Detecting depression stigma on social media: A linguistic analysis

Highlights

• This study reveals the ways in which depression stigma presents on social media.

• Linguistic analysis can be helpful to detect depression stigma on social media.

• 6.09% of relevant social media posts reflected depression stigma.

• The accuracy of online detection of stigma reached 75.2% and 86.2% (F-Measure).

Abstract

Background

Efficient detection of depression stigma in mass media is important for designing effective stigma reduction strategies. Using linguistic analysis methods, this paper aims to build computational models for detecting stigma expressions in Chinese social media posts (Sina Weibo).

Methods

A total of 15,879 Weibo posts with keywords were collected and analyzed. First, a content analysis was conducted on all 15,879 posts to determine whether each of them reflected depression stigma or not. Second, using four algorithms (Simple Logistic Regression, Multilayer Perceptron Neural Networks, Support Vector Machine, and Random Forest), two groups of classification models were built based on selected linguistic features; one for differentiating between posts with and without depression stigma, and one for differentiating among posts with three specific types of depression stigma.

Results

First, 967 of 15,879 posts (6.09%) indicated depression stigma. 39.30%, 15.82%, and 14.99% of them endorsed the stigmatizing view that “People with depression are unpredictable”, “Depression is a sign of personal weakness”, and “Depression is not a real medical illness”, respectively. Second, the highest F-Measure value for differentiating between stigma and non-stigma reached 75.2%. The highest F-Measure value for differentiating among three specific types of stigma reached 86.2%.

Limitations

Due to the limited and imbalanced dataset of Chinese Weibo posts, the findings of this study might have limited generalizability.

Conclusions

This paper confirms that incorporating linguistic analysis methods into online detection of stigma can be beneficial to improve the performance of stigma reduction programs.

Introduction

Depression is a common and serious mental disorder that affects people negatively. According to a report of the WHO, in 2015, the proportion of the global population with depression is estimated to be 4.4% (World Health Organization, 2017). Getting professional treatment can be helpful for most people with depression (Wang et al., 2003). However, the existence of stigma associated with depression hinders help-seeking behaviors and delays treatment of mental illness (Barney et al., 2006, Schomerus et al., 2009). Therefore, lowering depression stigma would improve mental health outcomes.

Stigma reduction campaigns work best when targeted. Information distributed through mass media is an important contributor to the dissemination of mental-health-related material that may increase stigma (Corrigan et al., 2005, Dietrich et al., 2006, Morgan and Jorm, 2009, McGinty et al., 2014, Niederkrotenthaler et al., 2014, Koike et al., 2015, Maiorano et al., 2017). If the characteristics of such stigma information can be understood, it may be possible to design interventions that counter its negative influence. However, the majority of previous studies investigating the impact of media coverage of mental illness mainly focused on traditional media platforms (e.g. newspapers and TV programs), with relatively little attention paid to new media platforms (i.e. social media). Recently, a few studies used of social media data to monitor mental-illness-related stigma information, including depression stigma information (Fu et al., 2013, Reavley and Pilkington, 2014, Li et al., 2015). These studies employed human coders to manually process potential stigma information, and established a series of schemes for coding stigma information on social media, which may provide us an evaluation method to determine whether a piece of social media post reflects mental-illness-related stigma or not. However, the sheer volume of social media data makes it difficult to identify stigma information by human coders. Therefore, there is a dire need for automatic detection of stigma, which may give us an efficient way to target stigma information automatically among massive information or at least minimize the information searching scope quickly. However, computational methods for detecting depression stigma have not yet been established.

The words that people use in their language expressions can provide clues into their psychological processes (Pennebaker et al., 2003). Evidence confirms that health-related stigma, including mental-health-related stigma, can be conveyed through different patterns of language use (Anthony et al., 2008, Downie and Black, 2010, Puhl et al., 2013, Slebioda, 2013). This suggests that incorporating linguistic analysis methods into online detection of stigma may improve our ability to screen for depression stigma information automatically on social media.

By analyzing Chinese language expressions in Sina Weibo posts, a famous microblogging service in China, this study aims to build computational models for detecting depression stigma information on social media.