Cultural Differences 1

Running head: CULTURAL DIFFERENCES IN PAIN EXPRESSION

Cultural Differences in Pain Expression in a Cold Pressor Task

Nancy Alvarado

California State Polytechnic University, Pomona

Ralph B. Jester

University of California, Irvine

Christine R. Harris

University of California, San Diego

Julia F. Whitaker

University of Utah, Health Sciences Center

Author Contact:
Nancy Alvarado, Ph.D.
Department of Psychology and Sociology
California State Polytechnic University
3800 W. Temple Ave.
Pomona, CA 91768
Phone: (909) 869-3896
Email: nalvarado@csupomona.edu

Key words: acute pain; pain assessment; racial and ethnic differences; under-treatment of pain,

cold pressor, pain facial expression

 Cultural Differences 2

Abstract
 Cultural Differences 3

Cultural Differences in Pain Expression in a Cold Pressor Task

Numerous studies have documented that members of ethnic and racial minority groups,

and women, are more likely to be under-treated for pain (Bonham, 2001; Weisse, Sorum &

Dominguez, 2003; Hoffman & Tarzian, 2001). Under-treatment has been reported in emergency

treatment of bone fractures (Todd, Deaton, D’Adamo, & Goe, 2000; Todd, Samaroo &

Hoffman, 1993), cancer pain (Cleeland, Gonin, Baez, Loehrer & Pandya, 1997; Cleeland,

Gonin, Hatfield, Edmonson, Blum, Stewart & Pandya, 1994), and postsurgical care (McDonald,

1994). Further, even when treated, they are less likely to be given stronger opioid drugs

(Pletcher, Kertesz, Cohn & Gonzales, 2008). Williams (2002; Kappesser & Williams, 2002;

Kappesser, Williams & Prkachin, 2007) suggested that this under-treatment may arise from

physician bias, a tendency to underestimate a person’s pain and thus under-prescribe pain

medication, or even misdiagnose the underlying condition. Prkachin (Prkachin, Berzins &

Mercer, 1994; Prkachin, Mass & Mercer, 2004; Prkachin, Solomon, Hwang & Mercer, 2001)

documented a systematic underestimation of pain that was greater among experienced health

care providers than untrained observers. He and his colleagues demonstrated that the actual

facial movements of a pain expression increased with greater self-reported pain, providing a

read-out that was accurate at high intensities of pain, less so at lower intensities. However,

despite this perceptual information, physicians displayed an underestimation bias when physician

pain ratings were compared to sufferer pain ratings on the same rating scale. This suggests that

physicians may observe pain expressions but discount them, placing less reliance on them than

untrained observers do. In this research, we explored reasons for the observed under-treatment of

pain and accompanying physician underestimation bias. Specifically, we investigated whether

members of racial and ethnic minority groups (here called subcultures) were likely to show

differences in their pain expressivity, self-report or physiology that might be misinterpreted by

physicians and result in under-treatment for pain.

 Cultural Differences 4

The tradeoffs between detection of malingering and accurate appraisal of pain may

contribute to physician underestimation bias (Williams, 2002). When physicians become too

suspicious, their failure to accurately assess pain may contribute to a vicious circle in which

chronic pain sufferers exaggerate their pain in order to be adequately treated while physicians

respond to the exaggeration by increasingly discounting their expressions of pain. Studies of

chronic pain sufferers suggest that individuals may over time exaggerate their expressions in

order to more effectively communicate with health care deliverers, even showing pain

expressions when no pain is felt (Prkachin, Berzins & Mercer, 1994; Prkachin, 2005). Chronic

pain sufferers are most often suspected of malingering (Mendelson & Mendelson, 2004).

Historically, physicians have been suspicious of higher pain reports. Mendelson & Mendelson

(2004) review historical attitudes toward malingering. They cite Hackett (1971), who described

a prejudiced physician attitude toward patients in pain, and Miller & Cartlidge (1974) who

defined malingering as not only simulation of disease or disability not present, but also, a much

more frequent gross exaggeration of minor disability. Mendelson & Mendelson (2004) stated,

“Given the subjective nature of pain, it therefore becomes problematic to determine what would

be the “expected” extent of pain associated with a particular physical lesion…” (p. 425).

Physicians are reported to become less sympathetic, even angry, if they believe that patients are

exaggerating their pain symptoms or reports (Poole & Craig, 1992; Prkachin & Craig, 1995).

Keefe & Dunsmore (1992) noted, “Conscious efforts to communicate pain through guarded

movements, facial expressions, or extreme ratings of pain upset and even enrage clinicians.” (p.

97, quoted by Prkachin & Craig, 1995, p. 202).

Thus, difficulties can arise for members of minority subcultures if their experience of

pain is divergent from that of the majority of patients, or if their expressivity is different. Such

differences may lead physicians to suspect dissimulation and give rise to discounting of self

report or facial expression or even suspicion of malingering. The actual experience of pain
 Cultural Differences 5

among minority group members is not as well studied as among non-minorities. For example,

opioid-induced hyperalgesia (a greater sensitivity to pain-inducing experience) may be expected

to affect minority group members differentially if their exposure to such drugs differs from that

of other groups. Recent investigations suggest that real differences exist in pain experiences

among African American; Hispanics, and Asians (Lipton & Marbach, 1984; Edwards &

Fillingim, 1999; Breitbart & McDonald, 1996; Sheffield, Biles, Orom, Maixner & Sheps, 2000;

Bates, Edwards & Anderson, 1993; Brena, Sanders & Motoyama, 1990; McCracken, Matthews,

Tang & Cuba, 2001; Weisenberg, Kreindler, Schacht & Werboff, 1975; Barak, Weisenberg,

1988; Faucett, Gordon & Levine, 1994; Sternbach & Tursky, 1965). Greater awareness of sex

differences in pain experience has led to a consequent correction in physician expectations

(Weisse, Sorum & Dominguez, 2003). Differences in emotional response to clinical situations,

including pain-related anxiety, have also been documented in African American patients

(Fillingim, Edwards, & Doleys, 2002; McCracken, Matthews, Tang & Cuba, 2001). These

studies suggest that increased ratings of pain can result from an increased experience of pain, not

solely from a tendency to describe pain differently or to use rating scales differently.

Stereotypes about the expression of pain complicate actual differences in pain experience

and report. Early research in pain expression had no method for objectively and systematically

describing facial action, as now exists with Ekman and Friesen’s Facial Action Coding System

(FACS) (1978). As with early research on emotional expressions, studies of pain expression

were often impressionistic rather than scientific. Studies of “Yankee” (Northeastern American)

and Jewish subgroups (Tursky & Sternbach, 1967; Sternbach & Tursky, 1965) contributed to

today’s general belief that ethnic groups may have characteristic styles of pain expression.

Yankees were considered stoic whereas Mediterranean people were described as “dramatic.”

With the advent of better methods for characterizing facial activity, the issue of styles or dialects

of expression can be more rigorously addressed. For example, Asians are stereotypically thought
 Cultural Differences 6

to be more stoic than European-Americans (Chang, 2003). Recent attempts to study this using

infant facial expression showed mixed results for Japanese and Chinese or Chinese-Canadian

infants compared to non-Asian infants (Rosmus et al., 2000). Further, it is unclear whether

culture can strongly modify facial expression. Learning-based models of facial expressivity

attribute cultural differences to socialization, yet attempts to modify facial expression were

unsuccessful for adults modeling tolerance and intolerance of pain (Craig & Patrick, 1985). A

conflicting study in which mothers modeled pain behavior to their children (Goodman &

McGrath, 2003) confounded heritability with modeling. More recently, the ability to both

enhance and suppress expression has been associated with successful coping with adversity

(Bonanno et al., 2004). As with many characteristics, larger variability in expressivity exists

within a cultural group than between groups (Prkachin, 1992; Prkachin & Craig, 1995).

Whether cultural differences in expression exist or not, wide individual variability

implies that an entirely genuine pain response may be inappropriately treated if the physician

holds mistaken expectations about the amount of pain a patient should be expressing. A mistaken

expectation may arise from a misunderstanding of real differences in pain experience coupled

with triggered stereotypic beliefs about minority pain. To complicate matters, stereotypes differ

for males and females due to different gender roles in some subcultures. For example, Hispanic

patients are stereotypically judged as histrionic or overly expressive, especially when female. An

expressive Hispanic man may be judged as malingering because his expressions are inconsistent

with expectations about macho stoicism. Or he, himself, may conceal and under-rate his own

pain because he considers it unmanly to reveal pain. African-American and Asian stereotypes

have historically included insensitivity to pain as part of the justification for mistreatment during

enslavement or “Coolie” indentured servitude (Chang, 2003). Such stereotype-driven

expectations conflict with the observed greater sensitivity to pain reviewed above. Physician

attitudes about malingering or misuse of emergency room services by immigrant groups such as
 Cultural Differences 7

Hispanics and Asians with low socioeconomic status may contribute to resentment and lessened

sympathy among physicians. Further, immigrant patients may believe that the way to be a “good

patient” is to suffer without complaint, leading to misdiagnosis or under-treatment.

Studies of pain deception show that adults and children can more readily exaggerate than

suppress pain facial expressions (Poole & Craig, 1992; Larochette, Chambers & Craig, 2006).

When there is a discrepancy between self-report of pain and facial expression, physicians give

greater credence to the facial expression (Prkachin & Craig, 1995; Poole & Craig, 1992). They

believe that self-reports are easier to fake than facial expressions. This reliance on facial

expression is problematic because of the large individual differences in expressivity. Prkachin &

Craig report “At severities at which the pain was reported to be substantial, 13-50% of subjects

displayed no facial evidence.” (p. 194; Wilkie, 1995). How are facially inexpressive patients

who self-report strong pain treated when they are members of an ethnic minority? It seems likely

that there is a difference in the potential for under-treatment when inexpressivity is stereotype-

congruent rather than stereotype-incongruent. These questions have not been investigated.

In the face of such complexities, it would be useful to identify a “gold standard” for

assessment of pain. Pain is typically assessed using multiple indicators, including physical

responses such as tachycardia, blood pressure reactivity, muscle rigidity, and behavioral

measures such as self-report (verbal and “Oucher” or “Face” scales), guarding, and facial

expression (Jensen & Karoly, 2001). However, use of indicators such as tachycardia and blood

pressure reactivity is frequently prevented by advanced age, hypertension (which is anti-

nociceptive), beta blocking or hypertension medication, or presence of narcotics, leaving the

physician reliant on facial expression and self-report. Self-report is given strong weight except

where patients are suspected of medication seeking or malingering (McCaffery, 1979; Mersky,

1979), but Poole & Craig (1992) note that self-report is affected by situational factors and

incentives. Further, self-report may be unreliable when patients are infants, children or elderly,
 Cultural Differences 8

when conscious introspection is impaired, or when patients cannot speak or communicate

effectively, increasing reliance on facial expression of pain. Further, self-report scales are

subject to a variety of anchoring, retrospective and other scale biases, only now being

investigated. Some researchers suggest that under-treatment of pain might be reduced if facial

expression were preferred to other indicators (Williams, 2002; Prkachin, 2007).

Accurate estimation of pain may be accomplished using facial expression, but only if

physicians use the cues actually present in the face instead of relying upon beliefs about pain

(Prkachin, Berzins & Mercer, 1994). Poole & Craig (1992) demonstrated that observers tended

to give more weight to a person’s pain facial expression than to their pain statements, even when

the expressions were faked or suppressed and thus inaccurate. Hill & Craig (2004) suggested that

individuals can be trained with feedback to more accurately assess pain facial expressions,

whereas information-based training was unhelpful. Unfortunately, many individuals are facially

inexpressive or show incongruence between facial expression and other pain indicators, making

the use of facial expression as a gold standard problematic for many sufferers.

Although physicians routinely deal with pain, they may not be pain experts. Because

assessment of pain is generally a holistic judgment with several inputs, physicians may be

unaware of the extent to which their own cultural stereotypes and beliefs can affect their

interpretation of pain indicators. As Prkachin & Craig (1995) noted, even when beliefs about

pain are explicit, “Stereotypes fail to recognize tremendous within-group differences and small

between-group differences that call into question their utility.” (p. 198). Thus, understanding the

nature of real and assumed differences among groups is crucial to: (a) accurate interpretation of

pain self-report and facial expression; and (b) accurate identification of medication-seeking,

malingering and other deception. Optimal treatment for members of minority subcultures relies

upon accurate estimation of pain whichever indicators are used.

This research used multiple measures to assess pain expressivity on a cold pressor task
 Cultural Differences 9

across four subcultures within the student population: (1) African Americans; (2) Asian

Americans and Asian immigrants; (3) Hispanic; and (4) European Americans. The dependent

variables included: (1) autonomic measures; (2) facial expressivity; (3) self report using several

types of scales; (4) measurement of pain attitudes by questionnaire; and (5) measurement of

acculturation by questionnaire. Our goal was to examine similarities and differences on these

measures across the four groups, to examine coherence among the measures within groups, and

to identify any differences that might be linked to stereotypes or the systematic under-treatment

among physicians and other health care professionals. We predicted that cultural beliefs would

mediate pain expressivity (facial expression) and self report and that we would find greater

stoicism within the Hispanic and Asian groups and greater expressivity among African

Americans. We further predicted that women would be more expressive than men, based on their

greater expressivity in other facial expression studies and the greater acceptance of complaint

about physical discomfort permitted of women within many cultures. Specific hypotheses are

noted where appropriate in the Results section.

Method

Participants

Our intention in this study was to recruit students in order to limit prior experience and

increase the likelihood of finding healthy subjects without chronic pain or drug use. Participants

are listed in Table 1 by sex and ethnicity. All subjects were students at the California State

Polytechnic University, Pomona, ranging in age from 18 to 53 (mean = 22.1, sd = 4.0). Subjects

were included in the data analysis only if they had complete data on all measures. Most subjects

were recruited using an online human subjects pool signup system and were given course credit.

Due to difficulty recruiting sufficient participants via the subject pool, some African American

and Asian American male subjects were recruited from campus locations via flyers and were

paid $10 for their participation. Subjects were screened for health problems that might interfere
 Cultural Differences 10

with their ability to experience pain or produce facial expressions (e.g., high blood pressure,

schizophrenia or depression). Individuals currently taking pain medication or medications that

might alter autonomic response were also excluded.

Materials

The cold pressor task was administered using a Jeio Tech RW-0525G refrigerated

circulating bath which maintained water temperature at 3o Centigrade within two tenths of a

degree. A Biopac MP100WS psychophysiological recording system was used to record heart

rate, blood pressure and electrodermal activity. A Vasotrac system was used to continuously

measure blood pressure via a wrist cuff. A Sony 900 digital video camera was used to record

facial expressivity during a 10 minute baseline and throughout the cold pressor task.

To assess pain attitudes, a review of the literature on cultural beliefs about pain was

conducted. This resulted in a series of questions that were presented to four focus groups, asking

subjects about their family attitudes and beliefs about pain. From this, an exploratory pain

attitudes questionnaire was developed, consisting of 63 questions assessing different aspects of

belief about pain and its appropriate expression. A more detailed description of the focus group

and questionnaire results is reported by Englert, Jester, Alvarado, Harris and Whitaker (2009).

Self-report rating scales included a series of seven-point scales anchored by emotion

terms, the McGill Pain Inventory, and a seven-point pain rating scale. Acculturation was

measured using an inventory appropriate to each subject’s self-described cultural background.

For Asian Americans, we presented the SL-ASIA (Suinn, Rickard-Figueroa, Lew, & Vigil,

1987). For Hispanics, we presented the Short Acculturation Scale for Hispanics (Marin, Sabogal,

Marin, Otero-Sabogal & Perez-Stable, 1987). For African Americans, we presented the revised

African-American Acculturation Scale AAAS-R (Klonoff & Landrine, 2000). FIX Two scales

were used to assess African American acculturation, one based on adherence to cultural practices

within the African American subculture, the other based on attitudes toward European
 Cultural Differences 11

Americans. The European American scale was the counterpart to the latter and largely concerned

attitudes toward African Americans. The Hispanic and Asian American acculturation scales were

largely based on languages spoken and participation in activities within the subculture.

Procedures

After informed consent and screening for inclusion, subjects were asked to wash their

hands and remove any jewelry or watches. They then were seated inside a cubicle in a chair

facing a screened video camera (to minimize its salience) while an experimenter affixed sensors

to monitor heart rate at ankles and wrist. The Vasotrac wrist cuff and a finger sensor for

electrodermal activity were attached. Following that, the subject was left alone for a 10 min

videorecorded baseline period to permit the subject to accustomed used to the camera and

presence of the sensors. The experimenter then returned. A doorbell-type signal was placed in

the participant’s right hand and the experimenter asked the subject to press it, to demonstrate its

operation. The subject was instructed to press the signal button at the first sign of pain and again

when the pain became intolerable and they wished to remove the hand from the water. Subjects

were asked to tolerate the water as long as possible, but were also told they could remove the

hand whenever they wished. The experimenter left the cubicle and from outside told the subject

to “go” (start the task) by placing his or her hand in the cold water up to the wrist, with fingers

open but not touching the sides or bottom of the tank. The experimenter was seated behind a

screen outside the cubicle and recorded the start, first signal and second signal on the Biopac

record. When the signal button was pressed for the second time (or the subject removed the

hand), the experimenter returned and dried the hand with a towel. If the subject did not press the

second button by the end of 3 minutes, the experimenter returned and ended the task.

Immediately following the task, the subject was asked to complete the McGill Inventory on a

clipboard and rate the pain. Following that, the subject was moved to a computer to complete the

other rating scales and inventories. Pain and emotion ratings were presented first, followed by
 Cultural Differences 12

the pain attitudes questionnaire and then the acculturation questionnaire, followed by a

debriefing. If the subject tolerated the pain for 3 minutes (timed out), the experimenter asked

follow up questions about that during the debriefing.

Results

Cold Threshold and Tolerance

The cold threshold was measured as the time from first immersion in cold water to the

time the signal button was pressed for the first time, to signal the experience of pain. Cold

tolerance was measured as the time from the first button press to signal pain until the second

button press to request removal of the hand from the water. A time-out was recorded when the

subject left the hand in the water for 3 minutes, at which time the hand was removed by the

experimenter. No significant differences were found across the four subcultures for threshold,

tolerance, or total time in water. However, 12 timeouts were observed in the European American

group compared to 6 for Asian American and 4 for the other groups, which contributed to a

nearly significant difference in total time in water, F(3,179)=2.126, p=.099, but did not affect

tolerance or threshold. Examining the debriefing responses, this higher number of timeouts

appeared to be related to the number of athletes in the European American group. They reported

surfing and swimming in cold water or using cold water hydrotherapy to treat sports injuries.

They may also have been more likely to regard the pain manipulation as a challenge or

competition. With all timeouts removed, there was no significant difference in mean pain

threshold, tolerance or total time in water across groups. A consistent sex difference was found

for threshold, t(181)=4.188, p=.000 and total time in water, F(1,155)=7.715, p=.006, but no

significant difference was found for tolerance, t(181)=1.265, p=.207.

Biophysiological Measures

Physiological measures were used for two purposes: (1) to verify that participants

experienced pain during the cold pressor task; and (2) to determine whether any group showed
 Cultural Differences 13

significant differences in physiological response. Christine please add results here

Pain and Emotion Ratings

Participants rated pain and a series of emotions using seven-point rating scales,

including: happiness, anger, fear, anxiety, frustration, embarrassment, calmness, sadness and

surprise. No significant differences were found across the four subcultures on any of these scales

(p < .05). Significant gender differences for happiness, excitement, pain and embarrassment

were found (see Table 2). When ratings were normalized within sex, no significant differences

were found across subcultures. Similarly, when ratings were normalized within subculture, no

significant differences were found by group, but the sex differences remained significant.

Although no other significant differences were found after normalization, a trend was observed

in which Asian American participants showed smaller sex differences in their ratings compared

to the other three groups, while Hispanic and African American groups showed consistently

larger sex differences, suggesting that sex roles affected ratings more for those groups than for

Asian Americans and to a lesser extent, European Americans, as illustrated in Figure 2.

Participants also rated the quality of their experience using the McGill Pain Inventory. It

includes rating scales anchored by a series of descriptive terms including: throbbing, shooting,

stabbing, sharp, cramping, gnawing, hot or burning, aching, heavy, tender, splitting, exhausting,

tiring, sickening, punishing, cruel, fearful and painful. The inventory is scored by combining

subsets of these scales to produce a sensory subscale and an affective subscale. Significant

differences were found on the sensory subscale, by group, with African American and Hispanic

participants of both sexes reporting higher scores, F(3,159)=2.878, p=.038. Significant

differences were also found by gender, with female participants consistently reporting higher

scores than males, F(1,159)=12.399, p=.001. There was no significant interaction between sex

and ethnicity, F(3,159)=.403, p=.751. For the affective subscale, no significant main effects

were found, but a significant interaction between gender and subculture was found,
 Cultural Differences 14

F(3,172)=2.958, p=.034. This occurred because Asian American and European American female

participants gave higher ratings than males whereas Hispanic and African American males gave

higher ratings than females. Thus the pattern of results for the sensory subscale was considerably

different than the affective subscale, as shown in Figure 3.

In medical contexts, patients are frequently asked to estimate the amount of pain they are

willing to withstand using a ten-point scale. Our participants were asked to do the same. After

the pain manipulation, they then rated their pain using the same ten-point scale, permitting a

comparison of the anticipated and actual experience. The scale, as in medical contexts, was

labeled discomfort. The estimates were significantly correlated with the experienced discomfort,

r=.50, p=.000 but there was a significant difference between the estimated and experienced

ratings, F(1,137)=8.413, p=.005. All participants tended to overestimate the amount of

“discomfort” they would be willing to withstand. There were no significant differences between

subcultures and no significant interactions. There was a significant sex difference, with males

producing higher estimates than females and a greater discrepancy between their estimated and

actual ratings, F(1,137)=14.107, p=.000. Ratings by sex and subculture are compared in Figure

4. Interestingly, both the discomfort estimates and the actual discomfort ratings are uncorrelated

with the ratings made using a scale labeled “pain” immediately after the task, r=-.095 for

estimated discomfort and r=-.003 for actual discomfort.

Pain Attitudes Questionnaire

Although space does not permit a detailed analysis of the responses to the Pain Attitudes

questionnaire developed for this study (see Englert et al., 2009), an analysis of the responses for

each question showed that seven questions produced significant differences across groups. These

are listed in Table 2. To determine whether culture affected responding, a discriminant analysis

was used to classify subjects into subcultures based upon their questionnaire responses. Because

the group sizes were unequal, prior probabilities were computed from the group sizes during the
 Cultural Differences 15

analysis. Those who had timed out were excluded. The discriminant analysis classified

participants into their self-identified subcultures with 81.1% accuracy. When sex was used as the

grouping variable, subjects were classified with 86.7% accuracy. To explore the impact of

culture on responses, subjects were divided into high and low acculturation groups using a

median split of their acculturation inventory scores. Those with high acculturation scores were

considered to be closer to the mainstream culture and those with low scores were considered to

be closer to their subculture. The discriminant analysis was then run again on each of these two

subsets of participants. The two separate groups of subjects classified by acculturation both were

correctly classified with 98.6% accuracy, however, the groupings are tightly clustered (more

cohesive) for the less acculturated subjects, as can be seen in Figure 5a. The improvement of

classification accuracy when acculturation is used to partition the sample demonstrates that

cultural attitudes toward pain do differ in ways that characterize each group. The questions that

were most important as discriminators are generally those that produced significant differences

between scores (see Table 2). However, in addition to the items in Table 2, the following

questions were important to the discriminant analysis: (1) People who have suffered great pain

are to be admired; (2) I avoid the doctor because I am concerned that they might find something

truly wrong with me; and (3) There is no point to complaining when in pain. These questions

had significance levels of .07 using one-way ANOVA. Overall, groups differed along

dimensions of expressivity and stoicism, with Asian Americans reporting the least expressivity

and the most stoicism, and African Americans showing the greatest expressivity and the

strongest belief that pain is important during medical treatment (see Table 2).

Facial Expression

Facial activity during a ten minute baseline and during the cold pressor task was

videorecorded. Facial expressions were coded during the entire duration of the cold pressor task.

A comparable baseline period of equal duration was also coded, immediately before the end of
 Cultural Differences 16

the baseline during the time when the subject was assumed to be most accustomed to the

presence of the camera. Ekman and Friesen’s (1978) Facial Action Coding System (FACS) was

used to identify facial movements associated with pain in previous research. These movements,

called action units (AUs) were compared in two ways: (1) mean number of each AU per subject;

and (2) number of subjects showing the presence or absence of each specific AU. Overall facial

activity was also measured by coding the number of events (co-occurring patterns of facial

activity) per subject (Ekman & Rosenberg, 1997). Representative pain expressions are shown in

Figure 6. No differences across groups were found for the AUs most frequently associated with

pain (e.g., AU 20, 9, 4+7, Craig & Patrick, 1985). No differences in the mean number of events

were found across groups, F(3,173)=0.66, p=.580 or by sex, F(1,173)=2.56, p=.110, suggesting

similar levels of expressivity. A discriminant analysis using all AUs as input variables correctly

classified subject by group with 45.3% accuracy (25% is chance) and by sex with 71.3%

accuracy (50% is chance). Further analysis showed that AUs related to affect, not pain,

accounted for group differences, with an interaction between sex and subcultural group. While

the results were not significant, African American women were more likely to show AU 1+4

(characterized as a distress expression), Hispanic and African American women were more

likely to smile when in pain (AU 1+6), and Hispanic men were more likely to frown (AU 4+7).

The frequency of smiling during pain was higher for women than men, F(1,173)=3.893, p=.050.

Rates of smiling and frowning by group are shown in Figure 7. As can be seen, Hispanic males

and females showed the greatest sex difference in expressive behavior, while Asian American

males and females showed the greatest similarity. There was a slight, nonsignificant trend

toward less expressivity for Asian American subjects across all types of AUs. No significant

gender/ethnicity interactions were found.

Discussion

The multiple measures used in our study present a complex picture in which some
 Cultural Differences 17

measures show significant differences while others do not, but a pattern emerges. The high

acculturation of our college student sample may have prevented finding strong cultural

differences that may exist in more isolated communities and older immigrant groups.

Nevertheless, we found significant differences related to culture in the attitudes and beliefs about

pain and in the self-report and expressive behavior related to affect but not pain. No

physiological differences were found across subcultures, suggesting that the pain experience was

similar from a physical standpoint but different emotionally.

Based on our physiological measures, we are confident that our subjects experienced

strong pain. Excluding the timeouts which occurred when subjects approached the time-limited

cold pressor task as a competitive exercise, no group appeared more or less willing to tolerate

pain and thresholds for feeling pain were the same across groups. We did find the sex

differences noted in previous studies, present within each of our cultural groups. The cultural

differences observed were found in facial expression related to affect (smiling, frowning,

distress), as classified by Ekman and Friesen (1978). The significant differences across

subcultures were also found in the self report of emotional states (happiness, embarrassment,

anger) and in the McGill affective subscale (not the sensory subscale). We found significant

differences in attitudes and beliefs about pain, largely along dimensions related to expressivity,

expectations for physicians, and the value of stoicism.

We found some support for the stereotype that Asian American are more stoic, but

believe this is related to attitudes and cultural norms about expressing pain. Asian subjects did

not choose swearing as an appropriate expression of pain for anyone (e.g., self, mother, father),

although all three other groups did. They were the only group to select “shake it off” as a

mother’s response to a cut hand. Although the difference was not significant, Asian American

subjects were consistently the least expressive facially (see Figure 7). Asian American subjects

were more likely to disagree that pain should be important to physicians and less likely to expect
 Cultural Differences 18

pain treatment (see Table 2). Sex differences in the use of pain and affect ratings scales were

smallest for Asian American subjects.

In contrast, African American subjects seemed to value pain expression more and were

more expressive during the task. They agreed most strongly with the importance of physicians

treating pain, were more likely to express pain even in contexts where a job might be lost, and

tended to rate pain higher using the rating scales. They also showed more facial expressivity,

although this was a non-significant difference. African American women showed more frequent

distress expressions (AU 1+4) than any other group. African American subjects were more likely

to expect pain relief as part of medical treatment, despite also stating that doctors were more

likely to suspect them of medication seeking.

Hispanic subjects showed large sex differences, both in rating scale usage (see Figure 3)

and in facial expressivity (see Figure 7). This suggests that culture-related gender roles may

dictate differences reflected in expectations and behavior for Hispanic subjects. The impact of

stereotypes about machismo did not produce suppression of expression, as expected. Instead,

there appeared to be suppression of positive affect and greater frowning, rather than in greater

stoicism overall. Hispanic subjects were nearly as expressive as African Americans and more

expressive than European Americans. There was no observed tendency to tolerate pain longer or

to rate pain as less severe, as might be expected given macho stereotypes. There may have been

a tendency to exaggerate pain, perhaps to compensate for inability to withstand the cold for as

long as desired. Greater embarrassment was reported for Hispanic and African American males,

perhaps related to their failure to tolerate the cold for the entire three minutes.

Although the difference was not significant, we observed that Hispanic and African

American women were more likely to smile when in pain. Smiling may be related to

embarrassment or shame, emotions that may arise in medical settings (Miller, 1996; Harris,

2006). This is important because physicians may believe that someone who is in pain would not
 Cultural Differences 19

smile and thus discount their self-reported intensity of pain. Women are also an undertreated

group (Weisse, Sorum & Dominguez, 2003) and it may be that the increased smiling among

women in general, and minority women in particular, may account for much of the observed

under-treatment. This would not account for the under-treatment among Asian patients reported

in the literature, but it may be that different explanations apply to different groups.

The presence of smiling while in pain by both males and females during this task was

unexpected. Such smiling may be related to embarrassment or shame, emotions that may arise in

medical settings. In particular, males seemed to smile right before removing their hands from the

water, perhaps as an acknowledgement that they had to give up (Keltner, 1995). Smiling may be

a coping mechanism when feeling embarrassment or shame (Keltner & Anderson, 2000; Keltner

& Buswell, 1997). Subjects were seated alone in a cubicle during the cold pressor manipulation

to minimize social interaction, so they were not smiling at anyone or in supplication at any

person. If physicians believe that smiling always indicates pleasure or the absence of pain, this is

a serious misunderstanding of the nature of pain expression.

Several factors worked against finding significant differences across groups. First, we

kept water temperature at 3o centigrade. Lower water temperature tends to reduce sex differences

(CITE) by making it more difficult to tolerate pain in service of other goals. We believe that

may have impacted cultural self-expectations in similar ways. Second, although diverse, our

subjects are well-acculturated and many are psychology majors. They have been well-educated

about culture and gender roles and that may have influenced their responses. The differences

observed in this study are most likely the hardiest remaining after socialization into the

mainstream culture. We would predict that greater differences would be observed in studies of

less acculturated samples. Even so, attitudes and beliefs about pain were sufficiently different

that subjects could be classified by their responses into cultural groups with high accuracy. That

suggests that culture is an important mediator of pain experience and behavior in medical
 Cultural Differences 20

settings. Beliefs and attitudes do influence emotional response and it is well known that emotion,

in turn, influences pain experience. If the strongest impact of culture is on emotion rather than

pain, it is still important to address such differences in medical practice.

Our findings suggest that it would be worthwhile for physicians to take into account

cultural differences when assessing pain. African American patients may be more expressive as a

group, without being suspected of intentionally exaggerating their behavior. Asian American

patients may be less expressive while feeling stronger pain than would be indicated by the same

behavior in someone from a different culture. Hispanic males may frown when in pain while

Hispanic females may smile and show little negative affect despite pain. European Americans

are more likely to tell you about their pain. Because cultural differences appeared to have less

impact on pain expression (grimacing, nose wrinkling, eye tightening), it may be worthwhile for

physicians to learn to attend to these independent of affective expression (smiling, frowning).

Finally, the 10-point discomfort scale used by many hospitals showed little relation to the

various pain ratings and facial expressions in our study. All groups tended to overestimate the

amount of pain they were willing to tolerate, with greater accuracy for females than males. More

seriously, subjects tended to interpret the label “discomfort” to mean something different than

“pain.” If physicians are using the term discomfort as a synonym for pain our results suggest

there may be serious miscommunication.

The need for improved pain assessment is especially pressing today as the Joint

Commission on Accreditation of Healthcare Organizations (2000) has called for the monitoring

of pain as a fifth vital sign and made effective pain control part of hospital accreditation

standards. Consideration of these, and similar, cultural influences on pain expression can and

should be incorporated into physician training to improve accuracy of pain assessment.

 Cultural Differences 21

References

xBarak, E. & Weisenberg, M. (1988). Anxiety and attitudes toward pain as a function of

ethnic grouping and socioeconomic status. Clinical Journal of Pain, 3, 189-196.

xBonanno, G.A., Papa, A., Lalande, K., Westphal, M & Coifman, K. (2004). The

importance of being flexible: The ability to both enhance and suppress emotional expression

predicts long-term adjustment. Psychological Science, 15, 482-487.

xBonham, V.L. (2001). Race, ethnicity, and pain treatment: Striving to understand the

causes and solutions to the disparities in pain treatment. Journal of Law and Medical Ethics, 29,

52-68.

xBreitbart, W., McDonald, M.V., Rosenfeld, B., Passik, S.D., Hewitt, D., Thaler, H. &

Portenoy, R.K. (1996). Pain in ambulatory AIDS patients. I: Pain characteristics and medical

correlates. Pain, 68, 315-321.

xBrena, S.F., Sanders, S.H. & Motoyama, H. (1990). American and Japanese chronic

low back pain patients: Cross-cultural similarities and differences. Clinical Journal of Pain, 6,

118-124.

xChang, I. (2003). The Chinese in America: A narrative history. Viking.

xCraig, K. & Patrick, C. (1985). Facial expressions during induced pain. Journal of

Personality and Social Psychology, 48, 1080-1091.

xEkman, P. & Friesen, W. (1978). Facial Action Coding System Manual. Palo Alto, CA:

Consulting Psychologists Press, Inc.

xEkman, P. & Rosenberg, E. (1997). What the face reveals: Basic and applied studies of

spontaneous expression using the Facial Action Coding System (FACS). New York, NY: Oxford

University Press.
 Cultural Differences 22

xCleeland, C.S., Gonin, R., Baez, L., Loehrer, P., Pandya, K. (1997). Pain and treatment

of pain in minority patients with cancer: The Eastern Cooperative Oncology Group Minority

Outpatient Pain Study. Annals of Internal Medicine, 127, 813-816.

xCleeland, C.S., Gonin, R., Hatfield, A.K., Edmonson, J.H., Blum, R.H., Stewart, J.A.,

& Pandya, K.J. (1994). Pain and its treatment in outpatients with metastatic cancer. New

England Journal of Medicine, 330, 592-596.

xEdwards, R.R. & Fillingim, R.B. (1999). Ethnic differences in thermal pain responses.

Psychosomatic Medicine, 61, 346-354.

Ekman, P. (1977). Biological and cultural contributions to body and facial movement. In

J. Blacking (Ed.), The anthropology of the body (pp. 38-84). London: Academic Press.

xEkman, P. & Friesen, W. (1978). Facial Action Coding System Manual. Palo Alto, CA:

Consulting Psychologists Press, Inc.

xEkman, P. & Rosenberg, E. (1997). What the face reveals: Basic and applied studies of

spontaneous expression using the Facial Action Coding System (FACS). New York, NY: Oxford

University Press.

xEnglert, P.A., Jester, R.B., Alvarado, N., Harris, C.R., and Whitaker, J.F. (2009). ADD

xFaucett, J., Gordon, N. & Levine, J. (1994). Differences in postoperative pain severity

among four ethnic groups. Journal of Pain Symptom Management, 9, 383-389.

xFillingim, R., Edwards, R., & Doleys, D. (2002). Letter to the Editor (Commentary on

McCracken et al.), The Clinical Journal of Pain, 18, 136-137.

xGoodman, J. & McGrath, P. (2003). Mothers’ modeling influences children’s pain

during a cold pressor task. Pain, 104, 559-565.

xHarris, C.R. (2006). Embarrassment: A form of social pain. American Scientist, 94,

524-533.
 Cultural Differences 23

xHill, M. & Craig, K. (2004). Detecting deception in facial expressions of pain. Clinical

Journal of Pain, 20, 415-422.

xHoffman, D.E. & Tarzian, A.J. (2001). The girl who cried pain: A bias against women

in the treatment of pain. Journal of Law & Medical Ethics, 29, 13-27.

xJensen, M.P. & Karoly, P. (2001). Self-report scales and procedures for assessing pain

in adults. In: D. Turk & R. Melzack (Eds), Handbook of Pain Assessment, Second Edition

(Chapter 2, pgs 15-34). New York: The Guilford Press.

xJoint Commission on Accreditation of Healthcare Organizations (2000), Implementing

the new pain management standards. Oakbrook Terrace, Ill.: JCAHO.

xKappesser, J. & Williams, A.C. (2002). Pain and negative emotions in the face:

judgments by health care professionals. Pain, 99, 197-206.

xKappesser, J., Williams, A.C. & Prkachin, K. (2006). Testing two accounts of pain

underestimation. Pain, 124, 109-116.

xKeefe, F.J. & Dunsmore, J. (1992). Pain behavior: Concepts and controversies.

American Pain Society Journal, 1, 92-100.

xKeltner, D. (1995). Signs of appeasement: Evidence for the distinct displays of

embarrassment, amusement, and shame. Journal of Personality and Social Psychology, 68, 441–

454.

xKeltner, D., & Anderson, C. (2000). Saving face for Darwin: The functions and uses of

embarrassment. Current Directions in Psychological Science, 9, 187–192.

xKeltner, D., & Buswell, B.N. (1997). Embarrassment: Its distinct form and appeasement

functions. Psychological Bulletin, 122, 250–270.

xKlonoff, E. & Landrine, H. (2000). Revising and improving the African American

Acculturation Scale. Journal of Black Psychology, 26, 235-261.

 Cultural Differences 24

xLarochette, A.C., Chambers, C.T. & Craig, K.D. (2006). Genuine, suppressed and faked

facial expressions of pain in children. Pain, 126, 64-71.

xLipton, J.A. & Marbach, J.J. (1984). Ethnicity and the pain experience. Social Science

& Medicine, 19, 1279-1298.

xMarin, G., Sabogal, F., Marin, B.V., Otero-Sabogal, R. & Perez-Stable, E. (1987).

Development of a short acculturation scale for Hispanics. Hispanic Journal of Behavioral

Sciences, 9, 183-205.

xMcCaffery, M. (1979). Current misconceptions about the relief of acute pain. In B.L.

Crue Jr. (Ed.), Chronic pain. Jamaica, NY: Spectrum.

xMcCracken, L., Matthews, A., Tang, T. & Cuba, S. (2001). A comparison of blacks and

whites seeking treatment for chronic pain. The Clinical Journal of Pain, 17, 249-255.

xMcDonald, D. (1994). Gender and ethnic stereotyping and analgesic administration.

Research in Nursing & Health, 17, 5-49.

xMendelson, G. & Mendelson, D. (2004). Malingering pain in the medicolegal context.

The Clinical Journal of Pain, , 423-432.

xMersky, H. (1979). Pain terms: A list with definitions and notes on usage. Pain, 6, 249-

252.

xMiller, R. S. (1996). Embarrassment: Poise and peril in everyday life. New York:

Guilford Press.

xPletcher, M.J., Kertesz, S.G., Kohn, M.A. & Gonzales, R. (2008). Trends in opioid

prescribing by race/ethnicity for patients seeking care in US emergency departments. JAMA,

299,70-78.

xPoole, G. & Craig, K. (1992). Judgments of genuine, suppressed, and faked facial

expressions of pain. Journal of Personality and Social Psychology, 63, 797-805.

 Cultural Differences 25

xPrkachin, K. (1992). The consistency of facial expressions of pain: a comparison across

modalities. Pain, 51, 297-306.

xPrkachin, K. (2007). The coming of age of pain expression. Pain, 133, 3-4.

xPrkachin, K., Berzins, S., & Mercer, S. (1994). Encoding and decoding of pain

expressions: a judgment study. Pain, 58, 253-259.

xPrkachin, K. & Craig, K. (1995). Expressing pain: The communication and

interpretation of facial pain signals. Journal of Nonverbal Behavior, 19, 191-205.

xPrkachin, K. Mass, H. & Mercer, S. (2004). Effects of exposure on perception of pain

expression. Pain, 111, 8-12.

xPrkachin, K.M., Solomon, P., Hwang, T., & Mercer, S.R. (2001). Does experience

affect judgments of pain Behavior? Evidence from relatives of pain patients and health-care

providers. Pain Research and Management, 6, 105-112.

xRosmus, C., Johnston, C., Chan-Yip, A. & Yang, F. (2000). Pain response in Chinese

and non-Chinese Canadian infants: is there a difference? Social Science & Medicine, 51, 175-

184.

xSheffield, D., Biles, P.L., Orom, H. Maixner, W., & Sheps, D.S. (2000). Race and sex

differences in cutaneous pain perception. Psychosomatic Medicine, 62, 517-523.

xSternbach, R.A. & Tursky, B. (1965). Ethnic differences among housewives in

psychophysiological and skin potential responses to electric shock. Psychophysiology, 1, 241-

246.

xSuinn, R., Rickard-Figueroa, K., Lew, S. & Vigil, P. (1987). The Suinn-Lew Asian

Self-Identity Acculturation Scale: An initial report. Educational and Psychological

Measurement, 47, 401-407.

xTodd, K.H., Deaton, C, D’Adamo, A, & Goe, L. (2000). Ethnicity and analgesic

practice. Annals of Emergency Medicine, 35, 11-16.

 Cultural Differences 26

xTodd, K.H., Samaroo, N., & Hoffman, J.R. (1993). Ethnicity as a risk factor for

inadequate emergency department analgesia. Journal of the American Medical Association, 296,

1537-1539.

xTursky, B. & Sternbach, R.A. (1967). Further physiological correlates of ethnic

differences in responses to shock. Psychophysiology, 4, 67-74.

xWeisenberg, M., Kreindler, M.L., Schacht, R., Werboff, J. (1975). Pain, anxiety and

attitudes in black, white, and Puerto Rican patients. Psychosomatic Medicine, 37, 123-135.

xWeisse, C., Sorum, P. & Dominguez, R. (2003). The influence of gender and race on

physician’s pain management decisions. The Journal of Pain, 4, 505-510.

xWilkie, D.J. (1995). Facial expressions of pain in lung cancer. Analgesia, 1, 91-99.

xWilliams, A.C. (2002). Facial expression of pain: An evolutionary account. Behavioral

and Brain Sciences, 25, 439-488.

 Cultural Differences 27

Author Notes

Send correspondence to: Nancy Alvarado, Department of Psychology and Sociology,

California State Polytechnic University, Pomona, 3801 W. Temple Avenue, Pomona, CA 91768.

We thank Phyllis Ann Englert, Noriko Coburn, Yvonne Burgos, Ruben Hoyos, and the many

undergraduate students who assisted with this research. This research was supported by NIH

MBRS/Score Grant S06 GM053933.

 Cultural Differences 28

Table 1. Participants included in data analysis.

Female Male Total

African American 17 15 32

Asian American 25 25 50

European American 22 26 48

Hispanic 29 24 53

All Groups 93 90 183

 Cultural Differences 29

Table 2. Pain Attitudes Questionnaire items showing significant differences across subcultures.

Item African Asian Hispanic European F or 2 p

American American American

If you were walking and 3.69 4.42 3.68 3.92 2.80 .041
you came across a person (1.36) (1.47) (1.60) (1.22)
who had just bumped into a
very sharp object, and they
showed no reaction at all,
what would you think of
them? (1=disapprove,
7=approve)
If your mother cut her hand Bandage it Bandage it Bandage Bandage it 51.83 .042
while working around the and go and go it and go and go
house, what would she be back to back to back to back to
most likely to do? work, work, work, work,
scream or scream or swear, swear,
yell, tell yell, shake scream or scream or
you how it off yell, cry, yell, tell
much it ask for you how
hurt help much it
hurt
I expect the doctor to listen 6.66 (.70) 5.96 6.5 (.98) 6.23 (.99) 4.04 .008
to me when I talk about my (1.18)
pain. (1=disagree, 7=agree)
Pain relief is a major part of 5.59 4.74 5.25 4.94 2.90 .036
medical treatment. (1.29) (1.59) (1.29) (1.31)
(1=disagree, 7=agree)
Doctors assume that 2.91 2.08 2.42 1.58 (.90) 6.04 .001
everyone from my ethnic (1.63) (1.32) (1.77)
background is seeking pain
medicine to sell on the
 Cultural Differences 30

street. (1=disagree, 7=agree)

There is a difference 6.03 (.97) 5.46 5.79 5.25 3.33 .021
between being in pain and (1.30) (1.20) (1.30)
being injured. (1=disagree,
7=agree)
I do not tell others when I 2.69 3.30 2.62 3.37 2.77 .043
am in pain because I might (1.64) (1.75) (1.48) (1.63)
lose my job or my place on
an athletic team.
(1=disagree, 7=agree)
 Cultural Differences 31

Figure Captions

Figure 1. Pain experience results (total time in water with timeouts excluded) showed no

difference by ethnicity but a consistent sex difference for all four groups.

Figure 2. Standardized self-report ratings within subculture show that African American and

Hispanic males report greater embarrassment during the cold pressor task.

Figure 3. McGill ratings showed a different pattern of results for the sensory and affective

subscales.

Figure 4. Males and females in all four groups predict they can withstand more discomfort than

they are willing to tolerate.

Figure 5. Less acculturated subjects (top) are more readily classified by subculture based on their

pain attitudes than highly acculturated subjects (bottom).

Figure 6. Representative facial expressions during the cold pressor task for the four subcultures

(top to bottom: African American, Hispanic, Asian American and European American).

Figure 7. Males frowned more frequently whereas females smiled more frequently when in pain,

with larger sex differences for Hispanic subjects.

Cultural Differences 32
Cultural Differences 33
Cultural Differences 34
Cultural Differences 35
Cultural Differences 36
Cultural Differences 37
Cultural Differences 38