Both attitude outcomes variables were standardised according to the Z score. We used unadjusted and adjusted linear regression models to examine factors associated with (1) prejudice and exclusion and (2) tolerance and support for community care. The trend is based on earlier timepoints continuing along the same trend. We also extrapolated a linear trend line for the years 2009–13, which projects what population attitude scores might have been without the Time to Change campaign. We calculated the annual mean item score of the English population for attitude items related to (1) prejudice and exclusion and (2) tolerance and support for community care.
Based on this index, we calculated mean regional familiarity with the Time to Change campaign on the basis of the mean score by region. We also created an index of familiarity with the campaign (0=no awareness, 1=seen or heard advertising once or twice, 2=seen or heard advertising three to five times, and 3=seen or heard advertising six times or more) to investigate the relation between campaign awareness and intended reduction in stigma. We then calculated regional prevalence of respondents reporting overall awareness of the Time to Change campaign. Respondents who reported some campaign awareness were then asked: “How many times, before this interview, have you seen or heard ANY of the advertising in the pictures?” Response options were (1) once or twice, (2) three to five times, and (3) six times or more. Overall awareness was then calculated by asking the following question whilst showing the mental health advertising stills: “Do you think you have seen this advertising, or similar, during the last year?” Individuals who reported “yes” or “yes, seen similar” were categorised as campaign aware, whereas those who responded “no” or “don't know” were categorised as not campaign aware. Participants were first shown several pictures from different Time to Change advertisements that appeared on television, on the radio, in magazines, or on the internet. Prompted awareness of the Time to Change campaign was assessed in 20. Additional information regarding the survey methods can be found elsewhere. Interviews are done in the participants' homes by fully trained personnel with computer-assisted personal interviews, and demographic information is collected at the end of the interview. Census small-area statistics and the Postcode Address File are used to define sample points, which are stratified by Government Office Region and social grade. The Attitudes to Mental Illness survey is done using a quota sample, with sample points selected by a random location methodology. Phase 1 campaign activity officially ended in 2011 however, the assessment of phase 1 continued into 2012. Notably, although Time to Change received funding in October, 2007, the social marketing campaign did not begin until after the survey was run in 2009 and hence we use 2009 as a baseline for the assessment. Mental-health-related attitudes were assessed each year however, assessment of mental-health-related knowledge and intended behaviour did not begin until 2009, in line with the phase 1 assessment of the Time to Change anti-stigma campaign. The Attitudes to Mental Illness survey is an omnibus survey commissioned by the Department of Health and done among a nationally representative sample of adults residing in England. The survey was not done during the years 2004–06 and so data are not available for these years. We used data from the 20–13 national Attitudes to Mental Illness survey.