When Is A Request for Assisted Suicide Legitimate? Factors Influencing Public Attitudes Toward Euthanasia
A mail survey of 810 Greater Vancouver Area residents investigated how the public's acceptance of a request for euthanasia was influenced by the method of death (e.g., lethal injection vs. withdrawal of life - support) and by the identity of the patient featured in a vignette (e.g., stranger vs. oneself). The study also identified considerations people found most important in deciding whether a patient's request for euthanasia was legitimate (e.g., patient's pain, chance for recovery). Life - support withdrawal was found significantly more acceptable (90% support) than a lethal injection (79% support), yet the identity of the person involved did not affect the acceptability of euthanasia. However, a factor analysis suggested that the decisions about oneself may be more complex and more closely scrutinized than decisions about others. The considerations rated most important by participants paralleled legal guidelines from the Netherlands and Oregon.
Euthanasia certainly is not a recent concept, but the paradoxical stand of modern medicine, torn between an ever increasing capacity to prolong life and an inability to cure a number of debilitating diseases, has given the ending of life an entirely new meaning. In recent years, a number of highly publicized euthanasia cases have occurred in Canada (N.B. v Hotel - Dieu de Quebec, 1992; Rodriguez v. British Columbia, 1993; R. v. Latimer, 1995), stirring public controversy and heating up a sensitive debate between proponents and opponents of the right to die. This debate revolves around key arguments such as the difference between withdrawing treatment and actively inducing death, patients' rights to autonomy and privacy, the quality of life of the terminally ill, and the risk of misuse and abuse for the diseased, the elderly, and the handicapped (Chipeur & Maxwell, 1994; Hollander, 1989; Pellegrino, 1989; Rodriguez v. British Columbia, 1993; Yarnell & Battin, 1988).
ACTIVE VERSUS PASSIVE EUTHANASIA
A great source of controversy in the right to die debate pertains to the distinction between active and passive euthanasia. Active euthanasia involves "a deliberate act undertaken by one person with the intention of ending the life of another person to relieve that person's suffering" (Senate of Canada, 1995, p.14) (e.g., lethal injection), whereas passive euthanasia is "not starting ... or stopping treatment that has the potential to sustain the life of a patient" (Senate of Canada, 1995, p.14). Assisted - suicide, which involves providing someone with the means, advice, or any other form of help in committing suicide, fits into the active category.
Opinions among the general public and health - care professionals generally suggest that "inducing the death" of a patient is perceived to be worse than "doing nothing" to prevent a patient from dying (Bosmann, Kay, & Conter, 1987; Rachels, 1975; Sugarman, 1986; Winkler, 1985). The American, British, Canadian and World Medical Associations' official position is that the deliberate ending of life is unethical, regardless of the patient's request. These associations, however, grant that it is permissible to honour a terminally ill patient's request to let the disease follow its course (see Rodriguez v. British Columbia, 1993; Williams, 1991). In many jurisdictions, health - care providers who withdraw life - sustaining devices and issue "do not resuscitate" orders are no longer condemned by either civil or criminal courts, whereas those who participate in assisted suicide could face criminal and civil liability (Latimer, 1991; Lynn, 1988; Rachels, 1975; Wanzer, Federman, Adelstein, Cassel, Cassem, Cranford et al. 1989).
Many argue, however, that the distinction between active and passive euthanasia is morally irrelevant if the intrinsic goal of both is to eliminate human suffering (Lynn, 1988; O'Rourke, 1991; Rachels, 1975; Winkler, 1985). Some further argue that once the decision has been made not to prolong life, circumstances may even justify resorting to active rather than passive euthanasia to avoid prolonging agony (Rachels, 1975).
Regardless of the debate, a clear line between passive and active euthanasia may be difficult to draw (Lynn, 1988; Winkler, 1985). Aggressive treatment of the terminally ill, which may involve administration of lethal doses of sedatives that depress respiratory functions and precipitate death, is not considered killing because the primary intent is to provide pain relief (Klagsbrun, 1991; Latimer, 1991; Lynn, 1988; O'Rourke, 1991; Rachels, 1975; Roscam - Abbing, 1988). Nonetheless, such aggressive treatment may not only "kill" but also violate the right to privacy if performed without specific considerations for the patient's wish and needs (Latimer, 1991).
THE RIGHT TO AUTONOMY AND PRIVACY
The right to autonomy advocated by our society recognizes people's access to accurate information and their capacity to make health and treatment decisions (Klagsbrun, 1991; Latimer, 1991). The right to privacy further insures protection against undesired, invasive treatment (Wanzer et al., 1989). In recognition of these rights, both the Law Reform Commission of Canada (1982) and the U.S. Presidential Commission (see Winkler, 1985) recommend that competent patients be given the right to refuse life - sustaining treatment. Many states have enacted laws providing for living wills by which patients refuse the resort to heroic measures for their care in the event of terminal illness, and laws protecting physicians from possible action by the patient's family when the physicians comply such wills (Jackson & Youngner, 1979; Wanzer et al., 1989; Zucker, 1977).
End - of - life decisions are highly personal, and previous research highlights possible differences in attitudes toward assisted suicide and euthanasia for oneself and for others. Kinsella and Verhoef (1993) found that physicians were more uncertain about considering euthanasia for a close relative than for themselves or for an immediate parent. Blendon, Szalay and Knox (1992) suggested that Americans think about right - to - die issues in two very distinct ways: as a public policy issue (options and legal rights that should be available to all) and as an issue of personal choice (options individuals would chose for themselves if they were terminally ill or irreversibly comatose). Wade and Anglin (1987) attempted to identify situations that would lead people to endorse euthanasia for themselves and for their parents and found that different decisional criteria were used for self than for parents.
Nevertheless, personal autonomy has its limits: Courts and practitioners agree that personal autonomy sometimes needs to be restricted for the greater good of society (Latimer, 1991; Pellegrino, 1989; Rodriguez v. British Columbia, 1993). Although most health care professionals feel that living wills should be legally acceptable, only a slight majority of them agree that they should be legally binding (Bosmann, Kay, & Conter, 1987). Many fear that the legalization of active euthanasia would result in a "slippery slope," leading to the possible misuse of euthanasia against vulnerable members of our society (Chipeur & Maxwell, 1994; Hollander, 1989; Pellegrino, 1989; Rodriguez v. British Columbia, 1993; Yarnell & Battin, 1988). The so - called "right to die" should not become a "duty to die" (Van Der Sluis, 1988, p. 108).
ESTABLISHMENT OF GUIDELINES FOR ASSISTED SUICIDE AND EUTHANASIA
A patient's request for euthanasia, even when clearly expressed, does not automatically render euthanasia ethical or acceptable (Brescia, 1991; O'Rourke, 1991). When the wish to die is expressed, questions arise regarding whether this request stems from potentially treatable depression, pain or disease symptoms (Chochinov, Wilson, Enns, Mowchun, Lander, Levitt, & Clinch, 1995; Jackson & Youngner, 1979; Klagsbrun, 1991; Quill, 1991; Rabkin, Gillerman, & Rice, 1976; Wanzer et al., 1989), from an attempt to mask other, less socially acceptable problems, or from a fear of treatment based on misperception or misinformation (Jackson & Youngner, 1979; Wanzer et al., 1989). Many agree that physicians ought to investigate such motives closely before deciding what action to take (Quill, 1991; Rabkin et al., 1976).
Health care professionals generally believe that the establishment of clear guidelines for euthanasia would facilitate decisions about whether life should be prolonged, especially in cases involving incompetent patients (Bosmann et al., 1987; Cassel & Meier, 1990; Yarnell & Battin, 1988). Such guidelines exist in the Netherlands, where assisted - suicide has been sanctioned by the state in cases meeting specific criteria since 1986 (Angell, 1988; Wanzer et al., 1989). These criteria, revised in 1993 (Griffiths, 1994), are as follows: (a) the patient's request must be voluntary, stable and enduring; (b) the patient must be experiencing suffering that is, in the patient's view, intolerable; (c) all alternatives acceptable to the patient for relieving the suffering must be tried first; (d) the patient must be competent and fully informed; (e) the physician must consult with a second physician, whose judgement can be expected to be independent, and with the patients' relatives. These guidelines include an expectation that physicians who become involved in instances of active euthanasia or assisted suicide report to the judicial authorities (Van Der Maas, Van Delden, Pijnenborg & Looman, 1991).
In the United States, propositions for the legalization of physician- assisted suicide were defeated in 1991 in the state of Washington and in 1992 in California, while a more recent attempt was approved in Oregon, in November 1994. The successful approval of the Oregon proposal suggests that adequate safeguards might be the key to making such alternative acceptable to the public. Among the guidelines proposed, the three pivotal ones are: (a) the patient must have a life expectancy of less than six months; (b) the patient must request a lethal dose of drug on three separate occasions; and (c) the patient must inject the dose him/herself (Annas, 1994). A recent decision by the Ninth Circuit Court of Appeals further strengthened the right of mentally competent, terminally ill patients to seek a physician's assistance in dying by striking down the statute prohibiting physician - assisted suicide in the State of Washington (Compassion in Dying v. State of Washington, 1996).
The Present Study
In light of the arguments reviewed above, how does the public decide whether euthanasia can legitimately be requested by a patient? The current study investigated how public attitudes toward euthanasia were influenced by (1) the method of death to be employed (e.g., life - support withdrawal vs. lethal injection); and (2) the identity of the patient making the request (e.g., stranger vs. self). Most of the studies reviewed did not explicitly compare acceptance of passive versus active euthanasia (Adams, Bueche, & Schvaneveldt, 1978; Bosmann et al., 1987; Ho & Penney, 1991; Jorgenson & Neubecker, 1980/1981; Shuman, Fournet, Zelhart, Roland & Estes, 1992), and many failed to provide a clear definition of both (Devins, 1980/1981; Klopfer & Price, 1979; Sawyer, 1982; Singh, 1979; Wade & Anglin, 1987). No previous work has focussed on the self versus unknown other distinction.
The current study also examined some of the circumstances that are taken into consideration when individuals are asked to judge the acceptability of a request for euthanasia. To identify the range of such considerations, a pilot study was conducted which included an extensive literature review and a series of semi - structured interviews with 40 volunteers recruited among students and staff on campus.
Other than voluntariness of the request and euthanasia method, 14 situational considerations were identified from the interviews. They included: the severity of the physical painthe patient is in; the patient's degree of physical dependency (e.g., being bedridden, being fed); the patient's mental alertness or capacity for informed consent; the patient's chances for recovery and life expectancy; current laws prohibiting euthanasia and assisted suicide; the cost of life - support treatment; the severity of the patient's psychological suffering (e.g., depression, anxiety); the age of the patient; the etiology of the disease (e.g., hereditary vs. lifestyle); considerations given to alternative treatments; considerations given to other pain - relieving treatments; the stability over time of the patient's request (e.g., not a spur of the moment decision); the diagnosis and how much is know about the patient's disease; and the opinions of friends, family members, or people emotionally close to the patient.
Finally, in keeping with previous research, this study documented some demographic predictors of attitudes toward euthanasia. Past studies found support for the influence of religiosity (i.e., the extent of one's commitment to a religious organization) (Adams et al., 1978; Anderson & Caddell, 1993; Jorgensen & Neubecker, 1980/1981; Ostheimer & Moore, 1981; Shuman et al., 1992; Singh, 1979; Wade & Anglin, 1987); religious denomination (Ostheimer & Moore, 1981; Singh, 1979; Wade & Anglin, 1987); age (Adams et al., 1978; Devins, 1980/1981; Haug, 1978; Klopfer & Price, 1979; Slezak, 1982); and education (Ho & Penney, 1991; Pollard, 1994).
It was hypothesized that: (a) passive euthanasia would be judged more acceptable than active euthanasia; and that (b) religiosity would be the best demographic predictor of euthanasia opinions and inversely related to the acceptability of euthanasia. In addition to the formal hypotheses, it was speculated that: (c) the acceptability of euthanasia might differ as a function of the identity of the patient; and that (d) some situational considerations would systematically emerge as being most important in judging the acceptability of a request for euthanasia, and might cluster differently depending on the identity of the person involved.
Questionnaires were mailed to 2000 people selected randomly from a directory of households in the Greater Vancouver Area. Each respondent was mailed a cover letter, a questionnaire, and an addressed, postage - paid return envelope. One month later, another package was mailed to all respondents to maximize the response rate. Twenty - two (2.65%) of the returned questionnaire were discarded for the following reasons: 15 because more than half the questions were left unanswered; two because the respondent was under 18 years of age; and five because the questionnaire was returned after the deadline for data collection. The final sample used for statistical computation comprised 810 questionnaires, representing 43% of the deliverable questionnaires.
Fifty - five percent of the 810 respondents were men, 44% were women, and 1% did not indicate gender. Participants ranged in age from 19 to 95 years (M = 48.16 years, SD = 16.52 years).
Four different versions of a questionnaire were developed by crossing both levels of the method variable (e.g., "active" versus "passive") with both levels of the person featured (e.g., "self" versus "other") and the euthanasia method used. In two scenarios, the patient was described simply as "Chris"(f.1) (e.g., "other" condition). In the other two scenarios, respondents were asked to imagine themselves as terminally ill (e.g., "self" condition). Both levels of this "person" variable were crossed with both levels of euthanasia method, namely a lethal injection (e.g., "active") or the withdrawal of life - support (e.g., "passive").
An example of one of these scenarios is as follows (i.e., other/active euthanasia condition):
Chris is a single parent of one. Chris suffers from a terminal illness which will likely result in the loss of the ability to swallow, speak, move, or even breathe without assistance, and which will almost certainly result in Chris being confined to bed and being maintained on life - support before dying. Chris wishes to remain alive only so long as life can be enjoyed. When Chris reaches the point when life cannot be enjoyed anymore, Chris also will be unable to commit suicide alone, and Chris wishes to have a physician's assistance in getting an injection of a lethal drug which will terminate Chris' life within a few hours of its administration."
After reading the brief scenario, respondents were asked to rate the acceptability of the request for euthanasia described, on a scale from 1 (not at all acceptable) to 7 (absolutely acceptable).
Following this opinion question, respondents were asked to rate, on a scale from 1 (not at all important) to 7 (absolutely important), the importance of the 14 considerations identified during the pilot work. A Cronbach alpha reliability analysis revealed a high level of internal consistency between these considerations (r = .885, n = 787).
The last section of the questionnaire documented demographic characteristics of the respondents including religiosity, religious affiliation, age and level of education. Questions about gender and occupation were included to allow discussion of the generalizability of the findings and comparisons with previous research.
RESULTS Demographic composition of the four subsamples
A 2 (method) x 2 (person) ANOVA revealed that the groups differed significantly on religiosity (two - way interaction; F(1,781) = 5.40, MS'Symbol not transcribed'e = 44.54, p < .05, n = 781), and chi - square analyses revealed that the group differed significantly on gender (X'Symbol not transcribed'2(1) = 5.54, p < .05, n = 801) and level of education, 'Symbol not transcribed'2(6,797) = 12.09, p < .05. However, since oneway ANOVAS revealed that neither gender nor education significantly influenced opinions toward euthanasia, all demographic characteristics except for religiosity are reported for the sample as a whole in Table 1.
Influence of Demographics on Acceptability of Euthanasia
A Pearson product moment correlation performed on the sample as a whole revealed a significant negative relationship between religiosity and acceptability of euthanasia (r = - .40, p < .01). Since groups differed in mean degree of religiosity, scenarios were then analyzed separately. In all four conditions, religiosity was found to be significantly negatively correlated with acceptability for euthanasia (p < .01; active/other: r = - .48; passive/other: r = - .30; active/self: r = - .51; passive/self: r = .33; all effects sizes medium to large, Cohen, 1992). Age, however, was not found to be significantly correlated with acceptability of euthanasia.
Oneway ANOVAs on opinion ratings were performed for religion on the sample as a whole. Religion was coded into four categories: No Religion, Roman Catholic,(f.2) Protestant, and Other.(f.3) Religion was found to have a significant effect on acceptability of euthanasia (F(3,750) = 19.74, p < .0001). Post - hoc Neuman - Keuls pairwise comparisons revealed that Roman Catholics (M = 5.01, SD = 2.08, n = 85) and respondents committed to "Other" religions (M = 5.13, SD = 2.11, n = 55) found euthanasia significantly less acceptable than Protestants (M = 5.62, SD = 1.86, n = 302). Furthermore, respondents committed to any religion (Roman Catholic, Protestant and Other) all were found to be significantly less accepting of euthanasia than non - committed respondents (M = 6.31, SD = 1.24, n = 304).
Difference in Acceptability of Euthanasia as a Function of Method and Person
A 2 (method) x 2 (person) ANOVA was performed on the acceptability ratings of euthanasia requests. Only the main effect for method of euthanasia was significant, F(1,786) = 21.266, p < .001, MS'Symbol not transcribed'e = 62.69. The mean acceptability of passive euthanasia (M = 6.07, SD = 1.44, n = 408) was significantly higher than mean acceptability for active euthanasia (M = 5.51, SD = 1.97, n = 379).
Difference in Importance of Situational Considerations as a Function of Method and Person
A 2 (method) x 2 (person) MANOVA was performed on the importance ratings given on a 7 - point scale to each of the 14 situational considerations listed on the questionnaires. The analysis revealed a main effect for the person variable only (Hotelling T'Symbol not transcribed'2 (770) = 0.351, p < .001).
This MANOVA was followed up by a series of 2 (method) x 2 (person) ANOVAs performed on the 9 considerations for which a significant main effect emerged. These ANOVAs revealed that the following nine considerations were rated as significantly more important when making a decision about "self" than "other": degree of physical pain (other: M = 4.95, SD = 2.17, n = 385; self: M = 5.52, SD = 1.75, n = 424); degree of physical dependency (other: M = 4.98, SD = 2.10, n = 386; self: M = 5.78, SD = 1.60, n = 424); chance of recovery (other: M = 5.76, SD = 1.82, n = 384; self: M = 6.08, SD = 1.53, n = 423); cost of treatment (other: M = 3.79, SD = 2.15, n = 386; self: M = 5.14, SD = 1.80, n = 422); age (other: M = 3.04, SD = 1.92, n = 386; self: M = 3.95, SD = 2.03, n = 424); opinions of friends and relatives (other: M = 3.92, SD = 1.97, n = 386; self: M = 4.85, SD = 1.73, n = 423); etiology (other: M = 2.99, SD = 2.16, n = 386; self: M = 4.01, SD = 2.16, n = 423); diagnosis (other: M = 4.28, SD = 2.29, n = 384; self: M = 5.67, SD = 1.82, n = 423); and pain treatment (other: M = 5.28, SD = 1.87, n = 384; self: M = 5.94, SD = 1.30, n = 424).
Clustering of Situational Considerations
A series of exploratory principle component factor analyses with oblique rotations(f.4) were performed on the importance ratings of the 14 situational considerations for each scenario individually. This was done to see if decisions were arrived at by a similar process, regardless of the method of euthanasia and the person involved. A factor loading criterion of .5 or higher was used to include a variable in a factor. None of the variables loaded highly on more than one factor. The models that best fit the data and explained most of the variance were a two - factor model for scenarios involving a stranger, and a three - factor model for scenarios involving self.
In the two - factor model that emerged for the "other" scenarios, the first factor was mostly a grouping of symptoms experienced by the patient and treatment related considerations, and was therefore labelled "internal/symptoms/treatment" (see Table 2 for factor loadings). The second factor was mostly a grouping of societal considerations and was therefore labelled "external/society." At first glance, diagnosis and etiology, which are disease - related, look like they would better belong to Factor 1. However, they constitute external labels that identify an illness and its origin, and represent a social convention rather than a physical experience, which may explain why they load on to the second factor (see Table 2).
For the three - factor model that emerged in the "self" scenarios, the first and second factors were comparable to the two factors which emerged in the "other" conditions. However, three variables, namely considerations for the treatment of pain, considerations for alternative treatment and diagnosis, loaded onto a third factor. The first factor, therefore, was labelled "internal/symptoms", the second was labelled "external/society," and the third was labelled "treatment" (see Table 3 for loadings).
Consistent with previous surveys (Bosmann et al., 1987; Ho & Penney, 1991), and with the stance of many medical associations (e.g., American, British, Canadian and World Medical Associations) and legal commissions (Canadian Law Reform Commission and U.S. Presidential Commission), the present study suggests that in the eyes of the public, ending life by means of a lethal injection is significantly less acceptable than withdrawing life - support, regardless of who is requesting such procedure. The difference in acceptability between active and passive euthanasia suggests the public still perceives an important distinction between "inducing death" and "letting die" (Latimer, 1991; O'Rourke, 1989; Rachels, 1975; Winkler, 1985).
This study also speaks to the highly personal nature of end - of - life decisions as results suggest that decisions about oneself are more closely scrutinized than decisions about others. The majority of considerations listed in the questionnaire were judged significantly more important when making a decision about "self" than "other," and the factor analyses revealed a more complex factor structure for "self" scenarios. Although the latter stands in contrast to previous findings by Wade and Anglin (1987), which suggested a four - factor model best fit endorsement of euthanasia for both "self" and "other," it is noteworthy that Wade and Anglin's (1987) "other" condition involved parents rather than a stranger. In the current study, it is not surprising that people may have appraised the acceptability of euthanasia for a complete stranger differently than people asked to appraise the acceptability of euthanasia for their parents. This distinction between decisions involving oneself versus those involving a stranger is consistent with conclusions of previous surveys, which suggested that attitudes toward euthanasia as a personal option were often different, and influenced by different factors, than attitudes toward euthanasia as a public policy issue (Blendon et al., 1992; Kinsella & Verhoef, 1993).
The most important considerations in judging the acceptability of a request for euthanasia were, in order: chance for recovery, mental alertness, considerations for alternative treatments and for pain - relieving treatment, psychological suffering and stability over time of the patient's request for euthanasia. The current study was the first to look at such ordering, which closely parallels the guidelines used in the Netherlands to regulate assisted suicide (see Angell, 1988). The Dutch guidelines stress the importance of insuring that the patient's request is voluntary, stable and enduring, that the patient is competent and capable of making an informed decision, and that all other treatment alternatives acceptable to the patient have been tried. However, the Dutch guidelines do not specifically state that the patient should be suffering from a terminal illness, a provision which was however included among the guidelines approved in Oregon (Annas, 1994) and emphasized in the recent 9th Circuit Decision (Compassion in Dying v. State of Washington, 1996).
The emphasis respondents placed on considering the psychological well - being of patients requesting euthanasia echoes an ongoing debate in the Netherlands pertaining to the mandatory involvement of a psychiatrist to assess psychiatric problems which may affect decision - making capacities, or the desire to die when patients request treatment, withdrawal or assistance in dying (Hendin, 1994; Huyse & Hengelveldt, 1989; Moldawski, 1993). The importance placed on a psychological evaluation also reflect arguments in favour of euthanasia which stress that the loss of a sense of "selfhood" or of a sense of "purpose" would be an important trigger for many in considering requesting euthanasia (Mullens, 1996; Seale & Addington - Hall, 1994).
As hypothesized, the best demographic predictor of attitudes toward euthanasia was religiosity, which was inversely related to the acceptability of euthanasia. This finding was consistent with findings from a number of previous surveys (Adams et al., 1978; Jorgenson & Neubecker, 1980; Ostheimer & Moore, 1981; Shuman et al., 1992; Singh, 1979; Wade & Anglin, 1987). A self - description as Roman Catholic also was found to reduce support for euthanasia, as suggested previously (Ostheimer & Moore, 1980; Pollard, 1994). Other studies may have failed to detect such differences due to methodological problems such as a greater degree of within - group than between - group variance (Wade & Anglin, 1987), or such as a dichotomization of the religion variable into religious versus non - religious (Singh, 1979), which may have masked possible differences between various religious affiliations.
Among the study's limitations, the most notable is, first, the representativeness of the sample. It is noteworthy that a response rate of 43% is high for a mail survey, considering the controversial nature of the topic and the lack of specificity of the sample consulted. However, the few phone calls and written comments received by the investigators suggested that those who strongly opposed euthanasia were more reluctant to complete the survey than those who were in favour of it. Some respondents indicated that encouraging research on the issue equated showing support for it and, therefore, refused to participate. This perception also may reflect that many people have not yet resolved this issue for themselves.
In terms of demographics, respondents tended to be older, more educated, and more often professionals than the population of British Columbia (Statistics Canada, 1994). Furthermore, residents of British Columbia may have had greater exposure to the issue than Canadians living in other provinces, via the extensive media coverage of the case of Sue Rodriguez and of the vocal interest of a local MP, Svend Robinson.
Second, the use of specific definitions for active and passive euthanasia restricts the generalizability of the current findings to other forms of euthanasia. However, the pilot study highlighted the need to use specific definitions and examples of euthanasia to avoid confusion.
Third, the distinction between decision - making for "self" and "other" may have been artificial. People responding to the "self" scenarios may have found it difficult to imagine themselves as terminally ill, as suggested by a number of people interviewed during the pilot study. It is possible that both respondents who read the "other" and "self" scenarios based their responses on their personal experience with death or euthanasia involving someone close to them.
Further research is needed to explore the influence of variables such as experience with terminal illness and death among friends and family members and ethnicity on attitudes toward euthanasia. Experience with death and dying patients has been found to influence to some extent attitudes toward, and decisions about, euthanasia (Anderson & Caddell, 1993; Breitbart, Rosenfeld, & Passik, 1996; Shuman et al., 1992; Slezak, 1982). However, questions about death and euthanasia experience were excluded from the mail questionnaire because of the lack of opportunity for individual debriefing and the emotional responses such questions elicited during the interviews. Similarly, although the need to explore the influence of cultural beliefs on attitudes toward the termination of life has been increasingly recognized (Breitbart et al., 1996; Seale & Addington - Hall, 1994), questions about ethnicity and culture were not included as the brevity of a mail survey did not allow for an in - depth exploration of various cultural beliefs and of the extent to which respondents may or may not endorse such beliefs. The influence of cultural beliefs may be of particular interest in the context of the…