Background The press’ representation of mental illness often includes images of people as dangerous, and there is evidence that this contributes to stigmatising understandings about mental illness. Little is known about how newspapers portray mental health on their Twitter feeds. Aims To explore the representation of mental health in the UK national press’ Twitter feeds. Method Content analysis was used to code the Tweets produced by UK national press in two time periods, 2014 and 2017. Chi-square analysis was used to identify trends. Results The analysis identified a significant reduction in the proportion of tweets that were characterised as Bad News between 2014 and 2017 (χ2 = 14.476, d.f. = 1, p < .001) and a significant increase in the tweets characterised as Understanding (χ2 = 9.398, d.f. = 1, p = .002). However, in 2017, 24% of the tweets were still characterised as Bad News. Readers did not retweet Bad News stories significantly more frequently than they were produced. Conclusions There is a positive direction of travel in the representations of mental health in the Twitter feeds of the UK press, but the level of Bad News stories remains a concern.

Aims. Media representations of mental health problems may influence readers’ understanding of, and attitude towards, people who have received psychiatric diagnoses. Negative beliefs and attitudes may then lead to discriminatory behaviour, which is understood as stigma. This study explored the language used in popular national newspapers when writing about schizophrenia and considered how this may have contributed to the processes of stigmatisation towards people with this diagnosis. Methods. Using corpus linguistic methods, a sample of newspaper articles over a 24 month period that mentioned the word ‘schizophrenia’ was compared with a similar sample of articles about diabetes. This enabled a theory-driven exploration of linguistic characteristics to explore stigmatising messages, whilst supported by statistical tests (Log-Likelihood) to compare the data sets and identify words with a high relative frequency. Results. Analysis of the ‘schizophrenia’ data set identified that overtly stigmatising language (e.g. “schizo”) was relatively infrequent, but that there was frequent use of linguistic signatures of violence. Articles frequently used graphic language referring to: acts of violence, descriptions of violent acts, implements used in violence, identity labels and exemplars of well-known individuals who had committed violent acts. The word ‘schizophrenic’ was used with a high frequency (n=108) and most commonly to name individuals who had committed acts of violence. Discussion. The study suggests that whilst the press have largely avoided the use of words that press guidance has steered them away from (e.g. “schizo” and “psycho”) that they still use a range of graphic language to present people with a diagnosis of schizophrenia as frighteningly ‘other’ and as prone to violence. This repetition of negative stereotypical messages may well contribute to the processes of stigmatisation many people who experience psychosis have to contend.