ONLINE ADDENDUM to ‘Freezing Deliberation through Public Expert Advice’
Section 1: Methodology for text database searches and coding
As described in the main text’s section 3, Infomedia in Denmark and Retriever Research Sweden were searched for relevant texts. Each database was searched for texts from April 15 to December 31 2009. The search terms combined ‘pandemic’, ‘H1N1’ and ‘vaccination’. The search frame in Retriever included all texts in city newspapers, domestic newswires, and national professional print media (storstadspress, nyhetsbyrå, tidskrifter, fackpress). The search frame in Infomedia included all texts in nationally distributed newspapers (landsdækkende aviser). The difference in frames was due to differences in the databases’ delimitable parameters. The searches yielded 673 Swedish and 425 Danish print media texts.
All texts were coded to enable quantitative measurements. Any mention in one text of one person’s or organization’s support for one policy option was defined as a ‘claim’ and became a separate unit of analysis. For each claim, coded characteristics included: advocated vaccination coverage (claim); claim-maker’s name, position and title; text type and word length; name of medium; coders’ subjective scores for detail of justification for claim and for text author’s attitude.
In addition:
- If multiple people or organizations were cited in the same text, one claim was coded for each (e.g., one article citing three sources that support the same policy option was coded as three claims).
- People were primary, organizations secondary. That is, a separate claim was only coded for an organization if no person from the organization was named.
- Op-eds, columns, editorials and letters to the editor that supported a policy option were coded as one claim by each author (e.g., an op-ed with three authors was coded as three claims).
- Multiple mentions in the same article of the same person (or organization) were not coded as separate claims.
- The analyses used attributable claims, sub-set of all claims identified (107 out of 119 for Sweden, 127 out of 128 for Denmark). Attributable means that a text names or describes an individual (or organization) as the source of the claim.
Claims were sorted into three general categories:
- Support for general mass vaccination (‘all’ should be vaccinated).
- Support for targeted or risk-group only vaccination (‘some’ should be vaccinated).
- Support for no vaccination (‘none’ should be vaccinated).
The analyses focused primarily on the sources that contributed to the discourse (claim-makers). Coding sorted claim-makers into six categories:
- Certified experts. Public health experts (physicians, epidemiologists, virologists and other university trained professionals specializing in health issues, influenza or vaccines), who are civil servants and tasked with public health/pandemic flu management by the national or a local government, e.g., a state epidemiologist at a national or regional health government agency.
- Other experts. Public health experts who are not employed as civil servants in government public health/pandemic management.
- Other civil servants. Civil servants who are not public health experts.
- Politicians. Government ministers, party officials, members of parliament and local elected officials.
- Journalists. Journalists and editors employed by print media.
- Civil society. Members of the general public and non-expert representatives of interest organizations.
Section 2: Figures
Figure 1. Frequency of Claims and Claim-makers
Figure 2. Who Claimed What
Figure 3. Timing of Policy Positions Claimed in 2009
Figure 4. Pro- and Counter-Policy Claims by Text Type