Full text of DOI:10.1207/S15327752JPA7701_12

This article was downloaded by: [University of Calgary]
On: 07 October 2014, At: 12:50
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,
UK
Journal of Personality
Assessment
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/hjpa20
Psychometric Properties of the
Observer Alexithymia Scale in a
Clinical Sample
Mark G. Haviland , W. Louise Warren , Matt L. Riggs
& Molly Gallacher
Published online: 10 Jun 2010.
To cite this article: Mark G. Haviland , W. Louise Warren , Matt L. Riggs & Molly
Gallacher (2001) Psychometric Properties of the Observer Alexithymia Scale in a
Clinical Sample, Journal of Personality Assessment, 77:1, 176-186, DOI: 10.1207/
S15327752JPA7701_12
To link to this article: http://dx.doi.org/10.1207/S15327752JPA7701_12
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the
information (the “Content”) contained in the publications on our platform.
However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness,
or suitability for any purpose of the Content. Any opinions and views
expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the
Content should not be relied upon and should be independently verified with
primary sources of information. Taylor and Francis shall not be liable for any
losses, actions, claims, proceedings, demands, costs, expenses, damages,
and other liabilities whatsoever or howsoever caused arising directly or
indirectly in connection with, in relation to or arising out of the use of the
Content.

This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan,
sub-licensing, systematic supply, or distribution in any form to anyone is
expressly forbidden. Terms & Conditions of access and use can be found at
http://www.tandfonline.com/page/terms-and-conditions

JOURNAL OF PERSONALITY ASSESSMENT, 77(1), 176–186
Copyright © 2001, Lawrence Erlbaum Associates, Inc.
Psychometric Properties of the Observer
Alexithymia Scale in a Clinical Sample
Mark G. Haviland
Department of Psychiatry
Loma Linda University
W. Louise Warren
Western Psychological Services
Los Angeles, California
Matt L. Riggs
Department of Psychology
Loma Linda University
Molly Gallacher
Department of Social Work
Loma Linda University
The purpose of this study was to evaluate the psychometric properties of the Observer
Alexithymia Scale (OAS; Haviland, Warren, & Riggs, 2000) in a clinical setting.
Clinical and counseling psychologists used the OAS to rate outpatients (n = 192) with
various Diagnostic and Statistical Manual of Mental Disorders (American Psychiat-
ric Association, 1994) diagnoses. Reliability and validity data are similar to the initial
nonclinical data (n = 819): OAS scores are reliable (coefficient α = .90), and the five-
factor structure—Distant, Uninsightful, Somatizing, Humorless, and Rigid—was
confirmed. Moreover, the OAS does a relatively good job of differentiating clinical
from nonclinical cases. The OAS is psychometrically sound, and it appears to be a
useful tool for collecting and evaluating observer data on the clinically relevant, ev-
eryday expressions of alexithymia.
Alexithymia is a personality trait that appears to reflect deficits in the cognitive pro-
cessing and regulation of emotions (Lane, Ahern, Schwartz, & Kaszniak, 1997;
Taylor, 2000; Taylor, Bagby, & Parker, 1997). A unique trait (Haviland & Reise,

OBSERVER ALEXITHYMIA SCALE
177
1996; Luminet, Bagby, Wagner, Taylor, & Parker, 1999), alexithymia is associated
with various of serious mental and physical illnesses (e.g., substance dependence
and abuse, panic, posttraumatic stress, somatoform, and eating disorders; Taylor et
al., 1997). As such, researchers have found this construct useful as they study the
roles that personality and emotions play in the development and treatment of these
often intractable and costly diseases (Taylor, 2000).
The most popular alexithymia assessment tool is the Twenty-Item Toronto
Alexithymia Scale (Bagby, Parker, & Taylor, 1994). This self-report scale contin-
ues to be used most often by itself, however, despite the originators’ recognition
that scores are best evaluated in the context of other relevant observer and self-re-
port information (Taylor et al., 1997). As in personality and social psychology re-
search, generally (Funder, 1999), observer reports have been used far less
commonly in alexithymia research than have self-report measures, largely because
they are harder to obtain. Thus, conclusions about the alexithymia construct are
based almost entirely on the administration and correlation of self-report scales
(Taylor, 2000).
Two promising observer alexithymia measures—the Twelve-Item Modified
Beth Israel Hospital Psychosomatic Questionnaire (Taylor et al., 1997) and the
California Q-set Alexithymia Prototype (CAQ–AP; Haviland, 1998; Haviland &
Reise, 1996)—are time consuming and, for some raters, difficult to complete.
Thus, they have seen very limited use. A third observer measure—the Observer
Alexithymia Scale (OAS; Haviland et al., 2000)—is the subject of this study.
The OAS is a 33-item scale that purportedly measures alexithymia. Item con-
tent was taken from the CAQ–AP definition of alexithymia (Haviland & Reise,
1996). In CAQ–AP terms, the prototypic alexithymic person has difficulties expe-
riencing and expressing emotion; lacks imagination; and is literal, socially con-
forming, and utilitarian. He or she is not insightful, is humorless, has not found
personal meaning in life, and anxiety and tension find outlet in bodily symptoms.
This definition is consistent with the original construct formulations (Nemiah,
Freyberger, & Sifneos, 1976; Nemiah & Sifneos, 1970) and with the more recent,
comprehensive descriptions (Taylor, 2000; Taylor et al., 1997).
The OAS taps five alexithymic features: (a) distant (unskilled in interpersonal
matters and relationships), (b) uninsightful (lacking good stress tolerance and in-
sight or self-understanding), somatizing (having health worries and physical prob-
lems), (d) humorless (colorless and uninteresting), and (e) rigid (too self-
controlled). Among the OAS’s greatest strengths is that it is an observer report
with ordinary-language items accessible to both clinical judges and a (target) per-
son’s acquaintances or relatives.
The reliability and validity of the OAS have been examined for people in
nonclinical settings. Scores are reliable (coefficient α for the total score = .88 in
one sample, and .89 in a second; test–retest reliability = .87; coefficient α for the

178
HAVILAND, RIGGS, WARREN, GALLACHER
subscale scores range from .69 to .86 in two samples), and the instrument’s five-
factor structure is replicable (Haviland et al., 2000). Given these encouraging pre-
liminary findings, this study was designed to evaluate the OAS’s psychometric
properties in a clinical setting. Our primary purposes were (a) to evaluate the
psychometric properties (reliability and factorial validity) of the OAS in a clinical
sample using clinicians as judges and (b) to determine whether OAS scores distin-
guish clinical from nonclinical cases.
METHOD
Survey
We mailed the OAS to a random sample of 1,000 doctoral-level clinical and coun-
seling psychologists and distributed it to 20 local clinical psychologists (all were
American Psychological Association members providing patient care). Reminder/
thank-you postcards were sent approximately 4 weeks after the initial mailing and
distribution. Psychologists were asked to think of a patient whom they knew very
well and to rate the patient on the 33 OAS items. Our first preference was for each to
choose a patient with a Diagnostic and Statistical Manual of Mental Disorders
(DSM–IV; American Psychiatric Association, 1994) Axis I diagnosis of substance
dependence or abuse, panic, posttraumatic stress, somatoform, or eating disorder
(disorders associated with alexithymia). Our second preference was for them to
choose a patient with depressive or dysthymic disorder (disorders not necessarily
associated with alexithymia). We asked them not to choose patients with delirium
or dementia, bipolar disorder, or schizophrenia (because the data would be difficult
to interpret) or with an Axis II diagnosis of schizoid or antisocial personality disor-
der (disorders distinct from, but sharing features with, alexithymia). The psycholo-
gists also were asked to supply information about themselves (sex, age, and number
of years in practice) and their target patients (sex, age, race/ethnicity, educational
level, psychiatric diagnoses, and trauma experience).
Alexithymia is associated with our “first preference” disorders, being in poor
physical health (Taylor et al., 1997), and having experienced traumatic stress
(Krystal, 1988). Our goal was to have a good range of alexithymia severity in the
final sample and for the sample to include truly alexithymic individuals. We col-
lected diagnosis and trauma experience data to estimate the likelihood of achiev-
ing that goal.
Psychometric Properties
Internal consistency and scale–subscale relations. To evaluate inter-
nal consistency, we used coefficient alpha. To evaluate scale–subscale interrela-
tions, we used Pearson correlation.

OBSERVER ALEXITHYMIA SCALE
179
Factor structure (confirmatory factor analysis). We tested a hierarchical,
second-order model using EQS for Windows, Version 5.6 (Bentler & Wu, 1995).
The first-order factors—Distant, Uninsightful, Somatizing, Humorless, and
Rigid—were hypothesized to be products of the single, second-order construct,
alexithymia. Before testing the model, however, we created within-dimension item
parcels (two to four randomly selected items in each parcel and two or three parcels
for each dimension).
To evaluate model fit, we used the conventional chi-square test (Bollen, 1989)
and comparative fit index (CFI; Bentler, 1990). The chi-square statistic (and its
corresponding p value) tests the null hypothesis that the actual covariance matrix
among the variables is the same as the one implied by the specified model. A
nonsignificant (p > .05) chi-square, thus, is desirable and suggests that the model is
an adequate representation of the data. The CFI is an estimate of the proportion of
sample information that has been explained by the model, and it can range from 0
to 1 (perfect fit). Values above 0.90 are considered adequate; values in the 0.80 to
0.89 range are marginal. The distribution of CFI indexes (and other commonly
used fit indexes) is not known, so probabilities are not associated with their values.
OAS Total and Subscale Correlates
To compare total and subscale differences between scores obtained for women and
men, we used effect sizes (d = the mean difference divided by the within-population
standard deviation; Cohen, 1992). Values of .20, .50, and .80 correspond to small,
medium, and large effects, respectively. To evaluate total and subscale score differ-
ences related to age and education, we used correlation. Values of .10, .30, and .50
correspond to small, medium, and large effects, respectively (Cohen, 1992).
Clinical–Nonclinical Group Comparisons
T scores and effect sizes. We calculated T scores (M = 50, SD = 10) based
on the ratings of 819 nonclinical (people-in-general) targets (Haviland et al., 2000).
Nonclinical raterswere undergraduate andgraduate or professional school students
who rated parents, spouses, girlfriends or boyfriends, friends, adult children, sib-
lings, aunts or uncles, cousins, and in-laws. To evaluate group differences (clinical
vs. nonclinical subgroups), we used effect size (d).
Receiver operating characteristic analyses. We used receiver operating
characteristic (ROC) analyses to determine OAS total and subscale score thresh-
olds for differentiating the clinical from the nonclinical group members. ROC
curves are plots of sensitivity against false-positive rates, and each ROC analysis
gives an area under the curve (AUC). The AUC index can range from 0.5 (no pre-
dictive power) to 1 (perfect prediction). Optimal thresholds give the best balance

180
HAVILAND, RIGGS, WARREN, GALLACHER
between identifying the highest proportion of true clinical cases and the lowest pro-
portion of nonclinical cases incorrectly identified as clinical cases.
Item-group correlations. To determine which OAS items were the best
(and worst) clinical–nonclinical discriminators, we correlated each item with
group. Items were coded 0 (never, not at all like the person), 1 (sometimes, a little
like the person), 2 (usually, very much like the person), and 3 (all of the time, com-
pletely like the person). Group was coded 0 (nonclinical) and 1 (clinical).
RESULTS
Response Rate
Of the 1,000 surveys, 976 were deliverable. Of the deliverable surveys, 200 were
returned (20.5%), and 180 of those (90.0%) were usable. Of the 20 surveys distrib-
uted to local psychologists, 12 (60%) were returned and all were usable. Total N for
the study, thus, is 192. Of the 192 psychologist respondents, 54.2% were women,
and 45.8% were men. Average age was 49.2 years (SD = 7.6), and average years in
practice was 18.0 (SD = 7.8).
Patient Characteristics
Of the 192 patients, 69.8% were women, and 30.2% were men; mean age = 39.2
years (SD = 10.4). Most were White (88.0%), and 51.1% had bachelor’s or post-
graduate degrees. A DSM–IV Axis I diagnosis was provided for 190 participants
(one for 140 patients, two for 38, three for 10, and two for 4). Among the 190, the
most common first-listed diagnoses were anxiety/posttraumatic stress (32.6%),
anxiety/other than posttraumatic stress (20.5%), mood/major depressive (15.3%),
mood/dysthymic (11.1%), eating (8.9%), substance dependence (7.4%), and
somatoform (2.6%) disorders. Axis II diagnoses were listed for 53 participants, and
the most common disorders were borderline (41.5% of the 53; 11.6% of the 190 for
whom DSM–IV diagnoses were given) and dependent (17.0% and 4.7%, respec-
tively). Axis III diagnoses were listed for 78 participants. Trauma experiences were
noted for 133 (69.3%) participants. Sixty-three had experienced sexual abuse
(47.4% of the 133, and 32.5% all participants), and 49 (36.8% and 25.5%, respec-
tively) had experienced physical abuse. Although we cannot be certain that the
sample included individuals with severe/trait alexithymia, these figures increased
our confidence that such was the case.
OAS Scores
OAS and subscale means and standard deviations are shown in Table 1. OAS items
are rated on the previously described 0 to 3 scale, and, thus, total scores can range

OBSERVER ALEXITHYMIA SCALE
181
TABLE 1
OAS and Subscale Scores: Means and Standard Deviations
All Patientsa
Womenb
Menc
Scale
No. of Items
M
SD
M
SD
M
SD
OAS
33
10
8
5
5
47.6
14.5
13.1
7.3
6.0
6.8
14.6
5.6
4.3
4.4
3.1
3.5
47.7
14.5
13.2
7.3
6.0
6.6
14.1
5.5
4.3
4.2
3.1
3.3
47.3
14.4
12.7
7.2
5.9
7.0
15.6
5.9
4.2
4.9
3.3
4.0
1. Distant
2. Uninsightful
3. Somatizing
4. Humorless
5. Rigid
5
Note. OAS = Observer Alexithymia Scale.
aN = 192. bn = 134. cn = 58.
TABLE 2
OAS and Subscale Correlations
Scale
OAS
1
2
3
4
5
OAS
1. Distant
2. Uninsightful
3. Somatizing
4. Humorless
5. Rigid
.79
.73
.56
.72
.65
.41
.10
.64
.45
.45
.34
.26
.17
.20
.44
Note. OAS = Observer Alexithymia Scale.
from 0 to 99. Among all participants (N = 192), OAS scores ranged from 15 to 76
among women (n = 134) and 17 to 76 among men (n = 58). The mean OAS score for
all patients was 47.6 (SD = 14.6).
Psychometric Properties
Internal consistency and scale–subscale relations. Coefficient alpha
for the OAS was .90. Subscale alphas were .87 for Distant, .78 for Uninsightful, .87
for Somatizing, .82 for Humorless, and .75 for Rigid. The scale–subscale
intercorrelations matrix is shown in Table 2. Total subscale correlations ranged
from .56 (Somatizing) to .79 (Distant), and the median was .72. Subscale
intercorrelations ranged from .10 (Distant–Somatizing) to .64 (Distant–Humor-
less); the median was .38.
Factor structure (confirmatory factor analysis). After an initial run, we
fixedtheparameterssignificantlycontributingtomodelmisfit(LagrangeMultiplier
test; Bentler & Wu, 1995) and then ran a relaxed model (all fixed covariances were

182
HAVILAND, RIGGS, WARREN, GALLACHER
eitherparcel-to-parcelorfirst-orderfactortofirst-orderfactor).Thefinalmodelhad
anexcellentfittothedata:χ2(34,N=192)=47.50,p=.06(CFI=.988).Thepathsbe-
tweenallfirst-orderfactorsandtheiritemparcelswerelargeandsignificant(range=
.61 to .98; p < .05). Moreover, all five paths between the second-order construct and
the first-order dimensions were significant (p < .05), and four of the five were large
(range=.61to.93).Somatizing(.21)wastheexception.Moreimportant,noneofthe
parameter estimates between parcels and unrelated first-order factors resulted in
largestandardizedresiduals(i.e.,nosignificantcrossloadingofparcelstosecondary
subscales), which supports the distinctiveness of the five subscales.
OAS Total and Subscale Correlates
The mean differences between women and men on total and subscale scores
(shown in Table 1) were small (all effect sizes < .20). The correlations between age
and OAS and subscale scores also were small, r < .10. The correlation between edu-
cation level and OAS total score was in the small-to-medium effect size range, r =
–.24. The correlations between education level and Distant, r = –.30, Uninsightful,
r = –.25, and Humorless, r = –.21, also were in the small-to-medium effect size
range. We did not evaluate differences in OAS performance for groups based on
ethnicitybecauseonly20ofthepatientswerefromminorityethnicbackgrounds.
Clinical–Nonclinical Comparisons
T scores and effect sizes. The OAS total score difference between clini-
cal and nonclinical groups was substantial (13.4, d = 1.3), as were all subscale dif-
ferences, range = 7.8 (Somatizing) to 11.0 (Humorless); d range = 0.8 to 1.1.
ROC analyses. A summary of the ROC analyses is given in Table 3. For
OAS total score, the AUC was .82, and the best threshold score was 40 (sensitivity
= .73, specificity = .80). Subscales 1, 2, and 4 had better predictive power (AUCs of
TABLE 3
Receiver Operating Characteristic Analyses
Scale
AUC
Cutoff
Sensitivity
Specificity
OAS
.82
.76
.77
.66
.78
.71
40
12
11
6
4
5
.73
.70
.70
.63
.73
.72
.80
.69
.71
.64
.66
.60
1. Distant
2. Uninsightful
3. Somatizing
4. Humorless
5. Rigid
Note. AUC = area under the curve; OAS = Observer Alexithymia Scale.

OBSERVER ALEXITHYMIA SCALE
183
TABLE 4
Item-Group (Clinical–Nonclinical) Correlations
Correlated Items
r
Most highly: He or she …
is playfula
.39
.39
.38
.36
.36
.35
.34
.33
.33
.32
.30
knows him or herself wella
is flexiblea
has physical problems that are hard to treat
tells jokes and makes funny remarksa
understands his or her needs very wella
likes to touch or be toucheda
is imaginative; creativea
has a good sense of humora
likes to have close friendsa
has strong emotions that he or she cannot explain
Least: he or she …
is sensitive to other peoplea
talks about physical pain or discomfort
must “go by the book”
has trouble finding the right words for his or her feelings
likes to explore his or her feelingsa
.09
.08
.06
.06
.00
aItems are reverse scored.
.76, .77, and .78, respectively) than did Subscales 3 and 5 (respective AUCs = .66
and .71).
Item-group correlations. The items that were most highly, r > .30, and least
correlated, r < .10, with clinical–nonclinical group are shown in Table 4. Not sur-
prisingly, given the results of the ROC analyses, items from the Distant,
Uninsightful, and Humorless subscales were the best discriminators. Interestingly,
none of the least correlated items were negatively related to group.
DISCUSSION
The primary purpose of this study was to evaluate the OAS’s psychometric proper-
ties in a clinical sample. Clinical reliability and validity data are similar to the initial
data for those in nonclinical settings (Haviland et al., 2000)—total and subscale
scores are reliable, and the five-factor structure is stable. In light of these encourag-
ing findings, we recommend that researchers and clinical investigators use both to-
tal OAS and subscale scores (Distant, Uninsightful, Somatizing, Humorless, and
Rigid). We recommend also that, when possible, researchers obtain and compare
OAS scores from multiple, independent judges. Judge–judge agreement is a good

184
HAVILAND, RIGGS, WARREN, GALLACHER
proxy for accuracy (Funder, 1999), and consensus scores, generally, are more reli-
able (and valid) than individual scores (Cheek, 1982). Block (1961/1978), Funder
(1999), and McCrae (1994) have made several useful suggestions for calculating
consensus scores (and for evaluating discrepant reports).
Our secondary purpose was to determine how OAS scores differ for clinical
versus nonclinical cases. First, all OAS scores for clinical participants (a group
largely composed of patients with disorders and experiences associated with
alexithymia), on the average, were considerably higher than the scores for
nonclinical participants, with large effect sizes of 1.3 for total raw score and 0.8 to
1.1 for the subscale scores. In fact, using an OAS total score cutoff of 40, one can
distinguish between the two groups reasonably well. It is important to underscore,
however, that it remains to be demonstrated that OAS scores can make the distinc-
tion between cases in which the presence or absence of alexithymia has been estab-
lished independently. At present, however, a gold standard for identifying
alexithymic cases does not exist. Moreover, it is not known whether alexithymia is
appropriately thought of as “present or absent” as opposed to “present along a con-
tinuum.” Given these uncertainties (and the desirability of using both observer rat-
ings and self-reports in personality research; McCrae, 1994), OAS scores clearly
are best evaluated with other direct and indirect observer and self-report
alexithymia assessments.
This study is not without its limits, and chief among them is the relatively low
response rate. This raises two important questions. First, was the sample size large
enough for the confirmatory factor analysis? Although, no firm guidelines have
been established for minimum number of participants per variable (item) in ex-
ploratory factor analyses or confirmatory factor analysis, we met the 5-to-1 partici-
pant-to-variable rule of thumb, and fell just shy of the generally preferred sample
size of 200 (Floyd & Widaman, 1995).
Second, what were the consequences of nonresponse bias on mean OAS total
and subscale scores? Unfortunately, we can only speculate about the possible ef-
fects. Nonresponding psychologists, for example, might have chosen patients with
different levels of alexithymia severity, and thus shifting mean scores up or down.
It seems unlikely, however, that such shifts (even several points in either direction)
would have materially affected the results of the confirmatory factor analysis. On
the other hand, for several reasons, one must exercise caution in interpreting the
various OAS mean total and subscale scores: (a) the effects of nonresponse bias
are unknown; (b) we intentionally biased the sample by asking clinicians to choose
patients with problems associated with alexithymia; and (c) alexithymia is associ-
ated with older age, male sex, lower socioeconomic status, and fewer years of edu-
cation (Lane, Sechrest, & Riedel, 1998), and our sample was disproportionately
female and highly educated.
Despite these limits, the results of this study support the use of the OAS, partic-
ularly in research protocols when an observer measure of alexithymia is desired. It

OBSERVER ALEXITHYMIA SCALE
185
is based on a comprehensive, consensus definition of the construct (Haviland &
Reise, 1996); it is brief and accessible to clinical and lay judges; and it taps every-
day expressions of alexithymia that are observable and relevant to diagnosis, treat-
ment planning, and outcomes assessments.
ACKNOWLEDGMENT
This study was supported, in part, by Grant 0316–6955–00 from the Loma Linda
University School of Medicine Research Resources Committee.
REFERENCES
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th
ed.). Washington, DC: Author.
Bagby, R. M., Parker, J. D. A., & Taylor, G. J. (1994). The Twenty-Item Toronto Alexithymia Scale I:
Item selection and cross-validation of the factor structure. Journal of Psychosomatic Research, 38,
23–32.
Bentler, P. M. (1990). Comparative fit indices in structural models. Psychological Bulletin, 107,
238–246.
Bentler, P. M., & Wu, E. J. C. (1995). EQS for windows user’s guide. Encino, CA: Multivariate
Software.
Block, J. (1978). The Q-sort method in personality assessment and psychiatric research. Palo Alto, CA:
Consulting Psychologists Press. (Original work published in 1961)
Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley.
Cheek, J. M. (1982). Aggregation, moderator variables, and the validity of personality tests: A peer-rat-
ing study. Journal of Personality and Social Psychology, 43, 1254–1269.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159.
Floyd, F. J., & Widaman, K. F. (1995). Factor analysis in the development and refinement of clinical as-
sessment instruments. Psychological Assessment, 7, 286–299.
Funder, D. C. (1999). Personality judgment: A realistic approach to person perception. San Diego, CA:
Academic.
Haviland, M. G. (1998). The validity of the California Q-set alexithymia prototype. Psychosomatics, 39,
536–539.
Haviland, M. G., & Reise, S. P. (1996). A California Q-set Alexithymia Prototype and its relationship to
ego-control and ego-resiliency. Journal of Psychosomatic Research, 41, 597–608.
Haviland, M. G., Warren, W. L., & Riggs, M. L. (2000). An observer scale to measure alexithymia.
Psychosomatics, 41, 385–392.
Krystal, H. (1988). Integration and self-healing: Affect, trauma, and alexithymia. Hillsdale, NJ: The
Analytic Press.
Lane, R. D., Ahern, G. L., Schwartz, G. E., & Kaszniak, A. W. (1997). Is alexithymia the emotional
equivalent of blindsight? Biological Psychiatry, 42, 834–844.
Lane, R. D., Sechrest, L., & Riedel, R. (1998). Sociodemographic correlates of alexithymia. Compre-
hensive Psychiatry, 39, 377–385.
Luminet, O., Bagby, R. M., Wagner, H., Taylor, G. J., & Parker, J. D. A. (1999). Relation between
alexithymia and the five-factor model of personality. Journal of Personality Assessment, 73,
345–358.
McCrae, R. R. (1994). The counterpoint of personality assessment: Self-reports and observer ratings.
Assessment, 1, 159–172.

186
HAVILAND, RIGGS, WARREN, GALLACHER
Nemiah, J. C., Freyberger, H., & Sifneos, P. E. (1976). Alexithymia: A view of the psychosomatic pro-
cess. In O. W. Hill (Ed.), Modern trends in psychosomatic medicine (Vol. 3, pp. 430–439). London:
Butterworths.
Nemiah, J. C., & Sifneos, P. E. (1970). Affect and fantasy in patients with psychosomatic disorders. In
O. W. Hill (Ed.), Modern trends in psychosomatic medicine (Vol. 2, pp. 26–34). London:
Butterworths.
Taylor, G. J. (2000). Recent developments in alexithymia theory and research. Canadian Journal of
Psychiatry, 45, 134–142.
Taylor, G. J., Bagby, R. M., & Parker, J. D. A. (1997). Disorders of affect regulation: Alexithymia in
medical and psychiatric illness. Cambridge, England: Cambridge University Press.
Mark G. Haviland
Department of Psychiatry
LLUSM
11374 Mountain View Avenue
Loma Linda, CA 92354–3842
E-mail: haviland@ix.netcom.com
Received June 23, 2000
December 7, 2000