Introduction

Depression represents a spectrum of mood disturbances that substantially impacts global health. As one of the most prevalent chronic mental disorders worldwide, depressive disorders - particularly major depressive disorder (MDD) and dysthymia - account for 37.3% of the total disability-adjusted life years (DALYs) attributed to mental disorders1. This condition manifests through a range of symptoms including depressed mood, diminished interest, cognitive slowing, and disturbances in sleep and appetite, with severe cases potentially leading to suicidal ideation2. The World Health Organization identifies depression as the leading contributor to global disability, affecting over 300 million people worldwide3. Studies have shown that only a small percentage of people with depression receive effective treatment and achieve complete remission. More seriously, the majority of people with depression never seek professional help. These low rates of remission and seeking help reflect serious shortcomings in the prevention, diagnosis and treatment of depression4,5.

A key factor in the undertreatment and poor uptake of depression is the stigma attached to the condition by society and patients themselves6,7,8. The theoretical understanding of stigma was first systematically developed by Goffman9, who defined it as a “deeply discrediting attribute” that reduces an individual’s social acceptance. Building upon this foundation, Link and Phelan10 further conceptualized stigma as a process involving labeling, stereotyping, and social differentiation occurring within power relationships. These general stigma theories have been extensively applied to mental health research, particularly in understanding depression stigma. For instance, Corrigan and Watson11 demonstrated how Goffman’s framework explains the devaluation and social rejection faced by individuals with depression, while Angermeyer and Matschinger12 empirically validated Link and Phelan’s process model in depression contexts, revealing specific stereotypes of unpredictability and incompetence associated with depression. The application of these theoretical frameworks to depression has shown how stigmatization processes lead to reduced social acceptance and treatment-seeking behavior13. With the growing global awareness of mental health concerns, exploring strategies to reduce depression-related stigma has emerged as a critical research priority in both academia and public health, given its substantial impact on treatment-seeking behaviors and healthcare outcomes14.

In this context, the Continuum Beliefs Approach (CBA) has received increasing interest as an emerging anti-stigma strategy. This approach aims to reduce the formation and spread of stigmatizing attitudes by removing artificial boundaries between “us” and “them” and reducing the public’s psychological distance from people with mental disorders. It has been shown that continuity beliefs can be associated with a reduction in public stigma towards mental disorders15,16,17,18,19.

The measurement of stigma requires a comprehensive framework that captures its multiple facets. Building on established theoretical models11,20, stigma can be understood through three key dimensions: stereotypes, prejudice, and discrimination. Stereotypes represent the cognitive component, reflecting society’s characteristic negative beliefs about individuals with mental disorders21. Prejudice, manifesting as negative affective reactions, serves as a crucial mediator between stereotypical beliefs and discriminatory behavior22. Discrimination, as the behavioral component often measured through social distance acceptance, represents the actual tendency to exclude individuals with mental disorder from various social domains23. This integrated approach enables a more complete understanding of stigma’s manifestations and provides clear targets for intervention strategies.

Although some studies have explored the impact of continuum beliefs on public stigma towards depression, these studies have primarily focused on explicit attitudes. However, for sensitive topics such as mental health stigmatization, individuals may adjust their explicit responses due to social desirability or self-presentation concerns. In contrast, implicit attitudes often more accurately reflect individuals’ true thoughts, as they are less susceptible to social desirability bias. Nevertheless, Investigations on implicit public attitudes remains relatively scarce24. Given the limitations of existing study, the present study aims to comprehensively investigate the influence of continuum beliefs on both explicit and implicit attitudes towards public stigma of depression by combining explicit and implicit measurement methods. This approach not only reveals consciously controlled responses but also captures potential unconscious biases, thereby providing a more comprehensive understanding of the role of continuum beliefs in reducing depression stigmatization.

In order to achieve this aim, firstly, this study guided individuals to establish continuity beliefs and investigated individual attitudes using a scale to verify whether the guided continuity beliefs would have an impact on the public stigma of depression. We hypothesized that exposure to a continuum belief text would reduce both explicit and implicit public stigma attitudes towards depression across three key dimensions: cognitive components (negative stereotypes about depression), affective components (prejudicial affective reactions), and behavioral components (discriminatory tendencies measured through social distance acceptance). For explicit attitudes, we expected the continuum belief group to demonstrate lower negative stereotypes, prejudicial affective reactions and higher social distance acceptance compared to the control and categorical belief groups. Similarly, for implicit attitudes, as measured by the Single Category Implicit Association Test (SC-IAT), we anticipated that the continuum belief group would show weaker automatic negative associations across these same three dimensions. By examining both explicit and implicit attitudes, we aimed to capture a comprehensive picture of how continuum beliefs might influence different aspects of stigma at both conscious and unconscious levels.

Method

Participants

This study was approved by the Ethics Committee of Hunan Normal University (No. 344 of 2023) and carried out following the Declaration of Helsinki. The sample consisted entirely of Chinese participants who were recruited and completed the experiment via online platforms in December 2023, with electronic informed consent obtained from all participants. To control for potential carryover effects, we implemented a between-subjects experimental design across both studies.

Participants were recruited separately for Study 1 (explicit measurement experiment) and Study 2 (implicit measurement experiment). Within each study, participants were systematically assigned to one of three experimental conditions: (a) continuum beliefs group, (b) categorical beliefs group, or (c) control group.

According to G*Power (version 3.1.9.7) calculations, for the present study’s one-way three-level ANCOVA, the total sample size required to predict a statistical power of 80% at the significance level of α = 0.05 with a medium effect size of f = 0.25 was a minimum of 159 participants. A total of 366 participants were initially recruited for Study 1, and 277 participants were recruited for Study 2. To ensure data quality, we implemented strict exclusion criteria for all participants. General exclusion criteria for both experiments included: (1) participants who were currently diagnosed with depression or had a history of depression, (2) psychology majors, whose academic background might influence responses, and (3) participants who failed to fully watch the priming text materials. For Study 1, additional exclusion criteria included: (1) participants who selected the same response option for all items on the scale, and (2) those with response times below 30 s, indicating inattentive or rushed answers. For Study 2, participants with D-scores beyond ± 3 SD from the sample mean were excluded as statistical outliers. Following the application of predetermined exclusion criteria, the final sample comprised 323 participants in Study 1 (66.9% female; Mage = 25.68, SD = 7.07), and 252 participants in Study 2 (66.7% female; Mage = 26.19, SD = 6.99). The detailed demographic characteristics are presented in Tables 1 and 2.

Table 1 Explicit measurement experiment of experimental demographic data.
Full size table
Table 2 Implicit measurement experiment of experimental demographic data.
Full size table

Procedure and materials

Procedure

This present investigation comprised two parallel yet independent experimental designs: In Study 1 (explicit measurement experiment), the procedure was as follows: First, participants completed a demographic questionnaire assessing sociodemographic characteristics (gender, age, educational level) and depression-related experiences (indirect depression contact experience). Next, all participants read a vignette describing “Wang Hua,” a fictional character exhibiting prototypical symptoms of MDD, with diagnostic labels intentionally omitted. Following the vignette, participants were randomly allocated to one of three experimental conditions: (a) continuum beliefs group, (b) categorical beliefs group, or (c) control group. After exposure to the corresponding theoretical materials, the Continuum Beliefs Scale was administered as a manipulation check. Finally, for the assessment of explicit stigma toward depression, participants completed standardized measures evaluating three fundamental dimensions: (a) cognitive components (negative stereotypes), (b) affective components (prejudicial affective reactions), and (c) behavioral components (discriminatory tendencies (measured through social distance acceptance). In Study 2 (implicit measurement experiment), the steps prior to the manipulation check were identical to those in the explicit measurement experiment, including demographic information collection, reading the description of Wang Hua’s depressive symptoms, and random assignment to one of the three theoretical conditions with corresponding intervention materials. Instead of completing stigma-related questionnaires, participants in the implicit measurement experiment underwent the Single Category Implicit Association Test (SC-IAT), which assessed automatic, non-conscious attitudinal tendencies toward individuals with depression.

Upon completion of the survey, participants were debriefed about the nature of the study, thanked, and provided with appropriate compensation. They were also reminded that reducing stigmatization is a collective effort requiring ongoing commitment from society as a whole.

Materials

(1) Vignette Describing Symptoms of a Fictional Character: This study adopted a case description adapted from Buckwitz et al.25, using the gender-neutral name “Wang Hua” to replace “Shane” in the original material. The vignette depicted “Wang Hua” experiencing core symptoms of depression: persistent low mood, anhedonia, sleep problems, fatigue and cognitive impairment. Diagnostic labels (“depression” or “MDD”) were intentionally omitted to minimize potential labeling effects. This standardized presentation ensured that all participants received identical symptom information while controlling for potential stigma activation associated with diagnostic labels.

(2) Experimental Manipulation of Continuum and Categorical Beliefs: The manipulation materials were adapted from established paradigms of Buckwitz et al.25 and Schomerus et al.17. All three groups presented the results of a fictitious study in the form of short articles, with the following content: (a) The Continuum Beliefs group began with “Is there a clear boundary between depressed individuals and healthy individuals? No, it’s just a matter of degree,” emphasizing the prevalence of depressive symptoms in the general population and the continuous variation in their severity. (b) The Categorical Beliefs group began with “How do we distinguish the boundary between depression and health? There is a clear distinction between the two,” emphasizing the qualitative differences between depression and normal mood fluctuations, and the clear boundaries between healthy and affected populations. (c) The Control group provided factual descriptions about depressive disorders, including their classification, core symptoms, and characteristics, but did not contain any information conceptualizing depression as either continuum or categorical beliefs. The three manipulation materials maintained high consistency in word count, format, structure, and linguistic style, differing systematically only in their theoretical framework orientation.

(3) Continuum Beliefs. Participants rated the degree to which they believe depression exists on a continuum, using a 7-item scale from Schomerus et al.17. Respondents indicated their agreement with statements such as “Sometimes we are all at least a little like Wang Hua, it is only the question how pronounced this state is” and “To some extent, most persons will experience problems that are similar to those of Wang Hua” on a 7-point Likert scale from 1 (strongly disagree) to 7 (strongly agree). The scale showed good internal consistency; higher scores represented stronger continuum beliefs (α = 0.725).

(4) Explicit Measures of Stigma. Participants’ explicit attitudes toward stigma were assessed across three dimensions: cognitive components (negative stereotypes about depression), affective components (prejudicial affective reactions), and behavioral components (discriminatory tendencies measured through social distance acceptance). For cognitive components, participants rated their agreement with statements related to blame, dangerousness, and unpredictability, such as “Wang Hua is to blame for her problems” and “Wang Hua is unpredictable,” using a 7-item scale from Schomerus et al.17. This scale, which had an internal consistency of α = 0.65, indicated that higher scores corresponded to stronger public stigmatization. For the affective components, affective responses were assessed using a 10-item scale adapted from Angermeyer et al.26, assessing feelings such as fear, anger, and pro-social reactions, with statements like “I feel uncomfortable” and “I am amused.” This scale showed good internal consistency (α = 0.72), with higher scores reflecting stronger stigmatization. For behavioral components, social distance acceptance was assessed using a 5-item scale (α = 0.82) that measured participants’ willingness to interact with Wang Hua across different social situations (e.g., ‘How willing would you be to spend an evening socializing with Wang Hua?’). Higher scores indicate greater social acceptance and interaction willingness, reflecting lower levels of public stigmatization.

(5) Implicit Measures of Stigma. The Single Category Implicit Association Test (SC-IAT) was designed to assess implicit stigmas related to depression (Fig. 1). The experimental procedure consisted of two phases: a practice phase (24 trials) and a test phase (72 trials). Only data from the test phase were included in the analyses. Briefly, each complete trial consisted of a 200 milliseconds (ms) fixation point, followed by stimulus presentation that remained on screen until response, and ended with a 200 ms feedback display. The SC-IAT material was adapted from Wang et al.27, comprising the target concept “depression” and 24 attribute words that assessed cognitive appraisals, affective responses, and behavioral tendencies. The attribute words were evenly distributed across three categories, with each category containing four positive and four negative words. Examples included cognitive words (e.g., “dangerous,” “abnormal” vs. “competent,” “strong”), affective words (e.g., “disgust,” “nervous” vs. “joyful,” “pleasant”), and behavioral words (e.g., “avoid,” “reject” vs. “approach,” “care”). Following the principles of implicit association testing, all trial were divided into compatible and incompatible conditions. In the compatible condition, participants associated depression with negative attributes, while in the incompatible task, they associated depression with positive attributes. To control for response bias, stimulus presentation frequencies were carefully balanced: in the compatible condition, “depressive disorder,” “negative word,” and “positive word” appeared in a 1:1:2 ratio, while in the incompatible condition, the ratio was 2:1:1 for “negative word,” “positive word,” and “depressive disorder.” This counterbalanced design ensured an equal distribution of left and right key responses. The experimental procedure was coded in PsychoPy (version 2021.2.3).

Fig. 1
figure 1

Schematic figure of the single-category implicit association test (SC-IAT). Interstimulus interval was set at 1500ms. If subjects make an error, a red “x” is presented in the center, and they are required to correct their answer. All words were presented in Chinese.

Full size image

Data analysis

All data in this study were processed using R software (version 4.3.2), with demographic information (age, gender, educational level and indirect depression contact experience) included as a covariate in all analyses. First, we conducted a one-way analysis of covariance (ANCOVA) to assess the effectiveness of the “Continuum” and “Categorical” manipulations across different groups. Next, for explicit measures experiment, between-group differences in depression stigma dimensions were examined using analysis of covariance, followed by post-hoc multiple pairwise comparisons with Bonferroni correction. Finally, for the implicit measures experiment, we applied the reaction time preprocessing method proposed by Karpinski and Steinman28. Reaction times greater than 10,000 ms and less than 350 ms were excluded, and incorrect responses were replaced with the average correct reaction time of the corresponding block plus a penalty of 400 ms. We then calculated the difference between the average reaction times for incompatible and compatible condition, dividing this difference by the standard deviation of all reaction times across both tasks to obtain the implicit effect value (D-score). Higher D-score indicated stronger negative implicit Stigma toward depression. Between-group differences in D-scores were analyzed using one-way analysis of covariance (ANCOVA), followed by post-hoc pairwise comparisons with Bonferroni correction.

Results

Estimation of manipulation effectiveness

For the explicit measures experiment, ANCOVA with age, gender, educational level and indirect depression contact experience as covariates revealed significant differences between the three groups, F (2, 310) = 10.615, p < 0.001, ηp² = 0.064. Post hoc comparisons using Bonferroni correction showed significant differences between the continuum beliefs group (M = 4.777, SD = 1.274) and the categorical beliefs group (M = 3.817, SD = 1.523, p < 0.001), and between the continuum beliefs group and the control group (M = 3.987, SD = 1.588, p = 0.001). No significant difference was found between the categorical beliefs group and the control group (p = 0.894), indicating that the manipulation effectively elevated continuum beliefs in the continuum condition while the categorical condition did not significantly differ from baseline beliefs reflected in the control condition (Fig. 2).

Fig. 2
figure 2

Comparison of the effectiveness of continuum belief manipulation across groups. All analyses employed ANCOVA, adjusting for the covariates of age, gender, educational level, and indirect depression contact experience. Significance levels from post-hoc tests are indicated: * p < 0.05, ** p < 0.01, *** p < 0.001.

Full size image

For the implicit measures experiment, ANCOVA controlling for age, gender, educational level, and indirect depression contact experience revealed significant differences between the three groups, F (2, 240) = 5.683, p = 0.004, ηp² = 0.045. Post hoc comparisons with Bonferroni correction showed that participants in the continuum beliefs group (M = 4.726, SD = 1.340) scored significantly higher than those in the categorical beliefs group (M = 3.973, SD = 1.509, p = 0.005) and the control group (M = 4.123, SD = 1.567, p = 0.041). No significant difference was found between the categorical beliefs group and the control group (p = 1.000). These results suggest that the continuum intervention effectively enhanced implicit continuum beliefs compared to both categorical and control conditions, while the categorical condition did not significantly differ from baseline beliefs reflected in the control condition.

Public explicit stigma

With the three stigma dimensions as dependent variables: cognitive dimension (negative stereotypes), affective dimension (prejudicial affective reactions), and behavioral dimension (social distance acceptance), and gender, age, educational level and indirect depression contact experience as control variables, a one-way ANCOVA was conducted on the three groups, and the results revealed significant differences among the three groups, as shown in Table 3.

Table 3 Mean scores and group differences in explicit (Study 1) and implicit (Study 2) stigma measures.
Full size table

In particular, on cognitive dimension (negative stereotypes), the three groups differed significantly, F (2, 310) = 8.579, p < 0.001, ηp2 = 0.052. Post hoc comparisons using Bonferroni correction showed that the continuum beliefs group (M = 2.724, SD = 0.480) reported significantly lower negative stereotypes than both the categorical beliefs group (M = 2.991, SD = 0.508, p < 0.001) and the control group (M = 2.918, SD = 0.419, p = 0.018). No significant difference was found between the categorical beliefs group and the control group (p = 0.353) (Fig. 3).

Fig. 3
figure 3

Three-dimensional comparison of stigma across group. The top three panels display explicit stigma results from Study 1, while the bottom three panels present implicit stigma results from Study 2. All analyses employed ANCOVA, adjusting for the covariates of age, gender, educational level, and indirect depression contact experience. Significance levels from post-hoc tests are indicated: * p < 0.05, ** p < 0.01, *** p < 0.001.

Full size image

On affective dimension (prejudicial affective reactions), the three groups differed significantly, F (2, 310) = 3.434, p = 0.033, ηp² = 0.022. Post hoc comparisons using Bonferroni correction showed that the continuum beliefs group (M = 1.977, SD = 0.523) reported significantly lower affective reactions than the categorical beliefs group (M = 2.163, SD = 0.472, p = 0.033). However, no significant differences were found between the control group (M = 2.138, SD = 0.484) and either the categorical beliefs group (p = 1.000) or the continuum beliefs group (p = 0.178) (Fig. 3).

On behavioral dimension (social distance acceptance), the three groups differed significantly, F (2, 310) = 4.702, p = 0.010, ηp² = 0.029. Post hoc comparisons using Bonferroni correction showed that the continuum beliefs group (M = 3.481, SD = 0.796) reported significantly higher social distance acceptance than the categorical beliefs group (M = 3.180, SD = 0.769, p = 0.010). The control group (M = 3.376, SD = 0.686) did not differ significantly from either the categorical beliefs group (p = 0.088) or the continuum beliefs group (p = 0.980) (Fig. 3).

Public implicit stigma

One-way ANCOVAs were conducted with cognitive, affective and behavioral D-scores as dependent variables, controlling for age, gender, educational level, and indirect depression contact experience as the control variables, and the specific statistical results are shown in Table 3. The results showed no significant differences between the three groups in any of the implicit measures: cognitive (F (2, 240) = 1.425, p = 0.243, ηp² = 0.012), affective (F (2, 240) = 0.512, p = 0.600, ηp² = 0.004) and behavioral (F (2, 240) = 0.239, p = 0.787, ηp² = 0.002) (Fig. 3).

While no differences emerged between groups, we further examined whether implicit bias was consistently present across all groups by comparing incompatible and compatible trials (Table 4). Paired-sample t-tests revealed that response times in the incompatible trials were significantly longer than in the compatible trials across all dimensions: cognitive (t (251) = 4.510, p < 0.001, d = 0.284), affective (t (251) = 9.554, p < 0.001, d = 0.602), behavioral (t (251) = 2.477, p = 0.014, d = 0.156) and the combined (t (251) = 12.017, p < 0.001, d = 0.757). These results indicate a robust implicit Stigma, suggesting that participants more readily associated depressive disorders with negative rather than positive attributes.

Table 4 Mean reaction times for all subjects on incompatible and compatible condition.
Full size table

Discussion

The present study examined the impact of continuum beliefs on both explicit and implicit public stigma towards depression. For explicit measures experiment, the continuum beliefs intervention demonstrated significant effectiveness across all three stigma dimensions. Specifically, participants who received the continuum intervention showed reduced negative stereotypes and prejudicial affective reactions, along with increased social distance acceptance, compared to categorical group. However, while the intervention successfully enhanced participants’ continuum beliefs, it did not significantly impact implicit stigma, as measured by response latencies across cognitive, affective, and behavioral dimensions. Notably, the presence of implicit stigma was consistently observed across all groups, with participants demonstrating stronger automatic associations between depression and negative attributes.

There appears to be a tension between a continuum model of mental health and a medical approach aimed at definitively diagnosing mental disorders, after all, the distinction between a “healthy person” and a “patient” is one of the core public expectations of psychiatrists. Continuum beliefs emphasize mental health conditions as a continuum rather than a dichotomy of “normal” or “abnormal”. Thus, the integration of the mental health-disorder continuum into the concept of mental disorders is a continuing challenge to academic psychopathology. In contrast to traditional categorization methods, this continuum approach focuses on the continuum of symptoms, viewing health and disorder as a continuum. This is not proposed to trivialize the suffering of people with mental disorders, including major depressive disorder, but rather to take a more inclusive view of mental disorders, which helps to better reflect the progressive relationship between mental health and mental disorders. At the same time, this viewpoint helps individuals to some extent to establish a belief in the continuum of mental health and mental disorders, eliminates the labeling and discrimination against mental disorders, and provides a new perspective for re-conceptualizing mental disorders.

The present study explored the effect of continuum beliefs on the public stigma of depressive disorders and found that after emphasizing the continuum view of depressive disorders, the continuum beliefs increased and the public stigma decreased in support of our hypothesis. This is in line with the findings of a previous study8. However, the relatively small effect size observed may be attributed to limitations in our intervention approach. According to common in-group identity theory29, a key mechanism for reducing intergroup bias involves recategorizing distinct group members into a single, inclusive group. Continuum beliefs facilitate this recategorization by blurring the boundaries between “people with depression” and “people without depression,” encouraging individuals to perceive all people as positioned along a shared mental health continuum rather than as categorically separate groups. Nevertheless, this approach primarily reshapes cognitive frameworks, yielding rapid but limited changes in perception, while emotional and behavioral shifts - requiring deeper internalization - remain weaker or delayed. Our brief textual intervention, though effective in increasing continuum beliefs and reducing stigma in the short term, likely lacked the depth and intensity needed to fully activate this recategorization or sustain broader attitudinal shifts. As Brewer30 suggests, the effective formation of a common in-group identity requires subgroup members to develop a shared perception of the superordinate group, with this group holding consistent psychological significance across individuals. When subgroup members fail to coalesce around a unified understanding of the superordinate group’s identity, the common in-group identity struggles to take root31. A single-session, text-based format may not provide sufficient time or engagement for participants to deeply process these new beliefs.

Previous research has underscored the multifaceted nature of mental disorder stigma, revealing strong negative associations with terms like “destructive” and “dangerous,” as well as automatic affective responses such as shame, fear, and anxiety toward affected individuals32,33,34. In Study 2, participants exhibited significantly faster response times in compatible tasks (pairing depression-related words with negative words) compared to incompatible tasks, indicating a robust, automatic cognitive and affective link between depression and negativity. While our findings demonstrate that evoking continuum beliefs significantly reduces explicit stigma toward depressive disorders (e.g., overt negative stereotypes), the intervention had no significant effect on implicit stigma (e.g., latent prejudice and negativity reflected in automated association tests). This divergence between explicit and implicit outcomes may stem from a combination of cultural influences and the limitations of our intervention approach.

One plausible explanation lies in the cultural context of our Chinese sample. Traditional Chinese values emphasizing “modesty and prudence” often encourage individuals to conceal their true opinions, particularly on sensitive topics that could invite negative judgment or offend others. This “face-saving” mentality may make respondents more inclined to give answers that conform to social expectations. In this study, when asked about relatively sensitive issues such as attitudes and values, respondents may be hesitant to give “correct” answers that conform to societal expectations, thus concealing their potentially negative thoughts. Consequently, while the intervention effectively altered explicit responses - aligning them with societal expectations - it may have failed to penetrate the deeper, less consciously controlled implicit attitudes. Additionally, the brevity of our single-session textual intervention may not have provided sufficient time or depth to reshape implicit stigma. Implicit attitudes, rooted in long-standing automatic associations, are notoriously resistant to change and often require prolonged or multifaceted strategies to shift35. While continuum beliefs can swiftly influence cognitive frameworks and explicit endorsements, a brief text-based approach likely lacks the intensity needed to disrupt entrenched implicit biases. This limitation aligns with the cognitive focus of our intervention, which prioritizes restructuring explicit perceptions over dismantling automatic affective responses. Moreover, the use of a Single Category Implicit Association Test (IAT) to measure implicit attitudes introduces significant limitations that may contribute to the non-significant findings. A key concern in social psychology is whether the IAT’s reaction-time-based associations truly reflect implicit bias or merely stereotype recognition (i.e., awareness of cultural stereotypes rather than personal endorsement)36,37. This debate questions the test’s construct validity, as factors such as task familiarity, attention variability, or cognitive processing speed can confound results, potentially misrepresenting individual attitudes38. Additionally, the IAT’s sensitivity to procedural variations and contextual influences may further obscure genuine implicit effects, raising doubts about what it actually measures- personal bias or broader societal associations. Although we implemented controls like randomized trial sequences and D-score standardization to enhance reliability, these measures cannot fully mitigate the test’s inherent methodological and theoretical shortcomings. These limitations suggest that the lack of significant implicit attitude change may partly reflect the IAT’s constraints.

Drawing from the experimental findings and methodological challenges, several limitations of the study merit attention. Firstly, although participants were randomly assigned to experimental conditions, we used only post-intervention assessments. To minimize demand characteristics, participants were initially told they were participating in a general survey on mental health attitudes, with the study’s purpose and hypotheses revealed after data collection. While this approach reduced expectancy effects, the absence of pre-intervention baselines hinders individual-level change assessment. Moreover, given that stigma is unlikely to shift significantly in the short term, this design may not fully capture the intervention’s potential impact. Future studies should incorporate pre- and post-intervention measures and explore the long-term effectiveness of continuum beliefs, potentially through longitudinal designs, to better understand the mechanisms underlying stigmatizing attitudes and their modification over time. Secondly, a critical limitation lies in our use of a Single Category Implicit Association Test (IAT) to measure implicit attitudes. Future research should address these issues by refining implicit measures, such as developing multi-method approaches combining the IAT with physiological or behavioral indicators (e.g., eye-tracking or response latency tasks), to enhance construct validity and better distinguish personal bias from cultural associations, providing a more robust assessment of implicit stigma. Finally, applying a measure developed in other context17,26 to a Chinese population poses cultural challenges. Chinese values like collectivism and social harmony reshape stigma perceptions - downplaying blame or dangerousness compared to Western norms - suggesting that the tool may not fully align with local nuances, compromising its validity. Future research should prioritize culturally tailored instruments validated for the Chinese context.

Our study demonstrates that endorsing continuum beliefs about depression effectively reduces explicit stigma, as evidenced by decreased negative stereotypes, emotional reactions, and social distance. However, it exerts no significant impact on implicit stigma, suggesting that implicit attitudes may resist modification through short-term interventions. Participants consistently exhibited an implicit bias linking depression with negative attributes, particularly in the affective domain. These findings underscore the multifaceted nature of stigma, with explicit and implicit attitudes exhibiting divergent responses to interventions. While continuum beliefs hold promise for mitigating explicit stigma, addressing entrenched implicit biases likely necessitates more robust strategies, highlighting the need for further investigation into comprehensive approaches.