16
Springer is collaborating with JSTOR to digitize, preserve and extend access to Ethical Theory and Moral Practice. http://www.jstor.org Springer Consent and the Problem of Framing Effects Author(s): Jason Hanna Source: Ethical Theory and Moral Practice, Vol. 14, No. 5 (November 2011), pp. 517-531 Published by: Springer Stable URL: http://www.jstor.org/stable/41472621 Accessed: 28-10-2015 06:14 UTC Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTC All use subject to JSTOR Terms and Conditions

Hypothetical Consent Hannah

Embed Size (px)

DESCRIPTION

Hypothetical Consent Hannah

Citation preview

Page 1: Hypothetical Consent Hannah

Springer is collaborating with JSTOR to digitize, preserve and extend access to Ethical Theory and Moral Practice.

http://www.jstor.org

Springer

Consent and the Problem of Framing Effects Author(s): Jason Hanna Source: Ethical Theory and Moral Practice, Vol. 14, No. 5 (November 2011), pp. 517-531Published by: SpringerStable URL: http://www.jstor.org/stable/41472621Accessed: 28-10-2015 06:14 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 2: Hypothetical Consent Hannah

Ethic Theory Moral Prac (2011) 14:517-531 DOI 1 0. 1 007/s 1 0677-0 1 1 -9266-y

Consent and the Problem of Framing Effects

Jason Hanna

Accepted: 1 February 2011 /Published online: 18 February 2011 © Springer Science+Business Media B.V. 2011

Abstract Our decision-making is often subject to framing effects: alternative but equally informative descriptions of the same options elicit different choices. When a decision- maker is vulnerable to framing, she may consent under one description of the act, which suggests that she has waived her right, yet be disposed to dissent under an equally informative description of the act, which suggests that she has not waived her right. I argue that in such a case the decision-maker's consent is simply irrelevant to the permissibility of proceeding. I then consider two alternative views. According to the first, people susceptible to framing are able to give valid consent so long as they are sufficiently informed. This suggestion, I argue, maintains an overly narrow focus on mere quantity of information to the exclusion of other choice-affecting factors. A second response, which appeals to hypothetical consent, is likewise of little use in resolving the moral problem posed by framing effects. I conclude that if susceptibility to framing undermines the validity of consent, we may have good reason to reconsider whether consent has the rights-waiving function commonly attributed to it.

Keywords Consent • Informed consent • Rights • Waiver • Framing effects • Biases

Our decision-making is often subject to framing effects: alternative but equally informative descriptions of the same options elicit different choices. Indeed, framing impacts some of the most important decisions we make. To illustrate, consider one classic study, which asked respondents to indicate their hypothetical preferences about two treatments for lung cancer - surgery and radiation. Some respondents were provided with the following description of the outcomes for each treatment: "Of 100 people having surgery 10 will die during treatment, 32 will have died by 1 year and 66 will have died by 5 years. Of 100 people having radiation therapy, none will die during treatment, 23 will die by 1 year and 78 will die by 5 years" (McNeil et al. 1982, p. 1260). Call this "the mortality rate description." Other respondents were provided with the survival rate description, in which the same outcomes were expressed

J. Hanna (El) Department of Philosophy, Northern Illinois University, Zulauf 915, DeKalb, IL 60115, USA e-mail: [email protected]

£) Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 3: Hypothetical Consent Hannah

518 J. Hanna

as the likelihood of survival ("Of 100 people having surgery, 90 will survive the treatment period," and so on). The presentation format had a significant influence on the respondents' preferences. While radiation was preferred by forty-four percent of those who received the mortality rate description, it was favored by just eighteen percent of those who received the survival rate description (McNeil et al. 1982, p. 1261). Perhaps surprisingly, this effect was most pronounced among physician respondents, suggesting that it was not a product of misunderstanding.1 Instead, it seems that the risk of death during treatment "looms larger" for people given the mortality rate description than for people given the survival rate description (McNeil et al. 1982, p. 1262; Shafir et al. 2000, p. 338; Tversky and Kahneman 1981, p. 456). Since the increased risk of death during treatment is the major disadvantage of surgery relative to radiation, surgery appears less attractive under the mortality rate description. This sort of effect is hardly rare. For example, framing seems to influence decisions about whether to take therapeutic drugs (Halvorsen et al. 2007), seek preventative health screening (Schneider et al. 2001; Salovey and Williams-Piehota 2004), and continue life support for premature infants (Haward et al. 2008).

Framing has received a great deal of attention in psychology and economics. Its significance has been underappreciated by moral philosophers, however.2 In this paper, I attempt to partly rectify this philosophical neglect. I argue that when a person is susceptible to framing, her consent is simply irrelevant to how she may be treated. If so, the prevalence of framing effects may pose a problem for the prevailing anti-consequentialist and anti- paternalist view of consent. My strategy will be as follows. In Section 1, I describe two views of consent and indicate which I shall be discussing. I then argue in Section 2 that, according to this view, there is a strong case that decision-makers vulnerable to framing are unable to give valid consent. In Sections 3 and 4, I consider and reject responses to this argument that appeal to informed consent, hypothetical consent, and autonomy. If none of these responses succeeds, I conclude in Section 5, consent may not be relevant in the way that most ethicists suppose.

Throughout the discussion, I shall focus on examples from medicine, for two reasons: it is a context in which consent seems especially important and in which framing effects have been well documented. The arguments I offer, however, apply to any case in which informed consent is generally thought necessary to waive rights. I leave open the possibility that the following arguments do not apply to contexts (if there are any) in which valid consent need not be (very) informed.3

1 The Proprietary Gate Model

Most moral philosophers believe that consent - or at least informed, free, and competent consent - functions as a "proprietary gate" which opens the moral boundaries of personal

1 As noted above, the respondents were asked about their hypothetical preferences; they were not actually deciding on a course of treatment to be administered to them. But this does not seem to undermine the force of the example, since framing effects arise even when there are significant incentives to make a good choice: "Studies reported in the economic and psychological literature have shown that errors that are prevalent in responses to hypothetical questions persist even in the presence of significant monetary payoffs" (Tversky and Kahneman 2000, p. 221). Most of the discussion among philosophers has centered on the possibility of "libertarian paternalism,"

under which choices are structured so as to encourage prudent decision-making. This view - and the "libertarian paternalist" label - are due to Sunstein and Thaler (2006) and (2003). For example, some people hold that certain contracts should be enforced even if one of the parties docs not

have a robust understanding of the contract's terms. For discussion, see Bix (2010).

Ö Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 4: Hypothetical Consent Hannah

Consent and the Problem of Framing Effects 5 1 9

sovereignty to permit another's action (Kleinig 2010, p. 4). According to this proprietary gate (PG) model , the fact that В consents to A's performing some act shows that В has waived her moral right against A's performing that act. (Assume, unless otherwise noted, that the consent is informed, free, and competent.) As Judith Jarvis Thomson puts it, "The outcome of consent is that the consent-receiver has acquired a privilege, and thus that the consent-giver has ceased to have a claim" (1990, p. 348). For example, if В consents to A's giving her an ugly tattoo or having sexual relations with her, В has given A a permission to perform an act that would otherwise be a violation of B's right to bodily integrity.4 To be sure, even if В consents, A's act may be wrong for other reasons; it may be greedy, or unkind, or exploitative.5 But the act cannot be wrong because it infringes B's rights, and В can thus have no valid grounds for complaint about her treatment (Feinberg 1986, pp. 177-178).

The PG model can be contrasted with the evidentiary model , according to which consent is morally relevant only insofar as it provides evidence regarding the presence of some further feature (or set of features) in virtue of which the act is right or wrong. The evidentiary model of consent would obviously be favored by consequentialists. Consent might matter, on this view, insofar it provides (defeasible) evidence about the effects of an act on the subject's well-being: typically (though perhaps not always) the fact that a subject consents provides some reason to think that the act furthers her interests, while the fact that a subject dissents provides some reason to think that it does not. Of course, a defender of the evidentiary model need not claim that well-being is the only intrinsic value; perhaps there are others, and perhaps consent is evidentially relevant in relation to these other values, as well. Further, it might sometimes be practically important to obtain consent in order to secure the subject's cooperation in a joint undertaking, such as a regimen of medical treatment that requires active patient participation. According to the evidentiary model, however, consent possesses no moral force of its own. Both the evidentiary model and the PG model represent views about the moral, rather than the legal, importance of consent. A defender of the evidentiary model can with perfect consistency hold that some acts should be legally permitted only with the subject's consent on the grounds, for example, that any alternative policy would do more harm than good.

The differences between the PG model and the evidentiary model are especially salient in their treatment of paternalism. Defenders of the PG model would generally claim that it is impermissible for A to intervene in B's self-regarding affairs, without B's consent, even if doing so would greatly benefit B. Most defenders of the PG model would also hold that so long as В voluntarily consents to A's act, third parties are prohibited from interfering, even if A's act will seriously harm В (Feinberg 1986, pp. 98-101). If, on the other hand, consent is relevant only as evidence about the subject's well-being (or other values), then the absence of consent should not stand in our way when we are justifiably confident that intervention would benefit the subject (without compromising other relevant values).

For the remainder of this paper, I shall focus my critical attention primarily on the PG model. As we shall soon see, it is the view that seems to face the most significant challenge from the prevalence of framing effects.

4 This is what Heidi Hurd (1996) refers to as consent's "moral magic." See also Manson and O'Neill (2007, pp. 72-77). Kamm likewise claims that "[i]n ordinary morality, consent by the person to be harmed (when he is competent to give consent) can often justify acts that would otherwise be impermissible" (2007, p. 219). The PG model also seems to be presupposed by the maxim Volenti non fit injuria ("To one who consents no wrong is done"). For an endorsement of this maxim, see Feinberg (1984, pp. 115-117). 5 Thomson considers an example in which a boy is given a box of chocolates that he refuses to share with his younger brother. Although the younger brother has no right to the chocolates, the older brother may be "greedy, stingy, callous - but not unjust" (1971, p. 60).

Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 5: Hypothetical Consent Hannah

520 J. Hanna

2 The Problem of Framing Effects

To see why framing effects pose a problem for the PG model, consider a decision-maker vulnerable to the framing effect described at the beginning of this paper:

Framing-Induced Consent: Patient has lung cancer. Doctor explains the treatment options in terms of the survival rate description, and Patient consents to surgery. Had Patient been given the mortality rate description, however, she would not have consented to surgery. Call cases such as this, in which the subject's choice is affected by framing,

"framing cases." Of course, we often do not know whether a particular decision-maker is vulnerable to framing, but let us assume that we do in Framing-Induced Consent. (I consider the consequences of dropping this assumption in Section 5.) We can now ask whether a subject's consent is valid when she is subject to framing.6 As is standard, I shall say that a subject's consent is valid just in case it is sufficient to waive the relevant right. According to the evidentiary model, consent is never valid in this way, since consent does not function directly to waive rights. For this reason, framing effects do not seem to pose an especially great theoretical challenge to the evidentiary model. When a decision-maker is vulnerable to framing, her consent (dissent) presumably does not provide very good evidence that proceeding would (not) promote her well-being or other relevant values. The evidentiary model will then permit us to proceed (or not) directly on the basis of these values.7 But it may initially be unclear how the PG model would deal with Framing-Induced Consent.

A defender of the PG model may be tempted to offer the following resolution of this case: since Patient actually consents to surgery, it is permissible to proceed. Conversely, on this view, if Patient had actually dissented because she had been given the mortality rate description, then it would have been impermissible to proceed with surgery. In short, the proposal holds that treatment of Patient is morally governed by her actual choice. (A related proposal holds that Patient's actual choice is decisive only once she has been provided with both descriptions. I criticize this proposal further below, after setting out the problem I shall be discussing.) Framing effects may seem problematic because it is difficult to believe that the permissibility of an act could depend on something as trivial as how the question is asked. But if we are committed to the moral importance of consent, perhaps we should simply accept that any factor which affects the subject's choice to consent thereby also affects the permissibility of the act.

To see why framing cases cannot be resolved in this way, we must first recognize that consent is a de dicto propositional attitude (or, alternatively, some overt behavior that expresses such an attitude).8 One always consents to an act under a certain description, and one might consent

6 Although I shall generally speak in terms of the validity of consent , the arguments below also apply, with slight modification, to dissent. Most philosophers believe that dissent, like consent, can be invalid. Consider, for example, a case in which an ill-informed or intoxicated patient dissents to a beneficial medical procedure, there is no time to remedy her deliberative defect, and there is good reason to believe that she would consent if she were informed and unimpaired. 7 Of course, there may be epistemic problems that make it difficult to determine which course of action would benefit Patient. But these problems do not give rise to a paradox of the sort that I argue plagues the PG model. For my purposes, it is irrelevant whether consent is a "subjective state," such as an intention, or a

"performative." The subjective state view is adopted by Hurd (1996, pp. 124-126) and Alexander (1996, p. 165). The performative view is adopted by Wertheimer (2003, pp. 144-152) and Kleinig (2010, pp. 9-11). Regardless of how this debate is settled, consent is always given under an act description, as is demonstrated by the examples below.

Ö Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 6: Hypothetical Consent Hannah

Consent and the Problem of Framing Effects 521

under one act description but not another (Manson and O'Neill 2007, pp. 12-13; Westen 2004, pp. 40-42; Hurd 1996, p. 127; O'Neill 1985, p. 156). Consider the following case:

HIV Positive : В agrees to have sexual intercourse with А. В is unaware that A is HIV positive, and if she knew this she would not consent.

В consents under one description of the act (say, "intercourse with A") but not under another description ("intercourse with someone who is HIV positive"). Even though both descriptions are accurate, we cannot infer from the fact that В consents under the first that she consents under the second.9 Consider a second example:

Blood Transfusion : Doctor informs Jehovah's Witness (JW) that JW will soon die unless he receives a "standard liver transplant." Since JW is unaware that a "standard liver transplant" involves a blood transfusion, he signs a form consenting to the procedure, even though he is strongly opposed to blood transfusions on religious grounds.

Again, JW consents under one description of the act but is nonetheless disposed to dissent under a more readily understandable description.

If consent is always given under an act description, then we must determine which act descriptions matter for the purposes of determining whether the subject has waived her right. For brevity, I shall say that an act description is morally relevant in a given case if and only if a subject who consents under that description has given valid consent.10 In Blood Transfusion , virtually all defenders of the PG model would agree that "standard liver transplant" is not a morally relevant act description (assuming that JW does not know that a standard liver transplant includes a blood transfusion). Although JW consents under this description, he has not waived his right against surgery. 1 1

Return now to Framing-Induced Consent and consider the following question: which of the act descriptions - the survival rate description or the mortality rate description - is morally relevant? There are just three possibilities, two of which can be quickly dismissed. First, it may be that neither of the act descriptions is morally relevant. If so, then contrary to the proposal under consideration, Patient's actual choice is invalid. Second, it may be that just one of the act descriptions is morally relevant. This possibility is also inconsistent with the proposal suggested above, since the choices of those who decide under the other description would then be invalid. Further, given that the survival rate description and mortality rate description are equally informative, there is no non-arbitrary basis for claiming that just one is morally relevant.12

9 Cf. Westen (2004, p. 41), who discusses New York v. Hough , a case in which a man snuck into the bed of his identical twin brother in order to deceive the brother's girlfriend to have intercourse with him. 10 If some rights are inalienable, there will be cases in which no act description is morally relevant, in the sense in which I am using this terminology. One cannot provide valid consent to an act if he is incapable of waiving his right against it. 1 1 What should we say about HIV Positivei Is A guilty of sexual assault if he failed to inform В of his HIV status? Our answer to this question will depend on how informed we believe a person must be in order validly to consent to sexual relations. Notice, further, that one might hold that В has not given her morally valid consent, even though A should not be subject to legal sanction. For relevant discussion, see Wertheimer (2003, Chapter 9). As I have indicated, I will focus in the remainder of the paper on contexts in which it is generally agreed that valid consent must meet fairly demanding informational requirements. 12 To be sure, there may be some cases in which one way of presenting information is more confusing or more difficult for recipients to register. For example, many people are better able to answer statistical problems when data is presented in a frequency format (jc out of every у people have a certain problem) rather than a probability format (z percent of people have a certain problem). See Gigerenzer (2000, Chapter 4). But these concerns do not seem to arise in our example, in which the only difference is whether the risks are presented as mortality or survival rates.

Ö Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 7: Hypothetical Consent Hannah

522 J. Hanna

We are thus led to the conclusion that both act descriptions are morally relevant. Now by itself, there is nothing especially puzzling about the claim that two (or more) act descriptions are morally relevant in a given case. For example, a doctor could describe a particular procedure to a patient as, say, "arthroscopic knee surgery that poses such-and-such risks," or the doctor could describe the procedure in this way before adding the room number in which the surgery will take place. Both of these act descriptions - the one that provides the room number and the one that does not - are morally relevant, since a patient can validly consent to surgery regardless of whether she is provided with the room number.13 In cases such as Framing-Induced Consent , however, the claim that both differently-framed act descriptions are morally relevant yields an apparently intractable conflict. On the one hand, if Patient actually consents under a morally relevant act description, the PG model seems to imply that Patient has waived her right and that it is permissible for Doctor to proceed with the surgery. On the other hand, it generally seems impermissible to proceed when a person is disposed to dissent if provided with a morally relevant act description. In Blood Transfusion , for example, JW's disposition to dissent under the more informative description seems to explain why it is impermissible to perform the operation. It is also important to recognize that if Framing- Induced Consent is like other cases in which framing effects have been observed, Patient may be vulnerable to framing even after she gives her initial consent to the procedure. Kahneman and Tversky note that in many cases framing effects are "not eliminated even when the same respondents answer both [differently-framed] questions within a few minutes"; respondents stand by each choice when confronted with the respective presentation formats, even though they recognize the inconsistency and wish to avoid it (2000, p. 5).

14

It thus appears that Patient's actual choice does not determine the permissibility of proceeding in Framing-Induced Consent. There are strong grounds for claiming, paradoxically, that Patient has waived her right (since she has consented under a morally relevant act description) and that she retains her right (since she is disposed to dissent under a morally relevant act description). Call this "the problem of framing effects." To resolve the problem, the defender of the PG model needs to find some principled way to avoid one of the conflicting judgments above.15

Before proceeding, it will be useful to forestall one tempting response to this problem. Ruth Faden and Tom Beauchamp, among the few moral philosophers explicitly to address framing, suggest that it can be handled as follows:

13 I am indebted to Eric Chwang for this example. Kahneman and Tversky discuss a study in which participants were asked to choose between two public

programs for combating a disease. Many of the participants preferred different programs depending on whether the outcomes were framed as "lives saved" or "lives lost": "Respondents confronted with their conflicting answers are typically puzzled. Even after rereading the problems, they still wish to be risk averse in the 'lives saved' version; they wish to be risk seeking in the 'lives lost' version; and they also wish to obey invariance and give consistent answers in the two versions" (Kahneman and Tversky 2000, p. 5).

The argument I have offered may seem to have an unfortunate implication in the following case. Suppose that A chooses to perform some act after he is told that there is a ninety percent chance that it will not harm other (non-consenting) people. Suppose that A would not have chosen to perform this act, how „ /er, if he had been told that there is a ten percent chance that it will harm another person. It seems clear that if A's act docs harm an innocent bystander, we may hold A responsible by punishing him or at least requiring him to pay compensation. It seems to me that there is an important difference between this case and Framing-Induced Consent. As we have seen, most defenders of the PG model believe that consent (to medical treatment, at least) is valid only if it is informed and unimpaired. But few people believe that an agent must meet these standards before he can be held liable to pay compensation for injuries he has caused to others. For example, while most philosophers would agree that a person cannot validly consent to surgery while intoxicated, most also believe that a person may be held responsible for drunken behavior that harms others. As I have already indicated, I am dealing only with contexts in which it is generally believed that valid consent must be informed and (relatively) unimpaired. The original objection was brought to my attention by David Boonin.

Ô Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 8: Hypothetical Consent Hannah

Consent and the Problem of Framing Effects 523

[B]ecause the power and precise nature of framing effects are difficult to ascertain in concrete cases, the prudent course for professionals seeking... informed consent is to provide patients and [research] subjects with both sides of the story - the half-full and the half-empty presentations, the mortality and the survival frames - in the hopes of avoiding the gaps in understanding that framing effects may produce (1986, p. 321). This response holds that we can avoid any framing-related complications simply by

providing all alternative frames, or "both sides of the story." In some cases, this suggestion may be impractical. It may be overly demanding to require medical practitioners to provide all alternative frames, and it may tax our patience and attention if we had to listen to an explanation of a procedure's risks and then, moments later, listen to another explanation (and another?) in which the risks are framed differently. Nonetheless, there are two more substantial problems for Faden and Beauchamp's suggestion.

First, providing all frames is simply another way of framing the choice and has no claim to superiority over others. After all, people given only the survival rate description are not ignorant of mortality rates; any competent adult knows that a ninety percent chance of survival entails a ten percent risk of death. So we cannot argue in favor of providing both descriptions on the grounds that doing so renders the subject better informed or fills "gaps" in her "understanding." And if providing both descriptions does not render the subject better informed, it is unclear why doing so should be thought to confront the subject with a "neutral" frame or presentation format. Indeed, the problem of framing effects is precisely that there is no neutral way to present the alternatives.

Further, and just as importantly, even if it is conceded that we ought to provide "both sides of the story," the subject's choice may still be affected by other features of the frame, such as the "sequence in which options are presented" (Redelmeier et al. 1993, p. 74). Suppose, for example, that after Patient consents to surgery under the survival rate description, Doctor provides her with the mortality rate description. There are now two possibilities. Patient may dissent when provided with this new description, in which case the conflict in her choices merely leads us back to the original problem. Alternatively, now that she has actually consented, she may reaffirm her consent when given the mortality rate description. But given her original vulnerability to framing, such consent may only be an artifact of her having been given the survival rate description first. And surely there is no neutral temporal order in which to present the descriptions.

3 Informed Consent and the Disposition to Consent

There are two more promising ways in which defenders of the PG model might respond to the problem of framing effects. First, they may attempt to show that, on closer inspection, Patient's consent is valid. Second, they may attempt to resolve the problem by appealing to some variety of hypothetical consent. I consider the first of these strategies in this section, and the second in the following section.

A defender of the PG model might argue that while actual consent has moral force, the mere disposition to consent (or dissent) does not (see Thomson 1990, pp. 187-188; Brudney 1991, pp. 236-240). 16 In setting up the problem of framing effects, I claimed that Patient's disposition to dissent under a morally relevant act description suggests that she has not

16 Both of these authors argue against hypothetical consent. I shall later distinguish hypothetical consent, as commonly invoked in the literature, from the disposition to consent that is at issue here. But the arguments advanced by Thomson and Brudney may apply to both of these notions.

Ф Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 9: Hypothetical Consent Hannah

524 J. Hanna

waived her right. But if the disposition to dissent does not make a moral difference, this claim is mistaken, and we can thus avoid one of the judgments in the inconsistent pair that gave rise to the problem of framing effects. Now there are certainly some cases in which a subject's disposition to dissent seems to make a moral difference; in Blood Transfusion , JW's disposition to dissent under a more informative act description seems to account for the common judgment that he has not waived his right. But perhaps defenders of the PG model would claim that the impermissibility of proceeding in Blood Transfusion is explained not by JW's disposition to dissent, but instead by the fact that his actual consent is ill-informed. By contrast, we have been assuming that Patient in Framing-Induced Consent is fully informed about the risks (expressed in terms of survival rates) of the surgery to which she consents. In short, according to the reply we are now considering, what matters is not whether the subject is disposed tó choose differently under some alternative frame or act description, but instead whether her actual choice is sufficiently informed.

This line of argument runs into trouble when the critic attempts to explain how we can distinguish consent that is sufficiently informed from consent that is not. One obviously cannot argue that a person's consent is sufficiently informed only when she has complete information about the relevant act, since such a standard would be impossible to attain (Manson and O'Neill 2007, p. 28; Faden and Beauchamp 1986, p. 302). In order to give valid consent, it seems that one needs only information that is relevant to the choice. A patient can give informed consent to a medical procedure without knowing the number of the room in which the procedure will take place, since such knowledge is not relevant to her decision-making. But she cannot give informed consent unless she knows the prognosis of her condition under alternative treatment scenarios, since such information will likely be the basis for her decision.17 Information is relevant to a choice if knowledge of it would affect how the decision-maker chooses (or perhaps if such knowledge would be sufficiently likely to affect how the decision-maker chooses). This view supports the following informed consent principle : a subject's consent to an act is insufficiently informed if he is disposed to dissent (or it is sufficiently likely that he is disposed to dissent) under a more informative act description.18 This principle obviously explains why JW's consent is invalid in Blood Transfusion. Notice that if the informed consent principle is correct, the disposition to dissent does matter, insofar as it shows that the subject's actual consent is insufficiently informed.

In order to both accept the informed consent principle and maintain that Patient's consent is valid in Framing-Induced Consent , one would need to draw a sharp distinction between the following: (a) a subject's disposition to dissent under an act description that includes a greater quantity of information, and (b) a subject's disposition to dissent under an act description that includes something else - say, a frame or way of presenting information - that is not included in the description under which she consents. In particular, it would need to be shown that the former invalidates consent, while the latter does not. But it does not seem that the distinction between (a) and (b) is morally relevant in this way. To think otherwise fetishizes sheer quantity of information in a way that appears indefensible.

17 A similar view of relevance is defended in Chwang (2010). This account of relevance also has legal support. As Faden and Beauchamp point out, "[c]ase law hinges almost entirely on whether disclosures about undesirable outcomes or risks have been made, and, if not, whether they represent tolerable outcomes under the circumstances or whether knowledge of them would have altered the patients intentions or behavior" (1986, p. 247, emphasis added).

Notice that this principle provides a sufficient condition of one's being insufficiently informed; it does not provide necessaiy and sufficient conditions. I thus leave open the possibility that other informational limitations may compromise the validity of consent. I also leave it open just how likely it must be that some piece of information would change the subject's mind before that information becomes relevant.

Ö Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 10: Hypothetical Consent Hannah

Consent and the Problem of Framing Effects 525

This is so for at least two reasons. First, the view that (a) but not (b) invalidates consent may yield implications that many would reject. Consider the following (actual) example. A researcher is conducting a study on lysergic acid diethylamide, more commonly known as 'LSD." In explaining the study to potential subjects, the researcher uses the full chemical name, since she believes (let us assume correctly) that fewer people will enroll if she refers to the drug by its more common name. Suppose further that the researcher provides the subjects with all other relevant information about the drug's effects and the nature of the study. I suspect that many defenders of the PG model would claim that the research subjects have not given their valid consent, since (we are assuming) they would not have consented if the drug had been referred to as "LSD" (see Faden and Beauchamp 1986, p. 183; Manson and O'Neill 2007, p. 13). Nonetheless, a description in which the drug is called by its common name does not seem to provide a greater quantity of information than an equally full description in which it is called by its chemical name; each description merely uses a different name for the same substance.19 Since intuitions about the LSD case may differ, it may not provide a decisive counter-example to the view that there is a morally relevant distinction between (a) and (b). But the example does seem to jeopardize that view to the extent that we find something objectionable in the researcher's behavior.

Second, although the informed consent principle is intuitively plausible, proponents of the PG model will eventually need to defend it. Any such defense, however, is likely to suggest that the validity of consent is also undermined by (b), the disposition to dissent under an alternative frame. It might be argued, for example, that the importance of informed consent is explained by autonomy. Although this claim is often made, the notion of autonomy on which it relies is far less often explained.20 I further discuss appeals to autonomy at the end of Section 4. For present purposes, it may be plausible to think that a subject's choice fails to be autonomous when it does not reflect his settled values and preferences (Savulescu 1994; Scoccia 1990). If so, however, a subject's susceptibility to framing would also sometimes show that his choice fails to express autonomy; in framing cases, it is difficult to believe that both conflicting choices equally reflect the subject's settled values and preferences (Schwab 2006). Further, we do not care about information simply for its own sake. Information matters because, and to the extent that, it influences the subject's choice or alters her appreciation of the situation. The framing of options also influences choice and alters appreciation, making certain features (such as the chance of long-term survival or the risk of immediate death) more or less salient. It appears that any

19 One might point out that the common name description provides at least one piece of information not included in the chemical name description: that the substance in question is commonly called "LSD." Although this may be true, it does not challenge the conclusion I draw from the LSD case. For by the same token, the chemical name description provides information not included in the common name description: that the chemical name of the drug is "lysergic acid diethylamide." Although each description may provide different information about the name of the substance, neither provides more information than the other. It is thus unclear what reason we could have for claiming that one can give valid consent only under the common name description. (Perhaps it will be suggested that one can give truly informed consent only when one has been given both names. But I have already argued against this sort of proposal, and it at any rate seems rather difficult to believe that the researchers are required to provide participants with the chemical name in order to obtain valid consent.) Further, and most importantly, the use of the name "LSD" likely influences decision- making by playing on the subjects' emotions (Faden and Beauchamp 1986, p. 183) or changing the way they "think of' the test drug (Manson and O'Neill 2007, p. 13). Framing seems to affect decision-making in the same way. There thus appears to be no basis for sharply distinguishing the LSD case from the framing cases I have been discussing. I am indebted to a reviewer from this journal for feedback on this point.

For an exception, see Dworkin (1988, Chapter 7). On some views, autonomous consentyi/si is intormea, free, and competent consent (Beauchamp 2010). Such views seemingly cannot invoke autonomy to explain the importance of informed consent.

Ö Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 11: Hypothetical Consent Hannah

526 J. Hanna

reason we could have for thinking that information is relevant to the validity of consent is also a reason for thinking that framing or presentation format is relevant to the validity of consent. To claim that the presentation format does not matter simply because it is not information seems merely to resolve the problem by fiat.

In short, the moral considerations that support the informed consent principle suggest that a subject's consent is invalid when she is disposed to choose differently under an alternative (and equally informative) frame. Of course, we can rectify defects in the subject's knowledge by attempting to provide all relevant information. But as I have already argued, we cannot provide the subject with all frames; any attempt to do so would merely represent one particular way of framing the choice.

Before we turn to a further set of proposed solutions to the problem of framing effects, it is worth briefly discussing "libertarian paternalist" approaches that have recently become popular (Sunstein and Thaler 2006, 2003).21 Under libertarian paternalism, institutions harness the power of framing effects and other cognitive biases in order to improve individual decision-making (the "paternalist" component) without the use of coercion (the "libertarian" component). Libertarian paternalists would argue, for example, that Doctor should simply provide Patient with whichever frame is most likely to result in a choice conducive to Patient's own well-being. I shall not argue against this view. The important point is that it seems largely irrelevant to the issue that I have been discussing. I have been considering whether people can issue valid consent when they are vulnerable to framing effects. The libertarian paternalist proposal does not address this question. Instead, it seems to presuppose that Patient is able to give valid consent, since there is morally little to be gained by framing alternatives so that Patient invalidly consents to the act that best serves her interests.

At this point, philosophers sympathetic to libertarian paternalism may argue that it does suggest a way out of the problem of framing effects after all. It might be claimed that when two act descriptions would elicit different choices, only the description that would elicit the better choice is morally relevant. On this view, if Patient is better off receiving surgery, and the survival rate description is more likely to lead her to choose surgery, then the survival rate description is morally relevant, while the mortality rate description is not. This proposal seems arbitrary, however. The only reason given for favoring the survival rate description is that it induces prudent decision-making. But why would this show that Patient's decision is invalid (and not merely imprudent) if she chooses under the equally informative mortality rate description? I see no satisfactory way to answer this question consistent with the PG model. The most obvious answer is that a subject can give valid consent only when she is choosing prudently. But this proposal abandons the PG model, downplays the moral importance of consent, and seems to support paternalism that can hardly be described as libertarian.

4 Hypothetical Consent and Autonomy

If I have been correct so far, people are unable to give valid consent when they are vulnerable to framing. Many philosophers have argued that when a person is incapable of giving valid actual consent, we should look instead to hypothetical consent, and defenders of the PG model might be willing to argue that under certain circumstances hypothetical consent can waive rights. Arthur Kuflik, for example, argues that appeals to hypothetical consent can extend "respect for someone's right to decide what will happen in and to his or

21 Trout (2005) develops a very similar view, but he (wisely, I believe) refrains from describing it as "libertarian."

Ô Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 12: Hypothetical Consent Hannah

Consent and the Problem of Framing Effects 527

her own body" (2010, p. 140). Initially, hypothetical consent may appear unlikely to resolve the problem of framing effects; after all, whether a person would consent to an act may depend on how the options are framed. Yet most appeals to hypothetical consent require us to consider whether the subject would consent if she were fully informed and able to give valid consent (Kuflik 2010; VanDeVeer 1986, pp. 75-81). For example, we may ask whether a temporarily intoxicated person would consent to a certain intervention if she were sober. If susceptibility to framing prevents a subject from giving valid consent, then perhaps we should likewise consider whether the subject would consent if she were fully informed and immune to framing. J.D. Trout, for example, offers the following "warrant for intervention" in imprudent, framing-affected choices: "[I]f there were a debiasing option [viz., a way to eliminate the decision-maker's susceptibility to framing], or if the decision- maker were fully informed, they would not have made the decision they did" (2005, p. 430). To be sure, Trout seems to have in mind only "intervention" of the libertarian paternalist sort. But someone sympathetic to his suggestion might argue more generally that hypothetical consent determines the permissibility of proceeding in framing cases; on this view, it is permissible to perform the surgery in Framing-Induced Consent so long as Patient would consent if she were fully informed, competent, and immune to framing.

Even if we suppose that hypothetical consent sometimes has moral force, it does not help to resolve the problem of framing effects. This is so for two reasons. First, given that many people are consistently and predictably vulnerable to framing, there is no way to determine what they would choose if they were immune to framing. To be clear, the point here is not the obvious one that it is sometimes difficult in practice to determine how a subject would choose under some set of counterfactual conditions. Instead, the problem is that there is often simply no basis on which to claim that a person would (or would not) consent if she were unaffected by framing. Second, the hypothetical consent test seems compelling insofar as it asks whether the subject would consent if she had her normal capacities (VanDeVeer 1986, p. 75). But when a person is consistently vulnerable to framing, the view we are considering would have us ask whether she would consent if she had something other than her normal capacities. It is difficult to see how this form of hypothetical consent could be morally relevant.

An analogy will be helpful here. Many defenders of hypothetical consent argue that it cannot be applied to those who have always been incompetent, such as young children or people with permanent and severe cognitive disabilities. This is partly because, echoing the first point above, there is no basis for claiming that an always-incompetent person would (or would not) consent to some act if he were competent (Kuflik 20 10).22 Further, echoing the second point above, it would be odd if one were to defend one's treatment of a child by claiming that the child has "as good as consented" since he would consent if he were an adult. Even if we could somehow be confident that the child would consent if he were an adult, this point seems irrelevant to the permissibility of the act.23 Likewise, if a person has

22 Kuflik discusses the Saikewicz case, in which the court applied a "substituted judgment standard" to a man with a "mental age of approximately 2 years and 8 months" (2010, p. 143). Kuflik argues that the "best interests standard" should guide the treatment of such patients, and that this is a fundamentally different justification than that provided by the appeal to hypothetical consent (p. 144).

Gerald Dworkin has suggested that "[p]arental paternalism" towards children may be justified by "future- oriented consent," or the child's coming to "welcome" the interference when he is older (1971, p. 119). There are two points worth making about this suggestion. First, insofar as Dworkin appeals to the claim that the child actually will consent or come to welcome the parents' intervention, his argument does not appeal to hypothetical consent. Second, appeals to subsequent consent have been subjected to a barrage of criticisms, including those raised by VanDeVeer (1986, pp. 66-70) and Husak (1981, pp. 33-35).

£) Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 13: Hypothetical Consent Hannah

528 J. Hanna

always been vulnerable to a certain framing effect, it is difficult to see how we could justify our treatment of her by claiming that she would consent if she were now in a psychological condition (one rendering her immune to framing) that she has never actually been in.

Are there any other ways in which defenders of the PG model might respond to the problem of framing effects? A final proposal appeals to the value of autonomy. Autonomy is, after all, commonly thought to explain the moral importance of consent. Of course, philosophers have offered many different conceptions of autonomy, and I obviously cannot consider all of them here (see Christman 1988 for an overview). Some conceptions of autonomy seem irrelevant for our purposes, however. According to one view, autonomy is a right of "personal sovereignty" that prohibits intervention in a decision-maker's voluntary self-regarding choices; respect for autonomy in this sense requires that we secure the valid consent of autonomous adults before treating them in certain ways (Feinberg 1986; Christman 1988, p. 110). This proposal is not helpful if I have been correct that consent is invalid in framing cases.

Autonomy may be thought relevant to our problem for at least two reasons, however. First, it has been suggested that people lack "effective autonomy" when they are subject to framing effects and other cognitive biases: "Effective autonomy is fostered when biases resulting from bounded cognition are identified as risks in the decision processes of people and this is followed by 'appropriate debiasing'" (Schwab 2006, p. 578).24 The idea here seems to be that respect for autonomy requires that the person seeking informed consent eliminate the cognitive biases of the person giving consent. Although it may be nice if we were able to eliminate others' susceptibility to framing, this is not a reasonable strategy. As the psychological literature on heuristics and biases now attests, framing effects are "pervasive and robust" and cannot be easily eliminated (Kahneman and Tversky 2000, p. 5). Further, even the "most effective" debiasing strategies "are only a qualified success" (Trout 2005, p. 418). In light of this, we need a way to deal with people as the flawed choosers they are. One cannot successfully respond to the problem of framing effects by pointing out that it would not arise if people were immune to framing.

Second, many philosophers argue that a person is autonomous to the extent that he actually shapes his own life. This may require that he have the capacity to endorse his desires after critical reflection (Dworkin 1988; Frankfurt 1971), or that he form a life-plan or conception of the good that guides his choices (see Savulescu 1994; Young 1986, p. 8). To be sure, some philosophers, including Dworkin, hold that autonomy is a "global" concept that "evaluates" a whole life rather than a particular preference or episode (1988, pp. 15-16). But others have been concerned with the autonomy of preferences or choices (see Beauchamp 2010; Savulescu 1994; Elster 1982), and one might argue that we respect autonomy by helping agents to satisfy preferences that have been reflectively endorsed or appropriately formed. This view suggests that Doctor may proceed in Framing-Induced Consent so long as doing so serves Patient's autonomous aims. I will not argue against this proposal. The important point to recognize is that it seems to cohere with the evidentiary model of consent better than it coheres with the PG model. A defender of the evidentiary model, after all, is well-placed to claim that consent matters only insofar as it provides evidence that the relevant act serves the subject's goals or reflectively-endorsed desires; on this view, the act's relation to the subject's desires, rather than his consent, ultimately determines its permissibility. It is unclear, however, why the second autonomy-based proposal would appeal to defenders of the PG model. The fact that a certain act serves the subject's autonomous aims would not appear to show that the subject has somehow lost his

24 To be sure, Schwab (2006) does not claim that we can always rid people of their susceptibility to framing, and it is unclear how he believes we ought to proceed when we cannot do so.

Ф Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 14: Hypothetical Consent Hannah

Consent and the Problem of Framing Effects 529

right against the act. Indeed, to the extent that the satisfaction of the subject's reflectively- endorsed desires is good for him, the second autonomy-based proposal may not be much different from one according to which we may proceed when doing so benefits the subject.

5 The PG Model Revisited

Even if all of the arguments I have offered are sound, they show only that consent, whether actual or hypothetical, does not determine the permissibility of proceeding in framing cases. This conclusion may seem relatively modest, for two reasons. First, one might argue that consent usually functions in the way described by the PG model, even if consent is invalid in framing cases. If so, then while the problem of framing effects still requires a solution, perhaps it does not pose much of a challenge to the PG model. Second, in practice the problem of framing effects may seem less acute than I have been making it out to be, since we rarely know that a particular decision-maker is vulnerable to framing.

I believe that these responses underestimate both the theoretical and practical significance of the problem. Framing effects are "ubiquitous" (Tversky and Kahneman 2000, p. 220). Just as importantly, they operate in contexts - such as medicine - in which consent is supposed to be of paramount importance. Defenders of the PG model generally agree that the standards for valid consent must not be set so high that we often fail to meet them (Beauchamp 2010, p. 62; Manson and O'Neill 2007, p. 25). If I am correct, however, even generally competent, informed, calm, and sober decision-makers are often unable to give valid consent. In the study cited at the beginning of this paper, for example, there was a substantial shift in subjects' preferences depending on whether risks were framed in terms of mortality rate or survival rate. It may not be inconsistent for defenders of the PG model to admit that a substantial percentage of generally competent adults are unable validly to consent (or dissent) to surgical treatment for lung cancer. But such an admission may be devastating nonetheless, especially since the major selling point of the PG model is that it respects our actual choices about how we are to be treated. The less attainable valid consent is, the less will separate the PG model from the evidentiary model.

Turning to the second point mentioned above, the problem of framing effects cannot be easily brushed aside when we are able to determine that a particular decision-maker is vulnerable to framing, and such cases may not be all that uncommon. For example, doctors could actually confront their patients with alternatively-framed act descriptions in order to gauge their susceptibility to framing. If a patient's preferences are found to vary with the act description, the argument I have offered would imply that she is incapable of giving valid consent (or dissent).25 Indeed, it may not be unreasonable to test people for their susceptibility to framing before allowing them to give (or withhold) consent to acts likely to harm (or significantly benefit) them.

Even when we are unable conclusively to determine whether a particular decision-maker is vulnerable to framing, the problem of framing effects may still be of practical importance. Many philosophers believe that more harmful self-regarding choices must meet more stringent standards if they are to be protected from intervention (DeMarco 2002; Feinberg 1986, pp. 117-124).26 On this view, very harmftil self-regarding choices - such as the refusal of a life-

25 We cannot always test for susceptibility to framing in this way; as I noted earlier, the order in which preferences are presented can affect decision-making, and only one alternative can be presented first.

It is not entirely clear from Feinberg's discussion whether he believes that more harmful self-regarding choices must actually meet a higher standard of voluntariness, or whether he believes that we must merely be more confident that such choices meet a fixed standard. See Feinberg (1986, pp. 117 and 120).

Ö Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 15: Hypothetical Consent Hannah

530 J. Hanna

saving amputation - are morally protected from intervention only if we can be very confident that they meet the standards for valid consent. When a decision-maker is choosing in a context in which framing effects are known to operate, there is a non-trivial probability that she is vulnerable to framing. For instance, if one-fifth of otherwise competent adults are unable validly to choose among alternative treatments for a particular medical condition, we should be less confident (absent further evidence) that any particular person choosing in such circumstances has made a valid choice. In some cases, particularly when a great deal is at stake for the decision-maker and there is a very significant likelihood that she is vulnerable to framing, we may be justified in proceeding even without her actual consent. The arguments I have offered may thus have practical upshots unwelcome to defenders of the PG model.

Since framing effects are common and unavoidable, the plausibility of the PG model may be in jeopardy unless it can be shown that those who are vulnerable to framing can give valid consent. It does not appear that such an argument is forthcoming. In Section 3, 1 argued that if we accept a plausible principle of informed consent, we must either conclude that choices susceptible to framing are invalid or else maintain a narrow and fetishistic focus on mere quantity of information. When a subject is temporarily unable to give valid consent, many philosophers argue that our treatment of him should be guided by hypothetical consent. As I argued in Section 4, however, appeals to hypothetical consent do not resolve the problem of framing effects. Of course, there may be further responses available to defenders of the PG model, though they are somewhat difficult to anticipate given the scant attention that moral philosophers have paid to framing effects. But if I am correct that consent is irrelevant in framing cases, perhaps we should accept the evidentiary model. To be sure, a full defense of the evidentiary model lies outside the scope of this paper, and it has been rejected by most contemporary philosophers.27 The problem of framing effects should, however, lead us to re-examine the moral status of consent.28

27 As suggested above, opponents of paternalism generally reject the evidentiary model. This can be seen in their acceptance of the Volenti maxim described in n. 4 (Feinberg 1986, pp. 98-101, 176-178; VanDeVeer 1 986, pp. 42-44). The Volenti maxim would justify the anti-paternalist view that we cannot intervene in harmful conduct to which someone consents on the grounds that such conduct wrongly harms her. Feinberg, who provides perhaps the most thorough treatment of paternalism in the literature, explicitly rejects the evidentiary model and claims that "the right of self-determination" is "entirely undcrivative" of one's good (1986, p. 59). Thomson distinguishes claims against trespass, or "bodily intrusion," from claims against harm (1990, p. 209). Since she holds that claims against trespass are "fundamental" (p. 212) and can be waived through consent, consent is important on her view for reasons other than its evidentiary role.

I gratefully acknowledge support from a Northern Illinois University Research and Artistry Grant. For helpful feedback, I am indebted to an audience at the 2010 Rocky Mountain Ethics Congress.

References

Alexander L (1996) The moral magic of consent (II). Leg Theory 2:165-174 Beauchamp T (2010) Autonomy and consent. In: Miller F, Wcrthcimcr A (cds) The ethics of consent: theory

and practice. Oxford University Press, New York, pp 55-78 Bix В (2010) Contracts. In: Miller F, Wcrthcimer A (cds) The ethics of consent: theory and practice. Oxford

University Press, New York, pp 251-279 Brudney D (1991) Hypothetical consent and moral force. Law Philos 10:235-270 Christman J (1988) Constructing the inner citadel: recent work on the concept of autonomy. Ethics 99:109-124 Chwang E (2010) A puzzle about consent in research and in practice. J Appi Philos 27:258-272 DcMarco JP (2002) Competence and paternalism. Biocthics 16:231-245 Dworkin G (1971) Paternalism. In: Wasserstrom RA (ed) Morality and the law. Wadsworth, Belmont, pp

107-126 Dworkin G (1988) The theory and practice of autonomy. Cambridge University Press, New York

Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions

Page 16: Hypothetical Consent Hannah

Consent and the Problem of Framing Effects 531

Elster J (1982) Sour grapes - utilitarianism and the genesis of wants. In: Sen A, Williams В (eds) Utilitarianism and beyond. Cambridge University Press, New York, pp 219-238

Faden R, Beauchamp T (1986) A history and theory of informed consent. Oxford University Press, New York

Feinberg J (1984) Harm to others. Oxford University Press, New York Feinberg J (1986) Harm to self. Oxford University Press, New York Frankfurt H (1971) Freedom of the will and the concept of a person. J Philos 68:5-20 Gigerenzer G (2000) Adaptive thinking. Oxford University Press, New York Halvorsen PA, Selmer R, Kristiansen IS (2007) Different ways to describe the benefits of risk-reducing

treatments: a randomized trial. Ann Intern Med 146:848-856 Haward MF, Murphy RO, Lorenz JM (2008) Message framing and perinatal decisions. Pediatrics 122:109-

118 Hurd H (1996) The moral magic of consent. Leg Theory 2:121-146 Husak D (1981) Paternalism and autonomy. Philos Public Aff 10:27-46 Kahneman D, Tversky A (2000) Choices, values, and frames. In: Kahneman D, Tversky A (eds) Choices,

values, and frames. Cambridge University Press, New York, pp 1-16 Kamm F (2007) Intricate ethics. Oxford University Press, New York Klcinig J (2010) The nature of consent. In: Miller F, Wertheimer A (eds) The ethics of consent: theory and

practice. Oxford University Press, New York, pp 3-24 Kuflik A (2010) Hypothetical consent. In: Miller F, Wertheimer A (eds) The ethics of consent: theory and

practice. Oxford University Press, New York, pp 131-162 Manson NC, O'Neill О (2007) Rethinking informed consent. Cambridge University Press, New York McNeil B, Pauker S, Sox H, Tversky A (1982) On the elicitation of preferences for alternative therapies. N

Engl J Med 306:1259-1262 O'Neill О (1985) Between consenting adults. Philos Public Aff 14:252-277 Rcdelmeier DA, Rozin P, Kahneman D (1993) Understanding patients' decisions: cognitive and emotional

perspectives. JAMA 270:72-76 Salovey P, Williams-Piehota P (2004) Field experiments in social psychology: message framing and the

promotion of health protective behaviors. Am Behav Sci 47:488-505 Savulcscu J (1994) Rational desires and the limitation of life-sustaining treatment. Bioethics 8:191-222 Schneider TR, Salovey P, Apanovitch AM, Pizarro J, McCarthy D, Zullo J, Rothman AJ (2001) The effects

of message framing and ethnic targeting on mammography use among low-income women. Health Psychol 20:256-266

Schwab AP (2006) Formal and effective autonomy in health care. J Med Ethics 32:575-579 Scoccia D (1990) Paternalism and respect for autonomy. Ethics 100:318-334 Shafir E, Diamond P, Tversky A (2000) Money illusion. In: Kahneman D, Tversky A (eds) Choices, values,

and frames. Cambridge University Press, New York, pp 335-355 Sunstein C, Thaler R (2003) Libertarian paternalism is not an oxymoron. Univ Chic Law Rev 70:1 159-1202 Sunstein C, Thaler R (2006) Preferences, paternalism, and liberty. Philosophy 59(supp):233-264 Thomson JJ (1971) A defense of abortion. Philos Public Aff 1:47-66 Thomson JJ (1990) The realm of rights. Harvard University Press, Cambridge Trout JD (2005) Paternalism and cognitive bias. Law Philos 24:393-434 Tversky A, Kahneman D (1981) The framing of decisions and the psychology of choice. Science 211:453-

458 Tversky A, Kahneman D (2000) Rational choice and the framing of decisions. In: Kahneman D, Tversky A

(eds) Choices, values, and frames. Cambridge University Press, New York, pp 209-223 VanDeVeer D (1986) Paternalistic intervention. Princeton University Press, Princeton Wertheimer A (2003) Consent to sexual relations. Cambridge University Press, New York Westen P (2004) The logic of consent. Ashgate, Burlington (Vt.) Young R (1986) Personal autonomy: beyond negative and positive liberty. Croom Helm, London

Springer

This content downloaded from 216.17.119.80 on Wed, 28 Oct 2015 06:14:32 UTCAll use subject to JSTOR Terms and Conditions