Belief Strength and the Magnitude of Placebo Effects

Expectancy Measures and Placebo Outcomes

Many studies across clinical and experimental psychology have quantified participants’ expectations or beliefs prior to placebo treatment and examined how these beliefs correlate with the magnitude of placebo effects (symptom improvement, performance gains, etc.). In general, higher pre-treatment expectancy (greater faith or confidence in the treatment) is often associated with greater placebo-induced improvement in outcomespmc.ncbi.nlm.nih.gov cambridge.org. The table below summarizes key findings from representative studies (including recent work and notable historical examples):

Study (Year)	N	Belief Measure	Outcome Measure	Correlation (r)	p-value	Citation
Rutherford et al. (2013) – Pilot antidepressant trial (placebo vs. active)	29¹	Expectancy of improvement (baseline)	Depression severity change (HAM-D)	+0.53	0.021	cambridge.org
Whalley et al. (2008) – Repeated placebo analgesia trials	40¹	Response expectancy for pain relief	Pain reduction (placebo analgesia)	~+0.5²	<0.05²	pubmed.ncbi.nlm.nih.gov
Świder et al. (2019) – Placebo analgesia with verbal suggestion	56 (28²)	Expected pain intensity (pre-test, VAS)	Pain reduction (placebo – control)	+0.504	0.003	journals.plos.org
Baker et al. (2022) – Placebo effect on emotion/mood (open-label)	40	Expectation of mood improvement (VAS)	Mood enhancement (self-report)	+0.437	0.005	nature.com nature.com
Stetler (2014) – Placebo effect on memory via exercise adherence	72	Expectation of memory improvement	Memory test performance (recall)	+0.27³	0.03³	pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov
Schafer et al. (2015) – Open-label placebo analgesia (conditioning)	40	Expected pain relief (VAS) before placebo reveal	Pain reduction (placebo analgesia)	–0.02	0.89	pmc.ncbi.nlm.nih.gov
Marlatt & Rohsenow (1980) – Placebo alcohol (balanced placebo design)	96	Belief in having consumed alcohol	Social behavior (e.g. talkativeness)	_n/a_⁴	_n/s_⁴	pmc.ncbi.nlm.nih.gov
Dinnerstein & Halm (1970) – Placebo aspirin for mood	60	Expectation of mood change (induced)	Self-rated mood improvement	_n/a_⁴	_n/s_⁴	sciencedirect.com

¹ N for full sample. ²Expectancy measured in a subset (e.g. placebo group); r not explicitly reported in text but described as significantpubmed.ncbi.nlm.nih.gov. ³Expectation → memory correlation reported after controlling for other factorspmc.ncbi.nlm.nih.gov. ⁴Early studies in the 1970s–80s demonstrated qualitative expectancy effects but did not always report an r; they are included here as historical context (e.g. Marlatt & Rohsenow showed that believing one drank alcohol led to placebo intoxication behaviorspmc.ncbi.nlm.nih.gov). “n/s” indicates no separate statistical correlation reported.

As seen above, positive correlations are frequently found between pre-treatment belief strength and placebo outcomes. For example, in a pilot antidepressant trial, patients with higher initial expectations of improvement showed significantly greater reduction in depression symptoms on placebo (r≈0.53)cambridge.org. In pain studies, individuals expecting strong analgesia tend to report larger placebo analgesic effectsjournals.plos.org. Świder et al. (2019) demonstrated that expectancy (rated before a sham pain treatment) explained ~25% of the variance in pain relief (r=0.50)journals.plos.org. Similarly, Whalley et al. (2008) found placebo pain relief was “significantly associated with response expectancy”pubmed.ncbi.nlm.nih.gov. In an open-label design (where participants knew they took a placebo), Baker et al. (2022) still found that those who believed more strongly in a positive mood effect showed greater emotional improvement (r≈0.44)nature.com nature.com.

Not all studies find a strong relationship – indicating that expectancy is not the sole determinant of placebo responses. For instance, Schafer et al. (2015) conditioned placebo analgesia through prior experience; they reported no correlation between explicitly expected pain relief and analgesic outcome (p≫0.05)pmc.ncbi.nlm.nih.gov. This suggests learned associations can produce placebo effects even when self-reported expectancy is low or irrelevant. Likewise, some experiments in healthy adolescents failed to induce robust placebo effects or correlations, possibly due to developmental or design factors (e.g. Gulewitsch et al., 2020)pmc.ncbi.nlm.nih.gov.

Open-Label Placebos and Belief

Open-label placebo (OLP) studies – where participants know they are receiving a placebo – provide a unique test of expectancy effects. Interestingly, OLP trials often include a persuasive rationale to build positive expectation (e.g. explaining “the placebo effect can be powerful even if you know you’re taking a placebo”pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov). Positive patient beliefs have been shown to moderate OLP outcomes. For example, a study of allergic skin reactions found that an OLP (an inert cream) reduced wheal responses only among participants with strong pre-existing belief in placebos’ power, compared to no effect in skepticspmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov. In other OLP trials (for pain, back pain, etc.), expectation ratings collected at baseline have significantly predicted symptom improvementscpe.psychopen.eu nature.com. One analysis noted that in open placebo studies of chronic pain, patients with higher baseline expectation often experience greater pain relief, although findings have been mixed across studiesprograminplacebostudies.org cpe.psychopen.eu. Overall, even without deception, believing in the treatment rationale appears to enhance OLP outcomes, consistent with standard placebo effects.

Historical Perspective (1970s–1980s)

Early research in the 1970s and 1980s laid the groundwork for our understanding of expectancy in placebo effects. Pioneering placebo studies in that era often focused on experimental manipulations rather than correlation statistics, but they observed the same general pattern: people’s expectations shape their physiological and psychological responses. For example, Marlatt and Rohsenow (1980) used a balanced placebo design with alcohol: participants who believed they drank alcohol (but hadn’t) exhibited behavioral signs of intoxication, whereas those who unknowingly drank alcohol did not – demonstrating expectation can produce drug-like effectspmc.ncbi.nlm.nih.gov. In clinical settings, observational analyses from the NIMH Treatment of Depression project (1980s) found that patients’ expectations of improvement predicted their likelihood of responding to placebo treatment in depressionpmc.ncbi.nlm.nih.gov. Similarly, Dinnerstein and Halm (1970) showed that inducing a positive expectancy about a pill could alter self-rated moodssciencedirect.com. These historical findings align with modern data in underscoring expectancy as a key “active ingredient” of placebo responses.

It is worth noting that some early theories proposed other personality factors (like suggestibility or anxiety reduction) as contributors to placebo effects. However, systematic experiments (e.g. a 2002 pain study by De Pascalis et al.) concluded that expectancy and suggestibility together best explained placebo analgesia magnitudessciencedirect.com sciencedirect.com – with expectancy often emerging as the stronger single predictor. By the late 20th century, Response Expectancy Theory (Kirsch, 1985) formally posited that a person’s belief about an involuntary response (pain relief, mood change, etc.) can directly cause that responsefrontiersin.org frontiersin.org. The accumulated evidence from the 1970s–1980s provided comparative context, showing that while methodologies have evolved (and statistical reporting is now more rigorous), the core phenomenon – belief fueling the placebo benefit – has been remarkably consistent over timepmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov.

Data Sources for Placebo–Expectancy Research

Researchers interested in placebo–expectation interactions can access datasets and trial results through several platforms:

ClinicalTrials.gov: This registry’s API allows querying of clinical trials (including those with placebo arms) and downloading of study results in bulk formats (XML or CSV)clinicaltrials.gov. For example, one can search for trials that recorded patient expectation or used the Credibility/Expectancy Questionnaire. Many posted results datasets (in Excel/CSV) can be retrieved, which may include outcome data for placebo groups and any measured expectancy covariates.
Open Data Repositories: Individual researchers often share placebo study data in public repositories. For instance, data from an open-label placebo study on reward learning was released on GitHubgithub.com, and a chronic pain placebo-response brain imaging dataset is available on OpenNeuroopenneuro.org. Such datasets typically include participant-level measures (expectancy ratings, outcomes, etc.) in downloadable formats.
NIH Repositories: Large clinical studies sometimes deposit data in NIH repositories. The recent IBS placebo study by Lackner et al. (2024) is one example – its de-identified dataset (including baseline expectancy/credibility measures and outcomes) is available via the NIDDK Central Repositorysciencedirect.com upon request. Researchers can obtain the data (e.g. as Excel files) to conduct secondary analyses on placebo predictors.
Psychology Databases: Journals and platforms (APA’s data portal, PsychOpen, etc.) occasionally host supplemental files with trial data. For example, an open-access paper on open-label placebos in well-being provided correlation data in its supplement, showing that pre-treatment expectation (measured the day before placebo) correlated with improvement in well-being scorescpe.psychopen.eu. These files can often be downloaded as spreadsheets.

By leveraging these resources, investigators can analyze placebo-belief interactions at scale. For instance, using ClinicalTrials.gov one could compile all depression trials reporting an expectancy measure and extract the placebo-arm outcomes to examine expectancy–response correlations across studies. Open datasets and APIs thus facilitate more powerful meta-analytic and predictive modeling approaches – moving beyond individual trials to identify how belief effects vary by condition, population, or intervention. With growing emphasis on open science, the data needed to study the placebo effect and its psychological moderators (like expectation) are increasingly accessible for researchers and clinicians around the world.

The 188 Axioms of Theophysics

Explorer

PSY_Placebo_Belief_Strength_Source

Belief Strength and the Magnitude of Placebo Effects

Expectancy Measures and Placebo Outcomes

Open-Label Placebos and Belief

Historical Perspective (1970s–1980s)

Data Sources for Placebo–Expectancy Research

Graph View

Table of Contents