Do Implementation Details Matter? Experimental Evidence from Self-Affirmation Interventions and Productivity

Last registered on December 05, 2025

Pre-Trial

Trial Information

General Information

Title
Do Implementation Details Matter? Experimental Evidence from Self-Affirmation Interventions and Productivity
RCT ID
AEARCTR-0017288
Initial registration date
November 24, 2025

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
November 26, 2025, 6:59 AM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated
December 05, 2025, 8:13 AM EST

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Region

Primary Investigator

Affiliation
Max Planck Institute for Research on Collective Goods

Other Primary Investigator(s)

PI Affiliation
University of Cologne

Additional Trial Information

Status
In development
Start date
2025-11-26
End date
2025-12-10
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
Does the effectiveness of interventions depend on implementation details? We study this in the context of a psychological self-affirmation intervention on performance in a real-effort task. Self-affirmations are brief exercises that encourage individuals to reflect on personal values or experiences. Self-affirmations are promoted in the workplace, organizations, leadership programs, and schools worldwide. They have shown promise in alleviating mental health challenges such as anxiety, stress, and low self-esteem. However, despite these promising findings, critical questions remain about how to implement self-affirmation interventions in order to improve productivity. There exist many versions of self-affirmations, and the efficacy of such exercises might be highly sensitive to implementation details. We study this in a unified setting with high-powered sample sizes.
External Link(s)

Registration Citation

Citation
Schneiders, Sebastian and Sebastian Tonke. 2025. "Do Implementation Details Matter? Experimental Evidence from Self-Affirmation Interventions and Productivity." AEA RCT Registry. December 05. https://doi.org/10.1257/rct.17288-1.1
Experimental Details

Interventions

Intervention(s)
We investigate the role of 4 implementation dimensions in a 2x2x2x2 design (16 treatments). We test the (i) type of self-affirmation (personal experiences vs. values), (ii) whether participants have to write about one value or experience or as many as they like (one vs. many), (iii) whether their anonymously written content is shared with the researchers or not (public vs. private), and (iv) whether the purpose of the self-affirmation exercise is revealed or not (purpose vs. no purpose).
Subsequently, participants can work on a real-effort task with substantial piece-rate incentives, in which they have to count the number of zeros in a series of 3x10 matrices filled with randomly generated zeros and ones for a maximum of ten minutes.
Intervention (Hidden)
Intervention Start Date
2025-11-26
Intervention End Date
2025-12-10

Primary Outcomes

Primary Outcomes (end points)
- Number of correctly solved real-effort tasks.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
- Psychological ease
- Psychological engagement
- Self-perceptions during the real-effort-task
- Mistakes, attempts, time, time per grid, breaks, and early exits in real-effort task
Secondary Outcomes (explanation)
- Psychological ease: survey items (ease, difficulty, comfort of performing during writing exercise) and behavioral measures during the task (word count, time, valence of writing)
- Psychological engagement: survey items (meaningfulness and self-relevance of written content) and behavioral measures during the task (word count, time, valence of writing).
- Self-perceptions during the real-effort task: survey items (stress, focus, difficulty, confidence, mood)

Experimental Design

Experimental Design
We systematically vary four key implementation details of self-affirmation (SA) interventions using a between-subjects design with 16 treatments. In addition, we have two standard active controls (one for values affirmation, one for experiences). The study is implemented in Qualtrics and conducted with U.S. participants recruited on Prolific.

Theoretical Rationale:

Self-affirmation (SA) interventions are designed to help individuals maintain a coherent and positive self-perception when faced with psychological threats. Yet, the effectiveness of SA interventions varies dramatically across studies, with some studies finding null or even negative effects. We hypothesize that much of this heterogeneity arises from implementation details (specific design features) that can induce a trade-off between:

1. Psychological engagement: An extensive and more authentic self-reflection enhances self-perceptions and motivation.
2. Psychological ease: An extensive self-reflection protocol can overwhelm participants and lead to difficulties in answering with ease. If participants believe they failed at the self-affirmation task, this can lead to disaffirmation.

Optimal intervention design maximizes engagement and ease. Prioritizing engagement over ease, however, can lead to disaffirmation if participants struggle to answer. Prioritizing ease, on the other hand, may reduce meaningful self-reflection. We test this experimentally across four key implementation dimensions. We hypothesize that these implementation details can shift this balance.

Hypotheses:

H1. Private vs. Shared

Private writing may remove evaluative concerns, thereby fostering authenticity and psychological ease. Shared writing, however, may increase engagement and deeper reflection if participants believe others care about their reflections. We hypothesize that privacy increases ease, but reduces engagement with an unclear overall effect. Hence, we treat this as a theoretically ambiguous, two-sided hypothesis.

H1 (Private vs. Shared): The treatment effects of self-affirmations differ depending on whether participants complete the exercise privately or with the expectation that others may read it.


H2. One vs. Multiple Reflections

Asking participants to produce multiple reflections (“one or more”) instead of one reflection may increase engagement, but on the other hand may decrease ease. Prior work (Rockenbach et al., 2025) finds that requesting to write about too many experiences can even backfire. Allowing participants to decide to write about one or more values or experiences might induce more engagement at the cost of reduced ease. We hypothesize that writing a single affirmation increases ease, but reduces engagement with an unclear overall effect. Hence, we treat this as a theoretically ambiguous, two-sided hypothesis.

H2 (One vs. Multiple): The treatment effects of self-affirmations differ depending on whether participants are prompted to write about one or multiple values or experiences.


H3. Values vs. Experiences

A values affirmation asks individuals to reflect on and write about their core personal values (such as family, creativity, or honesty) and why these values are meaningful to them. In contrast, an experiences affirmation (e.g., recalling a time one felt successful or proud) focuses on specific positive memories or achievements through recollection of past competence. Values-based affirmations allow one to choose among multiple domains and to write more abstractly. It could therefore increase ease. Experience-based affirmations rely on the recall of concrete, personally meaningful events. These can elicit deeper emotional resonance, but may also increase discomfort, particularly when recalling experiences is difficult. We speculate that the values affirmation increases ease, but decreases engagement with an unclear overall effect. Hence, we treat this as a theoretically ambiguous, two-sided hypothesis.

H3 (Values vs. Experiences): The treatment effects of the SA interventions differ depending on whether participants write about abstract values or concrete personal experiences.


H4. Disclosed Purpose vs. Undisclosed Purpose

Revealing an intervention’s purpose can increase engagement as participants understand the goal of the exercise. On the other hand, it may also trigger reactance or stigma (“I must need this help”) or add evaluative pressure as participants might believe what they write is inadequate for an effective self-affirmation and thereby decrease engagement. We speculate that writing with an undisclosed purpose increases ease, as participants are naïve and might just write what comes to mind. Since the overall effect is unclear, we treat this as a theoretically ambiguous, two-sided hypothesis.

H4 (Purpose vs. Undisclosed Purpose): The treatment effects of the SA interventions differ depending on whether participants are informed about its purpose and given a choice to participate.

---------------------------

In addition, we introduce two treatments that vary the financial incentives for the real-effort task. Participants will receive either 0.01 GBP per correctly solved item (N = 150) or 0.07 GBP per correctly solved task (N = 150). These conditions provide a benchmark for interpreting the magnitude of our self-affirmation interventions to changes in monetary incentives.

We also implement an autonomy-based preparation condition (N=300), in which participants are given around 5 minutes to prepare for the upcoming challenging task, however they choose. These conditions provide a benchmark for interpreting the magnitude of our self-affirmation interventions to granting individuals discretion over how to prepare.
Experimental Design Details
Randomization Method
Participants will be randomly assigned to a treatment by Qualtrics. We stratify by sex.
Randomization Unit
Individual level.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
3200 subjects who complete the experiment.

In addition, we 300 subjects in the money benchmark treatment and 300 subjects in the autonomy-based preparation condition.
Sample size: planned number of observations
3200 subjects who complete the experiment. In addition, we 300 subjects in the money benchmark treatment and 300 subjects in the autonomy-based preparation condition.
Sample size (or number of clusters) by treatment arms
Active Controls: 400 subjects for each of the two controls (800 subjects).
SA variations: 150 subjects per treatment with 16 variations (2400 subjects).

In addition, we 300 subjects in the money benchmark treatment and 300 subjects in the autonomy-based preparation condition.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
• We assume a drop-out rate of 5% (after exposure to the treatment). • Exclusion criteria: attention check and bot detection (honeypot and automated text authenticity checks to identify forbidden LLM use). To test our main hypotheses, we have at least 1200 participants (ITT) in each dimension group. • We are able to detect an effect size of d=0.11 (alpha = 0.05) with 80% probability. • We are able to detect an effect size of d=0.21 (alpha = 0.05) with 99.9% probability. Against control conditions: • All SA-treated (2400 vs 800 subjects). We are able to detect an effect size of d=0.11 (alpha = 0.05) with 80% probability. • One dimension (1200 vs 800 subjects). We are able to detect an effect size of d=0.13 at (alpha = 0.05) with 80% probability. • One treatment (150 vs 400 subjects). We are able to detect an effect size of d=0.27 (alpha = 0.05) with 80% probability. Treatment vs. Treatment • One treatment comparison (150 vs 150 subjects). We are able to detect an effect size of d=0.32 (alpha = 0.05) with 80% probability.
IRB

Institutional Review Boards (IRBs)

IRB Name
German Association for Experimental Economic Research
IRB Approval Date
2025-11-21
IRB Approval Number
g9H4S6qX

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials