Australian attitudes to early and late abortion

Victorian abortion laws were reformed in October 2008 following a review by the Victorian Law Reform Commission.1 The laws in New South Wales, South Australia and Queensland are similar to the old Victorian laws which were criticised as being obsolete and unclear.1,2 Recently, there has been debate on abortion law reform in both Queensland and New South Wales.3,4

Previous surveys of community attitudes to abortion5-7 have several limitations.1 It is unclear what proportion of those who support the right to access abortion believe that it should be restricted on the basis of factors such as gestational age and women’s reasons for seeking abortion.1 Europeans tend to support women’s access to abortion;8 Americans are more likely to oppose it.9

Late abortion is especially controversial, although less than 2% of abortions occur at 20 weeks or later.10,11 Few data support the belief that Australians strongly oppose women’s access to late abortion,12 while surveys in the United Kingdom13 and United States14 do report opposition.

Given the limitations of existing data, community views may be misinterpreted in public debate about abortion law reform. We conducted a survey of Australian attitudes to abortion, including late abortion, during mid 2008.

An online survey of Australian adults was conducted between 28 and 31 July 2008. The target population was defined as all Australian residents aged 18 years or older. Oversampling was used in Victoria so that more statistically accurate results could be gained for Victoria, in light of its pending abortion law reforms. To provide representative national results, the oversample for Victoria was weighted to reflect (in addition to other demographic criteria) the proportion of the Australian population residing in Victoria.

The survey was conducted by a private market research organisation (Crosby Textor, Sydney, NSW) with experience in measuring public opinion on health and medical issues. An online panel of more than 500 000 recruited participants (PermissionCorp, Sydney, NSW) was used as a representative sample frame for this study. Participants on this panel elect to either receive awards points or participate in a prize draw in return for completing a survey. A random sample (stratified by age, sex and location) was invited to participate, provided with a short preamble regarding the subject matter of the survey, and then asked for their consent to participate.

Respondents were asked to confirm their background information, such as age, sex and location, and provided with factual information about the proposed law change in Victoria. They were then asked about their views on abortion. To avoid invalid or prompted answers, respondents were not permitted to revise previous answers once they had proceeded to a new question. The questionnaire was tested before full online launch, and standard quality control procedures were applied.

The survey data were weighted15,16 by sex, age and location according to the latest available data from the Australian Bureau of Statistics;17 notably, the oversample for Victoria was weighted to reflect actual population proportions of sex, age and location. The weighted population breakdown was as follows:

• Sex: male, 49%; female, 51%

• Age: 18–19 years, 4%; 20–24 years, 9%; 25–29 years, 10%; 30–34 years, 9%; 35–39 years, 11%; 40–44 years, 9%; 45–49 years, 10%; 50–54 years, 9%; 55–59 years, 8%; 60–64 years, 7%; ≥ 65 years, 14%

• Location: New South Wales, 33% (21% Sydney, 12% elsewhere); Australian Capital Territory, 2%; Victoria, 24% (18% Melbourne, 6% elsewhere); Queensland, 20% (9% Brisbane; 11% elsewhere); Western Australia, 10% (7% Perth, 3% elsewhere); South Australia, 8% (6% Adelaide, 2% elsewhere); Tasmania, 2%; Northern Territory, 1%.

Weighted frequencies and cross-tabulations were then used to produce results as percentages. A summary of the results has been reported elsewhere.18

In accordance with the National Health and Medical Research Council National statement on ethical conduct in human research,19 ethics approval was not sought because the survey was anonymous and followed procedures that are usual for public opinion surveys.

Four per cent of individuals who were invited to participate in the survey declined, and 15% of started surveys were not completed. Of the 1050 respondents who completed the survey, 526 resided in Victoria and 524 resided in other states and territories. After weighting the oversample for Victoria, the effective national sample size was 798.15 Weighting factors other than that relating to location fell within acceptable margins (0.96–1.04).

The maximum margin of error for this effective national sample size of 798 is a ± 3.5 percentage point confidence interval20 in 95 of 100 cases (at the 95% confidence level)21 with results of about 50%. Margins of error are smaller for results further from 50%, but may increase for analyses of subsamples. For example, the effective oversample for Victoria of 523 has a maximum margin of error of ± 4.3 percentage points in 95 of 100 cases with results of about 50%, and the effective sample for other states and territories of 505 has a maximum margin of error of ± 4.4 percentage points in 95 of 100 cases with results of about 50%.

Our survey shows a high level of support for access to early abortion: 87% of respondents indicated that abortion should be lawful in the first trimester (61% unconditionally and 26% depending on the circumstances). There was little support for professional sanctions against doctors for providing terminations after 24 weeks’ gestation. When asked to consider specific, realistic situations in which late abortion might be considered, many respondents opposed sanctions against doctors, particularly when abortion is sought because of maternal or fetal complications rather than personal reasons. Respondents with a religious affiliation were slightly less likely to oppose sanctions than those without a religious affiliation.

Public opinion research can have a major impact on government policy and, therefore, on access to quality medical care. Although results of such research are quoted widely in medical literature, they are not usually peer reviewed. To our knowledge, our study is the first detailed survey of Australian attitudes to late abortion that includes attitudes in various clinical and social situations. Our findings challenge the belief that Australians strongly oppose women accessing late abortion.12

Limitations of our study include that it is cross-sectional at a single time point and small. Although the margins of error were small, these apply only to measuring a proportion based on the total sample. However, the sample size was sufficient to enable a high level of confidence, which makes the data generalisable to the Australian population. The 4% opt-out rate in our study minimised potential sample bias, and the 15% drop-out rate was not unusual for this data collection method.

Online data collection might be replacing paper-and-pencil surveys in academic research.22 It enables more candid and considered answers to highly personal and potentially confronting issues. Potential problems associated with online survey research, many not unique to this approach,22 include the use of volunteer sampling rather than probability sampling, the possibility that respondents are not representative of the target population, and the fact that not everyone has internet access. However, there is no conclusive evidence that responses differ between mail and online surveys,23 and data collected online may be more complete.24

Previous surveys in Australia5-7 have not usually specified the gestational age at which an abortion is performed. In the 2003 Australian Survey of Social Attitudes, 81% of respondents agreed or strongly agreed that a woman should have the right to choose whether or not she has an abortion; 9.4% disagreed or strongly disagreed.5 In the 2004 Australian Election Study, 89% of respondents said that women should be allowed to have an abortion, either readily when they want one (54%) or in special circumstances (35%); only 4% said abortion should not be allowed under any circumstances.6 In contrast, the Australian Federation of Right to Life Associations survey7 found that only 6% of respondents agreed with allowing abortion after 20 weeks’ gestation. Problems with the latter survey may have resulted in bias.1

Our data suggest that single general questions, as used in previous surveys, provide a limited view of community sentiment, as has previously been noted.1 Simple yes/no polls do not allow people to accurately express the subtlety of their views in the complex range of clinical and social situations in which access to abortion might be sought. For example, although 48% of respondents indicated that abortion in the third trimester of pregnancy should be unlawful, less than a quarter indicated that a doctor should face professional sanctions for performing termination after 24 weeks’ gestation when there is a risk to the physical or mental health of the woman or baby.

Respondents were more equivocal about sanctions in scenarios regarding late termination for reasons relating to the preferences or social circumstances of the woman. But less than 50% of respondents indicated that a doctor should face professional sanctions for performing a termination after 24 weeks’ gestation, even when there is no medical reason for the termination.

One possible explanation for the difference in responses to questions about lawfulness and sanctions is that respondents drew a distinction between legalising abortion and removing sanctions. Respondents with ambivalent attitudes to abortion may favour removal of sanctions but retention of a legal bar on abortion as an intermediate position between full legalisation and full prohibition. This could reflect the prevailing semi-legal status (which was in place until recently in Victoria, and is still in place in most other Australian jurisdictions), where abortion is illegal under at least some laws but professional sanctions are rarely, if ever, enforced.

Another possibility is that respondents may have an in-principle objection to abortion but are more prepared to accept it when they understand the reason, especially if there is a medical reason for abortion. Support for abortion was generally higher when respondents were given greater information about timing and reasons. We speculate that this higher level of support could have been because respondents had a greater ability to identify with women seeking an abortion when contextual information was present, or because respondents thought that the contextual details were morally relevant factors — that is, that they provided moral reasons for allowing abortion. It is possible that, when no context is given, respondents tend to assume there are no significant moral reasons for abortion to proceed.

The sensitivity of Australians’ views on abortion to contextual details may have implications for other debates about ethics. The more permissive attitude elicited when context was provided in our study may, for example, carry over to debates about euthanasia, the use of medicine or technology for human enhancement, organ donation, and embryonic stem cell research. It is an open question whether policy formation should be informed by views about abstract propositions or about cases containing contextual details.

Simple yes/no polls, as used in previous surveys, may give a misleading picture of public opinion, as a simplistic division between “pro-choice” and “anti-abortion” does not accurately reflect the views of Australians. Individuals have nuanced views that depend on the reasons for which women seek abortion. Nonetheless, opinion surveys — no matter how robust — should not dictate policy or law. Policies and laws should be grounded on ethical arguments. We have attempted to provide such arguments, adding to previous discussions of abortion and the law.2,25,26 This study has shown that Australians are supportive of both access to abortion, including late abortion in many circumstances, and liberal abortion law reform.