Public preferences toward public health

This study was funded by the Department of Health Public Health Initiative and examines attitudes and preferences toward compulsory public health programmes. As well as providing descriptive information relating to attitudes and preferences (as measured by willingness to pay), methodological work examines the role of attitudes in preference formation and modelling motivations.

This is work is based on previous work undertaken by Simon Dixon and Phil Shackley examining preferences for fluoridation and fortification of food with folic acid. However, this framework is expanded to incorporate the use of an 'atttitude-behaviour framework' as used by Eija Pouta (University of Helsinki) and the use of attitudinal information to explain respondent motivations, including warm glow, as used by Paulo Nunes (Vrije Universiteit, Amsterdam).

Study methods

Focus groups
A series of focus groups was undertaken with members of the public to gather information on their views, atttitudes and beliefs relating to compulsory public health programmes. As part of the focus group, participants were asked about two specific compulsory health programmes; either fluoridation and 20mph residential speed limits, or fortification of food with folic acid and banning smoking in public places. Details of the process and outcomes of each of these programmes was provided.

Transcripts were produced for each group, and qualitative analysis undertaken using the framework approach of Richie and Spencer. A set of themes, and sub-themes were produced and formed the basis of the next stage of the project.

Attitudes questionnaire
A set of 38 statements relating to attitudes and beliefs identified in the focus groups was constructed. After pilotting, these statements were divided into three groups; 'general public health attitudes', 'beliefs about a specified programme' and 'attitudes/beliefs relating to non-programme characteristics'. This classification was necessary, as it determines how they are positioned within the willingness to pay survey.

Willingness to pay survey
A willingness to pay (WTP) survey was undertaken with 800 respondents. A single compulsory health programme was examined in each survey, with four programmes being examined across the full sample; fluoridation of community water supplies, 20mph residential speed limits, fortification of flour with folic acid and banning smoking in public places. The general format of the survey followed that used by the EuroWill Group.

Additionally, a split-sample design was used to examine two further issues. Firstly, the general attitude and specific belief questions were place before the WTP question for half the sample, and after the WTP for the other half. This allows a test of the impact of gathering attitude and belief information on WTP responses and self-perceived answer quality. This builds on the work of Eija Pouta.

Secondly, a between-samples test of sensitivity to scope was built into the survey by using two different levels of effect for each programme. Insensitivity to scope is to be tested for using the conventional one-sided hypothesis test of unadjusted WTP values, and using the attitude-adjusted WTP values using the approach of Paolo Nunes. In the work of Nunes, this approach was found to explain 'warm glow' motivations, such that the adjusted WTP estimates passed the scope test.


33 members of the public attended one of six focus groups. Seven themes were identified in the final coding framework; effects of policy, alternatives, role of government, uncertainties, coherence of policy, rights and responsibilities, and other issues. The themes that were identified also show quite sophisticated thinking about health issues. Clearly the effects of policies were frequently raised, which is to be expected, but other issues are also considered important. Participants considered alternative interventions, uncertainties, the role of government, policy coherence and rights and responsibilities.

Eight hundred and sixty four people were interviewed in the community survey. Eighteen separate questionnaires were used in the survey, two of which were confined to the pilot. Excluding the pilot surveys, approximately 200 responses were gathered for each of four policy questions; fluoridation of water, fortification of flour with folic acid, banning smoking in public places and 20 miles per hours speed limits in residential areas. Marked differences were seen between the four schemes in terms of their direction of preference (i.e. the response to the policy vote question), with p<0.001 using a Chi-squared test. Higher levels of support were seen for speed limits and banning smoking in public places, with the lowest level of support seen for folic acid.

The rate of protests across the four policies is high for those in favour of the policy, ranging from 17.6% for folic acid to 39.7% for banning smoking in public places. The rates are higher still for those opposed to the policy. For those in favour of the scheme, attitude questions preceding the WTP question increases the rate of protest from 31.9% to 40.2%, whilst for those opposed to the scheme, it rises from 51.5% to 65.9%.

When the regressions of WTP were undertaken, the coefficients relating to income are all positive, yet only one is statistically significant. The coefficients on timing of the attitude questions are all positive, but not statistically significant. The size of each programme´s effect is not statistically significant. Degree of belief in the policy description has a statistically significant effect on WTP for fluoridation. There is also strong evidence (p<0.001) that the interviewer influences WTP responses for two of the schemes. The interviewer effects represent the largest impact on reported WTP for three out of the four schemes, and greater than the scope effect in all schemes.

The factor analysis of the attitude statements produced a four factor solution which was robust to the method of rotation. These were labelled `common sense´, `government role´, `warm glow´ and `rights and responsibilities´. It is notable that each of the four factors are each made up exclusively of statements that either reflect general attitudes as with factor 2 (e.g. `The government can be trusted with its health policies´), or attitudes toward a specific scheme as in factor 1 (e.g. `This policy will improve the health of the public´). This seems to indicate that we can separate out a respondent´s evaluation of the scheme itself into overarching attitudes and attitudes relating to the specific policy.

There is evidence that the factors influence WTP. `Common sense´ is significantly associated with WTP for smoking, `government role´ is significantly associated with WTP for fluoridation, `warm glow´ is significantly associated with WTP for speed limits and smoking, and `rights and responsibilities´ are significantly associated with WTP for folic acid. Interviewer effects remain present, and the magnitude of the individual coefficients are in excess of all other explanatory variables within the regressions.


This work has added to our knowledge of the WTP approach by developing a body of research relating to attitude information. The results show how attitudes related to various features of a scheme have an explanatory effect on WTP, but that this does not necessarily produce WTP estimates that pass the scope test. However, the nature of compulsory health programmes appears to induce high rats of protest, and responses that are scope insensitive even with the use of attitude information as explanatory variables. Consequently, the use of WTP in this way should not be considered for the valuation of compulsory health programmes, although the approach should be developed in the ways mentioned in other areas of public policy.


Two HEDS discussion papers have been published based on this work, with amended versions of these being sent to journals.

Supplementary materials relating to the work, such as transcripts, are also available from my web page.