Structure of Current Psychopathology and its Associations with Daily Life Experiences using the HiTOP-SR in a Mixed Clinical/Community Sample

Abstract

The Hierarchical Taxonomy of Psychopathology (HiTOP) is a dimensional nosological system that addresses key limitations with categorical frameworks, including heterogeneity, boundary, and comorbidity issues. The HiTOP consortium recently developed a new self-report instrument, the HiTOP-Self-Report Measure (HiTOP-SR), designed to operationalize the HiTOP model for use in research and clinical practice. In a set of preregistered analyses with a sample of clinical/community participants (75% female, 81% white), we explored the hierarchical structure of the HiTOP-SR scales using exploratory factor analysis (n = 637) and examined their associations with behaviors and experiences assessed in daily life (n = 531), such as affect, stress, impulsivity, energy, sleep quality, and social interactions. Findings indicate a nine-factor model, closely aligned with the HiTOP’s current structure, best represented the measure. The hierarchical structure of the HiTOP-SR generally converges with the HiTOP model, with several key departures, particularly for historically understudied constructs. Furthermore, the HiTOP-SR facet scales and domains associated with individual differences in daily behavior and experiences as anticipated, highlighting the construct validity and the potential clinical utility of this new measure. Our results have implications not only for the structure, validity, and clinical utility of the HiTOP-SR but also raise broader questions about the underlying nature of psychopathology as represented by the HiTOP.

The Hierarchical Taxonomy of Psychopathology (HiTOP; Kotov et al., 2017) is a dimensional system for organizing psychopathology hierarchically based on the covariation of the features of mental illness. The HiTOP is the product of a large consortium of mental health researchers who have come together to build a quantitative empirical model that describes the structure of psychopathology. The HiTOP aims to address some of the key limitations with traditional categorical frameworks, such as the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; American Psychiatric Association, 2013) and the International Classification of Diseases (ICD-11; World Health Organization, 2022), including a) heterogeneity within diagnoses, b) arbitrary categorical boundaries between groups of symptoms, c) excessive comorbidity issues, d) poor diagnostic reliability for many diagnoses, and e) limited clinical application and utility (Clark et al., 2017; Krueger et al., 2018; Simms et al., 2022). Since the publication of the seminal manuscript describing the model in 2017, the HiTOP has generated a new wave of research on the quantitative empirical structure of psychopathology, with numerous studies exploring the validity, generalizability, and clinical utility of the HiTOP framework as an alternative to categorical systems (e.g., Kotov et al., 2021; Ruggero et al., 2019; Ringwald et al., 2023).

One limitation of previous research on the HiTOP is that it has largely relied on questionnaires and clinical interviews based on traditional categorical frameworks, such as the DSM and the ICD. To address this issue, the Measures Development Workgroup of the HiTOP consortium recently developed a new self-report instrument, the HiTOP Self-Report Measure (HiTOP-SR; Simms et al., 2023; https://www.3plab.org/hitop), that was created by surveying the universe of psychopathology content falling within the existing HiTOP model and beyond and developing scales to cover such content (Simms et al., 2023). The process of content generation and preliminary scale development has been detailed in a series of manuscripts authored by the HiTOP Measures Development Workgroup (Cicero et al., 2022; Mullins-Sweatt et al., 2022; Sauer-Zavala, 2022; Sellbom et al., 2022; Simms et al., 2023; Simms et al., 2022; Verona, 2022; Watson et al., 2022; Zimmermann et al., 2022). The expectation is that the resulting scales will conform to or approximate the structure of the HiTOP based on the existing literature. The goals of these efforts are ultimately to have a measure to use in future research on the structure of psychopathology and provide practicing clinicians with a tool to allow them to use the HiTOP with patients for clinical assessment and diagnosis.

The HiTOP model is hierarchical and describes psychopathology at several levels, with varying degrees of specificity, ranging from very narrow features (i.e., specific signs and symptoms) to increasingly broader domains, including the commonly observed positive association among all (or most) domains of psychopathology (i.e., the p-factor; Caspi et al., 2014). Our focus in the current study is on the postulated psychopathology spectra (e.g., Internalizing, Disinhibited Externalizing, Antagonistic Externalizing, Thought Disorder, Detachment, and Somatoform) and their subfactors. For example, the Internalizing spectrum currently has four subfactors: fear, distress, eating pathology, and sexual problems. Compared to the strong consensus on the spectra level structure, there is less certainty on the exact position of some subfactors; for example, mania is currently housed under both the Thought Disorder and Internalizing spectra, but this may change as more research is conducted on this model.

The new HiTOP-SR measure is a critical development as it allows researchers to study the full spectrum of psychopathology (at least as captured by the HiTOP) in a unified manner with a single measure in the same sample, an important step that was not feasible until now due to the use of disorder-specific measures. As a result, the HiTOP-SR measure will ultimately help clarify questions about the exact placement of subfactors and other issues related to the structure of psychopathology, refine the HiTOP model, as well as improve its practical utility through the availability of a clinical tool covering its content (Simms et al., 2022, 2023). At the current juncture, the HiTOP-SR is being refined and validated through investigations of its structure and correlates with external measures and criteria.

The Current Study

The current study is part of the Intensive Longitudinal Investigation of Alternative Diagnostic Dimensions (ILIADD) project and was undertaken to explore the structure of the HiTOP-SR facet scales and how they relate to individual differences in behaviors and experiences in daily life in a mixed clinical and community sample. Validating this new measure using daily life outcomes is an important step as it helps clarify its construct/content validity through cross-method associations with relevant daily behaviors in addition to shedding light on the real-life relevance and, therefore, the potential clinical and practical utility of the HiTOP-SR. In line with these aims, we first estimated exploratory factor analyses (EFAs) in a sample of 637 participants to evaluate the higher-order structure of these scales. In a subset of these individuals (n=531), we next sought to test the construct and content validity of the facet scale scores and resulting domains from the EFAs by examining their relationship with behaviors reported in daily life assessed with ecological momentary assessment (EMA). Specifically, we had two research questions: (1) What is the higher-order factor structure of the HiTOP-SR facet scales? (2) How does each factor-analysis identified HiTOP-SR domain and each facet scale of the HiTOP-SR relate to: a) average momentary behaviors and experiences (negative affect, positive affect, impulsivity, stress, and energy), b) average daily sleep quality, and c) average social interaction behaviors (dominant behavior, agreeable behavior, and empathy during social interactions)?

In terms of the structure of the HiTOP-SR scales, we expect to find a structure similar to the rational organization of facet scales at the spectra level provided by the HiTOP Measures Development Workgroup (Simms et al., 2023). In addition, we have provided our hypotheses for the associations between the HiTOP-SR scales and individual differences in daily outcomes in Table 1. At the time of analysis, no studies had examined the structure of the HiTOP-SR scales, and this is the first study to examine their associations with individual differences in daily outcomes. Consequently, we lacked a clear expectation of how this new measure would perform or which specific domains might emerge. Therefore, we offered only broad, general hypotheses.

Table 1.

Hypothesized Associations Between the Main HiTOP Domains and Daily Outcomes

	HiTOP Domains
EMA Outcomes	Internalizing	Externalizing	Mania	Thought Disorder	Detachment	Somatoform
Negative affect	+	+	+	+	+	+
Positive affect	-	-	+	-	-	-
Energy	-	?	+	+	-	-
Stress	+	+	+	+	+	+
Impulsivity	-	+	+	+	-	-
Sleep	-	-	-	-	?	-
Empathy	?	-	?	?	?	?
Dominance	-	+	+	?	?	-
Agreeableness	-	-	-	?	-	-

Methods

Participants

Participants were recruited through a public university’s research recruitment registry. The criteria for the present study were that participants be between 18–50 years old and own a study-compatible Android or iPhone smartphone. Additionally, the sample was enriched for participants with a history of recent mental health treatment. To accomplish this, only participants who indicated a current or recent history of treatment were initially eligible to participate in the study, and targeted advertisements were sent to participants in the university’s research recruitment registry who had a documented history of mental health treatment. Eligibility criteria were later broadened to allow participation regardless of mental health treatment history. Of the total sample, 80% indicated a history of mental health treatment, with 62% indicating that they were currently in treatment or had received mental health treatment in the past 12 months. The total sample size for those who completed the baseline questionnaires was N = 665, and the total sample size for those who completed both baseline and EMA questionnaires was n = 552. Since we removed observations that exceeded 6 standard deviations for the HiTOP-SR scales (more details in Discussion), 28 participants were dropped and the final baseline sample size used in analyses was n=637. Additionally, 21 participants were removed from the EMA dataset due to having fewer than 3 entries, resulting in a final EMA sample of n=531. Demographic information for the sample is provided in Supplemental Table S3.

Procedures

Participants who met inclusion criteria provided consent and then were directed to complete a series of baseline self-report questionnaires. Following the completion of the baseline questionnaires, participants completed an online training for the EMA portion of the study. The training included information on downloading the app needed to participate in the study, MetricWire (MetricWire Inc., 2019), and an overview of the EMA portion of the study. Following the training, participants were required to complete a brief, 8-item comprehension quiz to help ensure understanding of the EMA protocol before moving forward with their participation in the study.

During the EMA portion of the study, participants received a total of eight random prompts between 8:00 am-11:00 pm each day for 15 days. In addition to the eight random surveys per day, participants also completed one morning survey each day asking about their sleep quality. Finally, participants completed event-contingent surveys focused on interpersonal interactions, with the expectation that participants record at least four social interactions each day to reach 100% compliance for the day. Social interactions were defined as a direct, real-time communication between the participant and another person that lasted for at least five minutes.

Any participant who completed the baseline assessment was entered into a raffle to win one of five $100 Amazon gift cards. Compensation for the EMA portion of the study was based on compliance rates, with higher compliance resulting in higher payments. Participants could earn up to a total of $200 for completing EMA surveys (which included a $25 bonus if their compliance rate was 90% or higher). To further incentivize EMA compliance, for each survey completed, participants would receive one entry into a drawing for one of five Apple Watches. The number of entries for event-contingent surveys that could count towards the drawing for the Apple Watch was capped at five entries per day, so that participants did not try to increase their chances of winning the Apple Watch by submitting a high number of event-contingent surveys. Thus, the total possible compensation for the study was $200, with additional chances to win a $100 Amazon gift card and Apple Watch. To further increase compliance, participants were contacted by email every four to five days after starting the EMA portion of the study.

Given the use of fully remote data collection, several steps were taken to ensure data quality and mitigate risk of low-quality data and/or fraudulent responses. First, participants had to be registered members of the university’s research recruitment database, which requires an authentication of one’s identity. Second, compensation was provided through disposable credit cards, which requires an authentication process involving the use of an individual’s social security number. Four attention check items were included in the baseline assessment (e.g., “Please select the option, ‘Strongly agree’”), and participants who failed more than one attention check item were excluded from the study. Finally, only one person per household was eligible to participate in the ambulatory assessment period, which was screened through the collection of home addresses.

Measures

Baseline Measures

Hierarchical Taxonomy of Psychopathology Self-Report Measure (HiTOP-SR).

The HiTOP-SR (Simms et al., 2023; https://www.3plab.org/hitop) is a 405-item self-report instrument that assesses psychopathology symptoms as represented in the HiTOP model (Kotov et al., 2017). The HiTOP-SR was designed to be broad but comprehensive in scope and assesses a wide variety of psychopathology symptoms and behaviors. Participants were asked to rate how much each item applied to them in the past month. Each Likert-like item was rated using a 1 (“Not at all”) to 4 (“A lot”) scale. Eighty-six specific symptom scales (e.g., Affective Lability, Depressed Mood) and nine alcohol use subscales (e.g., Craving, Hazardous Use) were computed using the HiTOP-SR items. At the time of the study, the 405-item version of the HiTOP-SR had not been finalized, and we employed an earlier, 495-item version of the measure. The primary difference between these versions is that the final version of the HiTOP-SR includes fewer items, which are the ones included here. However, for the Hallucinations facet scale, new items were developed for the final version of the HiTOP-SR. Thus, the Hallucinations facet differs between the version of the HiTOP-SR in the present study and the final version of the measure. In addition, at the time of data collection, the Cleaning scale of the HiTOP-SR, which is thought to be a marker of Fear, was not finalized and therefore was not included in this study.

In addition, to avoid an overly-narrow factor of Alcohol Use (among the Harmful Alcohol Use scales, the minimum correlation was r = .85; further details in Supplemental Materials), we collapsed the 21 Harmful Alcohol Use items into one scale of Harmful Alcohol Use (e.g., “I craved a drink of alcohol.”, “I drank a lot more than others before feeling drunk”, “I could not function without alcohol”). We took the average of all items for each subscale (i.e., Craving, Hazardous Use, Impaired Control, Role Interference, Tolerance, Withdrawal), and then averaged the subscale scores to create the composite scale. Based on editorial feedback on the first draft of this manuscript, we have not included the three alcohol consumption scales (i.e., Drinks per Occasion, Frequency of Use, Frequency of Intoxication) in our final analyses. Therefore, we used a total of eighty-seven HiTOP-SR scales in our analyses (eighty-six symptom scales in addition to one alcohol use-specific composite scale).

EMA Measures

Sleep Diary.

Each morning participants rated how well they slept, using a 0 (“Poorly”) to 10 (“Very well”) sliding scale.

Random Surveys

Momentary Experiences and Behaviors.

During each random prompt, participants rated their mood using four items. Positive affect was assessed with one item (“In the past 15 minutes, how POSITIVE have you felt?”) and negative affect was captured with one item (“In the past 15 minutes, how NEGATIVE have you felt?”), both using a 1 (“Neutral”) to 10 (“Very Negative/Positive”) sliding scale. In addition to rating positive and negative affect, participants rated their momentary energy level (“In the past 15 minutes, how ENERGETIC/AWAKE have you felt?”) and their momentary stress (“In the past 15 minutes, how STRESSED have you felt?”). Energy level was rated using a 0 (“Not at all”) to 10 (“Very Energetic/Awake”) sliding scale. Stress was rated using a 0 (“Not at all”) to 10 (“Extremely”) sliding scale.

Four items were used to assess momentary impulsivity. For each item, participants were asked whether they had engaged in the impulsive behavior in the past hour using a 0 (“Neutral”) to 10 (“Very Much”) sliding scale. The four items were, “I did something on impulse”; “I did things without worrying about the consequences”; “I decided to put off something that I had to do”; “I avoided doing something despite the consequences”.

Event-contingent Surveys

Interpersonal Behavior.

Participants rated their own interpersonal behavior using the Social Behavior Inventory (SBI; Moskowitz, 2021), which assesses behaviors along the major axes of the interpersonal circumplex (Dominance/Submissiveness and Warmth/Coldness; Kiesler, 1983). The SBI is a checklist instrument that asks participants to indicate all behaviors that they engaged in during the interaction. Participants are presented with separate checklists of dominant (11 items; e.g., “I asked them to do something”), submissive (11 items; “I did not say what I wanted directly”), quarrelsome (12 items; e.g., “I confronted them about something I did not like”), and agreeable behaviors (12 items; e.g., “I showed sympathy”) and asked to choose all behaviors that applied. Participants were also able to select a “None of the above” option. The degree of dominant behavior was calculated by subtracting the number of submissive behaviors from dominant behaviors and level of agreeable behaviors was calculated by subtracting the number of quarrelsome behaviors from the number of agreeable behaviors. Thus, higher scores on the dominance/submissiveness SBI scale reflect a higher degree of dominant behaviors, and higher scores on the agreeableness/quarrelsomeness scale reflect a higher degree of agreeable behaviors and the inverse is true for low values on each scale.

Empathy.

Four items from an empathy measure (Ringwald & Wright, 2021) were used to assess empathy during interpersonal interactions: “I considered what the person(s) I interacted with was thinking”; “I considered what the person(s) I interacted with was feeling”; “When the person(s) I interacted with showed emotions, I felt their emotions inside of me”; and “I empathized with the other person(s) I interacted with.” The first two items were designed to index cognitive empathy, while the third item indexed emotional empathy. The fourth item was designed as a face valid index of global empathy. Responses were made using a sliding scale from 0 (“Neutral”) to 10 (“Very Much”).

Analytic Plan

For factor analyses and exploratory structural equation modeling (ESEM) we used Mplus (v8.10; Muthén & Muthén, 2023). For all other analyses, we used Base R (v4.2.3; R Core Team, 2023) and Rstudio (v2023.9.0.463; Posit Team, 2023), in addition to the following R packages: psych (Revelle, 2023), psychTools (Revelle, 2024), multilevelTools (Wiley, 2024), and dplyr (Wickham et al., 2023) to preprocess and clean datasets.

Our analyses deviated from the preregistration in several ways, in part in response to newly released results on the HiTOP-SR from another research team (Zimmermann et al., 2024), as well as based on preliminary results from the early stages of our analyses suggesting that modifications to some aspects of the approach would be necessary to achieve sensible results. The deviations are described in detail in Supplemental Materials.

In addition to establishing the hierarchical structure of the HiTOP-SR scales, we also tested its construct/content validity and potential clinical relevance by estimating Pearson correlations between individual averages in momentary functioning scales (affect, stress, energy, sleep, impulsivity, interpersonal behavior) and all HiTOP-SR facet scale scores. We also estimated Pearson correlations between average momentary functioning and the score estimates for each EFA HiTOP-SR domain. Given the large number of tests, we used a significance threshold of p <.01 when interpreting the results.

We have provided hypotheses focused on the six main domains in Table 1. A “+” indicates that a significant positive association between the domain and daily functioning variable is expected, a “−” indicates that a significant negative association is expected, and a “?” indicates no prediction for the effect. Some higher-order constructs in the HiTOP-SR include lower-order facets, such as Internalizing (fear, distress, eating pathology, etc.) and Externalizing (disinhibition, antisocial, etc.), whereas other higher-order constructs do not, such as Detachment, Somatoform, and Thought Disorder, as represented by the rational organization of the measure proposed by the HiTOP Measures Development Workgroup (Simms et al., 2023). We did not specify explicit hypotheses about the directions and significance of associations for each individual facet scale. However, as a blanket prediction, if a facet is a positive indicator of the domain, we expected the facet to correlate with EMA outcomes in the same direction as the domain. In contrast, if a facet is negatively related to a domain, such as the negative association between anankastia and the Externalizing domain, we expected the facet to correlate inversely with EMA outcomes. Additionally, we hypothesized that the main facet scales of the HiTOP-SR will closely replicate the HiTOP model structure and the rational organization at the spectra level, as proposed by the HiTOP Measures Development Workgroup.

We had the following hypotheses for the associations between the HiTOP-SR domains and facet scales. If a domain does not include facets (indicated with an asterisk *), we list some examples of individual scales that we hypothesize associations for. The sign next to each facet/scale (+/−) indicates the predicted direction of association:

• Internalizing: fear (+), distress (+), eating pathology (+), sexual problems (+)

• Externalizing: antisocial behavior (+), antagonism (+), disinhibition (+), anankastia (−)

• Mania¹: manic energy (+)

• Thought Disorder*: e.g., hallucinations (+), delusions (+), eccentricity (+)

• Detachment*: e.g., restricted affectivity (+), romantic disinterest (+), submissiveness (+)

• Somatoform*: e.g., somatic preoccupation (+), disease conviction (+), body dissatisfaction (+)

Transparency and Openness

Hypotheses and planned analyses were preregistered and are available on the Open Science Framework (https://osf.io/7ge6m/overview). The study design was not preregistered. Note that we deviated from the preregistered EFA plan in several ways, and all deviations are reported in Supplemental Materials. All data, code, and raw output are also publicly available on the OSF project at https://osf.io/6yuhg/overview. All study procedures were approved by the University of Pittsburgh’s institutional review board.

Results

Descriptive statistics and reliability for study measures are reported in Supplemental Tables 1–2 (S1–2).

Latent Structure of the HiTOP-SR Scales

We estimated EFAs with oblique CF-Parsimax rotation, requesting solutions with 1–9 factors. We relied on parallel analysis results (Horn, 1965) to determine the maximum number of factors to retain. The eigenvalues of the estimated correlation matrix exceeded the 95^th percentile of random data generated eigenvalues through the first ten factors (first eleven empirical eigenvalues are 22.89, 5.28, 3.26, 2.74, 2.63, 2.40, 2.04, 1.90, 1.73, 1.56, and 1.44; first eleven 95^th percentile eigenvalues for random data are 1.83, 1.77, 1.73, 1.69, 1.65, 1.63, 1.60, 1.57, 1.55, 1.52, and 1.50). We considered up to nine factors, because the ten-factor solution did not seem to add significant interpretive value beyond the nine-factor model. Although we decided against including the ten-factor model in our main analyses and discussion, we have included the model results in Supplemental Table S12.

In addition to parallel analysis results, we used patterns of factor loadings, factor loading strengths, and interpretability to evaluate model fit for the nine EFA models. We were reluctant to establish hard-and-fast cut-offs for primary factor loadings or cross-loading values given the wide breadth of content and the expectation that many scales will have non-negligible cross-loadings. To get a better sense of how each factor emerged, we applied the bass-ackwards technique (Goldberg, 2006). This method involves estimating a series of factor solutions with increasing number of factors at each iteration and correlating the factor scores estimates for all solutions. We estimated EFAs, requesting models with 1–9 factors based on the parallel analysis results discussed above, with oblique CF-Parsimax rotation. Since factor score estimates are indeterminate, which could significantly distort the hierarchical structure obtained through the bass-ackwards method, we used Mplus and verified that our factor score estimates for the 9-factor solution did, in fact, demonstrate high determinacy, with values ranging from .918 – .958. A visualization of the outcome with the nine models is depicted in Figure 1. The factor score estimate correlations across models represent paths across different levels of the hierarchy. Because of the large number of scales, we restrict our description to the strongest scale loadings at each level (r ≥ |.30|), and we have only visually depicted paths with factor score estimate correlations of r ≥ |.45| in Figure 1. We use the bass-ackwards method here in a purely descriptive manner to visualize the unfolding of different factors across levels and as a result, we did not use a particular cut-off to determine when a new factor emerged as opposed to the same factor carrying forward to the next level.

Figure 1.

Graphical representation of the estimated HiTOP-SR hierarchy.

Note. The arrow colors aim to aid with visibility and do not have any particular statistical meaning.

Starting with the 1-factor solution (see Figure 1), this factor represents shared features across all psychopathology symptoms, which would be a representation of the “p-factor” (Caspi et al., 2014). We see strong loadings for scales with diverse content capturing both Internalizing and Externalizing symptoms, including Cognitive Problems (.79), Bodily Distress (.79), Depressed Mood (.77), Shame/Guilt (.77), Non-Persistence (.70), Suspiciousness (.68), and Angry Hostility (.63). On the other hand, however, several scales also had loadings below the .40 threshold, such as Risk Taking, Dietary Restraint, Gambling, Hyperdeliberation, and Harmful Alcohol Use (loadings ≤ |.22|). The average loading on this factor across all eighty-seven scales was .46, with a range of |.01| to |.79|.

At the second level of the hierarchy, the general psychopathology factor split into an Internalizing factor, with the strongest loadings from scales tapping into low mood and distress, such as Anhedonia (.83), Depressed Mood (.83), and Lassitude (.79), and an Externalizing/Rarity factor, with the strongest loadings from scales capturing Manipulativeness (.59), Risk Taking (.58), and Non-Planfulness (.55), as well as more intense or rarer experiences, such as Manic Energy (.65), Eccentricity (.65), Dissociation (.56), and Affective Lability (.50) At the three-factor level, Distress/Detachment emerged from the Internalizing factor, covering content related to social withdrawal and mood disturbances. Externalizing/Rarity had the strongest loadings from Risk Taking (.74), Manipulativeness (.64), Oppositionality (.63), and Non-Planfulness (.62) scales at this level. A Fear factor emerged as well, relating to both Internalizing and Externalizing/Rarity, drawing content related to fear and anxiety symptoms. At the four-factor level, the content of the Fear factor changed slightly to also include scales related to experiences of psychosis, such as Dissociation (.47) and Delusions (.37). In addition, a factor of Domineering, Anankastia, and Body Focus emerged primarily from Fear, covering content related to Body Focus (.57), Entitlement (.50), Exhibitionism (.48), Perfectionism (.48), Checking (.47), and Body Dissatisfaction (.46), with notable correlations with the Externalizing/Rarity and Distress/Detachment factors at the previous level. At the fifth level, Sexual Problems emerged, capturing scales such as Low Sexual Arousal (.99), Low Sexual Interest (.79), and Sexual Distress (.69), also with correlations with Distress/Detachment, Fear, and Domineering/Anankastia/Body Focus factors from level four.

At the six-factor level, the Externalizing/Rarity factor split into distinct Externalizing and Rarity factors, with the Externalizing factor (also emerging to a lesser extent from Fear and Distress) indicated most strongly by Callousness (.68), Social Aggression (.57), and Angry Hostility (.53) scales. The Rarity factor also emerged with notable, but smaller, contribution from Fear from the fifth level, drawing content related to Dissociation (.60), Conversion Symptoms (.59), and Appetite Loss (.46). Furthermore, the content of the Domineering/Anankastia/Body Focus factor from level five changed at the sixth level, with weakened loadings from anankastic scales (Hyperdeliberation [.30], Rigidity [.19], Risk Aversion [.04]) and a narrower focus on domineering, grandiosity, and body-focused symptoms, hence renaming the factor to Domineering/Body Focus. The seventh level gave rise to a Disorganization factor, drawing content primarily from the Distress factor with some contribution from the Rarity and Externalizing factors at the previous level, with Non-Persistence (.67), Cognitive Problems (.66), Disorganization (.66), and Checking (.50) as the highest loadings. At this level, the content of the Externalizing factor narrowed to focus exclusively on antagonistic scales, such as Callousness (.65), Social Aggression (.58), and Cynicism (.50); therefore we renamed this factor to Antagonism from this point forward.

At the eighth level of the hierarchy, Perfectionism/Eating Problems emerged from Domineering/Eating/Body Focus at the previous level, covering content from Dietary Restraint (.66), Muscle Building (.61), Body Focus (.60), and Perfectionism (.41). A Risk Taking factor also arose from Fear and Rarity factors, tapping into experiences of Risk Taking (.55), Risky Sex (.46), and Sex-Related Substance Use (.42), as well as negative loadings from anankastia symptoms, such as Risk Aversion (−.59) and Hyperdeliberation (−.43). In addition, the Rarity factor from level seven shifted focus to capture antagonistic detachment experiences, with scales such as Restricted Affectivity (.58), Callousness (.56), Oppositionality (.43), and Social Aloofness (.40). This factor also covered some psychosis-specific content, though with less salient loadings, including Eccentricity (.44) and Hallucinations (.32), as well as secondary loadings from Dissociation (.34) and Fantasy Proneness (.32). Consequently, we renamed this factor to Detachment/Psychosis at level eight. At the ninth and final level of the hierarchy, the Domineering factor, which was dissolved at the previous level, re-emerged with strong loadings from Grandiosity (.72), Well-Being (.62), Exhibitionism (.55), and Domineering (.48), hence the renaming to Grandiosity/Domineering. The Perfectionism/Eating Problems factor from level eight narrowed in focus to cover primarily eating and body-related content, resulting in a specific factor of Eating Problems. In addition, Detachment/Psychosis broadened at this level to capture a wide array of intense or rare experiences (more details below), justifying the renaming to Rarity/Psychosis. Finally, a new factor of Inhibitory Overcontrol emerged from Risk Taking at the previous level, with some contribution from the Disorganization factor, capturing experiences related to constraint and overcontrol.

Considering the pattern of emergence of individual factors as well as interpretability and theory, we chose the nine-factor solution as it reflected an adequate level of differentiation between the factors and most closely resembled the original HiTOP model. The model results from the nine-factor solution are reported in Table 2. The results from EFAs 1–8 are reported in Supplemental Tables 4–11 (S4–11).

Table 2.

Standardized Factor Loadings for the Nine-Factor Solution of HiTOP-SR Scales

HiTOP-SR Scales	Factor Loadings
	Sexual Problems	Distress	Disorganization	Grandiosity/Domineering	Eating Problems	Inhibitory Overcontrol	Fear	Psychosis/Rarity	Antagonism
Low Sexual Arousal	.98	−.09	−.03	.00	−.04	−.03	−.02	−.06	.00
Low Sexual Interest	.69	−.09	−.06	−.21	.03	.17	−.01	−.10	.08
Difficulties Reaching
Orgasm	.69	.08	.07	.15	.00	−.12	−.06	−.01	−.10
Sexual Distress	.64	.13	−.02	.05	.05	.03	−.07	.02	−.06
Sexual Pain	.52	−.03	−.08	.02	.04	.02	.17	.18	−.10
Sex-Related Substance Use	.30	.11	.07	.13	−.04	−.28	.08	.23	.05
Depressed Mood	.12	.66	.09	−.24	.10	.08	−.04	.03	.11
Anxious Worry	.07	.53	.15	−.03	.04	.26	.15	.03	−.06
Shame/Guilt	.14	.47	.19	−.20	.25	.12	−.06	.04	.05
Suicidality	.02	.46	.01	−.06	.14	−.06	.08	.24	.05
Lassitude	.14	.42	.33	−.17	.09	.24	−.05	−.05	−.06
Irritability	.08	.41	.11	−.05	.06	.16	.16	−.12	.32
Anhedonia	.18	.39	.10	−.35	.08	.18	−.10	.07	.21
Affective Lability	.05	.34	.17	.06	.08	−.01	.20	.12	.25
Bodily Distress	.15	.33	.13	−.08	.27	.17	.28	.06	−.07
Disorganization	.08	.09	.69	−.05	−.03	−.03	.03	.08	−.01
Non-Persistence	.09	.11	.64	−.13	.01	.17	−.08	.06	.09
Hoarding	.06	−.03	.57	.01	.23	−.03	.21	−.15	.06
Cognitive Problems	.10	.30	.51	.00	−.03	.21	.01	.17	.02
Non-Planfulness	.01	.10	.47	.07	.04	−.18	.08	.14	.29
Problematic Shopping	.06	−.02	.45	.03	.28	−.10	.21	−.27	.05
Checking	.08	.12	.36	.11	.09	.34	.09	.21	−.01
Grandiosity	−.09	−.17	−.10	.72	.00	.10	.00	.03	.02
Well-being	−.08	−.36	−.05	.62	−.04	.11	−.06	.03	−.10
Exhibitionism	−.06	.20	.11	.55	.24	−.14	−.03	.01	.14
Domineering	.10	.04	.01	.48	.17	.08	.02	−.12	.26
Social Aloofness	.16	.09	.04	−.37	.04	.34	−.01	.03	.30
Dietary Restraint	.01	−.13	−.16	.06	.65	.13	−.08	.01	−.07
Body Dissatisfaction	.06	.08	.12	−.22	.64	.07	.06	−.23	−.01
Muscle Building	−.06	−.06	−.18	.22	.56	.00	−.16	.05	.04
Body Focus	−.06	.14	.07	.38	.49	.03	.02	−.01	−.03
Purging	.10	−.05	−.05	−.08	.46	−.14	.10	.10	−.05
Excessive Exercise	.07	−.01	−.22	.11	.40	−.10	.02	.27	.05
Binge Eating	.09	.01	.19	−.14	.39	−.02	.04	−.18	.08
Hyperdeliberation	.01	−.04	−.14	.27	.05	.66	−.09	.06	−.05
Risk Aversion	.03	−.01	−.02	−.01	.04	.60	.22	−.18	.01
Social Anxiety	.06	.03	.23	−.28	.05	.34	.10	.14	.12
Somatic Preoccupation	.06	.14	−.10	.14	.30	.33	.18	.04	.04
Perfectionism	.04	.11	.08	.20	.27	.32	−.08	.14	−.05
Situational Phobias	−.01	−.09	.06	−.02	.02	.15	.63	.00	.03
Agoraphobia	.07	.03	.01	−.14	.02	.12	.55	.14	.05
Animal-Insect Phobia	.06	−.14	.06	.02	.12	.10	.49	−.09	.09
Health Anxiety	.09	.21	.06	.03	.19	.12	.43	−.04	−.04
Panic	.11	.32	.07	.01	.08	.08	.43	.23	−.10
Disease Conviction	.07	.10	.00	−.10	.09	.14	.40	.11	.08
Hypervigilance	.08	.29	.05	.11	−.01	.32	.39	.06	.13
Counting	.12	−.12	.15	.11	.08	.12	.32	.21	.05
Blood-Injection Phobia	.00	−.04	.08	.02	.03	.10	.32	−.01	.00
Trauma Reactions	.10	.22	.02	−.03	.08	.13	.31	.28	.12
Dissociation	.03	.12	.12	−.06	.02	.05	.26	.52	.09
Conversion Symptoms	.10	−.04	−.10	−.11	.15	−.15	.36	.43	.03
Appetite Loss	.03	.21	.09	−.03	.15	.03	.09	.42	−.08
Hallucinations	.04	−.01	−.06	−.01	−.01	−.03	.19	.40	.19
Eccentricity	.05	−.02	.27	.11	.01	.07	.02	.40	.31
Restricted Eating	.06	.14	.10	.00	.29	−.04	.05	.38	−.04
Manic Energy	−.03	−.07	.15	.24	.03	.03	.12	.37	.18
NSSI	.08	.23	.04	−.02	.18	−.13	.07	.37	−.05
Risk Taking	.03	.09	.13	.28	.06	−.28	−.18	.37	.22
Restlessness	.05	−.09	.35	.01	.09	.11	.07	.36	.05
Fantasy Proneness	.04	.14	.23	.03	.12	.21	.00	.36	.12
Nightmares	.10	.17	.12	.04	−.01	.12	.27	.34	−.10
Callousness	.07	−.03	.00	−.13	.07	.05	−.02	.13	.65
Social Aggression	.04	−.02	.12	.09	.08	−.13	.19	−.01	.51
Cynicism	.12	.20	.00	.05	−.01	.18	.25	−.03	.47
Angry Hostility	.03	.35	.02	.02	−.02	.00	.28	−.13	.46
Oppositionality	.02	.04	.15	.07	.00	−.04	−.04	.32	.43
Entitlement	.07	.09	.22	.28	.18	.06	−.06	−.12	.43
Suspiciousness	.12	.20	.01	.05	−.01	.03	.39	.02	.40
Rigidity	.11	.07	.24	.08	.02	.27	.02	−.04	.40
Manipulativeness	.02	.04	.07	.14	.15	−.15	−.03	.26	.38
Romantic Disinterest	.14	−.02	−.12	−.29	.14	.12	−.01	.01	.37
Restricted Affectivity	.15	−.02	.03	−.28	.19	.15	−.16	.30	.36
Deceitfulness	.08	.05	.28	.12	.21	−.09	.07	.05	.26
Gaming	−.03	−.17	.13	−.11	.11	.05	−.04	.16	.21
Antisocial Behavior	−.03	−.01	−.02	−.08	.18	−.27	.18	.15	.21
Paraphilias	.13	−.04	−.03	.12	.08	−.17	.08	.18	.17
Premature Orgasm	.12	−.03	.00	.09	.01	−.02	.04	.15	.15
Delusions	−.01	−.13	.05	.06	.03	.04	.23	.28	.14
Gambling	.03	−.10	.09	−.04	−.02	−.14	.13	.06	.11
Risky Sex	.03	.23	.03	.17	.08	−.29	.08	.23	.06
Insomnia	.14	.21	.13	−.14	.15	.05	.15	.12	.05
Trichotillomania	.03	.03	.15	.01	.11	.04	.04	.07	.00
Workaholism	.13	−.08	.05	.19	.14	.16	−.07	.14	−.05
Harmful Alcohol Use	.03	.17	.10	.26	−.02	−.20	−.10	.06	−.10
Excoriation	−.02	−.04	.28	−.01	.14	.10	.13	.15	−.22
Food Selectivity	.00	−.16	.23	−.06	.20	.16	.28	.11	.04
Submissiveness	.14	.05	.17	−.23	.00	.22	.03	.15	.02

Note. Loadings ≥ |.30| are bolded.

Nine-Factor Model of the HiTOP-SR: Associations between Factor-analysis Derived HiTOP-SR Domain Estimates and Facet Scales

As reported in Figure 1 and Table 2, the nine-factor model included factors that could reasonably be labeled, from left to right, Sexual Problems, Distress, Disorganization, Grandiosity/Domineering, Eating Problems, Inhibitory Overcontrol, Fear, Rarity/Psychosis, and Antagonism. The domain names were informed by the content of the items and the hypothesized domains in the HiTOP model. The first factor in Figure 1, Sexual Problems, captured content related to Low Sexual Arousal (.98), Low Sexual Interest (.69), and Difficulty Reaching Orgasm (.69). Distress, the second factor, had its strongest loadings from Depressed Mood (.66), Anxious Worry (.53), Shame/Guilt (.47), and Suicidality (.46). The next factor, Disorganization, covered content from Disorganization (.69), Non-Persistence (.64), Hoarding (.57), and Cognitive Problems (.51) scales. The strongest loadings for the fourth factor in Figure 1, Grandiosity/Domineering, came from Grandiosity (.72), Well-Being (.62), Exhibitionism (.55), and Domineering (.48) scales, along with a negative loading from the Social Aloofness scale (−.37). The fifth factor, Eating Problems, tapped into content primarily related to Dietary Restraint (.65), Body Dissatisfaction (.64), and Muscle Building (.56) scales. The Inhibitory Overcontrol factor drew its strongest loadings from Hyperdeliberation (.66) and Risk Aversion (.60), in addition to weaker loadings from Social Anxiety (.34), Somatic Preoccupation (.33), and Perfectionism (.32). Fear, the seventh factor, had the highest contributions from Situational Phobias (.63), Agoraphobia (.55), and Animal-Insect Phobias (.49). The eighth factor, Rarity/Psychosis, was marked by content primarily from scales representing psychosis-like experiences, such as Dissociation (.52), Eccentricity (.40), Hallucinations (.40), and Fantasy Proneness (.36), as well as other rare and/or extreme experiences, such as Conversion Symptoms (.43), Appetite Loss (.42), Manic Energy (.37), NSSI (.37), Risk Taking (.37), and Restlessness (.36). Finally, Antagonism, the ninth factor, was marked by Callousness (.65), Social Aggression (.51), Cynicism (.47), and Angry Hostility (.46).

Table 3 shows intercorrelations between the nine factors with modest to moderate effects, except for the associations of Grandiosity/Domineering with Disorganization (r = −.07), Eating Problems (r = .08), Fear (r = −.05), Rarity/Psychosis (r = .07), and Antagonism (r = −.01), in addition to the correlation between Rarity/Psychosis and Inhibitory Overcontrol (r = .03), which were all small.

Table 3.

Factor Intercorrelations for Nine-Factor Model

	I	II	III	IV	V	VI	VII	VIII	IX
I. Sexual Problems	–
II. Distress	.31	–
III. Disorganization	.25	.43	–
IV. Grandiosity/Domineering	−.18	−.19	−.07	–
V. Eating Problems	.25	.28	.27	.08	–
VI. Inhibitory Overcontrol	.22	.23	.21	−.13	.19	–
VII. Fear	.22	.28	.28	−.05	.24	.25	–
VIII. Rarity/Psychosis	.18	.22	.25	.07	.21	.03	.26	–
IX. Antagonism	.23	.27	.32	−.01	.24	.11	.25	.26	–

Note. Effect sizes ≥ |.30| are bolded.

HiTOP-SR and Daily Outcomes

Correlations between HiTOP-SR factor-analysis derived domains (i.e., factor score estimates for the nine-factor model) and daily outcomes are reported in Table 4. Zero-order correlations among HiTOP-SR scales and daily outcomes are reported in Table S13 and zero-order correlations between different daily and social interaction outcomes are reported in Table S14 in Supplemental Materials.

Table 4.

Correlations between Factor-analysis Derived HiTOP-SR Domain Estimates and Daily Experience/Behavior and Social Interaction Outcomes

HiTOP-SR Domains (Nine-Factor Solution)	Daily Experience/Behavior Outcomes						Social Interaction Outcomes
	Negative affect	Positive affect	Energy	Stress	Impulsivity	Sleep quality	Empathy	Dominance - Submissiveness	Agreeable - Quarrelsome
Sexual Problems	.24 ***	−.19 ***	−.15 ***	.26 ***	.13 **	−.21 ***	−.07	−.11 *	.09 *
Distress	.48 ***	−.39 ***	−.31 ***	.42 ***	.11 *	−.32 ***	−.03	−.06	.05
Disorganization	.22 ***	−.21 ***	−.22 ***	.25 ***	.33 ***	−.20 ***	−.03	−.02	.11 *
Grandiosity/Domineering	−.24 *	.39 ***	.34 ***	−.18 *	.07	.28 ***	.12 *	.16 **	.24 ***
Eating Problems	.11 *	−.09	−.12 *	.09	.15 ***	−.17 **	−.10 *	.02	.03
Inhibitory Overcontrol	.17 ***	−.15 **	−.15 **	.22 ***	−.05	−.10	−.02	.10	.21 ***
Fear	.19 **	−.14 **	−.15 **	.20 ***	.05	−.29 ***	.05	−.17 **	−.14 **
Rarity/Psychosis	.10 *	−.12 *	−.07	.14 *	.15 *	−.14 *	−.08	−.09	.04
Antagonism	.27 ***	−.21 ***	−.08	.18 ***	.27 ***	−.17 ***	−.26 ***	.06	−.09

Note. Highlighted cells indicate a finding in support of hypotheses. We did not anticipate the emergence of Sexual Problems, Grandiosity/Domineering, Eating Problems, and Inhibitory Overcontrol domains and therefore did not have separate hypotheses for these domains. Additionally, we did not have a specific hypothesis for ten out of fifty-four total predicted associations in Table 1 and have not highlighted the cells for these associations.

^*p≤ .05

^**p≤ .01

^***p≤ .001.

As reported in Table 4, the pattern of associations between daily outcomes and factor-analysis derived HiTOP-SR domains was largely in line with the existing literature. In terms of effect size, daily experience and behavior outcomes were moderately correlated with most HiTOP-SR domain estimates (average r = |.20|), with weaker associations between social interaction outcomes and factor-analysis derived domains (average r = |.09|).

Discussion

This study sought to examine the hierarchical structure of the HiTOP-SR scales, a new measure developed by the HiTOP Measures Development Workgroup (Simms et al., 2023) aiming to provide a comprehensive assessment of the symptoms covered by the HiTOP model (Kotov et al., 2017). In a series of exploratory factor analyses in a clinical/community sample of 637 participants, we estimated models with one to nine factors following parallel analysis results. Examining the pattern of factor loadings, interpretability, and the hierarchical structure of the nine models, we decided that the nine-factor solution most closely resembled the current structure of the HiTOP and provided adequate differentiation between conceptually distinct domains. We then aimed to validate the HiTOP-SR facet scale scores and factor-analysis derived domain estimates by testing their association with daily life functioning. Findings overall suggested significant associations between HiTOP-SR factor-analysis derived domain estimates and daily life outcomes, such as positive and negative affect, stress, impulsivity, sleep quality, and social interaction outcomes, pointing to the validity and real-life relevance of the HiTOP-SR measure.

The nine-factor model of the HiTOP-SR is largely consistent with the original HiTOP model (Kotov et al., 2017), with broadband Internalizing and Externalizing spectra. Of the nine factors, five map quite cleanly onto the current HiTOP structure––Sexual problems, Distress, Eating Problems, Fear, and Antagonism. There are, however, several key deviations. The (mis)alignment of our nine-factor solution with the current HiTOP structure is depicted in Figure 2. The subfactor labeled Mania by Kotov and colleagues is best represented by the Rarity/Psychosis factor in our model, which captures Manic Energy in addition to other extreme symptoms broadly in line with experiences of psychosis, such as Dissociation, Eccentricity, Hallucinations, and Fantasy Proneness. It is noteworthy that in models with 6–7 factors, the Rarity domain captured a wider range of extreme but conceptually distinct symptom scales, spanning psychosis, mania, antagonism, detachment, eating pathology, somatoform, and distress/fear. At level eight, however, it primarily borrowed content from Antagonism and narrowed down significantly to experiences of antagonism, detachment, and, less markedly, psychosis, before broadening slightly in the nine-factor solution to cover mania, psychosis, and detachment symptoms. Even at the ninth level, this factor was still marked by heterogenous symptoms outside the psychosis cluster, including Fear (Nightmares [.34]), Distress (NSSI [.37]), Externalizing (Risk Taking [.37], Oppositionality [.32]), Eating Pathology (Appetite Loss [.42], Restricted Eating [.38]), and Somatoform (Conversion Symptoms [.43]), hence the name “Rarity”. This factor did not seem as theoretically coherent and may have emerged artificially because of response outliers. We attempted to address this issue by capping all responses at 6 standard deviations. This, however, ultimately only weakened but did not remove the Rarity factor. One possibility for the emergence of this factor in our study is that the measure was designed to assess past-year symptoms, thus included many low base rate experiences, but we asked about experiences in the past month. This narrower timescale may have resulted in certain experiences appearing especially “rare.” Consistent with this possibility, factor analysis of the German version of the measure using the past-year time frame did not find a rarity factor (Zimmerman et al., 2024).

Figure 2.

Graphical representation of the (mis)alignment between the factor-analysis derived HiTOP-SR domains (bottom) and the current HiTOP structure (top).

Note. The color-coded numbers under each factor-analysis derived domain represent the mapping onto the corresponding HiTOP subfactor/spectrum.

We did not observe a distinct Somatoform factor, which is another deviation from the original HiTOP model. Instead, most somatoform scales showed either a primary or secondary loading on the Fear factor—Health Anxiety (.43), Disease Conviction (.40), and Conversion Symptoms cross-loaded on Fear (.36) and Rarity/Psychosis (.43). Other somatoform scales were distributed across factors besides Fear––Bodily Distress loaded on Distress (.33), and Somatic Preoccupation cross-loaded on Inhibitory Overcontrol (.33) and Eating Problems (.30). The absence of a specific Somatoform factor in our study raises questions about the placement of this factor in the current HiTOP model. Indeed, it currently remains tentative as a separate spectrum pending more data that might speak to its placement. Our results suggest that a Somatoform psychopathology spectrum may not be warranted. Similarly, a distinct Detachment factor did not emerge in our data; rather, detachment scales were distributed across different factors. Well-being cross-loaded on Grandiosity/Domineering (.62) and Distress (−.36), Romantic Disinterest was captured by Antagonism (.37), Restricted Affectivity cross-loaded on Antagonism (.36) and Rarity/Psychosis (.30), Social Anxiety had a primary loading on Inhibitory Overcontrol (.36), while Social Aloofness was spread across Grandiosity/Domineering (−.37), Inhibitory Overcontrol (.34), and Antagonism (.30). The absence of a cohesive Detachment factor in our data supports its overlap with other domains of psychopathology, such as distress and antagonism (Clark et al., 2021). Further investigation is warranted to clearly distinguish Detachment in future iterations of the HiTOP-SR. On the other hand, in line with the original HiTOP structure, Sexual Problems and Eating Problems emerged as distinct factors in our data. While this replication supports the robustness of these dimensions, it also raises a challenging question: whether it is conceptually meaningful and clinically useful to place such narrow, content-specific factors at the same hierarchical level as broader, higher-order factors like Fear and Distress. This discrepancy in breadth is a critical issue for the model’s coherence and utility that must be tackled as the HiTOP Consortium works on clinical translations of the model. Additionally, it is noteworthy that the current HiTOP-SR includes relatively few scales for such broad spectra as Detachment and Psychosis (six and seven scales, respectively), whereas narrower domains like Sexual Problems and Eating Problems each contain nine and ten scales. This imbalance in domain coverage for key spectra relative to narrower domains partially explains the emergence of distinct factors for Eating Problems and Sexual Problems in our analyses, despite their circumscribed focus, while indicators for broader spectra like Psychosis and Detachment were distributed across other factors. The proportional distribution of scales per domain is an important issue that must be addressed as the HiTOP-SR undergoes structural validation and echoes discussions about discrepancy in domain breadth, as noted above.

Another departure from the HiTOP is the emergence of the Grandiosity/Domineering factor in our model. This factor was primarily marked by scales representing a grandiose view of oneself, including Grandiosity, Well-being, Domineering, Exhibitionism, and a secondary loading from Body Focus. In this sense, it is perhaps closest to the Antagonism factor in the HiTOP representation, at least in theory, though more blue-ribbon antagonism scales, such as Callousness and Social Aggression, loaded relatively weakly onto this factor (−.13 and .09 respectively). In addition, we observed that the Grandiosity/Domineering factor included social connection, as suggested by Exhibitionism’s strong loading (.55) and Social Aloofness’ moderate and negative loading (−.37). One question worth exploring further about the Grandiosity scale is whether–and to what extent–it assesses truly maladaptive levels of grandiose self-perception, as opposed to healthy self-confidence and assuredness. In terms of face validity, some of the items do not necessarily capture true psychopathology, such as “I was very confident in myself around other people.” and “I enjoyed thinking about my future and the great things it would bring.” This question is especially relevant in light of the factor’s association with less negative daily experiences, such as negative affect and stress, and more positive experiences, like positive affect and energy (Table 4). That said, because these findings are entirely based on self-report data, individuals with grandiose and narcissistic tendencies may overestimate their well-being, downplay problems, or present themselves in an overly positive light, making interpretation of these associations increasingly difficult.

Furthermore, we observed a distinct Disorganization factor, capturing content related to both disinhibition, specifically Non-Perseverance and Non-Planfulness, as well as experiences that fall under the fear and distress domains of psychopathology, including Hoarding, Cognitive Problems, Checking, and Lassitude. This factor reflects experiences that are associated with behavioral disorganization and therefore taps into both internalizing and externalizing psychopathology. The Disorganization factor is perhaps closest to the Disinhibited Externalizing spectrum in the current HiTOP model, though it is clearly missing constructs related to risky behaviors. In the current model, the Risk Taking scale is captured by Rarity/Psychosis (.37), perhaps due to its extreme and/or rare nature, while other risky behavior scales, including Harmful Alcohol Use, Risky Sex, Gambling, and Gaming do not show a privileged relationship to any one factor. Although we did not observe a separate Risk Taking factor, a distinct Inhibitory Overcontrol factor did emerge, in some ways reflecting the opposite of the Risk Taking domain, primarily capturing Hyperdeliberation, Risk Aversion, and Perfectionism. This factor is also marked by a theme of detachment and fear, capturing content related to Social Anxiety, Hypervigilance, and Checking. We refrained from identifying this factor as Anankastia due to the marked detachment/fear content and the absence of other key anankastic symptoms, such as Workaholism and Rigidity. It is therefore worth considering how to best capture and represent the full Disinhibited Externalizing domain, including risky behaviors, and whether a distinct overcontrol domain is warranted as the HiTOP model undergoes revisions.

Many of the deviations from the HiTOP model observed in our data, such as the Rarity/Psychosis, Grandiosity/Domineering, Inhibitory Overcontrol, and Disorganization factors, as well as the emergence of narrower domains like Sexual Problems and Eating Problems that align with current HiTOP subfactors, may be partly due to these constructs being historically understudied. This is likely related to their lower base rates in the general population and the elevated risks associated with some of these symptoms (e.g., mania, psychosis), which pose challenges for participant recruitment. Considering that the original HiTOP framework was based on the available literature, it is unsurprising that the lesser studied constructs would be more tentative and less clearly defined in the model. An advantage of a comprehensive measure like the HiTOP-SR is that it allows researchers to fully assess a wide range of symptoms and experiences, thereby better informing our understanding of psychopathology, including the less common or understudied experiences of mental illness.

Although our results support the structural validity of the HiTOP-SR, it is worth highlighting that a number of HiTOP-SR scales did not demonstrate a privileged relationship with a single factor relative to others and therefore lacked a primary loading on any one factor. The reasons varied. Some scales, such as Deceitfulness, loaded moderately on several theoretically relevant factors, including Disorganization (.28), Antagonism (.26), and Eating Problems (.21), but none met our threshold of .30 for primary loading. In other cases, such as Trichotillomania, the lack of a primary loading may be due to the absence of other conceptually related scales in the measure (e.g., scales assessing tic disorders). Further investigation into the performance and validity of these “orphan” scales may be a valuable direction for future research on the HiTOP-SR. Additionally, the presence of some of these “orphan” scales, particularly those with a narrower focus, potentially poses a challenge for the HiTOP model. On one hand, the inclusion of these scales seems important in a truly comprehensive model of psychopathology; on the other, many of these narrower “orphan” scales likely will not fit well within the broader model—at least in its current form—and present a structural problem.

To examine the construct and content validity of the HiTOP-SR domain estimates, we explored their associations with measures of daily experiences and social interaction outcomes. This approach allowed us to assess whether and to what extent the factor-analysis derived HiTOP-SR domains relate to external outcomes in theoretically consistent ways. Overall, the results supported construct and content validity. For example, most HiTOP-SR domain estimates were significantly and positively associated with negative daily experiences, including negative affect and stress, with moderate effect sizes (average r = |.22|), as one would expect for different forms of psychopathology. The strongest association for both daily negative affect and stress was, indeed, with the Distress domain (rs = .48 and .42 respectively). It should be noted, however, that momentary negative affect and stress were highly correlated (r = .83) as reported in Table S14, indicating substantial shared variance among these constructs. In addition, daily impulsivity was most strongly associated with Disorganization (r = .33), and empathy had its strongest link with Antagonism (r = −.26), in line with each domain’s thematic focus.

Supporting construct validity, HiTOP-SR domain estimates showed significant negative associations with positive daily experiences (e.g., positive affect, energy, sleep quality). These patterns suggest that the HiTOP-SR scales capture constructs closely tied to psychopathology in daily life (e.g., stress, negative affect), as expected. Further informing discriminant validity, the Grandiosity/Domineering domain was a clear exception to these patterns, as it showed the reverse pattern, with positive associations with positive experiences and negative associations with negative experiences given its focus on a grandiose and positive self-perception, as discussed earlier. Taken together, these cross-method associations offer promising support for the construct and content validity of the HiTOP-SR domains and highlight the measure’s potential utility in capturing individual differences in day-to-day functioning.

As shown in Tables 1 and 4, our findings offered mixed support for the predicted associations between HiTOP-SR domains and daily functioning. We did not have specific hypotheses for ten out of fifty-four predicted associations, as we did not expect every HiTOP domain to have strong associations with every EMA measure, which in fact supports discriminant validity. In addition, we did not anticipate the emergence of several factors, including Sexual Problems, Grandiosity/Domineering, Eating Problems, and Inhibitory Overcontrol, and therefore did not have specific hypotheses for these domains. Of our predicted associations in Table 1, perhaps those with social interaction outcomes were the least supported, with support for three of ten hypothesized associations. Additionally, we observed overall fewer significant associations between social interaction outcomes and the HiTOP-SR domain estimates as well as weaker effect sizes compared to the associations with daily behaviors and experiences. This inconsistency may be partially attributed to the event-contingent nature of social interaction surveys, such that the higher relevance of contextual factors for the social interaction prompts reduces the influence of individual differences of the participants. These contextual nuances may introduce additional variability that may obscure or weaken the associations between social interaction outcomes and experiences of psychopathology.

Constraints on Generality

Our analyses were conducted in a large mixed clinical/community sample (N = 637, 75.4% female, 80.5% white, M age 30.7), enhancing the generalizability of our findings to the general population (as opposed to a particular group, e.g., college students). On the other hand, one limitation of this study was that the sample was predominantly white and female, which is a concern for generalizability. Given the scope of this project, we did not conduct invariance analyses with demographic variables, though this could be a fruitful direction for future work focusing on the HiTOP-SR. Furthermore, since all participants were English-speaking residents of the United States, we were not able to examine the role of culture in our analyses. This is an important direction that is already being explored in several projects on the translation of the HiTOP-SR (e.g., Zimmermann et al., 2024).

Implications, Limitations, and Future Directions

This study, to our knowledge, is the first attempt to examine the structure of the English version of the HiTOP-SR scales and validate the measure with daily life functioning, yet it has several important limitations. The version of the HiTOP-SR used in the present study asked about experiences in the past month. This differs from the official measure published by Simms and colleagues (2023), which focuses on experiences over the past year. Although this methodological difference may be a limitation in some instances and constrain the comparison of our findings with those of future HiTOP-SR studies focusing on the past year, it is also a strength. The logic behind changing the instructions from “the past year” to “the past month” was increasing the clinical relevance of our findings, because in clinical settings, clinicians often care most about current or recent experiences of psychopathology, which may be better captured by focusing on the past month rather than the past year. Furthermore, one issue with the HiTOP model more generally is that it conflates states and traits in the same domains (DeYoung et al., 2022), and the switch from focusing on the past year to the past month is an effort to emphasize symptomatic expression as opposed to more trait-like manifestations of psychopathology.

Another limitation of this study is that a few of the HiTOP-SR scale scores had McDonald’s omega total (ω_t; Zinbarg, Revelle, Yovel, & Li, 2005) reliability values below .70 (Blood-Injection Phobia: .67, Hallucinations: .69, Premature Orgasm: .59, Purging: .60, Social Aggression: .65). It is possible that changing the measure’s standard timeframe to the past month may have affected the scale scores’ reliability. Another reason could be the nature of our sample enriched for experiences of psychopathology (indexed by self-reported mental health treatment), as we might expect slightly lower internal consistency estimates than in a general sample. Nevertheless, the findings for these scales should be interpreted with this limitation in mind and future work should further explore the reliability of these HiTOP-SR scale scores. Additionally, of the total 495 items included in this version of the HiTOP-SR, only one item was reverse-keyed. Although this is common in clinical self-report measures, it may raise concerns about “yeasaying” or acquiescence bias (Block, 1965), which may have implications for the validity and utility of these scales.

Furthermore, although our findings held across different methodologies (baseline and EMA measures), which is a significant strength, all measures were self-report, which can offer a limited and, at times biased, perspective. Future research could explore the validity of the HiTOP-SR using behavioral measures and passive sensing technologies (e.g., Ringwald et al., in press). Finally, it is important to keep in mind that our EMA measures were relatively broad. While such measures may be useful in understanding the associations with daily life functioning in a broad and general manner, they do not allow for ascertaining more nuanced relationships between closely-related constructs. It would be worthwhile for future research to explore the validity of the HiTOP-SR scores using more specific constructs of daily life functioning.

Conclusion

This study is the first to explore the structure of the English version of the HiTOP-SR facet scales and how they relate to individual differences in behavior expression and the experience of emotions in daily life. In a preregistered mixed clinical and community sample, we partially replicated the structure of the HiTOP model using the HiTOP-SR and found general validity for this novel measure based on associations with daily life functioning. Our findings have implications for the utility of the HiTOP-SR in clinical and research contexts and contribute to the large and growing body of work on the dimensional structure of psychopathology.

Public Significance Statement:

We explored the statistical properties of the HiTOP-SR, a new self-report measure for assessing different aspects of mental illness. We found that the HiTOP-SR is best represented by nine overall themes and that it is a valid measure for future use in research and clinical practice.