoughts, have not been studied in the context of a (single session) meditation intervention. For any observed benefits in well-being seen in the Primary Analyses, these two Secondary Variables will be assessed for their mediating or moderating effects. See full pre-registered protocol for details on planned measurement and validation (as best as possible) of internal experiences, as well as predictions and rationale for planned analyses.

Tertiary Variables
Tertiary Analyses will be mostly exploratory in nature, with the goal of further clarifying individual effects of condition on improved well-being by reducing noise and finding the best model, as well as determining moderating effects of variables that have previously shown to correlate with meditation effectiveness (e.g., trait mindfulness, loneliness, etc.). Although these variables should be balanced across conditions due to random assignment (which will be checked prior to analyses), slight differences may create noise in the outcome measures. Additionally, some variables may interact with condition, and provide insight into who may benefit most (or least) from meditation. See full pre-registered protocol for the list of Tertiary Variables, including measurement details and rationale.

Sample and Study Costs

Sample Size Power Analysis
Based on the pilot data, the effect size was calculated from the differences in mean improvement scores of state anxiety between a 30-minute meditation and a 30-minute nature movie control condition. Using a standard calculation based on the F-test of this comparison, the resulting effect size is d = 0.42, 95% CI [0.07, 0.77] (Wilson, 2019). Using a different calculation method (based on raw “pre” and “post” scores and standard deviations of state anxiety rather than the F-statistic on normalized mean difference scores), the resulting effect size is dppc2 = 0.283 (Morris, 2008). Therefore, due to the low sample size from the pilot study and a large confidence interval on the effect size, an average effect size of d = 0.35 (i.e. f = 0.175) was used to calculate the N required to detect an effect of meditation for each study stage. Therefore, for an a priori F-test, using f = 0.175, Bonferroni-adjusted alpha = 0.002, power = 0.8, and 2 covariates (Expectation & Thoughts), the required number of participants needed to detect an effect is N = 913, or n = 91 in each of the 10 conditions. We will aim for 100 participants in each of our conditions, resulting in a total of N = 1,000 participants for the study.

Based on Prolific’s recommended fair pay minimum of a $9.60 hourly rate, an average study completion time of 40 minutes, and a projected sample size of N=1,000 participants, the researchers are requesting a total of $10,000 (about £7,000).
[1,000 participants x $6.50 per participant ($9.75 hourly rate) + service charges = $9,607.38 + small buffer for additional or unanticipated costs to attain a complete, representative sample]

Open Science Details

This study is pre-registered on OSF, where you may find the full proposal with detailed descriptions of the design, measures, predictions, and planned analyses. This page will also include intervention recordings, raw data, and cleaning & analysis code (in r) upon publication. This work will be submitted for publication in an open access journal. The researchers are grateful for your consideration, and are dedicated to improving not only scientific methods, replicability, collaboration and communication, but also to understanding and improving the well-being and mental health of this world. Thank you, and please reach out to Taylor Bondi (tbondi@ucsd.edu) with any additional questions.