Our study shows that the use of medical language to describe a person with symptoms of severe mental illness has increased since 1990, for both schizophrenia and depression. For schizophrenia, the percentage of derogatory labels used by respondents remained overall constant at about 20%. For depression, we observed a slight decline in clearly derogatory labels from 11 to 8%, but an increase in potentially derogatory and trivializing terms.

Before discussing our findings, we need to evaluate the strengths and limitations of our study. Our investigation stands out as the only vignette-based study examining trends over 30 years. The application of the same open-ended question to elicit labels for identical case vignettes throughout this time is unique. However, the long study period also brings about some methodological challenges. For the 1990 interviews, the original answers were not preserved verbatim due to technical restrictions on the amount of data storage at the beginning of the 1990s; hence, we were unable to revise the codes assigned at that time in the case of newly added sub-categories. However, the coding system provided extensive verbatim examples for each code, so we are confident that the categories were applied consistently over the entire period. Further, the combination of the 1990 (West) and 1993 (East) surveys led to an overrepresentation of people living in East Germany for that wave. However, prevalence in categories did not differ meaningfully between East and West German respondents at that time; hence, this procedure should not have had an impact on our findings (for more details, see supplementary material in Table S6).

In 1993 and 2001, only the male version of the vignettes was used, which likely affected the results for those years. In both the earlier and later surveys, we observed slight variations in the frequency distributions of categories based on the presented gender of the vignette, particularly regarding the use of derogatory statements (refer to Table S7 in the supplementary material). However, these differences did not prove to be significant.

We used case vignettes describing someone with severe, untreated mental health symptoms. Thus, developments in attitudes towards milder courses of depression or schizophrenia, or towards people who are already in treatment, are not covered by our study. While focusing on unprompted label assignments for an unlabelled case vignette seems adequate to capture the ability to correctly recognize and label a mental health illness, other aspects of MHL, like knowledge about treatment options or actions that should be taken in a concrete situation, are not covered by our study. Across the survey waves, there was a decreasing response rate, from 70% in 1990 to 57% in 2020. Declining response rates are observed in many social science time-trend studies [26], although the face-to-face interview method still performs better than, for example, telephone surveys in this regard.

Finally, the use of an open-ended question poses a difficulty in quantitative analysis and constrains the statistical possibilities due to their nominal distribution. However, open-ended questions provide an unprompted, undistorted view of the individual's perception of the phenomenon described and thus may serve as a very direct indicator of MHL in the general population. At the same time, the willingness and ability of the interviewees to articulate themselves cannot be dismissed as a potential bias.

Considering those limitations, our results nonetheless indicate an overall increase in the prevalence of psychiatric terms in the general population, with the term “depression” being used more readily than “schizophrenia”. This corresponds to the higher lifetime prevalence of depression, estimated at 15–20% (point prevalence 5–8%) [27,28,29,30] compared to schizophrenia, with a lifetime prevalence of 2–4% (point prevalence around 0.4%) [30,31,32]. It also reflects public awareness and media coverage [33,34,35]. Depressive disorders are more frequently mentioned and receive generally more favourable and balanced media coverage compared to schizophrenia. A German newspaper analysis found that in 2019, coverage of schizophrenia in print media occurred about half as often as coverage of depression and was often solely concerned with crime and violence, without mentioning illness concepts, treatment modalities, or prognosis. The coverage of depression, in contrast, was much more balanced, with ample reference to treatment [36]. Additionally, an increasing number of celebrities who have publicly acknowledged their own or their close ones' experience of depression has likely contributed to the spread of psychiatric terminology cantered around depression and to a de-tabooing in terms of perceived normalization, getting used to explicitly naming and recognizing depressive symptoms [37, 38].

Several findings of our study illustrate normalization, particularly in dealing with depressive symptoms, and the term “depression”. Approximately one-third of the responses that were categorized as correct labelling used the adjective “depressive.” “Depressive”, however, has already permeated everyday language and is sometimes used synonymously with “down” or “low”, thus blurring the line between medical and everyday language. We also encountered colloquial variations of “depression” like “depri”, which we coded separately.

Our observation that the term “depression/depressive” was also used in approximately one-third of responses to the schizophrenia vignette over the years also illustrates that “depression” is a particularly widespread and possibly now a bit fuzzy term used about mental illness in general, which would partially explain the higher share of correct labelling in responding to the depression vignette. The increased occurrence of responses containing both correct and incorrect terminology for both vignettes since 1990 is consistent with a transition of psychiatric terminology from professional jargon to common usage but also reveals uncertainty among respondents regarding the appropriate use of such terminology. The normalization of depression is also reflected in the increase in trivializing comments like “They should just get some rest or take a holiday”. Together with a decrease in clearly stigmatizing language, this could be interpreted as a decrease in stigmatizing attitudes. However, trivialization of a severe disorder like major depression devalues the illness experiences of those affected. Our study also included quantitative measures of stigma, like the desire for social distance, the most widely used quantitative measure for discriminatory attitudes [39,40,41]. These results have been published elsewhere [22]. Emotional reactions got somewhat more compassionate and less uncomfortable and annoyed between 1990 and 2020 [22], and between 2011 and 2020, continuum beliefs concerning the depression vignette increased [42]. A study in the US, monitoring time trends in mental illness stigma from 1996 to 2018, found that in the most recent survey, the depression vignette elicited reactions similar to a vignette describing a “troubled person” without any psychiatric illness [43], clearly illustrating the normalization of depressive symptoms.

For the schizophrenia vignette, such a normalization cannot be found. A fifth of respondents consistently made stigmatizing statements, with a marked peak in 2001. The pronounced use of derogatory language in both schizophrenia and depression vignettes in 2001 reflects the dominance of a biogenetic model for mental illness aetiology, notably prevalent during the 1990s, termed the “Decade of the Brain.” [13, 44,45,46,47]. The quantitative results of the same surveys show an increase in the desire for social distance towards someone with schizophrenia and an increase in negative emotions like fear or feeling uncomfortable [22]. Continuum beliefs decreased between 2011 and 2020 [23, 48]. In the present study, respondents more frequently mislabelled or entirely overlooked schizophrenia, mistakenly identifying it as depression [22, 42, 49, 50]. Hence, taken together, our findings do not support the conclusion that increased illness recognition indicates a reduction in stigmatizing attitudes among the general population.

Overall, our study rather substantiates that mental health literacy in the German population regarding schizophrenia and major depression has markedly improved, yet stigmatizing or trivializing terminology has not decreased. Therefore, in addition to efforts to educate the public and raise awareness of mental health issues, dedicated efforts to reduce the stigma, particularly of severe mental disorders, are still urgently needed.