The methods of the evaluation have been described more fully in a published report 26 and summary paper 4. We acknowledge that our method adapted methods from the existing evaluation literature and was not a kaupapa Māori approach. Kaupapa Māori research has been defined as research that is culturally safe, relevant and appropriate, and is controlled by Māori, for Māori and with Māori. It addresses Māori needs and gives full recognition to Māori culture and value systems, upholds the interest and mana (dignity and respect) of the group, and embraces traditional beliefs and ethics which incorporate resistance strategies that embody the drive for tino rangatiratanga (unqualified indigenous authority and self-determination) 27. Such an approach may, for example, have focused more on the degree to which the legislation promoted Māori aspirations and participation.

We developed a logic model adapted from the evaluation model of the US Centers for Disease Control 28, and the logic model used for the evaluation of the Scottish smokefree legislation 29 for the evaluation 26. This incorporated process, and core (direct, anticipated impacts relating to primary objectives) and non-core (indirect, possible outcomes relating to secondary objectives) outcome indicators. The process and outcome information areas used to structure the evaluation were as follows:

For each of the process and outcome indicators, we identified data sources from literature searches, and from consulting members of the project team and other key informants. Data sources included existing papers, reports, policy documents, national health and demographic statistics. We summarised the evidence from each relevant data source. Where the data or existing analyses were judged insufficient for the purposes of the evaluation, and where practicable, we carried out additional analyses or studies. Additional studies and analyses reported in this paper are: further analysis of evaluation data from the Health Sponsorship Council (HSC) Monitor Surveys, an exploration of the experiences and attitudes of Māori stakeholders to the introduction and implementation of the SEAA, and an analysis of national Quitline data. The data presented is therefore a combination of summaries of relevant data and analyses carried out by other researchers, and key findings of additional quantitative analyses of existing data sources and additional qualitative research carried out by the authors.

Evidence on public and stakeholder knowledge, attitudes and support for smokefree policies; exposure to SHS in the workplace and at home; and the frequency of socially-cued smoking reported by smokers in pubs and bars, restaurant and nightclubs came from the HSC Monitor Surveys. Additional data came from the ASH Year 10 children's survey (for exposure to SHS in the home), and from National Research Bureau surveys, 1989–2001, for pre SEAA data on exposure to SHS in the workplace and home.

The HSC Monitor Surveys (2003–6) were nationally representative surveys with 2000–2500 participants aged 15 years or above, with a boosted Māori population sample of about 900 (500 in 2003). The surveys included questions that were developed specifically to monitor the SEAA. 30 Data collection for 2003–4 occurred prior to implementation of the main smokefree provisions of the SEAA and for 2005–6 after implementation. A general population sample was obtained through random digit dialling in 2003 and 2004, and by random sampling of private household numbers in 2005 and 2006. The person with the next birthday present at the time of the call (2003 or 2004) or within the household (2005 and 2006) was invited to participate. The Māori sample was made up of Māori participants from the general sample and from an additional random sample of households of individuals who self-identified as Māori on electoral rolls. Participants were interviewed using Computer Assisted Telephone Interviews which included questions on many topics related to smoking, SHS exposure and the SEAA, with many questions common to several or all of the surveys.

Response was defined as the estimated proportion of eligible people contacted who agreed to participate (consent rate) and the proportion of the eligible people identified for inclusion who agreed to participate (response rate). During 2003–6 (2006 figures are provisional) in the general survey these ranged between 29–38% for consent rates (38% 2003, 35% 2004, 29% 2005, and 34% 2006) and 29–38% for response rates (44% 2003, 41% 2004, 32% 2005, and 38% 2006). For the Māori sample, the range was 39–79% for consent (79% 2003, 69% 2004, 47% 2005, and 39% 2006) and 35–63% (62% 2003, 63% 2004, 39% 2005, and 35% 2006) for the response rate. There were no obvious trends in response and consent rates over time, except that the Māori over-sample consent and response rates were much lower in 2005–6 than 2003–4, possibly due to changed reporting practices between survey companies.

We carried out additional analyses of the 2003–2006 HSC Monitor Surveys, weighting the findings to the age, sex and ethnicity (proportion of Māori and non-Māori) distributions of the 2001 census 26. This ensured that changes in the distribution of survey responses were not the result of changes in the age, gender and ethnicity distribution of the populations participating in the surveys. Results were stratified by gender, ethnicity (Māori/non-Māori), smoking status (smoker/non-smoker) and by level of household income.

The ASH Year 10 Smoking survey is a national cross-sectional questionnaire survey of 14–15 year old children 31. Data are collected anonymously using self-completion questionnaires which have included questions on children's reports of parental smoking and SHS exposure in the home since 2001. Between 1999 and 2007 an average of over 30,000 students participated each year, representing more than 70% of all eligible schools and around half of all Year 10 students in New Zealand.

The National Research Bureau (NRB) studies were conducted in 1989, 1991, 1996 and 2001. 19213233 The surveys recruited a demographically representative sample of adults aged 15 years and over from households included in regional telephone directories, and collected data using telephone interviews. Response rates in the NRB surveys were around 60–70%, varying with the methods used to define non-responders. The interviews included questions on self-reported SHS exposure in the workplace. The 1989, 1991 and 1996 surveys also included questions on SHS exposures in the home and details on the workplace smoking policy. They therefore provided information on baseline data and trends in SHS exposure at work and in the home prior to the SEAA implementation.

The New Zealand Tobacco Use Survey was first carried out in 2006, and includes data on a wide range of topics including SHS exposure in the home and at work collected through face-to-face interviews 7. A complex systematic stratified sampling scheme was used to achieve a nationally representative sample of adults aged 15–64 years with over-sampling of Pacific Islanders, Māori and 15–24'year olds. The sample size was 5703 (response rate of 75%), of whom 1071 were Māori.

The New Zealand Health Survey was carried out using face-to-face interviews in 1996/7, 2002/3, and 2006/7. A systematic stratified sampling scheme was used to achieve a nationally representative sample of adults aged 15 years and over, with over-sampling of some ethnic groups including Māori. In the 2006/7 survey 12488 adults, including 3160 Māori. Response rates were 72% in 2002/3 and 68% in 2006/7 34.

Evidence about the experiences and attitudes of Māori stakeholders to the introduction and implementation of the SEAA came from eight in-depth interviews, six of which were conducted as face-to-face interviews by a Māori interviewer; and two by telephone interview by a research consultancy company. The Māori stakeholders were selected purposively to get a cross-section of opinions and experiences, and comprised four interviewees from Māori non-governmental organisations, a public health manager, a smoking cessation worker, a union official and a bar manager 26. The eight interviews with Māori stakeholders were part of a larger qualitative study involving 34 interviews which aimed to document enforcement activities, and investigate the attitudes, beliefs, experiences and perceptions of stakeholders about the process of implementation, enforcement and compliance with the smokefree aspects of the SEAA 26. The interviews with Māori stakeholders focused particularly on participants' hopes, aspirations and beliefs regarding the changes in legislation, their experiences as someone involved in implementing the changes and what they viewed as the wider impacts of the changes including the acceptability to Māori communities. We used content analysis to identify the main themes and illustrative quotes.

The main sources of information on quitting behaviour by ethnicity came from the monthly number of caller registrations and subsidised nicotine replacement therapy (NRT) vouchers issued by the national Quitline 35.

We found evidence of strong and growing support for the new smokefree legislation and its underlying principles among Māori and non-Māori (table 1). For example, data from the HSC surveys showed that the social unacceptability of smoking around children was high among Māori and non-Māori before and after implementation of the SEAA and was 96.2% (95% CI 94.7% to 97.2%) among Māori and 96.6% among non-Māori (95% CI 95.5% to 97.4%) in 2006. In the 2004–2006 surveys, respondents were asked whether they thought that the dangers of SHS have been exaggerated. The proportion disagreeing increased from 59.0% (95% CI 55.7% to 62.3%) to 69.9% (95% CI 66.1% to 72.0%) among Māori respondents and from 63.1% (95% CI 60.5% to 65.7%) to 71.7% (95% CI 69.2% to 74.4%) among non-Māori between 2004 and 2006. In the 2006 Tobacco Use Survey, Māori respondents were less likely (27.5%, 95% CI 23.4% to 31.5%) than non-Māori (40.1%, 95% CI 38.4% to 41.8%) to state they were bothered a lot by others smoking around them and to believe that SHS is harmful. However, these differences may be confounded by a greater proportion of smokers among Māori, as smokers are much less likely to be bothered by SHS and believe SHS causes harm. 7

Data from the HSC surveys showed that by 2006 there was overwhelming support (over 90% agreement, and 6% or less disagreement) for the right to live and work in a smokefree environment; and for indoor workers, including bar and restaurant workers, to work in a smokefree environment (table 1). Support among Māori respondents was slightly lower than non-Māori in 2006, but the absolute increase from levels of support in 2003 was greater among Māori than non-Māori. There were particularly large increases in support for the rights of hospitality workers to work in a smokefree environment among Māori between 2003 and 2006 (an absolute increase of 25–26%, compared with 8–9% among non-Māori).

Finally, the level of support for a ban on smoking in indoor workplaces increased steadily between 2004 and 2006, with similar increases among Māori and non-Māori (figure 1).

The Māori stakeholders interviewed were all largely supportive of the smokefree legislation and its implementation, though they noted that there was not universal support among Māori leaders and organisations at the time of developing and implementing the legislation. All participants stated that Māori have generally accepted this legislation and changes associated with it;

"I'm around Māori heaps and I don't hear anyone moaning; Māori have been generally compliant, Māori are impacted equally but not separately".

Several believed that the SEAA had enabled a new level of discussion about smokefree homes and getting rid of tobacco smoking entirely. The call by the Māori Party MP Hone Harawera in 2006 for legislation to ban all cigarette sales was cited as evidence for this. A more recent illustration is that some Māori tobacco control advocates and practitioners have recently supported a change in the social marketing message to Māori from Auahi Kore (smokefree) to Tupeka Kore (tobacco free) 3637.

The Māori health workers interviewed felt that sufficient information was available during implementation except for some confusion about what exactly constituted a non-enclosed outdoor area. Interviewees mostly reported that effective communication about the changes, timelines and potential impacts of the legislation was facilitated by existing networks of Māori tobacco control collectives, such as Auahi Kore groups, who met regularly at a regional and local level. In addition, interviewees reported that strong central leadership by groups such as Te Reo Mārama, a national Māori tobacco control advocacy group, and strong collaboration with advocacy organisations like ASH ensured that workers on the ground were kept well informed. Methods of dissemination mentioned included Te Reo Mārama posters, smokefree resources for use on marae, Māori radio advertising, and "loads of hui" (consultation and feedback meetings).

Not all of the Māori stakeholders thought that sufficient information was available to them to aid in implementing the new legislation. For example, one interviewee reported difficulties in working on implementation of the SEAA in Māori communities:

"Māori environments and Māori speaking environments required appropriate information and people to disseminate it. So it was easy for me as a Māori and Māori speaker, but difficult for those of my colleagues who weren't. [ There was ] not enough clarification for Māori communities; when it came to marae it was left right off the radar. The target group, as Māori, was not taken into consideration. Māori, although a minority community have a high percentage of smokers. We feel like we were targeted for change, but weren't getting the right information through to make the change."

Most stakeholders interviewed strongly supported the legislation and cited positive impacts for Māori communities and smokers. For example, some interviewees gave anecdotal evidence of increased uptake of cessation activities within District Health Boards following implementation, and increased work for Aukati Kai Paipa (a Māori smoking cessation programme).

Some of the Māori stakeholders raised concern about the new law. For example, one stakeholder was concerned that outside spaces were becoming normalised as places for smoking. One interviewee commented: "smokers are still role modelling to rangatahi (children and adolescents) about smoking" (when smoking outside). Another Māori health manager was disappointed that the legislation did not apply to some indoor areas, particularly marae. However, this was in contrast to other interviewees, who felt the approach was appropriate, and that smokefree marae should be introduced by Māori communities and not imposed through legislation.

In 2001 an estimated 25% of the Māori adult workforce was exposed to SHS in the workplace during worktime and 50% during work breaks 21. This compared with estimates of 17% of all adult workers exposed to SHS in the workplace during worktime and 33% during breaks.

Self-reported SHS exposure from others smoking indoors at work in the previous week was reported in HSC Monitor surveys from 2003–2006. Figure 2 shows that this fell from 27.2% (95% CI 22.8% to 32.1%) in 2003 to 8.9% (95% CI 6.9% to 11.4%) in 2006 among Māori workers compared with from 18.7% (95% CI 16.3% to 21.4%) to 7.5% (95% CI 5.9% to 9.4%) among all employed adults 26. By 2006 the difference in reported exposure between Māori and non-Māori was minimal. There were similar findings in the 2006 Tobacco Use Survey with 12.0% (95% CI 8.7% to 15.4%) of Māori and 7.6% (95% CI 6.3% to 8.8%) of non-Māori reporting SHS exposure (time period not specified) indoors in the workplace 7.

Data on self-reported SHS exposure in the home was available from surveys carried out by the NRB between 1989 and 1996. These revealed that exposure was common with the proportion of adults reporting regular smoking in the home being 26% in 1989, 25% in 1991 and 28% in 1996 193233, i.e. there was little evidence of change during this period. Exposure was higher for Māori (48% in 1996) and people of Pacific Island origin (34% in 1996) households compared with European/other households (24% in 1996) 19.

The HSC Monitor surveys included questions addressing SHS exposure in the home during the previous week and whether or not there were rules on smoking in the home (table 2). Self-reported SHS exposure in the home (any reported smoking by other people in the home in the last seven days) declined most steeply in absolute terms in Māori households (households where the respondent self-identified as Māori): from 31.0% (95% CI 27.3% to 34.9%) to 16.7% (95% CI 14.4% to 19.3%) between 2003 and 2006; and in Māori households with children: from 31.6% (95% CI 27.1% to 36.5%) to 13.6% (95% CI 11.0% to 16.8%). This compared to reductions from 18.0% (95% CI 16.0% to 20.1%) to 8.4% (95% CI 7.0% to 10.0%) in non-Māori households, and from 19.8% (95% CI 16.8% to 23.2%) to 8.3% (95% CI 6.2% to 11.0%) in non-Māori households with children. 26

There were similar trends seen with the largest absolute and relative reductions in reported smoking in the home in Māori households which contained one or more smokers in the HSC surveys (table 2). In households with one or more smokers SHS exposure reduced from 48.3% (95% CI 42.3% to 54.2%) to 29.3% (95% CI 25.2% to 33.8%) in Māori households, and from 40.7% (95% CI 36.3% to 45.4%) to 30.4% (95% CI 25.3% to 35.9%) in non-Māori households. There was also a reduction from 47.8% (95% CI 40.8% to 54.8%) to 23.0% (95% CI 18.4% to 28.2%) in Māori households with one or more smokers and one or more children. This compared to a decline from 38.2% (95% CI 32.1% to 44.7%) to 26.9% (95% CI 20.2% to 34.8%) in non-Māori households.

The proportion of homes reported as having smokefree policies (no smoking allowed anywhere inside) also increased during this time period. For example, the proportion of households reported with smokefree policies where there were one or more smokers and one or more children living in the household, increased from 58.7% (95% CI 51.6% to 65.4%) to 73.8% (95% CI 68.5% to 78.6%) among Māori households, and from 65.3% (95% CI 58.9% to 71.2%) to 68.4% (95% CI 60.4% to 75.5%) among non-Māori households.

Reductions in self-reported SHS exposure and increases in homes with no-smoking rules were more marked in Māori compared to non-Māori households, particularly among households with one or more smokers (table 2). However, SHS exposure in the home was still statistically significantly more common (16.7% [95% CI 14.0% to 19.3%] vs 8.4% [95% CI 7.0% to 10%]), and smokefree home policies statistically significantly less common (81.3% [95% CI 78.7% to 83.7%] vs 86.1% [95% CI 84.2% to 87.8%]), among Māori compared to non-Māori households in 2006. These figures were supported by findings from the 2006 Tobacco Use Survey, where 70.3% (95% CI 66.5% to 74.1%) of Māori and 85.1% (95% CI 83.5% to 86.6%) of non-Māori respondents reported no smoking occurring in their home 7.

However, in the HSC surveys, when the analysis was restricted to households with one or more smokers, the lower rate of reported smokefree policies in homes in Māori households in 2003 had disappeared or even reversed by 2006. For example, of households in 2003 with one or more smokers and one or more children, 65.3% (95% CI 58.9% to 71.2%) of non-Māori and 58.7% (95% CI 51.6% to 65.4%) of Māori had smokefree homes, but by 2006 the figures were 68.4% (95% CI 60.4% to 75.5%) for non-Māori and 73.8% (95% CI 68.5% to 78.6%) for Māori. These differences between Māori and non-Māori were not statistically significant.

In the ASH Year 10 survey, students were asked 'do people smoke in your home?' Compared with the findings from the HSC Monitor surveys, the downward trends in exposure using this measure were less marked. Levels of reported smoking in the home remained higher among Māori children throughout 2001–2007 (table 3). There was little change in reported exposure from 2001 to 2003, but since 2003 the proportion reporting smoking in the home has declined. The degree of decline has been similar among Māori and all children. There was a particularly marked fall in reported exposure between 2003 and 2004 among Māori 38. The fall in smoking in the home has occurred despite only minor decreases in the proportion of children reporting one or more parents smoke.

Another important source of SHS exposure is smoking in cars. However, there are no trend data available on SHS exposure in cars covering the period before and after implementation of the SEAA. In the 2006 NZ Tobacco Use Survey, the proportion of Māori reporting others smoking in the car was 30.1% (95% CI 26.6% to 33.6%) compared to 12.6% (95% CI 11.4% to 13.9%) among non-Māori 7.

The impact of the SEAA on smoking prevalence among Māori is unclear. There are three main sources of data to assess trends. Annual surveys conducted by AC Nielsen suggest that the prevalence of smoking among Māori changed little between 1990 and 2005, but may have been declining slowly since 2005 (figure 3) 39.

A question on smoking status had been included in the 1996 and 2006 New Zealand national population censuses. In 1996 smoking prevalence among all adults aged 15 years and over was 23.7%. This decreased to 20.7% in the 2006 census. The figures for Māori were 43.7% in the 1996 census and 42.2% in the 2006 census 39.

Both of these data sources suggest little change in smoking prevalence among Māori over the last decade. However, data from three recent New Zealand Health Surveys gives a different picture. Age-standardised daily smoking prevalence was 25.2% (95% CI 23.7% to 26.7%) in 1996/7, 23.4% (95% CI 22.2% to 24.7%) in 2002/3 and 18.7% (95% CI 17.7% to 19.7%) in 2006/7 in all adults aged 15 years and over. The figures for Māori adults were 46.0% (95% CI 41.8% to 50.2%) in 1996/7, 47.2% (95% CI 43.8% to 50.6%) in 2002/3, and 37.6% (95% CI 35.5% to 39.7%) in 2006/7 34. There was no statistically significant difference in smoking prevalence among Māori between 1996/7 and 2002/3, but there was a statistically significant fall in prevalence between 2002/3 and 2006/7. These data suggest a marked decline in smoking may have occurred during the period after the SEAA was implemented.

For a six-month period after implementation of the SEAA, there was evidence of increased quitting-related behaviour, with increases in caller registrations and in the issuing of NRT vouchers through Quitline 35. The increase was greater when a reduction in marketing of the Quitline in the period after implementation of the SEAA is taken into account, with Quitline caller registrations, per dollar of advertising directly linked to smoking cessation, at least a doubling in the six months after the law change compared to other six month periods between December 2002 and November 2005. The proportion of calls to the Quitline by Māori throughout the period from 2002–6, and including the 6 month period after implementation of the SEAA, was around 20% suggesting that the SEAA was an equal stimulus to quit among Māori and non-Māori 35.

Smokers were asked about the degree to which they smoked when visiting various hospitality industry venues in the 2003–6 HSC Monitor Surveys. Socially-cued smoking (smoking more frequently than normal in social settings) at bars and nightclubs declined substantially between 2003–4 and 2005–6. The decrease was present among Māori and non-Māori to a similar degree (figure 4). There were similar findings for socially-cued smoking in restaurants and cafés.