You are on page 1of 28

~ tt: 'A\tHI GHkIJ I ]) ~ .

I~c;,) C
IW£Ll ~t;::/NG I 'nts- lU lAw D4-tl()N~ CIt: ~Oo'vl<'" ~Syc..YlOL..OGy I
NeW yo~J(! t1AH1~-t. ~Mt"" ~ (.{/Vt)t\-n ON

sUl
Th
16 Hedonic Adaptation of
w]-

Shane Frederick and George Loewenstein ad


tra
ter
( St
IS
br,
Hedonic adaptation l·eftrs to a reduction in the affec-fects (perceptual, physiological, attentional, mo- tiv
tive intensity of favomble and tmfav01·able circum- tivational, hedonic, and so on) of a constant or by
repeated stimulus.' Adaptation can occur at several
stances. This chaptel' discusses the purposes, undel·lying Fe
mechanisms, and most common functionalrepresenta- different levels-from overt behaviors that reduce pL
tions of hedonic ndaptntion. We then examine some of exposure to a stimulus, to molecular changes at w:
the methodological pl·oblems that hamper research in the cellular level that diminish the perceived or ex- se i
this m'ea and l'eview the litemtul'e on adaptation in perienced intensity of an objective stimulus. For
fOltl' negative domaim, (noise, imprisonment, bel'eave-example, stepping from a dark building into sun-
ment, and disability) and four positive domaim (foods,shine induces a variety of behavioral and physi-
F1
ological responses to the increased light level-
el'otic images, increases in wealth, and impl·ovmtents in A<
appearance pl'odttced by cosmetic sUl;gery). Followingturning away from the sun, squinting, pupil con- Fi
traction, photochemical changes in the retina, and
this m1iew, we discuss several circumstances that Pl'O- n~

mote 01' impede hedonic adaptation. We conclude by neural changes in the areas of the brain that pro- Ill '
discussing the dark side of hedonic adaptation-the cess the retinal signal. Counterparts to these pro- ne
negative comequC1Ices for individuals and society. cesses occur when we step back into the building. w<
Although there are limits to what these adaptive en
Those surgical operations in the field, the sickening processes can achieve (as indicated by our need for b)
butchery that shook even the toughest of the na- welding helmets and night vision goggles), they th
tives, had gradually deadened our sensibilities; we do allow us to see normally over luminance inten- Ill'
were no longer able to judge the horror of it all ... sities that vary by a factor of over one million.' tic
blood flowing and spurting, the unbearable smell Hedonic adaptation is adaptation to stimuli that 11<
from suppurating wounds-all this left us unmoved. are affectively relevant. It relies on many of the h~
-Maurice Herzog, Annapurna (1952) same processes that underlie other types of adapta- d;1
tion and is often derivative of these processes. For e\'
MOST OF US are familiar with striking exanlples of example, hedonic adaptation to a foul odor is a VI '
people who seem to be adapting well to circum- direct consequence of sensory adaptation because d
stances that are extremely adverse . We may have an odor that is perceived as less intense will be m
seen footage of malnourished children playing hap- experienced as less unpleasant. Similarly, hedonic in
pily in garbage dumps or know of severely hand- adaptation to paraplegia may result in part from
icapped people who maintain a cheerful disposition physiological adaptations, such as increases in up - ar.
in spite of their disabilities. However, counterexam- per body muscle mass, that enable paraplegics to (h
ples come to mind as well: people who seem per- manipulate their wheelchairs more effectively. How- an
petually miserable, or those who were "never quite ever, many hedonic stimuli are cognitive rather til
the same" after experiencing some devastating than sensory. Thus, many of the processes involved ti\
event. This chapter examines both the extent and in hedonic adaptation involve cognitive changes- 11\
limits of hedonic adaptation-processes that attenu- in interests, values, goals, attention, or character- th
ate the long-term emotional or hedonic impact of ization of a situation. For example , the effects of l~
favorable and unfavorable circumstances. diminished mobility accompanying paraplegia will st
be less important if one develops new interests st
that do not demand as much mobility, such as ce
HEDONIC ADAPTATION playing Scrabble instead of tennis. 3 Hedonic adap - S:
tation may also involve consciously directing one's m
Adaptation, in its broadest sense, refers to any ac- attention away from troubling thoughts or engag-
tion, process, or ~echanism that reduces the ef- ing in actiyities likely to direct attention elsewhere, ti'
Hedonic Adaptation 303

such as playing sports or keeping busy at work. our objectiye circumstances. To illustrate this sec-
This strategy may diminish at least the "quantity" ond function of hedonic adaptation, consider a
of saddening cognitions, if not their intensity, man who has been incarcerated. The prisoner's sit-
when they do intrude into consciousness. Hedonic uation is illustrated in figures 16.1 and 16.2,
adaptation may also be facilitated by cognitive which specify hedonic intensity in terms of a Pros-
transformations of situations-for example, by in- pect Theory value function (Kahneman and Tver-
terpreting a trageqy as a "learning experience" sky 1979 ), in which utility is determined by the
(see, for example, Janoff-Bulman and Wortman difference between the prisoner's current state and
1977). Finally, neurochemical processes within the the state to which he has adapted. Before he has
brain may work to oppose persistent intense nega- adapted to his incarceration (see figure 16.1), he is
tive or positive affect (Solomon and Corbit 1974) miserable in his seven-foot cell. At this time, the
by desensitizing overstimulated hedonic circuitry. difference in value bel:\veen a seven-foot cell and a
For example, continual high-level cocaine or am- nine-foot cell seems insignificant, and he would
phetamine use may diminish the fimctioning of re- haye little motivation for acquiring the larger cell.
ward pathways in the brain (see, for example, Cas- However, after the prisoner has adapted (see fig-
sens et al. 1981; Wise and Munn 1995). ure 16.2 ), not only is he happier, but the differ-
ence in hedonic value bel:\veen the small and large
cells is much greater, and the prisoner would, cor-
Functions of Hedonic Adaptation
respondingly, be increasingly motivated to secure
Adaptive processes serve t",TO important functions . the larger cell.
First, they protect organisms by reducing the inter- Because the persistence of an aversive state is an
nal impact of external stimuli. (For example, sweat- indication that it cannot be changed, hedonic ad-
ing causes evaporative cooling, vvhich reduces or aptation may pre\'ent the continued expenditure
negates the increase in body temperature that of energy in futile attempts to change the un-
would otherwise accompany an increase in ambi- changeable and redirect motivation to changes
ent temperature.) Second, they enhance perception that can be made. To paraphrase a famous aphor-
by heightening the signal value of changes from ism, hedonic adaptation "provides the serenity to
the baseline level: The types of adaptive processes accept the things one cannot change, the courage
involved in visual perception illustrate this func- to change the things one can, and the wisdom to
tion . When we first walk indoors from the after- know the difference."
noon sun, we have difficulty seeing and would
have difficulty judging which of two dark rooms is
Shifting Adaptation Levels
darker. After we have been inside for a while, how-
Ve1"SUS Desensitization
ever, the adaptive processes that have restored our
vision have also restored our sensitivity to luminance Although we haye used the term "adaptation"
changes at the new, lower light level- enabling broadly to denote anything that reduces the sub-
us to detect, for example, a single lightbulb burn- jective intensity of a given stimulus, it is important
ing out in an auditorium lit by hundreds. to distinguish bel:\veen adaptive processes that di-
Hedonic adaptation may serve similar protective minish subjective intensity by altering the stimulus
and perception-enhancing functions. Hedonic states level that is experienced as neutral (shifting adap-
(hunger, thirst, pain, sleepiness, sexual excitement, tation levels) and adaptive processes that diminish
and so on) are necessary because they direct atten- the subjective intensity of the stimulus generally
tion to high-priority needs and provide the mo- (desensitization). Both processes diminish the sub-
tivation to engage in behaviors that satisfy those jective intensity of a given stimulus, but shifting
needs and to avoid behaviors that compromise adaptation levels preserve or enhance sensitivity to
them (Cabanac 1979; Damasio 1994; Pribram stimulus differences, whereas desensitization di-
1984, 2) . However, persistent strong hedonic minishes such sensitivity.
states (for example, fear or stress) can have de- A shifting adaptation level is illustrated by the
structive physiological concomitants, such as ul- prisoner example as described earlier. The condi~
cers, circulatory disease, and viral infections (see tions initially experienced as intensely negative are
Sapolsky, this volume). Thus, hedonic adaptation later experienced as hedonically neutral (Vo [7 -ft.
may help to protect us from these effects. cell] < 0, whereas Vt [7-ft. cell] = 0), and as the
Hedonic adaptation may also increase our sensi- prisoner's hedonic state improves, he also becomes
tivity to, and motivation to make, local changes in 11I0re sensitive to stimulus differences ([ Vt (9-ft.
304 Well-Being
FIGURE 16.1 Prisoner's Situation Prior to Adapting to Incarceration

Va [freedom]

Va [9-ft. cell] ==;;;;::::b""--


Va [7 -ft. cell]

cell) - Vt (7-ft. cell)] > [Vo (9-ft. cell) - Vo (7- creases over time (Va [7 -ft. cell] < Vt [7 -ft. cell]),
ft. cell)]). Desensitization, in contrast, involves a but unlike shifting adaptation levels, the prisoner's
change in the shape, rather than the position, of sensitivity to change also decreases ([ Vt (7-ft. cell)
the response function (see figure 16.3). Like a - V, (7ft. cell)] < [VO (9-ft. cell) - Vo (7-ft.
shifting adaptation level, the hedonic intensity de- cell )]). People who are hedonically desensitized-
th ;·
FIGURE 16.2 Prisoner's Situation Following Adaptation to Incarceration tJ'l
wi

At
V, [freedom] - - -- - - - - - + - -- - -- =..,..---
At
01:
V, [9-ft. cell] - - -- - - -- - i--7f' of
cc
IS
pc
V, [7-ft. cell] ----------,.if--+------f------------- Sc
freedom bl
111
tic
th
pe
fa
(L
b(
re
n,
l'
Hedonic Adaptation 305
'I
FIGURE 16.3 Sensitization and Desensitization

hedonic
intensity ""/,/""""""'" l,e",";",;OO

1 desensitization
. ... ... .. . ... . ... .. . ..... . ... .

stimulus intensity

that is, "hardened," "jaded," "jaundiced"- are Hall 1974).7 One might also consider mood-de-
typically unmotivated to make any kind of change , pendent memory as a form of sensitization, be-
whether local or global. 5 cause a depressed mood cues negative thoughts,
which intensify the depressed mood, which, in

Adaptation Versus Sensitization


turn, cues further negative thoughts, and so on
(Bower 1981).
>'-,
.. ,'-
: 13
; 'i3
\ '

Adaptation is not the only possible response to


ongoing stimuli. Sometimes the hedonic intensity Modeling the Functional Form of Adaptation
of a constant stimulus increases over time-a pro-
cess called "sensitization" (Groves and Thompson Helson (1947,1948,1964) was one of the first to
1973). The increasing irritation produced by ex- propose a quantitative model of adaptation. His
posure to a disliked roommate is a familiar example. model was an attempt to characterize formally the
Sensitization is rarely discussed in the literature effect of past stimuli on the subjective experience
but has been observed for some painful and other of a current stimulus. He introduced the notion of
intense stimuli (Thompson et al. 1973). Sensitiza- an adaptation level (AL)-the level of a stimulus
tion to pleasurable stimuli might be illustrated by that elicits no response (or, in the context of he-
the alleged increasing pleasure from successive ex- donic adaptation, the level that is affectively neu-
periences with marijuana, or high-quality wine, tral) . Helson proposed that AL is the average of
food, and culture.6 The "mere exposure effect" past stimulus levels (X = stimulus level, and t =
(Zajonc 1968) is another example of sensitization time ):8
because it is an inc1'easing hedonic response to
repetition of stimuli that are initially hedonically
neutral- such as pictures of male faces (Zajonc (16 .1)
1968) or Pakistani folk music (Heingartner and

d
306 Well-Being
He further proposed that an individual's hedonic A formulation that does explicitly include tlle enc<
state (u) at any time (t) is a function of the differ- temporal component and has been widely used to bad
ence between the current stimulus level, Xt'J and model adaptive processes (See for example, Har- S
AL: die, Johnson, and Fader 1993; March 1988; Ryder the
and Heal 1973) is: fon
(16.2 ) sati'
ALt = aXt - 1 + (1 - a)ALt _ b (16.3) abs.
wheref(O) = 0, and!, ;;:: 0, as is true of the value
function in figures 16.1, 16.2, and 16.3. Because with 1 ;;:: O. The a parameter determines tlle speed 1'at,
AL gradually converges to the value of any con- of adaptation. If a = 1, then the adaptation level is an d
stant stimulus level, X" the absolute difference be- equal to the last period's stimulus level, so tllat tlle 19~
tween X t and AL, gradually diminishes over time , individual's hedonic state will depend solely on the leyi
as does the absolute value of Un which is a func- difference in stimulus level between tlle last period hc e
tion of this difference. Thus, Helson's model cap- and the current period. If a = 0, tllen tlle adapta- hee
tures the essence of adaptation-that persistent tion level will not be influenced by past stimulus mc·
bad things gradually become less aversive, and per- levels at all. Equation 16.3 can be rewritten as: for
sistent good things gradually become progressively I
ALt - ALt - 1 = ALt - 1 ), (16.4)
a(Xt - 1 -
less pleasurable. of
Helson's specific formulation has, however, been which highlights the equation'S implicit assump- stir
criticized on a number of grounds. Sarris (1967 ) tion that the change in the adaptation level from me
demonstrated (in the context of judgments of the one time period to the next is proportional to the en(
heaviness of hefted weights) that extreme stimulus difference between the last period's stimulus level me
levels do not influence AL as much as Helson's and the last period's adaptation level. By applying th e
the01Y predicts. Parducci (1968) found that judg- equation 16.3 recursively, it is also possible to Mr
ments are influenced by factors other than the mean show that the adaptation level at any point in time fae
stimulus level and proposed "range-fi'equency" the- can be represented by a weighted average of past Or
my as an alternative to Helson's theory. Range- stimulus levels, with recently experienced stimuli te;'
frequency theory postulates that judgments of per- receiving greater weight than those experienced in SC I
ceptual stimuli (lengths ofiines, heaviness of weights) the more distant past.lO po
and hedonic stimuli (odors or rewards of money) To illustrate some of the implications of this for- los
are jointly determined by two separate contextual mulation, assume that f is the identity function, g t"~

features- the position and the rank of a stimulus that a = 0.5, and that X represents the objective of
witlun the stimuli included in some judgmental level of some positive stimulus occurring on seven lat
context. Suppose, for example, that a traveling consecutive days (for example, number of encyclo- le\
salesman earns $0, $95, $100, $100, and $90, re- pedias sold, hours of sunshine, minutes of atten- th :
spectively, on each of his first five days of work. tion from one's spouse). (n-
The range principle alone would cause $90 to be si z
evaluated favorably because $90 exceeds the mid- e\'<
point of the stimulus range ($ 50) . The frequency Mon. Tue. Wed. Thu . Fri. Sat. Sun. th .
principle alone would cause $90 to be experienced X 3 5 12 4 5 5 5 co
negatively because it ranks below the median stim- AL 3' 3 4 8 6 5.5 5.25 bl,
uli ($95). The aggregate judgment will depend on u 0 +2 +8 -4 -1 - 0.5 - 0.25 m.
the relative importance of these two principles (see Notes: X = stimulus level; AL = adaptation level; u = he- an
Parducci 1995). donic state. dr
Helson's and Parducci's models of adaptation 'For simplicity, we assume that the stimulus level had been co
share a common shortcoming-they do not ex- constant at a level of 3 in the past. ad
plicitly account for the role of time. For example, p~1

although both theories imply that the judged This example illustrates two principal features of ttl
quality of a dinner will be influenced by the qual- most models of hedonic adaptation: (1) the affec -
ity of past dinners, they make no explicit distinc- tive intensity and valence of a given stimulus level Is
tion between the effects of a fancy dinner experi- depend on past stimuli (note that Tuesday is a
enced last week and one enjoyed last year." Intu- good day, and Friday is a bad day, though the ob -
itively, it seems far more likely that recent events jective circumstances are the same on both days);
will have a larger impact on one's AL than events and (2) the affective intensity of a constant stim-
that are more reillote in time. ulus diminishes over time (note that the experi-
Hedonic Adaptation 307

ence of receiving five units of X gets better [less many experiences are inherently pleasant and do
bad] from Friday to Sunday). not require a positive comparison to make tl1em so
Several authors have proposed modifications to (see Kahneman, tl1is volume). An encyclopedia
the "weighted average" model of adaptation level salesman who likes his job can have a pleasant day
formation. Some have argued, for example, that in spite of average sales, just as an average pizza
satisfaction with income depends not just on the can taste good.
absolute improvement over past levels but on the A simple way of modifYing adaptation models
rate of improvement over time (Frank 1992; Hsee to account for this obvious truth is to add a posi-
and Abelson 1991; Hsee, Abelson, and Salovey tive or negative constant (u t = C + f[X t - ALt ])
1991; Loewenstein and Sicherman 1991). Strahi- to reflect the "inherent" goodness or badness of
levitz and Loewenstein (1998) have proposed that an experience- that is, the amount of pleasure or
hedonic adaptation to improvements is faster than pain one receives from an "average" experience
hedonic adaptation to deteriorations, an asym- (when X t = ALt ). This would reflect tl1e fact that
metry that implies a separate, larger a parameter celtain experiences continue to be pleasurable
for gains than for losses. even when the current level is at, or even below,
I t is difficult, however, to formally represent all AL. Experiences like pizza and sex that have a
of the factors that determine the influence of a large intrinsic pleasantness component "'ould re-
stimulus on adaptation level. Certain extreme or quire a large constant term, whereas a smaller con-
memorable events may have a much greater influ- stant would apply to experiences such as an exam
ence on adaptation level than most variants of the score, where experience is evaluated primarily by
models discussed here would predict. For example, comparative standards. This would capture the in-
the memory of a single romantic encounter with tuition that it takes a very bad pizza to make the
Mr. or Ms. Perfect may forever diminish the satis- experience unpleasant, but only a slightly lower
faction of subsequent interactions with Mr. or Ms. grade than you are used to getting to spoil a grade
Ordinary- even if that single encounter happened tl1at is good by other standards. I'
ten years ago and eyen if it has been succeeded by
scores of lesser experiences. Parducci (1995 ) re-
The Influence of Future Stimuli lj ,
ports that the memoirs of Russian emigres who
lost their wealth fleeing the 1917 revolution sug- Although adaptation level is typically assumed to
gest that they felt impoverished for the remainder be a function of past stimuli, it may also depend
of their lives because they continued to judge their on anticipation of future stimulus levels-a pro-
later material circumstances against the wealth cess called "feedforward." Feedforward was dra-
level they had enjoyed many years before, rather matically illustrated by Siegel and his colleagues G
than against some ayerage of their lifetime wealth (Siegel et al. 1982; Siegel, Krank, and Hinson .. Q
..;
(much less against a weighted average that empha- 1988). Rats injected with gradually increasing
sized the recent past). Conversely, even very recent doses of heroin came to tolerate dosages that
events may not influence our reference points if would have been lethal to them at the outset of
they do not seem relevant to some judgmental me study. However, the resistance to heroin was
context. ll For example, a Christmas bonus proba- observed only when the injection was administered
bly does not reduce our satisfaction with our nor- at a regular time in a regular location . The identi-
mal monthly salary in January, and resort owners cal dose of heroin administered at an unexpected
are probably not disappointed by the usual marked time or in a new location was often fatal. Appar-
drop in attendance following Labor Day. These ently, me rats' adaptive response to the heroin was
complexities pose serious challenges to theories of not triggered solely by ingestion of me drug, but
adaptation that attempt to model the impact of also by cues that became associated wim, and mere-
past stimuli as a simple function of their magni- fore predictive of, imminent heroin use . Feedfor-
tude and recency. ward processes have also been implicated as an im-
portant process underlying human drug addiction
Is an Average Stimulus Hedonically Neutral? (for reviews of mis literature, see Laibson 1997;
Loewenstein 1996).
According to a popular joke, sex is like pizza: Feedforward seems to apply to hedonic phe-
when it is good, it is really good, and when it is nomena as well. Van Praag and his colleagues (for
bad, it is still pretty good. This captures an impor- example, van Praag 1977; van Praag and van der
tant point that models of adaptation often miss: Sar 1988) found that tl1e income people view as
308 Well-Being
"sufficient" depends, in part, on their expectations tein 1998), their formation and relative weighting til
for the future.13 There is also evidence that the fi- has not been investigated empirically. ra
nal days of a prison sentence are often regarded as qt
the most frustrating-suggesting that prisoners tJ-
are "adapting" prematurely to the anticipated joys MEASUREMENT ISSUES tc
of freedom (Bukstel and Kilmann 1980, 482). m
Such adaptation may be responsible for the fre- A comprehensive treatment of the literature on if
quently noted anomaly that escape attempts are hedonic adaptation is impossible without consider- sr
often made near the end of a prison sentences. I4 ing tl1e many methodological problems tllat plague A
FeedfOlward processes may also help explain the this line of research. Sl

paradoxical observation that revolutions tend to tc


occur just after conditions begin to improve-the Scale Nom~ing
"revolution of rising expectations" (Gurr 1970). I'
Initial improvements may produce expectations of The literature on perceptual adaptation has hosted at
future improvement and create frustration when a long-standing debate about whether adaptation
the anticipated improvements do not occur or are is "real" or simply a matter of relabeling. Stevens n,
not realized quickly enough. ( 1958), for example, argues that some of the evi - 1Il

dence Helson cites to support his adaptation level si-


theory merely illustrates the judgmental relativity d,
of labels and not "adaptation" in its conventional tlt
Multiple Reference Points ti'
sense (see also Krantz and Campbell 1961). For
T he literature on sensOlY and perceptual adapta- example, when a given weight is called "very ai'
tion has commonly assumed that tile adaptation heavy" in the context of lighter weights and "very ((

level in a particular domain can be characterized light" in tile context of heavier weights, it seems
by a single summary number. As Helson (1947) more plausible that a different label is being given
comments, "For evety excitation-response con fig- to tile same (or a similar) perception than that the
til'ation, there is assumed a stimulus which repre- perception itself is changing. I6
sents the pooled effect of all tile stimuli to which Similar issues plague research on hedonic adapta- N
the organism may be said to be attuned or adapt- tion because answers to questions about well-being if
ed" (2). The assumption of a single adaptation inherently confound respondents' "true" happi- (r

level may be reasonable for many types of adapta- ness with their semantic conventions, reference a
tion. For example, the adaptation level of spiciness groups, and other factors tllat influence their in- us
marking the boundary bet\veen too spicy and too terpretation of the response scale on which well-
bland may be accurately modeled by a single num- being is reported. When asked for judgments of pr
ber representing the average spiciness of the foods well-being (for example, "How happy are you on tb
one is used to consuming. In other contexts, how- a scale from 0 to 100?"), respondents must usually sc
ever, tile effect of past stimuli cannot be summa- decide for tllemselves what tile endpoints of the 10
rized so simply. Consider the case of an individual response scale represent. Thus, one person may in- ,,·1
who earned $20,000 annually for her first six years terpret 100 as tile highest level of happiness tlley di
on me job, got a promotion that raised her salaty have thus far experienced, another may interpret it tJl
to $50,000, tllen was transferred t\vo years later to as the highest level of happiness they can imagine Se
a different department \\'here she earned $40,000. experiencing on earth, and a third might interpret ,Yt
What is her adaptation level income? It is possible it as a hypothetical state of ideal, heavenly bliss. etl
that she compares her current salary to a single ad- Similarly, one person might interpret 0 as the ab-
aptation level lying somewhere between $20,000 sence of salient good feelings, while another may
and $50,000. It seems more likely, however, that resetve that label to represent tile most intolerable
she has t\vo different adaptation levels-one at hell imaginable. .
$20,000 and tile otller at $50,OOO - that are in- An obvious problem arises if factors that affect
voked in ditferent situations and bOtll of which respondents' "true" happiness also affect their use
contribute to her satisfaction or dissatisfaction of scales. Suppose, for example, that a sample of
with her current salary I; Although the issue of quadriplegics and control subjects both rate tlleir
multiple reference points has been raised (see, for happiness as 80 on a 100-point scale. This number
example, Boles and Messick 1995; Kahneman may accurately represent tile true happiness levels me
1992; Schweitzer 1995; Strahilevitz and Loewens- of tl1e two groups. However, it may also overstate ha'
Hedonic Adaptation 309

the happiness of the quadriplegics if they implicitly efFects. For example, Kmpat (1974) found that
rate their own happiness relative to that of other prior exposure to tllreat not only lowered subjects'
quadriplegics (who may be much less happy than ratings of tlle "threateningness" of a given situa-
the control group )17 or elevate their current rating tion but also reduced their galvanic skin conduc-
to reflect the contrast to their extreme despair im- tance (a physiological measure of experienced
mediately following the onset of their disability, or threat)-suggesting that the difference in ratings
if they have adopted lower standards for the inten- was not purely a semantic phenomenon. Dar,
sity of positive affect that warrants the rating 80. Ariely, and Frenk (1995) used both verbal and be-
All of these forms of norming could lead re- havioral measures of pain in veterans who had suf-
searchers to overestimate the degree of adaptation fered injuries of varying severities in the past: pain
to paralysis. threshold-the length of time the subject could
Some researchers (for example, Diener et al. hold his index finger in hot (48 degrees Celsius)
1985) have suggested that frequency measures of water betore he classified the sensation as pain-
affect (the proportion of time an individual feels ful-and pain tolerance-the length of time be -
"good" or "bad") may be less susceptible to scale fore he pulled his finger out of the water. The au-
norming than intensity measures (how happy the thors found that veterans with more severe past
individual is at a given time, or how happy he or injuries not only held their finger in longer before
she is overall). Evidence from some judgmental classifying tlle sensation as· painful (10 .1 seconds
domains suggests that contextual effects do not in- versus 4.7 seconds) but also held their finger in
tluence the judged valence (positive versus nega- longer before terminating the experiment (5 8 sec-
tive) of a stimulus even if they do int1uence the onds versus 27 seconds).18 This is compelling, but
absolute value of the judgment. In other words, not conclusive, evidence that the veterans with
contextual factors may make things more or less more severe prior injuries actually felt less pain. It
good or more or less bad, but they do not make is possible that the more severely injured veterans
bad experiences good or good experiences bad. not only had difFerent semantic standards for the
(Changing the context of reference may turn a intensity of sensation tllat warrants tlle label "pain"
-7 into a - 2 but will not turn a -2 into a +2.) but also had different standards about how much l~ .
Kicking someone's dog is not rated positively even pain they should be willing to endure.
if it is evaluated in the context of very severe Not all studies, however, have observed parallel
crimes (Parducci 1968 ), and losing fifty dollars in results between physiological 3.nd self-report mea-
a card game is not considered good even if you sures. For example, Paulus, McCain, and Cox
usually lose more (Marsh and Parducci 1978). (1973) reported a positive relation between pris-
However, frequency measures have their own oner density and palmar sweat (a physiological
problems. First, valence does not capture every measure of stress), but not between density and
thing one might want to know about affect. A per- subjective appraisals of crowding. Ostfeld and his
son who experiences extreme highs and shallow colleagues (1987) found that prisoner density was
lows clearly enjoys a higher quality of life than one correlated with blood pressure, but not witll mea-
who experiences the opposite, even if the two in- sures of anxiety, hostility, or depression. Zisook,
dividuals spend identical proportions of time on Shuchter, and Lyons (1987) observed a dramatic
the positive and negative sides of the hedonic ledger. decrease in tearfulness in the first year following
Second, it is not obvious that even the valence of the death of a spouse, but no significant decline in
well-being judgments is immune to scale-norming reported depression . Divergence between mea-
effects. If you ask someone whether she is having a sures that are intended to measure tlle same con-
good day, she must still decide where to make the struct may indicate the presence of measurement
division between good and bad, and this choice error or bias, or it could point to the existence of
may be affected by tlle distribution of recent expe- multiple independent dimensions of hedonic status.
riences that set the context for that judgment.
Thus, a "no" response from a newlywed might in -
Context an.d Deman.d Effects
dicate a more positive level of affect than a "yes"
response from a person receiving chemotherapy. Some judgments requested from subjects in adap-
There have been occasional attempts to avoid tation studies are vulnerable to social-desirability
the problems created by scale norming by supple- effects. Adaptation may be overrepOlted if subjects
menting verbal responses with physiological or be- feel pressured to exaggerate how much they bene-
havioral measures "tllat should be immune to such fited from some intervention like a support group
310 Well-Being
(Conway and Ross 1984). Conversely, adaptation subgroups. For example, Janal and his colleagues
may be underreported if subjects fear appearing (1994) found that habitual joggers displayed a
callous if they truthfully report substantial or com- higher pain threshold for cold than nonjogging
plete emotional recovery from some negative event, control subjects. They noted, however, that this
such as the death of a spouse. Indeed, many psy- relative insensitivity could suggest either that jog-
chiatrists have considered the lack of affective reac- ging produces adaptation to pain or that people
tion to some negative event as a disorder. For who are more tolerant of discomfott are likely to
example, Wortman and Silver (1989) cite an Insti- select jogging as their form of exercise. 2l Longi-
tute of Medicine report by Osterweis, Solomon, tudinal studies (which compare the affect of peo-
and Green (1984) that stated: "The absence of ple at different points in time) are confounded by
grief phenomena following bereavement repre - the problem of history (other events that occur
sents some form of personality pathology.... Pro- during the observed time period), matul'ation (ef-
fessional help may be warranted for persons who fects that result from the aging of the observed
show no evidence of having begun grieving" (18). population), and regression to the mean if study
They also cited an article by Siller (1969) that participants are selected based on an extreme value
maintained: "[When] a newly disabled person of some hedonic measure (such as being coun-
does not seem to be particularly depressed, this seled for depression).
should be a matter of concern. [The person]
should be depressed because something relatively
significant has happened, and not to respond as
Measul'ement Ambiguities O'eated by a
such is denial" (292). Nonconstant Stil1tulus
Futthermore, apart from demand effects, simply
asking respondents about a particular event may Adaptation is well defined only when a response
temporarily raise its prominence. l9 This problem is diminishes or remains the same despite constant
noted by Lehman, Wortman, and Williams (1987) or increasing stimulus level; sensitization is well
in a paper about the long-term effects of losing a defined only when a response increases or stays the
spouse or child in a motor vehicle crash: "The same despite a constant or decreasing stimulus
most serious threat to the validity of these findings level. It is difficult to determine whether adapta-
concerns the possibility that respondents' distress tion or sensitization has occurred in situations in
scores were artificially intlated because ... by con- which both the stimulus and response levels move
tacting respondents and requesting an intetview, in the same direction. For example, suppose the
we raised a series of troubling issues that are not ainperage of an electric shock delivered to an ex-
normally on their minds" (228) . perimental subject is increased over a series of tri-
Asking about a troubling event may not only als, and the behavioral or subjective response mea-
raise its prominence but actually induce negative sure also increases-the subject yells louder, jumps
affect, a possibility that raises ethical as well as higher, or exhibits a greater elevation in blood
methodological issues. pressure. It isn't clear whether this observation
suppOtts adaptation (the response increase was less
Threats to Intel'11al Validity than it would have been without adaptation), sensi-
tization (the response increased more than would
Interpretation of study results from most empirical be expected from the stimulus increase), or nei-
research on hedonic adaptation is nonexperimen- ther (the response increase was commensurate
tal, and, thus, subject to the problem of con- with the stimulus increase). For another example,
founds. Retrospective studies (in which people are suppose that the concentration of some atmo-
asked to report their current subject state and to spheric pollutant doubles, and the subjective rat-
recall past states) are problematic because recollec - ing of air quality decreases from a 4 to a 3 (on a
tions are notoriously inaccurate and may reflect 5-point scale), or the average time spent outdoors
the individual's implicit theories about change decreased by twenty minutes. Would the changes
processes rather than a veridical recollection of in these dependent variables indicate adaptation or
prior hedonic states (Ross 1989).20 Cross-sectional sensitization to the increased air pollution? (Of
studies (which compare the affect of people who course, whether we choose to label the change in
have been exposed to a particular stimulus with the dependent variable "adaptation" or "sensitiza-
that of those who have not) suffer from the diffi- tion" may be less important than an evaluation of
culty of matchirfg the exposed and non exposed whether the effects are "big" or "small" relative to
Hedonic Adaptation 311

our expectations, or relative to the effects wrought interviewed first-year college students about their
by other changes.) reactions to dormitory noise in the first few weeks
of the school year and again at the end of the year.
Reported annoyance increased significantly. In a
Is It Possible to «Adapt to Depression»?
study on highway noise, Weinstein (1982) inter-
A second source of conceptual ambiguity regard- viewed a panel of residents four montl1s and six-
ing hedonic adaptation is illustrated by a col- teen months after the highway was opened. Re -
league's comment that people "can't adapt to de- ported irritation remained constant, and residents
pression," a speculation that struck us as strange, became incl'easingly pessimistic about their ability
after some reflection. Because depression is itself a to adjust to the noise (the proportion who said
subjective state, it is difficult to interpret what they would not be able to adjust increased from 30
adapting to depression could mean-that a person to 52 percent). Perhaps more tellingly, in two con-
doesn't feel bad when he feels bad? The concept of trol groups interviewed only once, fewer than one-
adapting to pain also seems problematic, though third of residents intelyiewed at t = four months
less so. While it still seems unusual to say, "Al- spontaneously mentioned highway noise as some-
though I'm still experiencing intense pain, I've tl1ing they disliked about the neighborhood, com -
completely adapted to it," the statement may pared to over half of those interviewed at t = six-
make sense if one separates the experience of pain teen months. Jonsson and Sorensen (1973) had
into sensa 1)' and affective components, as many earlier found comparable results, in a similar study.
pain researchers have done (Al1leS, Blanchard, and Cohen and his colleagues (1980, 1981 ) found lit-
Leventhal 1983; Fernandez and Turk 1992; Le - tle evidence of adaptation to aircraft noise among
venthal et al. 1979; Leventhal et al. 1989; Price, Los Angeles sc hoolchildren. Children in noisy
Harkins, and Baker 1987). For example, in a study schools did more poorly on a cognitive task and
by Price, Harkins, and Baker (1987), subjects ex- had shorter attention spans and higher systolic and
periencing pain separately rated its sensory and af- diastolic blood pressure than those from matched
fective components using two visual analog scales control schools. Neither longitudinal nor cross-
whose endpoints were either "the most intense sen- sectional analyses suggested that the differences in
I"':. .
sation imaginable" (the sensory component) or these outcomes diminished over time, with the ex-
"the most unpleasant feeling imaginable" (the af- ception of systolic blood pressure, where the dif- ~:
fective component). By considering the components ferences between noisy and quiet schools did de- '~ '

of pain separately, one could interpret "adapting crease slightly.


to pain" as a reduction in intensity of the affective
component, despite a constant or increasing sen- Incal'cel'ation Although incarceration is designed
sory component. 22 to be unpleasant, most of the research on adjust-
ment to prison life points to considerable adapta-
tion following a difficult initial adjustment period."
RESEARCH ON ADAPTATION IN In a study of British prisoners, Flanagan (1980)
SPECIFIC DOMAINS observed generally successful long-term adjust-
ment (although prisoners reported tl1at specific
We now turn to a review of the literature on he- stressors, such as the loss of relationships with
donic adaptation to specific experiences and con- people outside the prison, became increasingly dif-
di tions, both undesirable (such as noise) and desir- ficult to deal with as time passed). In a sample of
able (such as an increase in income). Our review is inmates whose sentences varied from one month
limited by the shOltage of high-quality empirical to ten years, Wormith (1984) observed significant
studies." We discuss only the small subset of do- improvement over time in deviance, attitude, and
mains where it seemed possible to draw at least personality measures. Mackenzie and Goodstein
tentative conclusions about the extent of hedonic (1985) recorded similar results. Zamble and Pro-
adaptation. porino (1990) and Zamble (1992) reported de-
clining dysphoria, a reduction in stress-related
Undesi,'able Experiences problems such as sleep disturbances, and decreas-
ing boredom over the course of prison sentences.
Noise Few studies have observed adaptation to Even inmates placed in solitary confinement for
noise, and some have even found sensitization (for a long periods adapt to their circumstances. Deaton
discussion, see Weinstein 1982). Weinstein (1978) and his colleagues (1977) report that American
312 Well-Being
soldiers placed in solitary confinement during the However, the progressive deterioration associated p
Vietnam War (for up to six years) were highly suc- with tllese diseases makes it difficult to measure t!
cessful in devising effective coping mechanisms. the degree of adaptation (see the section on non-
Suedfeld and his colleagues (1982) observed sim- constant stimulus). The progression of multiple
ilarly successful adaptation to solitary confinement sclerosis, for example, typically leads to increasing
in civilian prisons. Indeed, some found it diftlcult numbness, paralysis, spasticity, fatigue, vertigo, a
to adjust to release from solitary confinement. problems in bladder control, sexual dysfunction,
Although there appears to be considerable he- difficulty in communication, cognitive deteriora-
donic adaptation to imprisonment throughout tion, and visual impairments (Antonak and Livneh
most of the prison term, evidence suggests that 1995). Thus, in contrast to paralysis victims, whose \
prisoners find their incarceration less tolerable as they condition is likely to remain constant over time,
approach the end of their sentence-presumably sufferers of such debilitating diseases must cope
because they begin to compare the circumstances not only with the disabilities resulting from the
of incarceration with the freedom they are begin- cumulative deterioration they have thus far suf-
ning to anticipate (see section on feedf011Pard). fered but with new impairments as their disease .J
Bukstel and Kilmann (1980) reviewed thirty-one progresses. Even maintaining a constant hedonic
studies of adaptation to prison and concluded that state in the face of these deteriorating conditions
the hedonic response to incarceration shows a cur- would be impressive evidence of hedonic adapta-
vilinear pattern over time, with long-term im- tion. Thus, the hedonic deterioration that is com-
provement in functioning followed by a short- monly obsetved does not provide evidence that
term deterioration near the end of the sentence. adaptive processes are not occurring-only that
they are not occurring fast enough to keep pace
Disability/Disease Several studies have observed with the progression of the disease.
substantial adaptation to disability. The most fa -
mous is Brickman, Coates, and Janoft--Bulman's Loss (Bel'eavement) Studies of bereavement have
(1978) "Lottery Winners and Accident Victims: Is generally found that those who lose a child or
Happiness Relative?" Accident victims (those who spouse experience intense and prolonged grief
had become paraplegic or quadriplegic as a result (Dyregrov 1990; Lehman et al. 1987; Sanders
of an accident within the last year) rated their hap - 1980; Weiss 1987), and that such dramatic eftects
piness as 2.96 on a 5-point scale (above the mid- are much rarer in those losing fi·iends, colleagues,
point)-a result that is widely interpreted as evi - parents, or siblings who live in different house-
dence for remarkable adaptation to extreme mis- holds (Weiss 1988; Stroebe, Stroebe, and Hansson
fortune." Other researchers have come to similar 1993). Kaprio, Koskenvuo, and Rita (1987) noted
conclusions. Schulz and Decker (1985) interviewed that in tlle week following the death of a spouse,
one hundred middle-aged and elderly paraplegics suicide rates are elevated almost tenfold for wo-
and quadriplegics and found that reported well- men, and almost seventyfold for men. Wortman
being levels were only slightly lower than popula- and her colleagues (1992) found that it took al-
tion means of non disabled people of similar age .26 most one decade before widows and widowers ap-
Wortman and Silver (1987) found that quadri- proached a control group's scores on life satisfac-
plegics reported no greater frequency of negative tion, and nearly two decades before differences in
affect than control respondents . In a review of sev- depression were no longer significant.
eral empirical studies, Tyc (1992) found "no dif- While intense and prolonged grief is common,
ference in quality of life or psychiatric symptom- it is not universal. Wortman and Silver (1989 ,
atology" in young patients who had lost limbs to 1990) review evidence suggesting that a substan -
cancer compared with those who had not. 27 In a tial minority of individuals do not experience ex-
review of studies examining adaptation to burn in- treme grief at all. For example, Wortman and Sil-
juries, Patterson and his colleagues (1993) re- ver (1987) found that about 30 percent of parents
ported similarly high levels of psychosocial adapta- who had lost a child to sudden infant death syn-
tion by one year after the accident. drome (SIDS) showed no significant depression at
There is less evidence of adaptation to chronic any time following the infant's death. In the same
or progressive diseases such as chronic rheumatoid article, these authors contest the widely held belief
arthritis (Smitll and Wallston 1992), multiple scle- that the absence of grief shott1y after a death is a
rosis (Antonak an~d Livneh 1995), and other de - sign of repression or denial. They found, instead,
generative disorders (Livneh and Antonak 1994). that tlle absence of grief shortly after the loss is a
Hedonic Adaptation 313

posltlve indicator of long-term well-being, and and Diener 1995), wealth per se does not appear
that "delayed grieP' is, in fact, f.:lirly uncommon. to be driving the correlation. For example, Diener,
Similarly, there is little evidence supporting the Diener, and Diener (1995) conducted a massive
widespread belief that recovery fr0111 (adaptation cross-sectional study of reported well-being in one
to ) bereavement proceeds in "stages." Consider- hundred thousand people from fifty-five different
able variability is observed with respect to the countries whose per capita income ranged from
types of emotions that are experienced, their se- $120 (Tanzania) to $32,790 (Switzerland ). Al -
quence, and their interrsity at various times (Sil- though measures of per capita wealth and average
ver and Wortman 1980; Wortman et al. 1993). reported well-being were highly correlated (0.58),
While there are clearly large individual differences controlling for human rights eliminated the cor-
in styles of coping with bereavement, some of the relation. Fifth, Clark (1996) presents evidence
variability observed among people may be due to from British data that job satisfaction is strongly
the situational aspects of the loss. For example, related to changes in a worker's pay, but not to
gri ef seems to be particularly long-lasting if the levels of pay.
loss is unexpected (Lehman et al. 1987; \iVortman Scitovsky (1976) has even suggested that wealth
and Silver 1987). We return to this issue when may undermine happiness. He argues that pleasure
\\'e discuss forewarning as a moderator of adapta- results from incomplete and intermittent satisfac-
tion. tion of desires, and that the continuous comfOlts
made possible by substantial wealth may remove
the conditions necessaty for this experience. While
Desirable Experiences
consistent with the popular conception of the
JLI st as adaptive processes reduce the impact of "poor little rich girl," this speculation is inconsi-
negative changes, they may also diminish the plea- stent with the positive, albeit small, correlation be -
sure of positive events or improving circumstances. tween wealth and happiness that is observed in
In this section, we review evidence of adaptation cross-sectional empirical studies.
(or lack of adaptation) to two favorable changes,
increases in wealth and improved appearance follow- Cosmetic Surgn)' The small number of longi-
l-j ,
ing cosmetic surgery, and to the repetition of he- tudinal studies of people who ha\'e recei\'ed cos-
donically positive stimuli-erotic images and foods. metic surgety have generally not observed adapta- ~:
tion to increased attractiveness. The vast majority .~ .

Increases in Income Few studies have examined of patients repOlt satisfaction with the results of
the reb.tion between wealth levels and happiness cosmetic operations (Wengle 1986 ),29 which does
\\'ithin individuals across time. However, hedonic not appear to decrease over time . In a study of
adaptation is suggested by the following findings. nine hundred cosmetic and plastic surgery pa-
First, Brickman, Coates, and Janoff-Bulman (1978) tients, Reich (1982) found that satisfaction in-
studied twenty-two state lottery winners who had creased from 70 percent at three months after the
\mn between $50,000 and $1,000,000 within the operation to 85 percent four years later. Young,
previous year. These lottery winners reported only Nemecek, and Nemecek (1994 ) found that re-
slightly higher levels of life satisfaction than a con- ported satisfaction remained constant. Beale and
trol group (4.0 versus 3.8 on as-point scale).28 his colleagues (1985) did, hO\\'ever, observe a
Second, at least within the United States, there is slight decrease .3o
only a small positive correlation between wealth In addition to satisfaction with the surgety it-
and reported happiness. Diener and his colleagues self, several studies have indicated an o\'erall im-
(1993) estimated the correlation at 0.12. Third, provement in psychological health or life satisfac-
se\'eral studies have observed no change in well- tion as a whole. Ohlsen, Ponten, and Hamburt
being in a countly as its real income increases over (1978) noted that twenty-tlve of seventy-one
time (Campbell 1981; Duncan 1975; Easterlin women in their study were receiving psychiatric
1974, 1995). For example, between 1958 and treatment prior to a breast augmentation pro -
1987, real per capita income in Japan rose fivefold, cedure, whereas only three continued to do so
\\'hile subjective judgments of happiness did not after the operation. Klassen and his colleagues
increase (Easterlin 1995). Fourth, though moder- (1996) also found substantial reductions in psy-
ately strong correlations between wealth and well- chiatric symptomatology among people receiving
being have been found between countries (Vee- plastic surgety. Cole and his colleagues (1994)
nhoven 1991; Diener ~t al. 1993; Diener, Diener, reported that 73 percent of their patients re-
314 Well-Being
ported a higher quality of life after cosmetic sur- JUIces (Pliner 1982), novel fruits and cheeses b
gery, compared to only 6 percent who reported a (Birch and Marlin 1982), low-salt foods (Beau- G
lower quality of life. The largest gains were for champ, Bertino, and Engelman 1983), and plain
cosmetic breast surgery (both reductions and en- yogutt (Kahneman and Snell 1990).
largements), with slightly smaller gains for ab- The hedonic effects of repeated consumption f
dominoplasty (tummy tucks) and only slight may, of course, depend on the interval between
gains for rhinoplasty (nose jobs). All of these consumption events-one may like kiwi fruit more
studies should be treated with caution, however, and more every year, but not enjoy the fifth kiwi
because of the likelihood of strong demand ef- fruit consumed at dinner as much as the first .
fects in reporting satisfaction with cosmetic sur- There has been little research on the duration of
gery. 3l "sensory-specific satiety" (Rolls et al. 1981) for
different types of foods. Thus, conclusions about
Sexually Al'ousing Stimuli The evidence regard- the temporal dynamics of affect may depend on
ing adaptation to sexually arousing stimuli is mixed. tl1e length of the intelval between successive con-
O'Donohue and Geer (1985) and Koukounas and sumptions. For example, Kahneman and Snell
Over (1993 ) found that male sexual arousal (as (1990) found that people who consumed a small
measured by penile tumescence and subjective re- serving of their chosen flavor of ice cream over
POlt) diminished with repeated presentation of eight consecuti\'e days liked it less over time. Per-
erotic slides. Meuwissen and Over (1990) found haps an opposite effect would be found if the selV-
similar results for female sexual arousal to fantasy ings were consumed over eight consecutive weeks. 33
and sexual films. However, Smith and Over (1987)
found no decline in physiological or subjective
arousal for male subjects who were instructed to Explaining Domain-Specific Differences in
imagine the same sexual fantasy in eighteen re- Hedonic Adaptation
peated trials, and Laan and Everaerd (1995) found What distinguishes the things people can and can-
no evidence of reduction in genital arousal for fe- not adapt to hedonically? Why can people adapt to
males repeatedly presented with erotic slides-in prison but not to noise? Why are the pleasures of
fact, arousal increased somewhat over the ten trials increased income fleeting, while the pleasures of
despite returning to baseline levels during inter- foods and erotic stimuli are sustained despite re-
stimulus intervals.32 peated exposure? Part of the answer may lie in evo-
lutionalY pressures. Vve should not expect much
Foods Most adults enjoy every day at least one adaptation to stimuli that are necessary for contin-
substance they once found aversive, such as coffee, ued sUl\!ival and reproduction (for example, food
beer, tobacco, chili peppers, and other strong and sex), nor to stimuli that strongly compromise
spices (Rozin and Schiller 1980). Thus, for some sUl\!ival (for example, harmful chemicals, very hot
foods, repeated exposure not only reduces initial or cold temperatures) .3.
subjective unpleasantness (adaptation) but later Beyond this broad generalization, we may look
continues to increase subjective pleasantness (sen- for specific features of situations and stimuli that
sitization) . Torrence (1958) found that previous can help to explain differences in the degree of he-
consumption of pemmican (dried meat) by mili- donic adaptation across different domains. For ex-
tary aircrewmen led to increased liking when it was ample, the failure to adapt to highway noise may
later eaten during a survival exercise, even among reflect its high temporal variability. Perhaps people
the minority who initially found the taste pleasant. could adapt to a constant loud drone, but not to
Stevenson and Yeomans (1995) found that the he- occasional random bursts of noise from motorcycles
donic quality of the burning sensation of capsaicin with sawed-off muffiers. In the following section,
(the hot compound in chili peppers) was enhanced we discuss some of the moderators of adaptation.
by repeated exposure. Crandall (1984) gave the
employees of a remote Alaskan fish cannery access
to cake doughnuts not normally available in their
MODERATORS OF ADAPTATION
diet on thirteen occasions during a twenty-nine-
day period. He found a significant rise in con-
Social Support
sumption of doughnuts over time, which suggests
increased liking. ~Repeated exposure has also been Social contact with other people who have had
found to increase liking for unfamiliar tropical fruit similar experiences is generally found to facilitate
Hedonic Adaptation 315

hedonic adaptation. In their study of bereavement caregivers reported that they had already experi-
over the death of a spouse or child, Wortman and enced the loss of the person for whom they were
Lehman (1985) found that contact with a similar caring, and the greater the sense of loss during
other was frequently mentioned as the most help- caregiving, the lower the depression during be-
ful factor in assuaging grief, and it was never men- reavement" (681 ). Parkes and Weiss (1983) found
tioned as unhelpful. Mullan (1992) found similar that those who had "short forewarning" (less than
results. Regarding. adaptation to limb loss, Parkes two weeks' notice) that their spouse was going to
(1972) argues: "It is only when [handicapped per- die seemed less able to put the death behind them
sons] meets someone more seriously mutilated than those who had had a longer warning period,
than [themselves] who appears to be coping well and a large body of evidence revie\\-ed by Lehman,
and cheerfully with his disability that it becomes Wortman, and Williams (1987) is also consistent
possible for [them] to look to the future in an op - with the conclusion that preadaptation occurs, or
timistic and realistic manner."35 that the anticipation of loss accelerates the adapta-
People who have not had similar experiences, tion process .36
however, often give inadequate or inappropriate However, the benefits of advance notice of a
support. For example, Walker, MacBride, and Vac- negative outcome must be balanced against tl1e
hon (1977) concluded tlut even intimate friends aversiveness of waiting for that outcome to take
of widows did not support their need to mourn for place. Eyen if ad\'ance notice does improve post-
more than a few days after the death, and Shuch- outcome well-being, its o]Jcrall efrect on well-being
ter and Zisook (1993) cite the intolerance offam- is ambiguous since receipt of the bad news may
ily members to the bereaved one's continuing diminish the well -being of the person between the
grief as a major impediment to successful coping time the notice is received and the time the event
witl1 bereavement. Maddison and Walker (1967) actually occurs. Thus, the improyed post-outcome
reported that bel-eaved widows were often encour- adaptation doesn't necessarily compensate for tl1C
aged by friends to begin thinking about remar- earlier onset of misery associated with advance
riage within a few wee/,s of the husband's death. warning. In some cases, however, the net efrect
Similarly, Hughes, Good, and Candell (1993) seems clear. For example, in the study by Parkes
found that the psychological adjustment of recent and Weiss (1983 ), the tremendous reduction in
divorcees was impeded when friends continually long-term sufrering associated with advance notice
offered advice. Afflictions and disabilities with noof widowhood would seem to more than compen-
obvious manifestations (such as chronic fatigue sate for the reduction in affect from receiving the
syndrome) pose special problems as well, because bad news somewhat sooner.
sufferers are likely to face performance expecta- When forewarning does improye post-loss well-
tions that exceed their abilities (see Tyc 1992)_ being, how does it do it? Does it produce a type of
preadaptntion, or accelerate tl1e process of post-
Advance Notice loss adaptation? In principle, these processes could
be discriminated by a longitudinal study of people
Studies on forewarning of and anUCIpatory grief who sufrered a loss eitl1er witl1 or without pre-
over a loved one's death generally indicate tl1at it warning. 37 If forewarning advances the time course
improves affect and functioning after bereavement of adaptation (tl1at is, if tl1ere is preadaptation), as
(mough for a more equivocal interpretation of the illustrated by figure 16.4, tl1e immediate post-loss
evidence, see Stroebe and Stroebe 1987; and Hill, state would be more favorable among those who
Thompson, and Gallagher 1988). In a review of had forewarning. Ifforewarning accelerates the pro -
me literature on tl1e effects of forewarning on cop- cess of adaptation, however, as depicted in figure 16.5,
ing with the death ofa husband, O'Bryant (1991) then hedonic ratings would be similar immediately
concluded: "No investigation of conjugal bereave- after tl1e event but would improye more quickly.
ment . . . has found that adjustment to widow- Curiously, advance warning seems to predict
hood was poorer when the spouse had been fore- worse outcomes in other domains. Lazarus (1968)
warned" (229). In a study of the bereavement of reviews evidence suggesting tl1at subjects who
people who had cared for spouses or parents with were given a longer adyance warning that they
progressive dementia, Mullan (1992) states: "The would watch a gruesome film clip (about genital
progressive decline of the impaired family member subincision or wood -shop accidents) displayed
.. . affords ample opportunity to prepare psycho- more stress during tl1e film (as measured by skin
logically for the death"; moreover, "many of tl1ese conductance) than subjects who received no such
316 Well-Being
FIGURE 16.4 Hedonic Impact of Forewaming If It Advances the Start of
Adaptation r.
f
hedonic t;
response li
forewarning event s
occurs occurs y

l-
f
forewarni~g__ - - - - s

no forewarning a
(

advance warning. He concludes: "Very clearly, one is shown a \\'arning light that flashes ten sec-
when the anticipatOty times involved are relatively onds before the shock, the rats receiving the warn- (

brief, a longer period of anticipation results in a more ing exhibit lower stress and lower incidence of
marked stress reaction" (236). Breznitz (1967) stress-related disease than the rats who receive no F
made a similar observation, which he called the such warning. However, flashing the warning sig- (

"incubation of threat." Subjects were told they nal a half-hour before the shock produces the op- s
would receive a severe electric shock after either posite pattern-the warned rat displays greater
three, six, or twelve minutes. Those facing a lon- amounts of stress and stress-related disease than
ger time interval had faster heart rates immediately the non-\\'arned rats. These results make intuitive
preceding the shock. sense if one assumes that forewarning decreases
Sapolsky (this volume ) reports evidence from the aversiyeness of the event itself but increases
animal research suggesting that the effect of ad- the stressfulness of the \\'aiting period. Thus, the
vance warning may depend on the interval prior to hedonically optimal forewarning period may be
the aversive event. If 1:\\10 groups of rats are ex- one that is long enough to allow for the mobiliza-
posed to a random pattern of electric shock, but tion of defenses, but no longer.

FIGURE 16.5 Hedonic Impact of Forewarning If It Accelerates the Rate of


Adaptation

hedonic
response
forewarning event
occurs occurs

forewarning - - - -- 4 -...
~-~- ------ - -:...-

no forewarning
Hedonic Adaptation 317

1989; Wiggins et al. 1992). Those receiving ad-


Uncertainty verse test results e\'entually had better hedonic
Although folk wisdom often stresses the impor- outcomes than those who were not tested or those
tance of maintaining hope in the face of some who received an uninformative test result. Sap-
likely negative outcome, many researchers have olsky (this volume) reports that "During the blitz
suggested that successful adaptation to an ad- of London, there was more of an increase in ulcer
verse outcome is actually impeded by the possi- rates in the suburbs, which were bombed only in-
bility of relief. Herrmann and Wortman (1985 ), termittently, than in the city center, which was
for example, draw a sharp distinction between bom bed like clockwork each night." Schelling
stressors that "confront victims with a life situa- (1984) reports that addicts suffer less withdrawal
tion that is fundamentally altered" (168), such as pain when there is absol utely no possibility of ob-
the loss of their spouse or permanent paralysis, taining the drug.
and more ambiguous situations, such as that of a Adaptation to imprisonment provides further
couple who continue to fail to conceive a child anecdotal evidence for the conjecture that uncer-
for some unknown reason and who must contin- tain~' impedes hedofuc adaptation. Sapsford (1978)
ually decide between abandoning their hope to claims that uncertain~' about the length of their
conceive a child and intensifying their efforts at sentence is a source of misery for prisoners \\'ith a
conception-perhaps by seeking out new medical life sentence but also a chance of parole. Scitm'sky
specialists or procedures. The authors suggest (1976) cites Farber (1944) to the effect that "pris-
that the latter type of situation poses special ditli- oners who had hopes of parole suftered much
cui ties for adaptation. more than those \\'ho knew they would never be
Successful hedonic adaptation may require a released" (57 ). Roberts and Jackson (1991 ) cite
person to "take a hit"-to recognize, admit, and the case of a man released from prison after selV-
confront some loss. Even a small possibility that ing thirty-seven years under an indeterminate sen-
such actions will later prove to be premature or tence on the grounds that his continued imprison-
unnecessalY may prevent people from undertaking ment under these circumstances constituted cruel
them. Consistent with this argument, it is com- and unusual punishment. The stress of uncertai nty ~.
monly reported that concrete proof of a missing may be particularly acute for inmates on death row
person's death (such as physical remains) has a who cling to the small possibili~' of repeal. Thus,
large, long-term, positive impact on the coping of successful adaptation to death row may, paradox-
family members, and that, likewise, adaptation is ically, involye giving up hope (Gallemore and Pan-
poor for relatives of people who die without leav- ton 1972, 170).39
ing concrete evidence of their demise-such as
soldiers missing in action, the people who "disap-
Remindel's, Distractions, and
peared" from their homes during the dictatorships
Intrusive 17Joughts
in Argentina and Chile, or the victims of "ethnic
cleansing" in Bosnia. 3s Shuchter and Zisook (1993) comment that the
The notion that someone who is 100 percent bereaved find themselves trapped in a world of
sure that something bad has happened is better off continuous threats because of the triggers that cue
than someone who is 95 percent sure may seem painful thoughts and memories. For example,
counterintuitive. However, a small remaining hope among parents who have recently lost an infant,
may impede the onset of adaptive processes that the mere sight of other babies often stimulates in-
could eventually return one to normal hedonic tense grieving (Cornwell, Nurcombe, and Stevens
levels. Indeed, there is substantial evidence sug- 1977). The powerful influences of reminders raises
gesting such a pattern. Moulton and his colleagues t\vo issues for hedonic adaptation. First, does suc-
(1991) examined the impact of HIV testing in a cessful adaptation involve avoiding painful cues, or
sample of homosexual men and found that those confronting them, or some of each (for example,
testing positive showed less hopelessness and dis- by deliberately confronting grief-inducing cues in
tress than did a cohort of non-notified controls a supportive setting but otherwise avoiding them)?
(who agreed to be tested but did not wish to be Second, if cues cannot easily be avoided (as in tl1e
notified). A similar pattern has been observed in case of marital sex triggering memories of prior
studies examining the impact of notification of incest or rape), how can the emotional conse-
Huntington 's dise~se test results (Brandt et a1. quences be minimized?
318 Well-Being
Attributions ver lining"). In a study of seventy-eight women
with breast cancer, Taylor (1983) found that over (
Attribution theory predicts that (1) people will at-
half reported that the cancer experience had caused
tempt to assign responsibility for adverse events,
them to reappraise their lives in a favorable way.
and (2) those who hold themselves responsible
Similarly, ten of the twenty-nine accident victims
will feel worse than those who do not (see, for
in Janoff-Bulman and Wortman's (1977) study be-
example, Weiner 1982). The evidence we re-
lieved that God had intentionally (and benevo-
viewed on both of these points, however, was
lently!) selected them for victimization. Ho.wever,
mixed. Taylor (1983) found that 95 percent of a
in a study of people paralyzed in auto aCCIdents,
sample of women diagnosed with breast cancer
Lehman, Wortman , and Williams (1987) report
did tty to make some attribution for its. occur-
that three-fourths of their subjects were unable to
rence, such as stress, taking birth control pills, liv-
find any meaning in their loss.
ing near a chemical dump, hereditary factors, or diet.
From his experience in Auschwitz, Frankl (1963)
However, Downey, Silver, and \Vortman (1990 )
concluded that the ability to find meaning and
found that nearly half of the mothers in their
purpose in their suffering was essential for the sur-
study who lost their babies to SID? were. Ul:con-
vival of concentration camp prisoners. Less anec-
cerned about assigning responsibility. Simliarly,
dotally, Taylor (1983) found that women who
studies have found that less than half of all victims
found positive meaning in their breast cancer ex-
of accidents that led to severe burns (Kiecolt-
hibited "significantly better psychological adjust-
Glaser and Williams 1987) or physical disabilities
ment" (1163 ). Afr1eck and his colleagues (1987)
(Silver 1982) said they ever asked themselves ,
found that men who perceived benefits from a
"Why me?"
heart attack were less likely to have a subsequent
Regarding the second point, Wortman (1976 ),
attack and exhibited lower morbidity eight years
Janoff-Bulman (1979), and Tennen and Affieck
later. Of course, the causal direction between
(1990) have all argued that external attributions
meaning-making and adaptation may run in either
do not always facilitate coping. Janoff-Bulman and
or both directions. It is also possible that adapta-
Wortman (1977) found that victims of severe acci-
tion and meaning-making are spuriously corre-
dents who blamed themselves for the accident
lated through their common relation to a third
were coping more successfully eight to twelve
factor, such as innate happiness, which causes people
months afterward than those who did not, and
to both adjust successfully to tl1eir new condition and
that victims who blamed other people (as opposed
find meaning in it.
to some nonspecific external cause) displayed es-
pecially low coping scores. The authors postulate
that self-blame may help to maintain a sense of
personal control. However, Sapolsky (:I~s volun:e ) CounterJactuals
comments on the negative effects of ngldly ma1I1-
Counterfactuals- mental representations of things
taining an internal locus of control: "It is ben~fi­
that were not actually experienced-can some-
cial for individuals to interpret themselves as be111g
times influence the evaluation of (or adaptation
in control only if they live in relatively benign en-
to) a situation in the same way that actual experi-
vironments .. . . Such an attitude may be consid-
ences do. Thus, salient positive counterfactuals
ered pathological when it occurs in poor pe~ple
("Things could have turned out so much better")
with limited opportunities . . . . In such a setung,
should increase one's reference point and reduce
an internal locus of control, where one habitually
well-being, and salient negative counterfactuals
decides that those insurmountable odds could
("There but for the grace of God go I") should
have been surmounted if only they had worked
harder, is highly maladaptive. "40 reduce one's reference point and improve well-being.
Deliberate attempts to enhance affect by generat-
ing negative counterfactuals are not typically very
effective, however. We cannot usually make our-
Perception oj "Silver Linings)) selves happy by simply imagining more unfortu-
nate individuals or more unpleasant situations. To
The desire to explain an event is closely related to influence our hedonic state, counterfactuals must
the desire to find "meaning" in an experience or be plallsible, not just possible, alternatives to reality
to identif)r some positive consequence of it (a "sil- (see Kahneman and Miller 1986):1
Hedonic Adaptation 319

addictions, which are due in part to the decreas-


CONCLUSIONS ing pleasure from a given level of a good or activ-
ity and in part to the displeasure (craving) when
Limits of Adaptation consumption of the good or activity ceases (see,
for example, Koob et al. 1989; Loewenstein
The literature on adaptation has led some to the 1996).
view that long-run happiness and life satisfaction Second, the low sensitivity to difFerences in
do not depend much on one's objective circum- stimulus levels far from the adaptation level and
stances (Lykken and Telligen 1996 ). However, tlle high sensitivity to differences near the adapta-
anyone who has experienced extended periods of tion level (see figures 16.1 and 16.2) may degrade
happiness in certain environments and unhappi- the quality of decision-making, considering
ness in others should be skeptical of this extreme the experienced value of tlle relevant outcomes. For
conclusion. There are other reasons for skepticism example, fi:om tlle vantage point of a $10,000-
as well. First, as our review suggests, adaptation is a-year graduate fellowship, tlle difference between
not observed in all domains, and there may even a $60,000 or $65,000 starting salalY may seem in-
be some domains where sensitization is the rule. consequential and not even worth asking about,
Both of us can think of ongoing features of our much less bargaining over. Yet just a few years
environments, such as a good or bad roommate, later, when the student is a faculty member earn-
that were increasing sources of either pleasure or ing $60,000, she may be willing to pack up, move,
pain. Second, as argued earlier, some of the results and sever friendships to earn $5,000 more some-
indicating dramatic adaptation may be partly due where else.
to scale norming. Third, satisfaction with a partic-
Third, adaptation may erode our aesthetic stan-
ular outcome or outcome level is often influenced
dards . Dubos (1965) comments on adaptation to
by social norms, social comparisons, and other sa-
urban life:
lient referents that do not necessarily converge to
our average experience. For example, if a person This very adaptability enables [us] to become ad·
aspires to earn as much as his neighbor, the differ- justed to conditions and habits which will eventu·
ence between his income and his aspiration level ally destroy the values most characteristic of hu·
income may increase over time. Conversely, a man life. Millions . . . are so well adjusted to the
prosperous individual who grew up in poverty may urban and industrial environment that they no
remain perpetually delighted with her luxurious longer mind the stench of automobile exhaust, or ~,

lifestyle if her memory of her early days remains the ugliness generated by the urban sprawl; they
salient throughout her life. Fourth, and most obvi- regard it as normal to be trapped in automobile
ously, some things are just inherently good or traffic; to spend much of a sunny afternoon on con-
bad-they occupy only one side of a natural and crete highways among the dreariness of anonymous
immutable "zero point" (see Kahneman, this vol- and amorphous sn'ean1S of motor cars. (278-79)
ume).
Fourth, adaptation may work against moral
values. In his book Nazi Doctors, Robert J. Lifton
(1990) describes a process whereby German doc-
Costs of Adaptation
tors (who had taken the Hippocratic oath to do
Although hedonic adaptation confers enormous no harm) were gradually transformed into active
benefits by reducing the subjecti\'e effects of ad- killers. Doctors were first asked to be present
verse conditions, it has associated costs as well. when euthanasia took place, then to add their sig-
The most obvious cost of hedonic adaptation is nature to a document, then to supervise a mercy
that it occurs for goods as well as bads, creating killing, and so on, until they actually administered
what Brickman and Campbell (1971) have called the lethal injections to "eugenically undesirable"
the "hedonic treadmill"-the tendency for tran- persons. The famous Milgram experiment pro-
sitory satisfactions to eventually give way to indif- vides another example. Subjects were not asked to
ference or even dissatisfaction. Scitovsky (1976) administer a potentially lethal shock immediately
comments that "the attainment of a goal seems, but were given a series of requests to increase the
when the moment of triumph is over, almost like voltage slightly. Having just administered a one -
a let-down" (62). Adaptation to pleasurable ex- hundred-volt shock to someone, it may be difficult
periences may also be responsible for destructive to justifY stopping at that particular level rather
320 Well-Being
than complying with the experimenter's request to positive affect seems to be associated with activa-
increase the voltage to 105. tion of the left frontal hemisphere- see Ito and
Fifth, adaptation may reduce our empathy to- Cacioppo, this volume); and performance in af-
ward others in different situations. People indig- fect-mediated memory and decision-making tasks
nantly wonder: How can Donald Tmmp be so (see, for example, Arkes, Herren, and Isen 1988):'
rapacious when he already has so many millions? Assuming that future research provides a deeper
Why is that academic, already so renowned, greedy understanding of hedonic adaptation, is it likely
for still more recognition and so reluctant to ac- that such information would cause people to con-
knowledge the contributions of collaborators? How duct their lives differently? Would they stop wear-
could a few misplaced items provoke such an im- ing seatbelts with the assurance that tlley would
passioned response from our spouse when the get used to being paralyzed? Would they exploit an
apartment seems so immaculate by our former embezzlement opportunity knowing that prison
standards? All of these examples suggest the diffi- wouldn't be all that bad in the long mn? We sus-
culty of empathizing with the motivations and be- pect not. However, perhaps a few would be saved
haviors of others who have adapted to situations from the misel)' of renting a noisy apaltment they
different from our own. would never have gotten used to.
Sixth, just as adaptation causes us to fail to un-
derstand the behavior of others, it may cause us to V>le are grateful for comments and suggestions
mispredict our own emotions or behavior under from Robyn Dawes, Tonya Engstler-Schooler,
difterent future circumstances. For example, un- Bamch Fischhoft~ Donna Harsch, Daniel Kahne-
tenured academics overestimate how happy they man, Susan Nolan-Hoeksema, and Daniel Read,
will be if they get tenure or how unhappy they will and for financial support from the Center for Inte-
be if they fail to get tenure (Gilbert et al. 1997) . grated Study of tlle Human Dimensions of Global
People also typically underestimate how attached Change at Carnegie-Mellon University (NSF grant
they will become to objects that are not currently #SBR-9521914) .
in their possession (Loewenstein and Adler 1995).
Errors in predicting one's own feelings and behav-
ior are the focus of Loewenstein and Schkade's NOTES
chapter (this volume ).
1. Calling any behavior that reduces the effect of a stimulus
Final Comments an "adaptation" may be using the term too loosely. Sup-
pose that a person who moves to a hot climate repons
...
Existing research permits few general conclusions that she has gotten used to the heat. Does that repon
ii
constitute evidence of adaptation if we know that she
about hedonic adaptation . We have made an at- now wears lighter clothes, has purchased an air condi-
tempt to summarize results that recur in the small tioner, and ventures outdoors less during the middle of
subset of rigorous studies and to at least broach the day? Is it correct to say that she has "adapted" to the
some of the issues that remain unresolved. Clearly, heat or merely that she limited her exposure to it? Sim-
more research is required in specific domains and ilarly, suppose a bereaved man destroys all pictures of a
deceased loved one and moves to a difFerent city to
on specific moderators of hedonic adaptation. avoid contact with friends who might trigger disturbing
However, perhaps the single most pressing need is memories. If those steps do, in fact, reduce emotional
for richer and more diverse measures of hedonic pain, would it be con'ect to say that he has successfully
states. Along Witll subjecti\'e self-reports of sensa- "adapted" to his bereavement?
2. Adaptation is the primal)' biological property differen-
tion or affect, researchers should make greater use tiating physics and prychophysics (Shepard 1981). Com-
of alternati\'e hedonic indicators, including: global pare how a thermometer and a human hand register
measures of functioning, such as employment sta- temperature. Mercul)' has an invariant and unique re -
tus, frequency of hospital visits, and quantity of sponse-its volume is always the same at fifty-five de-
psychiatric help sought; behavioral measures, such grees, and always difTerent at other temperatures. It does
not matter whether the mercul)' was previously stored in
as tearfulness or facial expressions (see, for exam- an oven or an ice bath, or whether it was stored in either
ple, Husted Medvec, Madey, and Gilovich 1995), place for hours or days or years. Mercury has no memory
gross physiological correlates such as blood pres- for previous states. Humans and other organisms do not
sure; specific neurological correlates such as the behave this way. Our subjecth'e experience of, and re-
sponse to, a stimulus depends on more than its current
relative activation between brain regions (for ex-
physical intensity; it also depends on the strength, dura-
ample, negative aftect seems to be associated with tion, and recency of previously experienced stimuli. Our
activation of tlle right tl'ontal hemisphere, whereas experience of fifty-five-degree water is neither i111'ariant
Hedonic Adaptation 321

(fifty-fi"e feels wamlcr to a hand prcviously immersed in suggests that repeated exposure l'educes the affective in-
ice water than to one previously immcrscd in hot water) tensity of stimuli, and the "mere exposure" efFect, wh ich
nor collstlmt (fifty-five fecls different upon initial immer- shows the opposite . He speculated that repetition of
sion than it does late r) nor IIllique (fifty-five might feel "simple" stim uli decreases liking and repetition of "com-
the same to a hand previously immersed in icc water, as plex" stimuli increases liking. Rozin and Vollmecke (1986)
eighty does to a hand previously immersed in hot water). made a similar observation: "On the basis oftlle data on
3. Warr, Jackson, and Banks (1988) makc a distinction be- food and other research on exposure effects ... it seems
tween resigned adaptation and cOlZstmctive adaptation. that exposure is more likely to enhance liking when it
Resigned adaptation involves passivc acceptance of one's occurs at moderate frequcncy, and when the stimuli are
condition, or reducing one's aspirations. Constructive novel, relatively complex, or both. For many noyel
adaptation, on the othcr hand, involves active attempts items, there mav be enhanced liking at moderate levels
to mitigate one's loss or the adoption of different, but of exposure, followed by decreased liking when exposure
equally ambiti ous goals. becomes more frequent" (449). However, stimulus
4. Our eyes (just one of our five senses) can transmit be- "complexity" is difficult to measure, or even define, for
tween 1.6 and 3 million bits of information per second, many hedonically relevant stimuli like tastes or smells
thousands of times more information than our brains (Bornstein 1989).
can process (Scitovsky 1976, 52). Thus, some mecha- 8. In Hclson's model , AL is the arithmetic mean of the
nism must act as an informatio nal filter to select which psychological values or the geometl'ic mean of the objec-
perceptual information is processed . Adaptation serves tive stimulus le"el, because it was assumed that the psy-
this role by relegating constant stimu li to the perceptual chological magnitudes are a logarithmic fu nction of the
background and focusing attention on rapid changes in objecti,-e, physical stimulus magnitudes.
stimulus levels-perceptual signals most likely to require 9. Parducci (1995 ) does posit that stimuli experienced in
a behayioral response. the distant past drop out of the judgmental context,
5. For other examples of desensiti zation, imagine a com- though the model does not explicitly suggest the rate or
mercial fisherman who develops thi ck calluses on his degree to which they do.
hands. Such calluses are an adaptation because they re - 10. For the multiple -period case, the formula can be sim-
duce the abrasi,'c and laceratin g eflects of fishing nets plified to:
and fish spines. However, the calluses will probably also
cause a long-term reduction in overall tactile sensitivity,
such as the abilirv to discriminate between different ALt (16.5)
qualities of silk. S-imilarly, the destruction of cochlear
cilia caused by continued exposure to loud noises may
reduce the aversi,-eness of those loud noises but may also 11. Brown (1953 ) found tllat the efrects of previous stimuli
reduce sensiti"itv to changes in noise level at normal vol- on subjects' assessments of weights were influenced
umes (for example, when a baby's cIY upstairs is added much more when those stimuli were normal-looking
to the "olume of a television show downstairs). Cometto- laboratory brass weights than when they were metal
Muniz and Cain (1992) provide another example of de- trays of equal mass.
sensitization . Thev found that exposure to airborne irri- 12. It is not clear that eyen this modification would be ade-
tants like cigarette smoke increases mucus production, quate, because the form ul ation still implies that intro-
which physically impedes the transfer of molecules of air- ducing sufficiently excellent pizza into one's consump-
borne irritants fi-om the air to the free nelve endings. tion history could render the experience of average
This reduces the subjective intensity of airborne irritants, pizza unpleasant. Note that if f(X t - ALt ) is a mono-
but it also reduces perceptual sensitivity to (and there- tonically decreasing function of AL, and AL is a
fore the hedonic effect of) subtly pleasant smells. Linz weighted average of previous stimulus levels, X t - " . • .
and his colleagues (Linz, Donnerstein, and Penrod X t _ I, introducing a sufficiently extreme experience
1984, 1988; Linz, Donnerstein, and Adams 1989) have would raise AL enough to make the entire expression
argued that repeated exposure to graphic sexual vio- (c + f[X t - ALt)) negative.
lence, such as "slasher" films, causes a type of emotional 13. Van Praag and his colleagues have conducted some of
desensitization . They present evidence suggesti ng that tlle most careful studies focusing on the functional
viewers not only become less physiologically aroused and form of hedonic adaptation. In the late 1960s, van
emotionally responsive while watching such films but Praag developed a survey item tl1at asked respondents
less sympathetic to victims of violence in other, unre - to state the weekly, monthly, or yearly incomes tlley
lated contexts. associated Witll six different descriptors: "very bad,"
6. Despite reports of high-level pleasures from connois- "bad," "insufficient," ~'sufficient)" "good," and "very
seurs, there is little empirical evidence that total pleasure good." Van Praag defined the individual's adaptation
is increased. The refinement in the tastes of connoisseurs level income as the midpoint of the interval between
may increase their ability to discriminate and the plea- insufficient and sufficient. He and his colleagues have
sure they derive from small positive differences. How- conducted several studies to determine how the adap-
ever, such refinement may also increase the aversive ness tation level income changes over time as a function of
of negative difrerences. Moreover, a greater proportion past, current, and anticipated income. His estimates
of experiences may fall short of the connoisseur's ever- based on British data suggest tllat the current period's
increasing standards. It would be interesting to investi- reference point is an additive function of tllese tluee
gate tllis issue using objective measures of pleasure such variables, witll a weight of .16 for last year's income,
as facial encoding. .75 for this year's income, and .09 for next year's (an -
7. Berlyne (1%'0) oflered a possible explanation for the ap- ticipated) income. However, he notes, weight on the
parent conflict between the adaptation literature, which future is almost certainly underestimated owing to cr-
322 Well-Being
rors in variables, since he estimates the equation using • "Problems relating to deprivation of liberty . . . were
actual future income (identified by the researcher after ranked as the most severe problems among the
the fact) as a pro;..)' for expected future income. American long-term prisoners."
14. The second author has collected a large amount of data • "Parents generally suffer mood disturbances after the
on escape attempts in different states in an attempt to death of an infant. Depression and sadness is a very
test this anom aly. The data proved extremely difficult com mon reaction."
to analyze owing to a combination of reporting irregu- • "Major life changes make demands that are
larities, statistical complications, and other problems. larger and broader than those made by smaller
However, conversations with wardens revealed that ones."
they had not, in general, noticed any pattern of late- • "High self-esteem may contribute to the sense of
stage escape attempts . Several did mention, though, personal confidence."
that prisoners often attempt to escape afte r parole ap- • "Knowing that one experiences a positive event .
plications they had expected to be affirmed are re - yields the knowledge that something good has hap -
jected, a phenomenon that is consistent with a type of pened."
anticipatory adaptation. • " Enhanced motivation is associated with greater per-
15. Unless utility is a linear function of income, single and sistence at tasks."
multiple reference points yield different utilities. With a • "Almost all children of divorce regard their child-
single reference point formulation, 11 = f(540,000 - hood and adolescence as having taken place in the
[a $20,000 + ( I-a ) 550,000]), where a designates shadow of divorce."
the weighting of the two referents in determining the • "Each indi\'idual is constrained in his or her actions
average reference value. With a multiple reference point by capability constraints."
formulation, u = a f (54 0,000 - $20,000} + (l - • "When the processes of recovery are blocked, the re-
a ) f(540,000 - S50,000 }, where a designates the sult is a failure to recover."
fraction of time or attention devoted to each referent. • "Research on depression among people with Parkin-
16. This criticism is not applied to all of the phenomena son's disease ... suggest that between 4 percent and
Helson cites in support of his theory. Stevens does not, 90 percent of this population are clinically de-
for example, take issue with Helson's earlier work on pressed. "
changes in visual perceptions of color, which he admits
-' as "real" adaptation. In other cases, like the judged 24. The fac t that the initial shock of imprisonment gives
darkness of difrerent shades of gray, both real adapta- way to substantial adaptation is evident from the statistic
(

I, tion and semantic norn1ing may be occurring. that 50 percent of suicides in prisons occur within the
• 1,
17 . Of course, such social comparisons may have real he- first twenty-four hours of imprisonment (Hayes 1983 ).
donic consequences as well as influencing how people 25. The evidence of hedonic adaptation in the paper is not
," 1 interpret response scales. overwhelming- paraplegics rate their happiness as 2.96
1
It.
18. All but one of the sixteen subjects who had previously on a 5-point scale, compared to 3.82 for a control
:1:·r :·
sufrered severe injuries held their finger in the water for group. This seems to be a substantial difrerence, espe-
the entire experiment (sixty seconds), whereas none of cially considering the tendency for subjects to avoid ex-
.. the other twenty-four subjects did so. treme response categories (Poulton 1989).
.. i
19. A similar focusing efrect may occur when people pre- 26. However, the authors noted that one of the respon-
dict the hedonic impact of future events. Simply asking dents "frequently cried during the interview, and asked
people about how an event would afrect their well -being tl1e interviewer to get a gun and kill him."
seems temporarily to raise its prominence, leading to 27. Caine (1973 ) did not find similarly successful adapta-
an overprediction of its actual impact (Loewenstein and tion among adult amputees, however. Perhaps adults
Frederick 1997; Schkade and Kahneman 1998 ). suffer more "phantom limb" pain (Katz and Melzack
20. For exan1ple, Ross (1989) cited a study in which people 1990 ), which may be difficult to adapt to.
who completed a weight-loss program remembered 28. The authors suggested tlut the ecstasy accompanying
themselves as ha\ing been heavier at the outset of the the lottel)' win may have created a high reference point
program than they actually were, enabling them to against which the rest of the winner's life tl1en com-
maintain the (mistaken) belief that the program had pared unfavorably. They noted that the lottery winners
been effective. Likewise, women's memories of men- rated evel)'day activities (for example, eating breakfast,
strual pain bear little resemblance to their actual prior reading a magazine) as less pleasurable than did control
pain ratings. Instead, the pain ratings appear to be re- subjects. The potential costs of intense positive afrect
constructed ftom implicit (and inaccurate) theories are also discussed by Parducci (1984), Diener and his
about menstrual pain (McFarland, Ross, and DeCour- colleagues (1991), and Tversky and Griffin (1991).
ville 1989). 29. Over 95 percent of rhinoplasty (nose job) patients were
21. The authors did not, however, find differences in toler- satisfied with the results (Klabunde and Falces 1964).
ance with respect to other types of pain. The short-tern1 postoperative satisfaction rate of wo-
22. The set of experiential components we choose to call men receiving breast enlargements ranges from 72 per-
pain are themseh-es a subset of the aggregate experi- cent(Bealeetal.1985}t086percent(Youngetal.1994}
ence. People who are congenitally unable to feel pain to 89 percent (Ohlsen, Ponten, and Hambert 1978).
can nevertheless detect that they are being burned, 30. As noted by the authors , the reduced satisfaction may
pinched, or pricked- they simply do not experience have been due to objective changes in the shape or
those sensations as unpleasant (Cabanac 1971). softness of the breasts.
23. To illustrate some of the problems with the literature 31. Beale and his colleagues (1985) comment on tl1e
on adaptation, \VI: excerpt some lamentably characteris- strong demand effects to report satisfaction with the
tic passages below, without attribution: results of breast enlargements: "The women who have
Hedonic Adaptation 323

complied \\'ith an augmentation mammaplasty have cancer. If the group with no forewarning is adaptin g
suffered; they have gone through laborious interviews, more poorly, it could be because of the lack of fore-
waited for years, and finally they have taken the risk of warning o r because of tl1e particular characteristi cs of
being operated on. All this is done by free choice and deaths accompanied by little or no warning-for exam-
many times against the advice of relatives and doctors. ple, their violence o r their "preventability."
Against this background, it seems impossi ble to ad mit 38 . The difficulty of adapting to loss in the face of uncer-
to the interviewer, or even to oneself, that the opera- tainty is e\'ident in two passages from Sebastian Jun-
tion was a failure" .(484). ger's best-se ller T71C Peliect Stonn, which is about a
32 . Laan and Everaerd (1995) noted an interesting contra- storm in the Atlantic Ocean that that cost the lives of men
diction - that evidence of adaptation to sexual stimuli on the fishing boat Andrea Gail. As Sebastian recounts,
has been used to explain both inhibited and extl'eme
If the men on the Andrea Gail had simply died
sexualitv (inhibited because repeated behavior ceases to
and the bodies were lying in state somewhere,
be sufficiently arousing, and extnme because people
their loved ones could make their good byes and
engage in increasi ngly intense sex ual activities to main-
get o n with their lives. But th ey didn't die, they
tain prior b'els of gratification ).
disappeared oft' the face of the earth and , stri ctly
For both males and females, the literature consis-
speaking, it's just a matter of fa ith that these men
tently found that arousal diminished more slowly to
will never return. Such faith takes work, it takes
varied or novel stimu li than to constan t stimuli (the
efrort. T he People of Gloucester must willfully ex-
same slide repeatedl,' viewed ). In fact, novelty may not
tract tllese men from their lives and ban ish them
only diminish the effects of adaptation but reverse
to another world. (273 )
tl1em . Consider, for example , the "Coolid ge effect,"
which refers to the tendenc,' for new sexual partners to Writing abollt the \"ife of o ne of the missing men who
restore the sex drive of previously satiated male animals was lost at sea in the same storm, Sebastian relates "N-
(Bolles 19 75; De" 'sbury 1981 ). ter almost a montl1, Marianne Smitl1 is able to start ab-
33. Because rJp idly repeated consumption of a specific sorbin g the loss of her hu sband . As long as the planes
food temporarily reduces its subjective pleasantness are going out she holds on to some shred of hope, and
compared to other foods , access to a wide variety of that keeps her in a ghastly kind of limbo." (283)
foods seI\'es to maintain the pleasantness of continued 39. Unce rtai nty seems to impede adaptive behaviors in less
consumption . Indeed, increasi ng food variety has been weighty domains as well. For example , people seem far
shown to increase caloric intake in rats, cats, and hu - more reluctant to replace an item they have misplaced
mans (Rolls et al. 1981 ) and may partly explain the tl1an one they have broken, eve n tor items that cost
hi gher rates of obesity in cultures where a wide variety only a few dollars and which exhaustive searches have
of foods are readil" available (Rolls et al. 1981 ). failed to locate. It is very difficult to accept the fact that .~ .
34. Indeed, Cometto- ;\ \uni z and Cain (1992 ) found very tl1e ice cream scoop is "gone" if we do not know where
lirtle sensory adaptation to pungent, harm ful chemicals. it has gone to and this somehow seems to block or
35 . The recent rapid gro"Lb of specialized "support groups" greatly delay tl1e adaptive beha\~or of buying a new one.
may reflect a gro"'ing belief in the benefits of social 40. Weiss (1971 a, 1971b) found that laboratory rats who
contact " 'ith people who are experiencing similar prob- sit and endure electric shocks do not show as many
lems. H o\veve r, people's self-repo rts that they have stress symptoms as those who t1Y, unsuccessfully, to
been helped by a group are notoriously invalid mea- avoid the shocks-tlms obtai nin g feedback that tl1eir
sures of their actual hedonic change (see Ross 1989 ). coping responses have been ineffective.
For example, a recent study by Helgeson and her col- 41. On tl1is point, Parducci (1995, 179) comments: "Re-
leagues (1997 ) compared the effectiveness oftwo types minding well-fed children that tl1e starving waifs of N-
of support interyentions fo r women who had under- rica would be grateful for tl1e food they refuse to eat
gone breast surgery-an education program an d a sup- will not improve its taste. The well-fed child can hardly
port group- against a control condition in which believe that he or she could just as well be among the
women \\'ere simply referred to their ge neral practi- starvin g. But knowing full well how freezing it would
tioner. Although tl1ere was no difference in self-reported be outside, the same child might derive enhanced plea-
satisfaction with the two interventions, women who sure from snuggling beneath a down comforter o n a
participated in the support group actually came out cold winter night. There would have been ample expe-
worse on a variety of psychological and health measures rience with cold, perhaps as recently as when first slip-
than either the education or control groups at a six- ping down between the cold sheets."
month follow-up. 42. These ideas for alternative measures of happiness were
36. Preadaptation may occur for divorce as well as deatl1. stimulated by a discussion with Jo nathan Schooler.
Melichar and Chiriboga (1988 ) found that tl1e psycho-
logical adjustment among women who initiated di -
vorce was positively correlated with th e time since their
(private) decision to seek divorce, holding constant the REFERENCES
time since tl1e phYSical separation.
37. In practice, however, it is probably exceedingly difficult
Afrleck, G., Tennen, H. , Croog, S., and Levine, S.
to separate the effect of forewarning from the efrects of
other attrib utes that correlate with forewarning and (1987). Causal attribution, perceived benefits, and
may themselves intluence grief. For example, most of . morbidity after a heart attack: An eight-year study. Jo ur-
the deaths witl1 no forewarning are likely to occur from nal of Consulting and Clinical Psychology, 55, 29-35.
homicides and au to acciden ts, whereas deaths with Ahles, T. A., Blanchard, E. B., and Leventllal, H. (1983 ).
lon g fo rewarning are primarily from diseases such as Cognitive control of pain: Attention to the sensory as -
324 Well-Being
pects of the cold pressor stimulus. Cognitive T7?erapy Caine, D. (1973). Psychological considerations affecting b
and Reseal'ch, 7, 159-77. rehabi litation after amputation. Medical journal of 1-
Antonak, R. F., and Livneh, H . (1995) . Psychosocial ad- Australia, 2, 818-21. Die
aptation to disability and its investigation among per- Campbell, A. 1981. The sense of lVell-being in America: (
sons with multiple sclerosis . Social Science and Medi- Recent patten'ls and tl·cnds. New York: McGraw-Hill. 11

cine, 40, 1099 - 1108 . Cassens, G., Actor, c., Kling, M ., and Schildkraut, J. J.
Arkes, H . R., Herren, L. T., and Isen, A. M. (1988). (1981). Amphetamine withdrawal: Efiects on thresh - Di,
The role of potential loss in the influence of afrect on old of intracranial reinforcement. Ps),chopharmacology,
risk-taking behavior. Organizational Behavior and 73, 318 -22 .
Human Decision PI'ocesses, 47, 181-93. Clark, A. E. (1996). Are wages habit-forming? OECD
Beale, S., Hambelt, G., Lisper, H., Ohlsen, L., and working paper. DEELSA. Do
Palm, B. (1985) . Augmentation mammaplasty: The Cohen, S., Evans, G. W., Krantz, D. S., and Stokols, D.
surgical and psvchological effects of the operation and (1980). Physiological, motivational, and cogniti\"e ef-
prediction of the result. Annals of Plastic S1t1:gelY fects of aircraft noise on children: Moving ITom the lab-
14 (6 ), 473- 93. oratory to the field. Amel'ican Psychologist, 35, 231 - 43.
Beauchamp, G. K., Bertino, M., and Engelman, K. Cohen, S., Evans, G. W., Krantz, D. S., Stokols, D., and D\
(1983 ). Modification of salt taste. Annals of Internal Kelly, S. (1981). Aircraft noise and children: Longitudi-
Medicine, 98 (2 ), 763-69. nal and cross-sectional e\'idence on adaptation to noise D\
Beriyne, D. E. (1970). Novelty, complexity, and hedonic and the efiecti\"eness of noise abatement. j01l1"1lal of
value. PCI'ccption and Psychophysics, 8, 279- 86. Personality and Social Ps),chology, 40 (2), 331 - 45 .
Birch, L. L., and Marlin, D. W. (1982 ). "I don't like it; Cole, R. P., Shakespeare, V, Shakespeare, P., and Hobby,
I ne\'er tried it": Effects of exposure on two-year-old J. A. E. (1994). Measuring outcome in low-priority
children's food preferences . Appetite, 3, 353 - 60. plastic surgelY patients using quality-of-life indices .
Boles, T. L., and Messick, D . M. (1995). A reverse outcome BI'itisb jOl.lrnal of Plastic SUI;gC1Y, 47, 117-21 .
bias: The influence of multiple reference points on the Cometto-Muniz, J. E., and Cain, W. S. (1992). SensOlY
evaluation of outcomes and decisions. Ol;ganizational irritation: Relation to indoor air pollution. In \Y. G.
BehaJ7ior and H1Iman Decision Processes, 61, 262-75 . Tucker, B. P. Leaderer, L. Molhave, and W. S. Cain
Bolles, R. C. (1975). The them)' of motiJ1ation. 2nd ed . (Eds.), SOUI'CCS of indoor air contaminants: Chnrac -
New York: Harper and Row. tel'izing emissions and health impacts. Annals of the
Bornstein, R. (1989) . Exposure and afrect: Overview NeJV Yo,'k Acad(111)' of Sciences (vol. 641 ). New York:
and meta-analysis of research, 1968-1987. Psychologi- New York Academy of Sciences.
cal Bulletin, 106 (2), 265-89. Conway, M., and Ross, M. (1984). Getting what you
Bower, G. H. (1981). Emotional mood and memory. ",ant by revising what you had. journal of Pel'sonnlity
1 , Amel'ican Psychologist, 36, 129-48. and Social Ps),chology, 47, 738- 48 .
Brandt, J. , Quaid, K. A., Folstein, S. E., Garber, P., et al. Cornwell, J., Nurcombe, B., and Stevens, L. (1977).
( 1989). Presymptomatic diagnosis of delayed-onset Family response to loss of a child by sudden infant death
disease with linked DNA markers: The experience in syndrome. Medical joumal of Australia, 1, 656- 58. FI
Huntington's disease. journal of the American Medi- Crandall, C. (1984). The liking of foods as a result of
cal Association, 261, 3108-14. exposure: Eating doughnuts in Alaska. journal of So- F
Breznitz, S. (1967). Incubation of threat: Duration of cial Psycbology, 125 (2), 187- 94.
anticipation and false alarm as determinants of the fear Damasio, A. R. (1994). Descartes' en'01': Emotion, I'cason,
reaction to an unavoidable fi'ightening event. jounlal and the hU111a11 brain. New York. Putnam.
of ExpcI'imental Research in Pel'Sonality, 2, 173-79 . Dar, R., Ariely, D., and Frenk, H. (1995). The effect of past F
Brickman, P., and Campbell, D. (1971). Hedonic rela- injury on pain threshold and tolerance. Pain, 60, 189- 93.
tivism and planning the good society. In M. H. Ap- Deaton, J. E., Berg, S. W., Richlin, M., and Liuownick, (
pley (Ed.), Adaptation -ICl,el them)': A symposium (pp. A. J. (1977). Coping activities in solitary confinement
287-3 02) . New York. Academic Press. of U.S. Navy POWs in Vietnam. journal of Applied
Brickman, P., Coates, D. and Janoff-Bulman, R. (1978). Social Psychology, 7, 239- 56. (
Lottery winners and accident victims: Is happiness rel- Dewsbury, D. (198 1). Effects of noyelty on copulatory
ative? journal of Pel'sollnlity and Social Psychology, behavior: The Coolidge effect and related phenom-
36 (8), 917-27. ena. Psychological Bulletin, 89,464-82.
Brown, D. R. (1953). Stimulus similarity and the an- Diener, E., Colvin, R., Pavot, W., and Allman, A. (199 1).
choring of subjective scales. Anlerican jounlal of Psy- The psychic costs of intense positive affect. j0141'11al of
cholog)~ 66, 199-214. Pmonality and Social Ps)'cbolog)', 61 (3),492-503.
Bukstel, L. H., and Kilmann, P. R. (1980). Psychologi- Diener, E., Diener, M., and Diener, C. (1995). Factors
cal effects of imprisonment on confined individuals. predicting the subjective well-being of nations. jOllr-
Psychological BlIlletin, 88 (2),469-93. nal of Personality and Social Psycholog_\j 69 (5), 851 - 64.
Cabanac, M. (197't). Sensory pleasure. Q}tartn'ly ReJ1ielV Diener, E., and Fujita, F. (1996). Social comparisons
of Biology, 54 (1),1-29. and subjective well-being. In B. Buunk and R. Gib-
Hedonic Adaptation 325

bons (Eds.), Health, coping, (md socinl comparison. Mode ling loss aversion and reference dependence
Hillsdale, N.].: Erlbaum. effects on brand choice . Marketing Science, 12, 378-
Diener, E., Larsen, R. J., Levine, S., and Emmons, R. A. 94 .
(1985 ). Frequency and intensity: The underlying di- Hayes, L. M. (1983). "And darkness closed in": A na-
mensions of positive and negative affect. joumal of tional study of jail suicides. Criminal justice and Be-
Personality alld Social Psychology, 48, 1253-65. havi01, 10,461-84.
Diener, E., Sandyik, E., Seidlitz, L., and Diener, M. Heingartner, A., and Hall , J. V. (1974). AfFective conse-
(1993). The relationship ben"een income and subjec- quences in adults and children of repeated exposure to
tive well-being: Relative or absolute? Social Indicators auditory stimuli. joumal of Personality and Social Psy-
Reseal'ch, 28, 195-223. chology, 29 (6),719-23 .
Downey, G., and Sih-er, R. c., and 'vVortman, C. B. Helgeson, V. S. , Cohen, S., Schulz, R., and Yasko, J.
(1990 ). Reconsidering the attribution-adjustment re - (1997). Effects of education and peer discussion group
lation following a major negative event: Coping with interventions on six-month adjustment to stage I and
the loss of a child. j01ln1l11 of Penol1ality and Social II breast cancer. Working paper. Pittsburgh: Depart-
Psychology, 59, 925-40. ment of Psychology, Carnegie-Mellon University.
Dubos, R. (1965 ). Mall adnptillg. New Haven, Conn.: Helson, H. (1947) . Adaptation le"el as frame of refer-
Yale UniYersin' Press. ence for prediction of psychophysical data. American
Duncan, O. T. ( 1975 ). Does money buy satisfaction? So- j01tnzal of Psychology, 60 (1), 1- 29.
cial Indici1tors Re.rcfl/·ch, 2, 267-74. - -- . (1948) . Adaptation level as a basis for a quan -
Dyrcgrov, A. (1990 ). Parental reactions to the loss of an titative theOlY of frames of reference. Psychological Re-
infant child: A rcyiew. Scandinavian jOllmal of Psy- piew, 55 (6),297- 313.
chology, 31, 266-80. - - - . (1964). Adaptation-lepel theol)': An experimen-
Easterlin, R. A. (1974 ). Does economic growth improve tal and systematic approach to behapi01: New York:
the human lot' Some empirical evidence. In P. A. Harper and Row.
David and M. W. Reder (Eds .), Nations and house- Herrman, c., and Wortman, C. (1985) . Action control
holds ill economic growth (pp. 89-125) . New York: and the coping process. In J. Kuhl and J. Beckman
Academic Press. (Eds. ), Action control: From cogllition to bchavior (pp.
- - -. (1995 ). Will raising the incomes of all increase 151 -8 0 ). New York: Springer-Verlag.
the happiness of all' joumal of Economic Behapior and Hill, C. D., Thompson, L. W., and Gallagher, D. (1988).
Organization, 27, 35-47. The role of anticipator" bereayement in older women's
Farber, M. L. (1944). Suffering and the time perspective adjustment to widowhood. The Gerontologist, 28 (6), :5 .
of the prisoner. UlIipenit)' of Iowa Stltdies in Child 792-96.
Welja1'e, 20, 155-227. Hsee, C. K., and Abelson, R. P. (1991 ). The velocity
Fernandez , E., and Turk, D . C. (1992). Sensory and af- relation: Satisfaction as a function of the first deriva-
fective components of pain: Separation and synthesis. tive of outcome over time . jOllmal of Penonalit), and
Psychological Bulletin, 112 (2 ),205-7. Social Psychology, 60, 341-47.
Flanagan, T. ]. (1980 ). The pains oflong-teml imprison- Hsee, C. K., Abelson, R. P. , and Salovey, P. (1991 ). The
ment. British jOll1"1lnl of C,'iminology, 20, 148-56. relative weighting of position and velocity in satisfac-
Frank, R. H. (1992 ). Frames of reference and the inter- tion. Psychological Science, 2, 363 - 66.
temporal wage profile. In G. Loewenstein and J. Els- Hughes, R., Good, E. 5., and Candell, K. (1993). A
ter (Eds .), Choice OPel' Time (pp. 371-82). New York: longitudinal study of the effects of social support on \

Russell Sage Foundation . the psychological adjustment of divorced mothers. I


Frankl, V. E. (1963 ). Man's sear'ch for meaning. New journal of Divo rce and Reman'inge, 19 (1),37-56 .
York: \Vashington Square Press. Husted Medvec, v., Madey, S., and Gilovich, T. (1995).
Gallemore, ]. L., and Panton, J. H. (1972). Inmate re- 'vVhen less is more: Counterfactual thinking and satis-
sponses to lengthy death row confinement. Amel'ican faction among Olympic medalists. JOI/,.,ml of Person-
joumal of Ps),chiatl)', 129, 167-72. ality and Social Psychology, 69,603-10 .
Gilbert, D. T., Pinel, E. c., Wilson, T. D., and Blum- Janal, M . N ., Glusman, M ., Kuhl, J. P., and Clark, W. C.
berg, S. J. (1997). Affecti"e forecasting and durability (1994). Are runners stoical?: An examination of pain
bias: The problem of the invisible shield. Working pa- sensitivity in habitual runners and normal ly active con-
per. Canlbridge, Mass.: Department of Psychology, trols. Pain, 58, 109-16.
Harvard University. Janoff-B ulman, R. (1979). Characterological versus be-
Groves, P. M., and Thompson, R. F. (1973) . A dual - havioral self-blame: Inquiries into depression and
process theory of habituation: Neural mechanisms. In rape. joumal of Penollfllit)' alld Social Psychology, 37,
H. V. S. Peeke and M. J. Herz (Eds.), Habituation 1798- 1809.
(vol. 2, pp. 175-205 ). New York: Academic Press. Janoff-Bulman, R., and Wortman, C . (1977). Attribu-
Gurr, T. (1970). Why men rebel. Princeton, N.].: Prince- tions of blame and coping in the "real world": Severe
ton UniYersity Press. accident victims react to their lot. jOllnlnl of Person-
Hardie, B. G. 5., I.ohnson, E.]., and Fader, P. S. (1993). ality and Social Psychology, 35 (5) , 351-63.
326 Well-Being
Jonsson, E., and Sorensen, S. (1973). Adaptation to tual and empirical relations . Nebraska Symposium on
community noise: A case study. Joumal of Sound and Motivation, 16, 175-266.
Vibmtion, 26, 571-75. Lehman, D. R., Wortman, C. B., and Williams, A. F. M
Junger, S. (1997). The pelfect st01'm. New York: Harper- (1987). Long-term effects of losing a spouse or child
Collins. in a motor vehicle crash. joumal of Pel-sonality and
Kahneman, D. (1992). Reference points, anchors, nom1S, Social Psychology, 52 (1), 218-31. !v
and mixed feelings. Organizational BehaJlior and Hu- Leventhal, E. A., Leventhal, H., Shacham, S., and East-
man Decision Processes, 51, 296-312. erling, D. V. (1989). Active coping reduces reports of
Kahneman, D., and Miller, D. T. (1986). Norm theory: pain from childbirth. joumal of Consulting and Clini-
Comparing reality to its alternatives. Psychological Re- cal Psychology, 57, 365 - 71.
JlieH', 93 (2), 136-53. Leventhal, H., Brown, D., Shacham, S., and Enquist, G.
Kahneman, D., and Snell, J. (1990). Predicting utility. (1979 ). Effect of preparatory information about sensa-
In R. Hogarth (Ed.), Insights in decision lnaking: A tions, threat of pain, and attention on cold pressor
tl-ibute to Hillel J. Einhom (pp. 295-310). Chicago: distress. journal of Personality and Social Psychology,
University of Chicago Press. 37, 688-714.
Kahneman, D., and Tversky, A. (1979). Prospect theory: Lifton, R. J. (1990) . The Nazi doctors: Medical killings
An analysis of decision under risk. Econometrica, 47, and the psychology ofgenocide. New York: Basic Books.
263-91. Linz, D., Donnerstein, E., and Adams, S. M. (1989 ).
Kaprio, J" Koskenvuo, M., and Rita, H. (1987). Mortal- Physiological desensitization and judgments about fe -
ity after bereavement: A prospective study of 95,647 male victims of violence. Human Communication Re-
widowed persons. American Jot-wnal of Public Health, sea1'Ch, 15 (4 ), 509-22.
77,283 - 87. Linz, D., Donnerstein, E., and Penrod, S. (1984). The
Katz, J" and Melzack, R. (1990). Pain "memories" in effects of multiple exposures to filmed violence against
phantom limbs: Review and clinical observations. women. jou1'1lal ofComm1tnicatiol1, 34, 130-47.
Pain, 43, 319-36. - - - . (1988 ). The effects of long-term exposure to (
Kiecolt-Glaser, J. K., and Williams, D. A. (1987) . Self- violent and sexually degrading depictions of women.
blame, compliance, and distress among bum patients. joumal of Penonality and Social Psychology, 55, 758-
Joumal of Personality and Social Psychology, 53, 187- 68. (
93. Livneh, H., and Antonak, R. (1994 ). Review of research
Klabunde, E. H., and Falces, E. (1964). Incidence of on psvchosocial adaptation to neuromuscular disor-
complications in cosmetic rhinoplasties. Plastic and ders: I. Cerebral palsy, muscular dystrophy, and Par- (
Reconstl'uctiJle SU1;gC1)" 34 (2), 192-6. kinson's disease. Journal of Social Behavior and Per-
Klassen, A., Jenkinson, c., Fitzpatrick, R., and Good- sonality, 9 (5 ),201-30.
to:
if;
acre, T. (1996). Patients' health-related quality of life Loewenstein, G. (1996). A visceral account of addiction.
before and after aesthetic surgery. British Joumal of In J, Elster and O. J. Skog (Eds.), Getting hooked: Ra- (
Plastic Surgel)', 49, 433- 38. tionalit), and addiction (pp. 235-64). Cambridge:
Koob, G. F., Stinus, L., Le Moal, M., and Bloom, F. E. Cambridge University Press.
(1989). Opponent process theOlY of motivation: Neu- Loewenstein, G., and Adler, D. (1995). A bias in the
robiological evidence from studies of opiate depen - prediction of tastes. Economic jo 1t mal, 105, 929-
dence. Neul-oscimce and BiobehaJliol-al ReJlieH's, 13, 37.
135-40. Loewenstein, G., and Frederick, S. (1997) . Predicting
Koukounas, E., and Over, R. (1993). Habituation and reactions to environmental change. In M. Bazerman,
dishabituation of male sexual arousal. BehaJliour Re- D. Messick, A. Tenbrunsel, and K. Wade-Benzoni
search and Therapy, 31 (6), 575-85. (Eds.) , Psychological perspectiJles 011 the environment
Krantz, D. L., and Campbell, D. T. (1961). Separating (pp. 52-72). New York: Russell Sage Foundation.
perceptual and linguistic effects of context shifts upon Loewenstein, G., and Sichem1an, N. (1991). Do work-
absolute judgments. Joumal of Experimental Psychol- ers prefer increasing wage profiles? Journal of Labor
ogy, 62 (1), 35-42. Economics, 9,67-84.
Krupat, E. (1974). Context as a determinant of per- Lykken, D., and Telligen, A. (1996). Happiness is a
ceived threat: The role of prior experience. Joumal of stochastic phenomenon . Psychological Science, 7, 186-
Personality and Social Psychology, 29 (6), 731-36. 89.
Laan, E., and Everaerd, W. (1995). Habituation of fe- Mackenzie, D. L., and Goodstein, L. 1985. Long term
male sexual arousal to slides and film. ArchiJles of Sex- impacts and characteristics of long term offenders.
ual BehaJlior, 24 (5),517-41. O'iminal Justice and BeiJal'i01, 13,427- 47.
Laibson, D. 1. (1997). A cue-theory of consumption. Maddison, D ., and Walker, W. (1967) . Factors affecting
Working paper. Cambridge, Mass.: Economics De - the outcome of conjugal bereavement. British Journal
partment, Harvard University. of Ps),cbiatl)" 113, 1057- 67.
Lazarus, R. (1968). Emotions and adaptation: Concep- March, J. G. (1988). Variable risk preferences and adap-
Hedonic Adaptation 327

tive aspirations. jOlll-nal of Economic BehaJlior and Or- crowding. Bulletin of the Psychonomic Society, 6, 427-
ganization, 9, 5-24. 28 .
Marsh, H. W., and Parducci, A. (1978). Natural anchor- Pliner, P. (1982). The effects of mere exposure on liking
ing at the neutral point of categOIY rating scales. JOU1-- for edible substances. Appetite, 3, 283 - 90.
nal of Experimental Social Psychology, 14, 193-204. Poulton, E. C. (1989). Bias in quantifying judgments.
McFarland, C., Ross, M., and DeCourville, N. (1989). Hove, Eng.: Erlbaum.
Women's theories of menstruation and biases in recall Pribram, K H. (1984). Emotion: A neurobehavioral anal-
of menstrual symptoms. journal of Pel'sonality and So- ysis. In K R. Scherer and P. Ekman (Eds.), App1'Oaches to
cial Psychology, 57, 522-3I. emotion (pp. 13-38) . Hillsdale, N.J .: Erlbaum.
Melichar, J. F., and Chiriboga, D. A. (1988) . Sig- Price, D. , Harkins, S., and Baker, C. (1987). Sensory-
nificance of time in adjustment to marital separa- affective relationships among different types of clinical
tion. American journal of 01'thopsychiatlY, 58 (2), and experimental pain. Pain, 28, 297-307.
221-27. Reich, J. (1982). The interface of plastic surgery and
Meuwissen, 1., and Over, R. (1990). Habituation and psychiatry. Clinics in Plastic S1I1;gelY, 9 (3),367.
dishabituation of female sexual arousal. BehaJlioll1' Re- Roberts, J. V, and Jackson, M. (1991). Boats against
search and Therapy, 28, 217- 26. the current: A note on the effects of imprisonment.
Mou lton, J. M., Stempel, R. R., Bacchetti, P., Tem- LaJV and Human BehaJliol, 15, 557- 65.
oshok, L., and Moss, A. M. (1991). Results ofa one- Rolls, B. , Rowe, E., Rolls, E., Kingston , B., Megson, A.,
year longitudinal study of HIV antibody test notifica- and Gunary, R. (1981 ). Variery in a meal enhances
tion from the San Francisco General Hospital cohort. food intake in man. Physiology and Bel7flJlior, 26, 215-
journal of AIDS, 4, 787-94. 2I.
Mullan, J. T (1992). The bereaved caregiver: A pro- Ross, M . (1989). Relation of implicit theories to the
spective study of changes in well-being. The Ge1'01l- construction of personal histories. Psychological Re-
tologist, 32 (5), 673-83. view, 96, 341-57.
O'Bryant, S. L. (1991). Forewarning of a husband's Rozin, P., and Schiller, D. (1980). The nature and ac-
death: Does it make a difference for older widows~ quisition of a preference for chili pepper by humans.
Omega journal of D eath and Dying, 22,227- 39. MotiJlation and Emotion, 4 (1), 77-10 I.
O'Donohue, W. T, and Geer, J. H. (1985 ). The habitu- Rozin, P. , and Vollmecke, T. A. (1986). Food likes and
ation of sexual arousal. ArchiJles of Sexual Behavi01; 14, dislikes. Annual ReJliew of Nutrition, 6, 433-56.
233-46. Ryder, H. E., and Heal, G. M. (1973 ). Optimal growth
Ohlsen, L., Ponten, B., and Hambert, G. (1978 ). Aug- with intertemporally dependent preferences. R eJlielV of
mentation mammaplasty: A surgical and psychiatric Economic Studies, 40, 1-33.
evaluation of the results. Annals of Plastic Surgc!)" Sanders, C. M . (1980). A comparison of adult bereave-
2 (1), 42-52. ment in the death of a spouse, child, and parent.
Osterweis, M., Solomon, F., and Green, M . (Eds.) . Omega joumal of Death and Dying, 10, 303-22.
(1984). BereaJlement: Reactions, consequenccs, and Sapsford, R. J. (1978). Life-sentence prisoners: Psycho-
care. Washington, D.C.: National Academy Press. logical changes during sentence. Bl'itish journal of
Ostfeld, A., Kasl, S., D'Atri, D., and Fitzgerald, E. Criminology, 18, 128-45 .
(1987). Stl-ess, cl-olVding, and blood presSitre in prisol1. Sarris, V (1967). Adaptation-level theory: Two critical
Hillsdale, N.J.: Erlbaum . experiments on Helson's weighted-average model.
Parducci, A. (1968). The relativism of absolute judg- American joumal of Psychology, 80 (3),331-44.
ments. Scientific A1tte1-ican, 219, 84-90. Schelling, T (1984) . Self-command in practice, in pol-
- - -. (1984). Value judgments: Toward a relational icy, and in a theory of rational choice. Amn'ican Eco-
theory of happiness. In J. R. Eiser (Ed. ), Attit1ldinal nomic Review, 74, I-II.
judgment (pp. 3-21). New York. Springer-Verlag. Schkade, D., and Kahneman, D. (1998). Does living in
- --. (1995). Happiness, pleasure, andjudgmcllt: T77e Califomia make people happy? A focusing illusion in
contextual theory and its applications. Hove, Eng.: judgments of life satisfaction. Psychological Science, 9,
Erlbaum. 340-46.
Parkes, C. M. (1972). Components of the reaction to Schulz, R., and Decker, S. (1985). Long-ternl adjust-
loss of a limb, spouse, or home . joun1al of Psychosoma- ment to physical disability: The role of social support,
tic Resem-ch, 16, 343-49. perceiYed control, and self-blame. journal of Person-
Parkes, C . M ., and Weiss, R. S. (1983) . R ecoJle!)' fi'0111 ality alld Social Psycholog), 48 (5), 1162-72.
bereaJlement. New York: Basic Books. Schweitzer, M. (1995). Multiple reference points, fram-
Patterson, D. R., Everett, J. J., Bombardier, C. H., ing, and the status quo bias in health care financing
Questad, K A., Lee, V K, and Marvin, J. A. (1993 ). decisions. OI:ganizatiol1al BehaJlior and Human Deci -
Psychological effects of severe burn injuries. P,ycho- sion PI'OCcsses, 63, 69-72.
logical Bulletin, 113, 362-78. Scitovsky, T (1976). The joyless econ0111Y: The psychology of
Paulus, P., McCain, G., and Cox, V (1973). A note on h1l1na1l satisfaction. New York: Oxford Universiry
the use of prisops as environments for investigation of Press.
328 Well-Being
Shepard, R. (1981). Psychological relations and psycho- Tennen, H., and Afr1eck, G. (1990) . Blaming others for \'
physical scales: On the status of "direct" psychophysi- threatening events. Psychological Bulletin, 107, 209-
cal measurement. Journal of Mathematical Psychology, 32.
24,21-57. Thompson, R. F., Groves, P. M., Teyler, T. J., and
Shuchter, S. R., and Zisook, S. (1993). The course of Roemer, R. A. (1973) . In H. V. S. Peeke and M. J.
normal grief. In M. Stroebe, W. Stroebe, and R. Herz (Eds.), Habituation (vol. 1, pp. 239-71). New
Hansson (Eds.~, Handbook of bereavemmt: T71e01")" re- - York: Academic Press.
search, and intervention (pp. 23 - 43). Cambridge: Torrence, E. (1958) . Sensitization versus adaptation in
Cambridge University Press. preparation for emergencies: Prior experience with an
Siegel, S., Hinson, R. E., Krank, M. D. , and McCully, J. emergency ration and its acceptability in a simulated
(1982). Heroin "overdose" death: Contribution of survival situation. Jounlal of Applied Psychology,
drug-associated environmental cues. Science, 23, 436- 42 (1), 63 - 67.
37. Tversky, A., and Grifrin, D. (1991). Endowment and
Siegel, S., Krank, M. D., and Hinson, R. E. (1988). An- contrast in judgments of well-being. In F. Strack, M.
ticipation of pharmacological and nonpharmacological Arguyle, and N. Schwartz (Eds. ), SlIbjectil'e wcll-being:
events: Classical conditioning and addictiye behayior. An interdisciplin(7), perspective (vol. 21, pp. 101-18) .
In S. Peele (Ed.), Visions of Addiction. Lexington , Oxford: Pergamon Press.
Mass.: Lexington Books. Tyc, V. L. (1992). Psychosocial adaptation of children
Siller, J. (1969). Psychological situation of the disabled and adolescents with limb deficiencies: A review. Clin-
with spinal cord injuries. Rehabilitation Liw'ature, ical Psychology RClJiew, 2, 275-91.
30,290-96. van Praag, B. (1977). The welfare function of income in
Silver, R. L. (1982). Coping with an undesirable life Belgium : An empirical investigation . European Eco-
event: A study of early reactions to physical disabilitv. nomic Review, 337-69.
Ph.D. diss. Nonhwestern University, Evanston, Ill. van Praag, B. M. S., and van der Sar, N . L. ( 1988 ). Em-
Silver, R. L., and Wonman, C. B. (1980). Coping with pirical uses of subjective measures of well-being. Hlt-
undesirable life events. In J. Garber and M. E. P. Se- man Resoul'ces, 23, 193-210.
ligman (Eds.), Human helplessness: TheOl)' and appli- Veenhoven, R. (1991). Is happiness relative? Social Indi -
cations (pp. 279-340). New York: Academic Press . caton Research, 24, 1-34.
Smith, C. A., and Wallston, K. A. (1992). Adaptation in Walker, K. N., MacBride, A., and Vachon, M. L. S.
,,, patients with chronic rheumatoid arthritis: Application (1977). Social suppon networks and the crisis of
: 1.\ of a general model. Health Psychology, 11 (3), 151-62. bereavement. Social Science and Medicine, 11, 35-
I : .'
::! Smith, D., and Over, R. (1987). Does fantasy-induced 41.
sexual arousal habituate? Behaviour Research and Warr, P., Jackson, P., and Banks, M . (1988). Unemploy-
T71erapy, 25 (6),477- 85. ment and mental health: Some British studies. Journal
Solomon, R. L., and Corbit, J. D. (1974). An opponent- of Social Ism es, 44 (4), 47-68.
process theory of motivation: 1. Temporal dynamics of \-\feiner, B. (1982). The emotional consequences of
affect. Psychological Review, 81 (2), 119-45. causal attributions. In M. S. Clark and S. T. Fiske
Stevens, S. (1958). Adaptation-level versus the relativity (Eds.), Affect and cognition: T71e seventeenth annual
of judgment. Amel'ican JOltmal of Psychology, 71 (4), Canlegie symposium on cognition (pp. 185-210).
633- 46. Hillsdale, N.J. : Erlbaum.
Stevenson, R. J., and Yeomans, M. R. (1995) . Does ex- Weinstein, N. D. (1978). Individual differences in reac-
posure enhance liking for the chili burn? Appetite, 24, tions to noise: A longitudinal study in a college dor-
107- 20. mitory. Joumal of Applied Psychology, 63, 458-66.
Strahilevitz, M., and Loewenstein, G. (1998). The effect - - - . (1982). Community noise problems: Evidence
of ownership histOlY on the valuation of objects. J01l1'- against adaptation. Journal of Envil'onmental Psychol-
nal of Consumer Research, 25, 276-89. ogy, 2, 87- 97.
Stroebe, M., Stroebe, W., and Hansson, R. (Eds.). Weiss, J. M. (197la). Effects of coping behavior in dif-
(1993) . Handbook of bel'eavemellt: T71eOl)\ l'esem'ch, ferent warning signal conditions on stress pathology in
and inte71Je1ltion. Cambridge: Cambridge University rats. Journal of Comparative and Physiological Psychol-
Press. ogy, 77, 1- 13.
Stroebe, W., and Stroebe, M. (1987). Bereal'elllcnt a1Jd - - - . (1971b). Effects of punishing the coping re-
health. New York: Cambridge University Press. sponse (conflict) on stress pathology in rats. Joumal
Suedfeld, P., Ramirez, C ., Deaton, J., and Baker-Brown, of Compamtil'e and Physiological Psychology, 77, 14-
G. (1982). Reactions and attributes of prisoners in 21.
solitary confinement. C1'i1llinal Justice and BehaviOl; Weiss, R. S. (1987). Principles underlying a manual for
9,303- 40. parents whose children were ki lled by a drunk driver.
Taylor, S. (1983). Adjustment to threatening life events: American JOllmal of01'thops),chiM1Y, 57, 431-40.
A theory of cognitive adaptation. A1I1erica1J Psycholo- - - - . (1988 ). Loss and recovery. JOlu'nal of Social Is-
gist, 38, 1161 - 73 . sites, 44 (3), 37- 52.
Hedonic Adaptation 329

Wengle, H. (1986). The psychology of cosmetic sur- - - - . (1989). The myths of coping with loss. joumal
gel)': A critical overview of the literature 1960-1982. of Consulting and Clinical Psychology, 57, 349-57.
Part 1. Annals of Plastic SU1;gelY, 16 (5), 435-43. - - -. (1990). Successful mastety of bereavement and
Wiggins, S., Whyte, P. Huggins, M., Adam, S., et al. widowhood: A life course perspective. In P. B. Baltes
(1992). The psychological consequences of predictive and M. M. Baltes (Eds .), Successjitl aging: Perspectives
testing for Huntington's disease. New England JOU1-- ji-om the behavioral sciences (pp. 225-64). New York:
nal of Medicine, 327, 1401-5 . Cambridge University Press.
Wise, R. A., and Munn, E. (1995). Withdrawal from Wortman, C. B., Silver, R. c., and Kessler, R. C. (1993).
chronic amphetamine elevates baseline intracranial The meaning of loss and adjustment to bereavement.
self-stimulation thresholds. Psychopha1'1nacology, 117, In M. Stroebe, W. Stroebe, and R. Hansson (Eds.),
130-36. Handbooll of bCl-eavement: Theo1'Y, research, and inter-
Wormith, J. S. (1984). The controversy over the efrtcts vention (pp. 349-66). Cambridge: Cambridge Uni-
of long-term imprisonment. Canadian joumal of versity Press.
Criminology, 26, 423-37. Young, V. L., Nemecek, J. R., and Nemecek, D . A.
Wortman, C. B. (1976). Causal attributions and per- (1994). The efficacy of breast augmentation: Breast
sonal control. In J. H . Harvey, W. J. Ickes, and R. F. size increase, patient satisfaction, and psychological ef-
Kidd (Eds. ), New dil'cctiolls in attribution research fects. Plastic and Reconstructive SU1;gery, 94 (7), 958-
(vol. 1, pp. 23 - 51). Hillsdale, N.J.: Erlbaum. 69.
\VOltman, C. B. , Kessler, R. , Bolger, N ., House, J. Zajonc, R. (1968). Attitudinal effects of mere exposure.
(1992). The time course of adjustment to widow- jOllrnal of Personality and Social Psychology Monograph
hood: Evidence from a national probability sample. Supplement, 9 (2),2:1 - 32.
Unpublished manuscript, Duke University. Zamble, E. (1992). Behavior and adaptation in long-
Wortman, c., and Lehman, D. (1985 ). Reactions to vic· tem1 prison inmates: Descriptive longitudinal results.
tims of life crises: Support attempts that fail. In 1. G. O'iminal justice and Behavi01~ 19,409-25.
Sarason and B. R. Sarason (Eds. ), Social Sltpp01't: The- Zamble, E., and Proporino, F. (1990). Coping, imprison-
01')" research, and applications (pp. 463-89) . The ment, and rehabilitation : Some data and their implica-
Hague: Nijhof. tions. Criminal justice and Behrl1'i01, 17, 53-70.
WOltman, c., and Silver, R. (1987). Coping with irrevo- Zisook, S., Shuchter, S., and Lyons, L. (1987). Adjust-
cable loss. In Cataclysms, cl'iscs and catastrophes: Psy- ment to widowhood. In S. Zisook (Ed.), Biopsychoso-
chology in action. Master lecture series (\'01. 6, pp. cial aspects of bereavement (pp. 51-74). Washington,
189-235). Washington, D.C.: American Psychologi- D .C.: American Psychiatric Association Press.
cal Association.