This study targeted U.S. adults, aged 18 and older, and collected data on both those who use and those who do not use psychedelic substances. Data was collected via a voluntary and anonymous online survey as part of a larger study conducted from July 2024 to January 2025. The data analyzed here extends prior work; the present analysis incorporates additional participants and is cited alongside the initial PAEQ development paper (Hooper et al., 2025b) and the primary survey study (Hooper et al., 2025c) to maintain continuity. The study received approval from the University of Colorado Multiple Institutional Review Board, and participants provided informed consent electronically before proceeding with the survey.

Participants were recruited through social media advertisements, word-of-mouth, in-person outreach, and snowball sampling. Primary recruitment efforts focused on attracting participants residing locally in Colorado, with ads distributed via Meta and Google to eliminate bias inherent in recruitment from psychedelic-related forums. Individuals under 18 were excluded from participation, and no other specific inclusion or exclusion criteria were applied for enrollment. While mental health history, frequency of psychedelic use, and concurrent use of other psychoactive substances were recorded, they were not used as exclusion criteria. This inclusive approach aimed to maximize ecological validity.

The survey, hosted on the Research Electronic Data Capture platform (REDCap, www.redcap.com), gathered comprehensive information on participants' demographics, health status, and substance use patterns, including reporting on their acute psychedelic experiences. The present study included only complete data exclusively from current (past-year 1+ use) psilocybin users (n = 365), who were asked to describe their most typical experience. This phrasing was deliberately chosen to minimize bias by avoiding emphasis on either positive or negative experiences.

The survey was designed to assess various dimensions of participants' psychedelic experiences. Several validated instruments were used alongside the newly developed PAEQ to capture the full scope of these experiences and to establish convergent and divergent validity. Each instrument was originally developed and validated primarily in controlled experimental contexts, often within placebo-controlled psychedelic trials and administered during or immediately after the session to capture acute experiential content. In their original use, each instrument was designed to minimize recall bias and measure phenomenology in close temporal proximity to the psychedelic state. The present study administered these instruments retrospectively, outside of controlled settings, which introduces potential differences in construct expression and may influence reliability and validity estimates. While the core constructs measured by these scales are theoretically stable, their psychometric robustness in long-term retrospective contexts has been less frequently examined.

This study adhered to the ethical guidelines set forth by the American Psychological Association (APA). The survey was conducted anonymously, ensuring that no personal health information could be traced back to respondents or compromised. All data were securely managed and stored using the REDCap platform, which is designed to safeguard sensitive information. Access to the data was restricted to the research team through secure, password-protected logins.

Before participating, all individuals were fully informed about the study's objectives, and the voluntary nature of their involvement. Participants gave their informed consent electronically before starting the survey. Once data collection was complete, the data were exported into RStudio for analysis. Files were stored on the research team's password-protected computers to ensure the ongoing security of the information.

Data were exported from REDCap to excel. All data unnecessary for analysis were removed (incomplete or irrelevant measures). Categorical variables were recorded numerically and imputed into a codebook for reference. Variables were scored appropriately for each measure. Descriptive and standard psychometric analysis using the psych package was performed using R (R Core Team, 2022). Spearman correlations were computed using complete observations to identify monotonic relationships between variables of interest.

To ensure independent validation of the factor structure, the full cleaned dataset was randomly divided into two non-overlapping groups. Participants were assigned to either an exploratory group (EFA) or a confirmatory group (CFA) with approximately equal sample sizes (n = 182, 183). This random split was performed after initial descriptive and demographic analyses but prior to any factor analysis procedures. The EFA group was used to explore the latent structure of the scale, while the CFA group was used to independently test and validate the model identified in the EFA. A custom function was implemented to calculate a p-value matrix alongside the correlation matrix, enabling the evaluation of statistical significance for each observed association. CFA were conducted with the lavaan package (Rosseel, 2012). Maximum likelihood estimation was used in all CFA.

Dimensionality and item suitability for factor analysis were assessed using Bartlett's Test of Sphericity and the Kaiser-Meyer-Olkin (KMO) measure. Principal Components Analysis (PCA) was conducted to assess the underlying dimensions of the PAEQ, using eigenvalues and a scree plot to determine the appropriate number of factors. Subsequently, factor analysis using Principal Axis Factoring (PAF) was conducted to identify latent factors. Factor loadings were examined to interpret the structure and ensure construct validity.

Model fit was assessed using the comparative fit index (CFI) and the standardized root mean square residual (SRMR) (Bentler, 2006), with chi-square goodness-of-fit index changes reported between models. Following standard model fit criteria, a combination of fit indices was used, where values of SRMR <0.09 and CFI > 0.90 were considered indicative of an acceptable fit (Brown, 2006; Hu and Bentler, 1999). To account for potential non-normality in the data, chi-square values and CFI were computed using the Satorra-Bentler scaled test statistic (Satorra and Bentler, 2001).

Convergent validity was assessed by calculating Average Variance Extracted (AVE) for each latent factor using standardized loadings from the CFA model. AVE reflects the proportion of variance in observed indicators accounted for by the latent construct and was computed as the mean of the squared standardized loadings for each factor (Fornell and Larcker, 1981). Values exceeding 0.50 were interpreted as evidence of adequate convergent validity, indicating that the construct captures more variance than is attributable to measurement error.

To further assess convergent validity, Spearman's rank-order correlations were conducted between the total PAEQ score and theoretically related constructs. Given significant non-normality in the data as indicated by Shapiro-Wilk tests, Spearman's rho was used to identify monotonic associations. In addition to bivariate correlations, a multiple regression model was constructed to determine the predictive contribution of each external measure on PAEQ scores. This allows for a more precise examination of which components of the psychedelic experience most strongly relate to aesthetic engagement, while controlling for shared variance across predictors.

To evaluate the potential therapeutic relevance of aesthetic engagement, a separate series of linear regression models were conducted with the PAEQ total score as the sole predictor of various psychological and behavioral outcomes. These included self-reported changes in anxiety, depression, pain, sleep quality, alcohol use, opioid use, and overall quality of life as a direct result of a participant's typical (past year) psychedelic experience. Each model tested whether greater aesthetic engagement during a psychedelic experience was associated with improved wellbeing or behavioral change.

Survey respondents were primarily younger adults, with smaller proportions in older age groups. Gender representation was mostly male (38.1% female). Most respondents identified as White (85.2%), with smaller representations from other racial groups. Education levels varied, with about half reporting a degree from higher education (47.7%); see Table 1. Measurement scores are reported in Table 2.

The PAEQ Total score was scored by calculating the average score across all items for each respondent, with item 16 reverse coded. The average score among respondents was 4.69, with a standard deviation of 1.01, indicating moderate variability. Scores ranged from 0.76 to a maximum of 7. The 25th percentile was 4.10, the median was 4.76, and the 75th percentile was 5.38, suggesting that the majority of scores were clustered closer to the higher end of the scale.

An initial reliability analysis of the PAEQ items yielded a Cronbach's alpha of 0.89, indicating high internal consistency and suggesting that the items reliably measure a common underlying construct. Additional psychometric analyses revealed a squared multiple correlation (G6) value of 0.92, confirming the scale's robustness and reliability, as this metric provides a more conservative estimate of internal consistency (Guttman, 1945). The signal-to-noise ratio (S/N) was calculated as 8.3, demonstrating a strong ability to differentiate between true variance and measurement error.

Item-total correlation analysis identified item 16, “I experienced a heightened sense of ugliness,” as problematic due to a low corrected item-total correlation (0.19), indicating weak alignment with the overall construct. Despite being reverse-coded to align with other items on the scale, item 16 demonstrated high variability, suggesting it may not effectively measure the intended construct. Given these findings, item 16 was removed from the scale to enhance reliability. Following removal, the recalculated Cronbach's alpha increased slightly to 0.90, reflecting an improvement in internal consistency.

Bartlett's test of sphericity and KMO were analyzed to assess the suitability of the items for factor analysis. Bartlett's test was significant with χ² (190) = 1,596, p < 0.001, indicating that the correlation matrix was significantly different from an identity matrix and suitable for factor analysis. KMO yielded an overall Measure of Sampling Adequacy (MSA) of 0.86, which is considered meritorious, demonstrating that the sample was adequate for factor analysis. Individual item MSA values ranged from 0.77 to 0.93, further supporting the suitability of the dataset for factor analysis.

Dimensionality was assessed using PCA. Based on Kaiser's Criterion, which retains components with eigenvalues greater than 1, we selected the first four principal components based on eigenvalues of 6.18, 3.34, 2.03, and 1.70. Together, these components account for approximately 66% of the total variance in the dataset. Examination of the scree plot and factor loadings confirmed the retention of four factors, paralleling the intended four dimensions of the PAEQ; see Figure 1.

Based on the CFA results, modifications were made to the factor structure of the PAEQ to improve coherence, reliability, and construct validity. These changes aimed to refine the measure's ability to capture distinct dimensions of psychedelic aesthetic experience while maintaining accessibility for respondents.

We examined the validity and empirical consistency of the PAEQ by evaluating how well it measures its intended construct compared to related acute-experiential constructs that have already been validated within the psychedelic literature. Specifically, we aimed to establish convergent validity by comparing the PAEQ to existing measures known to assess overlapping dimensions of psychedelic experiences, such as mystical experiences (MEQ), emotional responses (EBI), and insight (PIS).

Given that Shapiro-Wilk tests indicated significant deviations from normality (p < 0.05) for all measures, Spearman's rank correlation was employed to assess the relationships between PAEQ and these external measures; see Table 5. Specifically, the Spearman correlation coefficient between PAEQ and MEQ Total was r = 0.693 (p < 0.001), with PIS Average it was r = 0.556 (p < 0.001), and with EBI Average it was r = 0.536 (p < 0.001). These results indicate that the PAEQ shares substantial, yet distinct (r < 0.85), variance with all three measures, supporting its convergent validity while demonstrating that it captures a unique aspect of the psychedelic experience.

To further assess the relationship between MEQ, PIS, and EBI as predictors of PAEQ, we conducted a multiple regression analysis. The model was statistically significant [F_{(3, 361)} = 134.3, p < 0.001] and explained 53% of the variance in PAEQ scores (R² = 0.53), indicating that these constructs are strong predictors of psychedelic aesthetic engagement. Among the predictors, MEQ was the strongest (β^* = 0.588, p < 0.001), suggesting that mystical experiences strongly correspond to heightened aesthetic engagement. EBI had a significant but smaller effect (β^* = 0.149, p = 0.004). In contrast, PIS was not a significant predictor of PAEQ (β^* = 0.072, p = 0.195).

We then conducted individual regression analyses to examine each PAEQ factor's predictive value on MEQ, PIS, EBI, and CEQ separately. In models predicting MEQ Total scores, Semantic engagement explained 56% of the variance (β^* = 0.75, p < 0.001), followed closely by Affective engagement, which explained 54% (β^* = 0.73, p < 0.001), and Flow, which accounted for 44% (β^* = 0.67, p < 0.001). Perceptual engagement explained a smaller but still meaningful portion of the variance (23%; β^* = 0.48, p < 0.001). These findings indicate that the semantic and affective dimensions of aesthetic experience most strongly predict mystical-type experiences.

For PIS Average scores, Affective engagement explained 36% of the variance (β^* = 0.60, p < 0.001), and Semantic engagement explained 35% (β^* = 0.59, p < 0.001), identifying them as the most influential predictors of psychological insight. Flow explained 18% of the variance (β^* = 0.43, p < 0.001), while Perceptual engagement accounted for 9% (β^* = 0.29, p < 0.001), suggesting that although immersive and perceptual features contribute, insight is primarily shaped by emotional and semantic content.

In predicting EBI Average scores, Affective engagement explained 34% of the variance (β^* = 0.58, p < 0.001), and Semantic engagement explained 32% (β^* = 0.57, p < 0.001), indicating strong contributions to emotional breakthroughs. Flow accounted for 17% (β^* = 0.41, p < 0.001), and Perceptual engagement explained 7% (β^* = 0.26, p < 0.001), again highlighting the primacy of emotionally and semantically rich aesthetic experiences in driving emotional breakthrough.

For CEQ Total scores, Flow explained the largest share of variance at 14% (β^* = 0.38, p < 0.001), followed by Semantic engagement (5%; β^* = 0.22, p < 0.001), Affective engagement (2%; β^* = 0.15, p = 0.003), and Perceptual engagement (2%; β^* = 0.15, p = 0.005). These results suggest that while all aesthetic factors are weakly positively associated with challenging psychedelic experiences, immersive absorption (Flow) is the strongest contributor.

To further investigate the relationship between psychedelic aesthetic engagement and perceived psychological and behavioral outcomes, we conducted regression analyses examining the PAEQ total score as a sole predictor of perceived changes in anxiety, depression, pain, sleep quality, opioid use, alcohol use, and overall quality of life. These models allowed us to assess whether individuals who reported greater overall aesthetic engagement during their psychedelic experience also experienced positive changes in wellbeing and behavior.

Higher PAEQ total predicted better sleep (β^* = 0.16, p = 0.002, R² = 0.03), lower pain (β^* = 0.13, p = 0.013, R² = 0.02), lower opioid use (β^* = 0.12, p = 0.021, R² = 0.02), lower alcohol use (β^* = 0.15, p = 0.004, R² = 0.02), reduced depression (β^* = 0.23, p < 0.001, R² = 0.05), reduced anxiety (β^* = 0.24, p < 0.001, R² = 0.06), and improved quality of life (β^* = 0.30, p < 0.001, R² = 0.09).

The current sample was predominantly white (85.2%) and male (61.9%), with nearly half of participants reporting at least a bachelor's degree (47.7%). These demographic characteristics reflect a largely Western, educated, industrialized, rich, and democratic (WEIRD) population (Henrich et al., 2010). Although recruitment strategies included public-facing social media and broad outreach beyond psychedelic-specific forums, the sample's demographic homogeneity constrains the generalizability of findings. The aesthetic experiences captured by the PAEQ may not reflect the experiences of individuals from more diverse racial, ethnic, or cultural backgrounds. Given evidence that cultural background shapes aesthetic perception (Masuda et al., 2008) and psychedelic phenomenology (Neitzke-Spruill, 2019), future research should prioritize diverse sampling and cross-cultural validation.

A key methodological limitation is that all measures were administered retrospectively, often months after the indexed psychedelic experience. While retrospective reports can capture enduring impressions, this design contrasts with their typical use in immediate post-session contexts and may introduce recall bias, mood-congruent memory distortions, or selective reconstruction of events. Future validation studies should incorporate real-time or next-day administration, both in controlled laboratory settings and in ecologically valid naturalistic contexts, to directly compare temporal stability and phenomenological fidelity.