November 26, 2023 (Updated on November 28, 2023)

On August 9, 2010, The New York Times published a profile on the US Army whistleblower, Chelsea Manning, the former soldier who was accused of releasing hundreds of thousands of classified military and diplomatic documents to Wikileaks, some of them exposing war crimes. At the end of the article, the journalist, Ginger Thompson, quoted something that Chelsea Manning had confided during an online chat with the disgraced hacker and Manning’s soon-to-be betrayer, Adrian Lamo. But Thompson only quoted part of what was in the chat transcript. According to Thompson, Manning said, “I wouldn’t mind going to prison for the rest of my life, or being executed so much, if it wasn’t for the possibility of having pictures of me plastered all over the world press.”

However, what Manning actually said was the following: “I wouldn’t mind going to prison for the rest of my life, or being executed so much, if it wasn’t for the possibility of having pictures of me plastered all over the world press as a boy.”

The full and correct quote was eventually used by the same journalist in another profile on Manning coauthored with John M. Broder three years later.

Although this is yesterday’s news, I believe no one has ever talked about this bit of narrative license by the “newspaper of record,” and it’s something that has always stayed with me. It’s not an insignificant omission in my opinion, and I don’t believe it was unintentional either, whether because the author judged it to be an incongruent piece of information and therefore disposable, or because the author was deliberately framing Manning’s story a certain way.

The first (and incorrect) quote suggests that Manning was worried about public opinion, i.e. that she was concerned about being portrayed by the press as a criminal, like someone in a Wanted poster. The second (and complete) quote suggests something quite different, namely that Manning was not second guessing her actions, and instead was anxious about being misgendered and deadnamed by the media and the public. Indeed, in interviews given after her pardon, Manning corroborates that she has never regretted her actions in Iraq.

As Chelsea Manning’s gender dysphoria became public, it provided the media another way to frame her story — as a troubled person whose whistleblowing was not so much a political war-cry as a personal cry for help. The media piled on Manning’s psychological history in their efforts to downplay the ideological and moral bases of her dissent, because… who in their “right mind” would do such a thing? In their 2013 New York Times profile on Manning, the authors conjecture that “the roots of Private Manning’s behavior may spring as much from [Manning’s] troubled youth as from [Manning’s] political views,” despite that Chelsea Manning had indicated in her own words that she was disillusioned by the atrocities she had witnessed in Iraq, such as secret footage of a US Apache helicopter killing journalists and civilians, and that she felt compelled to voice these war crimes.

A study

Although no formal content analysis has comprehensively examined the prevalence of mental illness cues in portrayals of political dissidents, it is easy to find cases of this phenomenon throughout history. For example, there are documented attempts from the 1950s to portray the Puerto Rican independence movement leader, Pedro Albizu Campos, as mentally ill, and in the 1970s, CIA agents attempted to steal Daniel Ellsberg’s psychiatric file in an attempt to find discrediting information about the Vietnam War whistleblower. A content analysis of US media coverage of the anti-corporate globalization movement in 1999-2000 revealed that protesters were often portrayed through what Jules Boykoff, professor of Politics and Government at Pacific University, called a “freak frame,” with descriptors including “fringe,” “crazy,” and “strange.”

Consistent with functional accounts of stigmatization, research has found that people are more inclined to punish a member of an out-group perceived as threatening, especially when doing so could help establish one’s belonging with the in-group. This suggests that when a dissident takes action against a foreign government, individuals may try to consider how the action was warranted and purposeful. But when the same act of dissent is targeted against the government that the individual identifies with, they may feel threatened and hence motivated to devalue the act as a sign of mental instability.

Several years ago, I conducted a little experiment under the supervision of Kathy Berenson, psychology professor at Gettysburg College, in which we hypothesized that media consumers would readily label political dissidents as mentally unstable, and that the use of mental illness framing in media portrayals of dissidents serves to suppress their message. (Although the results were interesting, we never published this study, because it didn’t add much to existing scientific knowledge. However, peer review found no fault in our methods.)

To examine the possible functions and consequences of this phenomenon, we assessed the attitudes of participants from the US about a fictitious dissident who was described either with or without explicit mental illness cues. We adapted a real news story to present information about the dissident while manipulating inclusion of information about mental illness symptoms and the country targeted by the dissent (either the United States or Russia) using a 2 x 2 design.


Our primary analyses are based on a sample of 392 participants from the United States (209 females, 183 males), collected in the summer of 2015. Participants had a mean age of 35.0 (SD = 11.20; range: 18 to 72). Their median level of education was some college (range: junior high school to graduate school). Participants identified as Asian (7.1%), Black person (6.9%), Hispanic (4.6%), Native American (0.5%), White (75.0%), Multiracial or other person (5.9%).

We recruited 517 potential participants through Mechanical Turk (MTurk), a system that pays workers to complete online tasks, and which has shown good test-retest reliability as a means of data collection. Compared to the US population, MTurk workers are younger, have higher education, lower income, and hold more liberal political views.

After providing informed consent, participants were randomly assigned to one of the four stimulus conditions. Conditions did not vary with respect to demographics or political ideology.

The stimulus article was adapted from two articles in The New York Times about the Manning case, but with the name of the protagonist changed to either Miller (US condition) or Petrovich (Russia condition). Notably, all text samples used to manipulate participants’ perceptions of mental illness were quoted directly from the source articles.

An example of one of three passages in the stimulus that included or excluded information about mental illness:

“[Barack Obama / Vladimir Putin] is going to have a heart attack,” [Miller / Petrovich] boasted. [But even as he professed a perhaps inflated sense of purpose, he called himself “emotionally fractured” and a “wreck” and said he was “self-medicating like crazy.”]

Source: The New York Times

Afterwards, participants answered questions regarding their attitudes about the dissident’s actions, intentions, and the role of perceived disability. An example of a set of questions assessing the extent to which participants attributed the dissident’s behavior to psychiatric disability:

[Miller/Petrovich] should have left politics to the experts and focused on fixing himself first; [Miller/Petrovich] would not have leaked secret information if he hadn’t had serious mental problems; [Miller/Petrovich]’s story is primarily a story about the consequences of mental problems; [Miller/Petrovich] revealed government secrets because he had mental problems and was not acting in his right mind.

Each item was rated on a six-point Likert scale with the following labels: 1 = strongly disagree; 2 = disagree; 3 = disagree a little; 4 = agree a little; 5 = agree; 6 = strongly agree. Internal consistency (α) for this scale was .91.

The participants then answered a set of questions about their political beliefs, and provided demographic information.

Finally, we assessed prior knowledge of the Manning case on which the stimulus article was based so that we could exclude participants familiar with the case. Excluding 119 participants who expressed recognition/prior knowledge of Chelsea Manning, left us with a sample size of 392.


As predicted, attitudes about the act of dissent were more negative when mental illness cues were present, especially among participants who endorsed politically conservative ideology.

Our first linear regression model (Table 1) shows a significant effect of mental illness framing on attributing dissent to psychiatric disability across all participants. Also as predicted, participants from this US sample were significantly more likely to attribute political dissent to psychiatric disability when it challenged their government. Model 2 shows that participants’ attributions of the dissent to psychiatric disability increased substantially with conservative political ideology. (See Figure 1.)

Table 1. Regression models predicting attributions to psychiatric disability (n = 392)

Figure 1. Attributions to psychiatric disability as predicted by condition and political ideology. Error bars represent standard errors.

Table 2 presents the effect of our manipulations on attributing dissent to positive intentions. Model 1 shows that acts of dissent against the US were significantly less likely to be attributed to positive intentions. Also as predicted, attributions to positive intent were significantly lower when the dissident was described with symptoms of mental illness. Model 2 shows that when the dissident was described as having mental illness symptoms, participants’ attribution of the dissent to positive intent was inversely related to conservative political ideology. (See Figure 2.)

Table 2. Regression models predicting attributions to positive intent (n = 392)

Figure 2. Attributions to positive intent as predicted by condition and political ideology. Error bars represent standard errors.

Thus far, our results show that as predicted, our manipulations influenced participants’ understanding of the dissident’s motivations. But did our manipulations also influence participants’ attitudes about the actual political issue related to the act of dissent? We tested this hypothesis in the linear regression analyses shown in Table 3. Although Model 1 shows no significant main effects of either manipulation, Model 2 again shows an effect of mental illness framing moderated by participants’ political ideology. Simple slope analyses indicate that when participants were exposed to a story about a dissident who was said to have mental illness symptoms, their agreement with an issue at the heart of the case (e.g. the importance of government transparency) had a significant inverse association with the extent to which they held a conservative political ideology. Describing a dissident as having symptoms of mental illness did promote dismissal of a political cause the dissident was associated with, but only among politically conservative individuals. (See Figure 3.)

Table 3. Regression models predicting attitudes about the political issue central to the dissent (n = 392)

Figure 3. Attitudes about government transparency as predicted by condition and political ideology. Error bars represent standard errors.


Media characterizations of political dissidents that include symptoms of mental illness influence the public to believe that the dissident’s personal problems explain their actions rather than a legitimate grievance or concern for the public interest, and such media effects are particularly strong for members of the public who hold conservative political beliefs. Regardless of media framing, however, US citizens are more likely to dismiss political dissent as a sign of mental illness when they perceive it as a threat to American society. Stigmatization appears to be driven not only by the characteristics of the messenger but also by people’s motivations to defend their group.

Although the effect sizes in this study are relatively small, it is not difficult to conceive that in a real world scenario where media consumers are bombarded by uniform messaging from multiple sources, the cumulative effects may substantially change public opinion on a major news story.

Limitations of this study

The stimulus we used was derived from real US news stories and thus had strong external validity. Although this is one of the strengths of our research design, it also raises several limitations.

The stimulus that we used in the mental illness condition was quite explicit in depicting mental illness, consistent with the original news articles that it was derived from. Yet as the public becomes more cognizant of mental illness stereotyping there may be an increasing use of code words, such as “bizarre” or “unhinged,” to imply rather than directly state that dissidents are mentally ill.

Another limitation of our study is that we didn’t assess other attitudes towards the dissident beyond perceptions of mental illness — such as overall likeability. We therefore cannot rule out the possibility that having a general negative impression of the dissident drove participants to denigrate them as mentally ill more than feeling threatened.


These two things can both be true: Chelsea Manning was motivated by her principles, and she was also experiencing gender dysphoria during the time that she released classified documents to Wikileaks. Manning’s gender dysphoria neither negates, nor explains, her act of political dissent.

Accusations of mental illness are often used in attempts to discredit activists and whistleblowers in the court of public opinion. The results of this study highlight the extent to which mental illness remains stigmatized, in that it is apparently difficult for many Americans to view those with psychiatric disabilities as capable of making rational decisions intended to serve the greater good.

Given that approximately 20% of the US adult population are living with mental illness, many political dissidents can certainly be expected to be experiencing mental illness as well. But even when a dissident’s psychiatric disability is clearly documented (and not just inferred, as is often the case), we still have a choice whether or not to frame the story in terms of mental illness. Focusing on dissidents’ psychiatric histories is a common way for the media to transform stories about complex and troubling political issues into human interest stories that are easier to trivialize.

Like it? Share with your friends!