partial and impartial ethicalreasoningin healthcare …in moral orientation. so, even if womenare...

J7ournal ofMedical Ethics 1997; 23: 226-232

Partial and impartial ethical reasoning inhealth care professionalsHelga Kuhse, Peter Singer, Maurice Rickard, Leslie Cannold and Jessica van Dyk

Centre for Human Bioethics, Monash University, Victoria, Australia

AbstractObjectives - To determine the relationship betweenethical reasoning and gender and occupation among agroup of male andfemale nurses and doctors.Design - Partialist and impartialistforms of ethicalreasoning were defined and singled out as being centralto the difference between what is known as the "care"moral orientation (Gilligan) and the 'justice"orientation (Kohlberg). A structured questionnairebased on four hypothetical moral dilemmas involvingcombinations of (health care) professional, non-professional, life-threatening and non-life-threateningsituations, was piloted and then mailed to a randomlyselected sample of doctors and nurses.Setting - 400 doctors from Victoria, and 200 doctorsand 400 nurses from New South Wales.Results - 178 doctors and 122 nurses returnedcompleted questionnaires. 115 doctors were male, 61female; 50 nurses were male and 72 were female. It washypothesised that there would be an association betweenfeminine subjects and partialist reasoning and masculinesubjects and impartialist reasoning. It was alsohypothesised that nurses would adopt a partialistapproach to reasoning and doctors an impartialistapproach. No relationship between any of thesevariables was observed.

IntroductionRecently, there has been considerable interest in therelationship between occupation and ethical reason-ing. This is particularly so with the health care pro-fessions, where doctors and nurses deal withsignificant moral problems each day. A large body ofwork has developed in the wake of LawrenceKohlberg's, and later, Carol Gilligan's research onmodes ofmoral reasoning, and it is of interest here toinvestigate how those modes are distributed amongdoctors and nurses.

In keeping with Piaget's model of cognitive devel-opment, Lawrence Kohlberg devised an influential

Key wordsEthical reasoning; health care ethics; justice and care;Kohlberg; Gilligan.

six-staged developmental theory of moral reasoning.While the cognitive stages postulated by Piaget wereintended to reflect advances in reasoning aboutlogical relations, Kohlberg takes his moral stages torepresent advances in reasoning about justice, arguingthat justice is the fundamental principle of morality.'Because Kohlberg conceived of moral thinking interms of justice, his experimental measure for an indi-vidual's level of moral reasoning - the MoralJudgement Interview - presented subjects with moraldilemmas that generally involved competing rights inorder to elicit from them reasoning relating tofairness, duty, equality and justice. Kohlberg definedand ordered his developmental stages according tothe idea that the more morally mature one is, themore one's moral thinking will be informed byabstract, "rational", objective, and universalisableconsiderations or principles. There were early indica-tions, though, that women generally scored lowerthan men on Kohlberg's measure, with fewer womendeveloping beyond Kohlberg's "conventionalist"Stage 3 of conformity to interpersonal expectations.23Rather than taking this as a reflection of women'slesser capacity for moral development, Carol Gilliganargued that women's lower scores reflected a differ-ence in the overall moral orientations of men andwomen. 456 Gilligan based this view partly on inter-view data gained from women freely discussing theirabortion decisions. It emerged from these interviewsthat women understood the moral problem con-fronting them not in terms of abstract considerationssuch as rights, duties and justice, but in the moreconcrete and contextual terms of their perceivedresponsibilities to care for and avoid hurting those inrelationship with them. Gilligan theorised that this"care" orientation in moral thinking is gender relatedand that it also admits of developmental stages linkedto the formation of women's self-concept whichcentres on interpersonal connectedness and mainte-nance of relationships, rather than on separation,individuality and autonomy that is typical of maturemale ego-identity. The care perspective is charac-terised by Gilligan in terms of a number of themeswhich are in direct opposition to Kohlberg's ethic ofjustice. Where mature moral reasoning on the justiceorientation is typified by reliance on principled,

on July 3, 2020 by guest. Protected by copyright.

http://jme.bm

j.com/

J Med E

thics: first published as 10.1136/jme.23.4.226 on 1 A

ugust 1997. Dow

nloaded from

http://jme.bmj.com/

Helga Kuhse, Peter Singer, Maurice Rickard, Leslie Cannold and _essica van Dyk 227

abstract, universalisable, impartial and non-affectivethinking, the care orientation involves an emotionallyinformed engagement with the concrete details of theparticular situation at hand. Although there is nowevidence that women do not consistently score lowerthan men on Kohlberg's test789 but sometimeshigher,' 11 there has nevertheless been some supportfor Gilligan's proposal of a gender-related differencein moral orientation. So, even if women are just ascapable of using justice concepts as men, they maystill prefer a care-oriented approach to moralthinking. Overall, though, the evidence for Gilligan'sproposal is inconsistent. While some studies show asignificant gender difference in orientation,'2 1314other researchers have observed little notable differ-ence.'5 16 17 18 There may be different explanations forthis variability, some relating to the possibility thatother variables are at work, and others focusing on theappropriateness of the particular reasoning contrastsbeing measured. Among the former is the view, advo-cated by Gilligan,6 that moral orientation is not simplydetermined by gender but by gender and experiencecombined. Experience-related factors such as educa-tion and occupation have been found by some tobe relevant to level and orientation of moral reason-ing. 14 19 20 One occupational area ofparticular interestto the current study is the health care profession,where reasoning differences have been identifiedbetween doctors and nurses. There is evidence tosuggest that the ethical reasoning of doctors isoriented towards the justice perspective, while that ofnurses is much more aligned with care.2' This differ-ence has been attributed not only to the traditionalgender divisions between those occupations (withnurses predominantly female and doctors male), butalso to the influence of established standards ofpractice and professional stereotypes: for example,doctors cure and nurses care; doctors are clinical,detached and justly administer healing, while nursesare responsive and sympathetic.22

Reasoning dimensionsBesides experience-based influences such as these,the overall variability in findings concerning moralorientation might be related to the particular reason-ing dimensions being measured. One interesting sug-gestion arising from a recent study conducted byGalotti et a123 is that the simple "justice/rights" and"care/responsibility" distinction is not sensitiveenough consistently to reflect gender differences inmoral thinking. Their study indicated that more fine-grained differences than the simple "justice/ rights"and "care/responsibility" distinction were at work inpeople's moral reasoning. It could be, then, that thevarying evidence for gender-related differences inmoral orientation is related to the fact that studies aremeasuring reasoning contrasts associated with thejustice and care perspectives that are either notspecific enough, or else, are not the most ethically or

philosophically salient contrasts between the two per-spectives. It is the latter view that motivates thepresent study.The ethics of justice and care are often charac-

terised in terms of their ostensibly opposedelements. While moral reasoning on the justiceperspective is taken, for example, to be rule-based,universalistic, abstract, and to concern rights andfairness, such reasoning is, on the care perspective,seen as inductive, responsive, contextual and con-cerned with preventing harm. If real differences inpeople's moral thinking are to be measured, though,it is crucial that the contrasts relied on to distinguishthe two perspectives be sound and pertinent. Someof those contrasts appear to be philosophically morecentral than others in distinguishing the two ethics.In the case of some, the extent to which they locate areal difference between the perspectives is unclear,either because a perspective does, in fact, implicitlysubscribe to some contrasted element it is deemednot to (this might be so with the justice perspectiveand the goal of non-violence and avoidance of harmthat is thought to be distinctive of the care ethic), orbecause a contrasted element attributed to a per-spective does not necessarily apply to it (this mightbe so with the legalism and strict adherence to rulessupposed essential to the justice ethic). There arealso broader philosophical reasons for questioningsome of the contrasts that have been emphasised,particularly when the care perspective is viewed as acritique of standard ethical reasoning. Kohlberg tookhimself to be measuring the fundamental aspects ofgenuine moral thinking, and Gilligan understoodherself to have uncovered a distinct (female) moralalternative to that standard (Kohlbergian and male)ethic. There are, however, sound philosophicalgrounds for supposing that both are mistaken inidentifying standard ethical reasoning exclusivelywith the Kantian tradition, where the concern forrightness, as opposed to personal or social good,defines genuine moral requirements and maturemoral judgment. Some of those properties thatGilligan attributes to the standard or traditional con-ception of ethics no longer apply once that traditionis construed more broadly to include defensible can-didates other than the purely Kantian one. Once thisis done, a narrower set of contrasts becomes relevantwhen distinguishing the care ethic from standard

24views.The view taken here is that these residual and

salient contrasts can be expressed in terms of thegeneral distinction between partial and impartialmoral reasoning.2425 The former, which charac-terises a central aspect of the care orientation,involves judgments that emphasise personal attach-ments and favour those with whom one is personallyconnected in situations where their interestscompete with others'. Impartial reasoning, on theother hand, involves judgments that are detachedand do not favour personal attachments, but reflect


http://jme.bm

j.com/

J Med E


ugust 1997. Dow

nloaded from

http://jme.bmj.com/

228 Partial and impartial ethical reasoning in health care professionals

a concern for what equal consideration and widerimpersonal responsibilities require. For example,when posed with a choice between, say, rewardingmy own son or another more deserving child, if Ibelieve that my son should be rewarded then I amjudging partially, and if I judge that the deservingchild should win, this is impartialist. Different sortsof reasons will underlie my judgment in each case. Imight favour my son because I value my attachmentto him, or am compelled by a special responsibilityto him, or believe that loved ones come first, etc. Imight reward the other child because I believe thatthe only objectively relevant moral factor in the situ-ation is desert, or because I believe that I must treatpeople equally and without favour, and so on.

It would be reasonable to expect that if there is anydifference in people's moral orientation, that differ-ence will be most apparent with respect to thecontrast in moral reasoning that most sharply anddeeply reflects the division between the care ethic andthe traditional moral views it opposes. In other words,we should expect it to be most apparent in relation topartial and impartial ethical reasoning. Furthermore,if the partial/impartial distinction does underlie thedifference in orientation represented by the care ethicand traditional moral views, then we should expectthis to be especially apparent in the ethical reasoningof nurses and doctors, which would typically bealigned with the care ethic and the traditional orienta-tion, respectively. In line with these expectations, thepresent study seeks to explore the distribution ofpartial and impartial ethical reasoning in a sample ofnurses and doctors. It was suggested earlier that moralorientation is influenced by the experience of beingmale or female. This experience, though, is not simplybased on biological sex, but is gender based. Weunderstand ourselves and others through the mediumofgender-based roles and characteristics, and developour ego-identity in the light of them. We thought itappropriate, therefore, to seek also to measure therelationship between gender-identification and moralorientation in our sample.

MethodPARTICIPANTSOne thousand questionnaires were mailed to 400doctors in Victoria, 200 doctors and 400 nurses inNew South Wales. These doctors and nurses wererandomly selected from databases from theVictorian and NSW Medical Boards, and theAustralian Nursing Federation. Three hundreddoctors and nurses (178 doctors and 122 nurses) ofmixed sex (115 male and 61 female doctors; 50 maleand 72 female nurses) returned completed question-naires. More than halfwere between 31 and 50 yearsof age, and 88% are currently employed within thehealth care field, with 78% of participants havingmore than 10 years experience. Thirty-eight per centof the nurses have, or were in the process of acquir-

ing, a tertiary qualification. Because respondentswere presented with ethical problems involving theirmothers, respondents' perceptions of the nature ofthat relationship were relevant. A large majority(77%) describe the relationship with their mother asbeing either "good" or "excellent", 15% claim it tobe "average", and only 8% of respondents spoke ofthe relationship as "below average" or "poor".

PROCEDUREUnlike Gilligan's care conception of morality whichwas derived empirically from women's reported expe-rience, the difference in moral reasoning investigatedhere is antecedently defined, and this will influencehow it is best measured. Moral dilemmas were used tofocus subjects' moral reasoning, and because theinterest is in subjects' orientation to predefined typesof moral reasons, standardised and hypothetical, asopposed to actual and self-selected, dilemmas werechosen. The former allow for much greater controlover content, and are more suited to use in quantita-tive tests than the latter, whose variable content andsubjective importance have been seen by some as aninfluence on reasoning orientation.261415 An open-ended interview or questionnaire that elicits freeresponse was seen as a less efficient means of testingfor orientation to predefined types of reasons than astructured questionnaire with predetermined stan-dardised responses which allow for quantitativeanalysis. Such a questionnaire was derived from twofocus groups and a pilot study. The focus groups wereeach composed of male and female doctors andnurses, and the relevant dilemmas asked participantsto choose, in various situations, between aiding theirmother (partialist) and aiding someone else who iseither very worthy from an impartial point of view orwhose being aided would have better overall conse-quences. The person-types depicted in these situa-tions as representing impartial worth or merit were ajudge, a medical practitioner, a saintly nun, and asportswoman. It was recognised that there was noplausible way of presenting an impartial choice alter-native in these situations without resorting to person-types of these sorts.

Given that the central concern of the study isthe relationship between profession and ethicalapproach, it was considered important to setdilemmas in both health care professional and non-professional contexts. Also, because our interest isthe approach to ethical reasoning that people mightconsistently take, the dilemma situations were variedin another significant respect - in terms of theseriousness or urgency of the problems involved. So,some dilemmas presented a problem where liveswere under threat, and others where mere comfortwas at stake. Focus group responses also stronglysupported the view that reactions to dilemmas mightbe influenced by whether they were posed in a healthcare professional context or not, and whether theyinvolved a life-threatening situation or not. So, for


http://jme.bm

j.com/

J Med E


ugust 1997. Dow

nloaded from

http://jme.bmj.com/

Helga Kuhse, Peter Singer, Maurice Rickard, Leslie Cannold andJessica van Dyk 229

the main study, four dilemmas were presented tocover the possible combinations of these factors.The first dilemma (Al), where the choice isbetween aiding the respondent's mother and a

judge, involved a non-life-threatening situation in a

non-professional context; the second (A2), posed a

choice between saving the respondent's mother anda medical specialist, and was a life-threatening situ-ation in a non-professional context; the third (A3)was a life-threatening situation in a professionalcontext where the life of the respondent's motherand that of a saintly nun are in danger; and thefourth (A4) involved a non-life-threatening situa-tion in a professional context, where the choice iswhether to benefit the respondent's mother or an

accomplished sportswoman. After further refine-ment through the pilot study, the final question-naire presented subjects with ten statementsfollowing each dilemma. The statements eitherexpressed a choice and a particular reason for it, forexample, "I should rescue the medical specialistfirst, because the medical specialist has an import-ant commitment on which others depend", or elsestated a general attitude or consideration, forexample, "I need to look at the situation from an

impersonal perspective". Five statements repre-

sented a partialist moral orientation and cited con-

siderations such as being attached to one's mother,having a special responsibility to one's mother, andloved ones coming first. The other five representedthe impartial orientation and cited considerationssuch as having a wider responsibility to society, nothaving a right to favour one's mother, and one'smother and x having equally valid reasons to beaided. Subjects were asked to rate their level ofagreement with each of the ten statements on thefollowing four-point scale: "agree", "somewhatagree", "somewhat disagree", and "disagree". Onthe basis of their overall levels of agreement withthe ten reason statements for each dilemma,subjects were asked to indicate finally if theybelieved they should aid their mother (partialistoverall), or aid the other worthy person representedin the dilemma (impartialist overall). Subjects'responses were classified as partialist or impartialistin accordance with this indication. An example ofone of the dilemmas has been included as an

appendix.

ResultsThe main concern of the survey was to discoverwhether gender or occupation (doctor or nurse)impacted on the approach (partialist or impartialist)that participants took to various moral dilemmas. Toaddress this question, we looked at the distributionof partial/impartial approaches in each of the fourdilemmas for the whole group. The results are pre-

sented in table 1.From these results it seems that in situations in

which lives are at stake, whether they are set in a pro-

fessional context or not, responses are overwhelm-ingly in favour of a partial approach. For thesituations in which no lives are at stake, and theoutcome is merely discomfort, there appears to be a

difference between the professional and the non-

professional situations. In the relevant professionalsituation (fourth dilemma), more respondents thinkthey should adopt an impartial approach, whereas,in the situation set outside a professional context thematter seems to be less clear-cut. Although a

majority of respondents lean towards an impartialapproach, the difference between those in favour ofan impartial approach and those that favour a partialapproach is not as big as in any of the other situa-tions.

Statistical analysisTo investigate whether thinking one should choosepartially or impartially in either of the four dilemmaswas dependent on (i) the sex of the respondents, (ii)their occupation, and (iii) their gender classification,cross-tabulations were generated and severalmeasures of association calculated. The Chi-square-based measures (Pearson, Phi and Cramer's V) are

problematic in the sense that they are difficult tointerpret. For samples that are too small, it won't bepossible to detect even large differences, while forlarge samples, even small differences can be statisti-cally significant. An alternative measure is theLambda statistic, based on the idea of proportionalreduction in error (PRE).27 The Lambda statisticindicates the proportion by which error can bereduced in predicting the dependent variable if theindependent variable is known. Lambda ranges from0 to 1, for example, a Lambda of 0-142 means thatby knowing the independent variable (for instance,

Table 1 Responses to dilemmas: overall

Dilemma Al Dilemma A2 Dilemma A3 Dilemma A4Non-profess * Non-profess Profess ProfessNon-life threat Life threat Life threat Non-life threat

Partial 122 (41-5%)** 263 (89/8%) 229 (8066%) 43 (15-0%/)Impartial 172 (5855%) 30 (10-2%) 55 (19-4%) 244 (85 0%)Unknown 6 7 16 13

Total 300 300 300 300

*Where the profession concerned is health care.**The percentage figures associated with each dilemma are percentages of the total number of known responses for that dilemma.


http://jme.bm

j.com/

J Med E


ugust 1997. Dow

nloaded from

http://jme.bmj.com/


knowing the sex of the respondents), the error inpredicting the dependent variable (for instance,partial/impartial approach to the dilemmas) can bereduced by 14-2%. However, even when Lambda is0, other measures of association may find relation-ships of a different kind.

Several Lambda measures were utilised.Symmetric Lambda assumes that there is no reasonto consider one of the variables dependent and theother independent. In our case, sex, occupation andgender-classification were assumed to be the inde-pendent variables. Subjects were classified as con-forming to either a masculine or feminine genderidentity according to the BEM Sex Role Inventoryquestionnaire. This questionnaire asked respondentsto classify themselves (on a seven-point scale rangingfrom "Never or almost never true" to "Always oralmost always true") with respect to 30 personalitycharacteristics, such as "willing to take risks", "eagerto soothe hurt feelings". Each personality character-istic has a predefined association (based on social

Table 2 Responses to dilemmas: occupation(doctor/nurse)

Dilemma A IDoctors Nurses Total

Partial 60 62 122(34.5%) (51-7%) (41.5%)

Impartial 114 58 172(65.5%) (48.3%) (58 5%)

Total 174 120 294(59-20/%) (40-80/%) (100%)

Unknown: 6Dilemma A2

Doctors Nurses Total

Partial 158 105 263(91 3%) (8755%) (89.8%)

Impartial 15 15 30(8.7%) (12-5%) (10-2%)

Total 173 120 293(59.0%) (41-0%) (100%)



Partial 136 93 229(81-4%) (79.50/%) (80 6%)

Impartial 31 24 55(18-6%) (20.5%) (19-4%)

Total 167 117 284(58-80/%) (41*20/%) (100%)

Unknown: 1 6Dilemma A4


Partial 23 20 43(13340/%) (17.40/a) (15.0%)

Impartial 149 95 244(86.6%) (82.6%) (85.0%)

Total 172 115 287(59.9%) (40.1%) (1000/a)

Unknown: 1 3

stereotypes) with one of four gender role descrip-tions: masculine, feminine, androgynous and undif-ferentiated. So, if a subject rates herself highlyoverall with respect to typically feminine characteris-tics, and lowly on masculine characteristics, she isclassified as in the "feminine" group.We looked at the appropriate Lambda to see how

much we could reduce our error in predicting theapproach of doctors' and nurses' to the fourdilemmas when knowing the values of the indepen-dent variables, sex, occupation and gender-classifi-cation. Subjects' responses classified according tooccupation and gender are presented in table 2, andtable 3 respectively. Our results showed that theapproach doctors and nurses adopt in all fourdilemmas is independent of either their sex, occupa-tion or gender classification. So, knowing the sex,occupation, or gender of the respondents does nothelp predict the reasoning approach they adoptwith the dilemmas. In addition, no relationshipwas observed between participants' approach to

Table 3 Responses to dilemmas: gender classification(according to BSRI)

Dilemma A IMascul Femin Androg Undiffer Total

Partial 29 26 30 36 121(34.5%) (51.7%) (41.5%) (42.4%) (41.4%)

Impartial 38 47 37 49 171(56.7%) (64.4%) (55.2%) (57.6%) (58.6%)

Total 67 73 67 85 292(22.9%) (25.0%) (22.9%) (29.1%) (100%)


Mascul Femin Androg Undiffer Total

Partial 60 63 60 78 261(90.90/) (86.3%) (89.6%) (91.8%) (89.7%)

Impartial 6 10 7 7 30(9.1%) (13.7%) (10.4%) (8.2%) (10.3%)

Total 66 73 67 85 291(22.7%) (25.1%) (23.0%) (29 2) 100%)



Partial 52 53 52 70 227(80.0%) (76.8%) (78.8%) (85.4%) (80.5%)

Impartial 13 16 14 12 55(20.0%) (23.2%) (21.-2%) (14-6%) (19-50/%)

Total 65 69 66 82 282(23.0%) (24.5%) (23.4%) (29-1%) (100%)

Unknown: 1 8Dilemma A4


Partial 15 10 10 8 43(22.4%) (13-7%) (15-6%) (9.8%) (15-00/a)

Impartial 52 63 54 74 243(776-06%) (86.3%) (84.4%) (90.2%) (85.0%)

Total 67 73 64 82 286(23.4%) (25.5%) (22.4%) (28.7%) (100%)

Unknown: 1 4


http://jme.bm

j.com/

J Med E


ugust 1997. Dow

nloaded from

http://jme.bmj.com/

Helga Kuhse, Peter Singer, Maurice Rickard, Leslie Cannold and Jessica van Dyk 231

dilemmas and several other variables (age, income,education, religion and respondent/mother relation-ship).

DiscussionsThe failure to find a significant relation betweenpartial/impartial moral reasoning and occupation orgender might be due to a number of factors. It ispossible that, because of the detail needed to conveythe desired ethical context for each dilemma, someconfusion may have arisen on the part of some par-ticipants, and responses may not have been as deter-minate as they could have been. Another possibilityrelates to the hypothetical manner in which theethical problems were presented. It may have beenthat this introduced an element of personal distanceinto the choice situation, which prevented genuineand immediate feelings of attachment that canmotivate a partialist response, from being effectivelyengaged with on the part of respondents. So, thosewho we would most expect to choose partially, forexample, nurses, may not have been given the appro-priate opportunity to demonstrate this. This sugges-tion, though, does not account for the resultsconcerning doctors, who were expected to respondsignificantly more impartially. The same hypo-thetical nature of the dilemmas that supposedlysuppresses a partial response in nurses, would alsobe expected to facilitate an impartial response inthose disposed to it. There are, however, otherpossible explanations for the present results. In thecase of occupation and orientation, the fact that nosignificant reasoning difference was observedbetween doctors and nurses might be an indicationof the weakening grip that stereotypes and ingrainedstandards are having on nurses' and doctors' con-ceptions of their professional roles. This might be soparticularly with nurses who are now, more thanever, consciously and critically reflecting on thenature of their professional status and activities. So,our findings, here, fail to support the views of thosewho see a specifically partialist "ethic of care" ascharacteristic of nurses, and an impartialist "justiceethic" as being characteristic of the medical profes-sion.

There is also another possible explanation forour results. The fact that a substantial majority ofsubjects, regardless of gender, occupation, etc,respond partially when lives are at stake indicatesthat there might be a relationship between theseriousness to subjects of the potential personal lossinvolved and their orientation to a partialistresponse. This tends to suggest that the dispositionto partialist and impartialist approaches might, inthat respect, be "problem" sensitive rather thangender or occupation related. In other words, itmight be that most people's ethical reasoningincludes both partialist and impartialist dimen-sions, and the one that dominates depends on the

level of personal costs involved in the problem situ-ation.The current study can be seen within a broader

theoretical context. The debate in many academicdisciplines over the care ethic versus standard ethicalviews has gained a lot of its momentum from beingembedded in a wider set of issues concerning genderrelations. Although this association has added to theprominence and immediacy of that debate, it has, tosome extent, also served to distract from someimportant philosophical issues about the relativemerits of these perspectives as ethical approachesand about where they overlap and differ. Such ques-tions about the philosophical status of the twoapproaches are of primary interest independently oftheir proposed associations with gender. Indeed,some of those questions had been on the philo-sophical agenda for some time before Gilliganproposed a gender connection.28 The fact that thepresent study finds no gender difference in moralorientation adds to the reasons for maintaining adegree of conceptual separation between the twosets of questions.

AcknowledgementThe authors thank the Australian Research Councilfor funding.

Helga Kuhse, PhD, is Director of the Centre for HumanBioethics, Monash University, Clayton, Victoria,Australia. Peter Singer, BPhil, is Deputy Director of thecentre. Maurice Rickard, PhD, Leslie Cannold,MBioeth, and Jessica van Dyk, BSc, were ResearchAssistants at the centre.

AppendixDILEMMA A2 (NON-PROFESSIONAL,LIFE-THREATENING)Person overboardIt is very late, and you are on the bow of an oceanliner far out at sea. Suddenly, on one side of theboat you hear cries of "help" and "rescue me" andrealise someone has fallen off the ship and isdrowning. You recognise the cries as coming from awoman to whom you were briefly introduced theother night. She is a universally well-regardedmedical specialist who has dedicated her life tohealth-issues in disadvantaged and indigenouscommunities. As you are about to rush to thewoman's aid, you hear cries of "help" and "rescueme" from the other side of the boat and realisesomeone else has fallen off the boat and isdrowning. You recognise these cries as comingfrom your mother, who you had believed was safelyin her cabin. Both your mother and the medicalspecialist are about the same age, and both appearto be in extreme distress. They are both aproxi-mately 20 metres from the ship. You realise that


http://jme.bm

j.com/

J Med E


ugust 1997. Dow

nloaded from

http://jme.bmj.com/


you will probably only be able to rescue one ofthem. Who should you attempt to rescue first?

References1 Kohlberg L. From is to ought: how to commit the nat-

uralistic fallacy and get away with it. In: Mischel T, ed.Cognitive development and epistemology. New York:Academic Press 1971: 79-97.

2 Kohlberg L, Kramer R. Continuities and discontinu-ities in childhood moral development. Human Develop-ment. 1969; 12: 93-120.

3 Holstein C. Irreversible, stepwise sequence in thedevelopment of moral judgement. Child Development1976; 47: 51-6.

4 Gilligan C. In a different voice: women's conception ofthe selfand morality. Harvard Educational Review 1977;47: 481-517.

5 Gilligan C. Women's place in man's life-cycle. HarvardEducational Review 1979; 49: 431-46.

6 Gilligan C. In a different voice: psychological theory andwomen's development. Cambridge MA: HarvardUniversity Press, 1982.

7 Colby A, Kohlberg L, Gibbs J, Liebermann M. A longi-tudinal study of moral development. Monographs of theSocietyfor Child Development. 1983; 48: 1-2 serial no 200.

8 Rest JR. Development in judging moral issues.Minneapolis: University of Minneapolis Press, 1979.

9 Walker LJ. Sex differences in the development of moralreasoning: a critical review. Child Development 1984;55: 677-91.

10 Thomas SJ. Estimating gender differences in the com-prehension and preference of moral issues.Developmental Review 1986; 6: 165-80.

11 Pratt MW, Golding G, Hunter W. Does morality havea gender? Merrill-Palmer Quarterly. 1984; 30: 321-40.

12 Gilligan C, Attanuci J. Two moral orientations: genderdifferences and similarities. Merrill-Palmer Quarterly1988; 34: 223-37.

13 Lyons NP. Two perspectives: on self, relationships, andmorality. Harvard Educational Review 1983; 53:125-44.

14 Pratt MW, Golding G, Hunter W, Sampson R. Sex dif-ferences in adult moral orientation. J7ournal ofPersonality 1988; 56: 373-91.

15 Ford MR, Lowery CR. Gender differences in moralreasoning: a comparison of the use of justice and careorientations. Journal of Personality and Social Psychology1986; 50: 777-83.

16 Friedman W, Robinson A, Friedman B. Sex differencesin moral judgements? A test of Gilligan's theory.Psychology of Women Quarterly 1987; 11: 37-46.

17 Walker LJ. A longitudinal study of moral reasoning.Child Development 1989; 60: 157-66.

18 Walker LJ, de Vries B, Trevethan SD. Moral stages andmoral orientations in real-life and hypotheticaldilemmas. Child Development 1987; 58: 842-58.

19 White CB. Age, education and sex effects in adultmoral reasoning. International J7ournal of Aging andHuman Development 1984; 27: 271-81.

20 Yacker N, Weinberg SL. Care and justice moral orien-tation: a scale for its assessment. Journal of PersonalityAssessment 1990; 55: 18-27.

21 Grundstein-Amado R. Differences in ethical decision-making processes among nurses and doctors. Journal ofAdvanced Nursing 1992; 17: 129-37.

22 Jecker NS, Self DJ. Separating care and cure: ananalysis of historical and contemporary images ofnursing and medicine. J7ournal of Medicine andPhilosophy 1991; 16: 285-306.

23 Galotti KM, Kozberg SF, Farmer MC. Gender anddevelopmental differences in adolescents' conceptionsof moral reasoning. Jrournal of Youth and Adolescence1991; 20: 56-85.

24 For a fuller discussion of whether caring and "connect-edness" can be properly understood in terms of partial-ist reasoning, see Cannold L, Singer P, Kuhse H,Gruen L. What is the justice/care debate really about?Midwest Studies in Philosophy 1996; 20: 357-77.

25 Blum LA. Gilligan and Kohlberg: implications formoral theory. Ethics 1988; 98: 472-91.

26 See, for example, Rothbart MK, Hanley D, Albert M.Gender differences in moral reasoning. Sex Roles 1986;15: 645-53.

27 This measurement technique was originally describedin Goodman L, Kruskal WH. Measures of associationfor cross-classifications, part 1. Journal of the AmericanStatistical Association 1954; 49: 732-64.

28 See Singer P, Cannold U, Kuhse H. William Godwinand the defence of impartialist ethics. Utilitas 1995; 7:67-86.

News and notes

Medical Ethics at the end of the 20th Century

The Ministry of Science in Israel is to sponsor an inter-national conference on Medical Ethics at the Close ofthe 20th Century. The conference will be held at TheVan Leer Jerusalem Institute, 5-8 January 1998, Israel.

Speakers will include: Baruch Brody, Baylor Collegeof Medicine; Tom Beauchamp, Georgetown; RaphaelCohen-Almagor, Haifa; Bernard Dickens, Toronto;Justice Dalya Dorner, The Supreme Court of Israel;Shimon Glick, Ben-Gurion; John Harris, Manchester;Govert den Hartogh, Amsterdam; Jan C Joerden,Europa-Universitat Viadrina; Eike-Henner Kluge,

Victoria; John Lantos, Chicago; Evert van Leeuwen,Vrije; Frederick Lowy, Concordia; John Robertson,Texas; Charles Sprung, Jerusalem; Avraham Steinberg,Jerusalem; Antonella Surbone, Memorial Sloan-Kettering Cancer Center, and Robert D Truog,Harvard.For more information please contact Ms Beki

Shimoni, Head, Conference Unit, The Israeli Ministryof Science, Building C, PO Box 18195, Jerusalem91181, Israel. Fax: 972-2-5824022. Phone: 972-2-5811220; 5847783.


http://jme.bm

j.com/

J Med E


ugust 1997. Dow

nloaded from

http://jme.bmj.com/

partial and impartial ethicalreasoningin healthcare …in moral orientation. so, even if womenare...

Documents