A PSYCHOMETRIC INVESTIGATION OF THE MULTIDIMENSIONAL

HEALTH LOCUS OF CONTROL SCALES WITH CIGARETTE SMOKERS

RICHARD J. COELHO
Michigan State University

This study examined the psychometric properties of the Multidimensional

Health Locus of Control (MHLC) Scales with chronic cigarette smokers (N
= 146) who volunteered for treatment through an at-large cessation clinic.
Results showed that expectancies for health control were not distributed
along three independent domains as implied by the format ofthe instrument.
Instead, health locus of control orientation was found to be bidimensional,
with domains defined by the Internal and Powerful Others measures. Several
alternative explanations for this finding are discussed.

Recognizing the necessity for a specialized measure of locus of control to assess

health-specific control expectancies, Wallston, Wallston, Kaplan, and Maides (1976)
developed a health-related locus of control scale. They claimed that their unidimensional
Health Locus of Control (HLC) scale would improve the prediction of health-related
behaviors in contrast to the more generalized measure, such as Rotter's (1966) InternalExternal Locus of Control Scale. However, largely due to its low internal consistency
(i.e., Lewis, Morisky, & Flynn, 1978) and evidence by Levenson (1975) ofthe utility of a
multidimensional locus of control construct, Wallston and his associates (Wallston,
Wallston, & DeVellis, 1978) revised the instrument. The new instrument, the
Multidimensional Health Locus of Control Scale (MHLC), was designed to measure
three separate dimensions of locus of control beliefs related to health behavior: Internality (IHLC); Powerful Others externality (PHLC); and Chance (CHLC) externality.
Two different f^orms (A and B) ofthe instrument are available; each contains three 6-item
scales. In addition, a third version that consists of 12-item scales can be obtained by combining Forms A and B.
According to the locus of control construct, which is derived from Rotter's (1954)
social learning theory, those with an internal health locus of control orientation believe
that they can control many aspects of their health through their own behavior. Persons
with a chance health orientation believe that the control of their health lies with luck,
fate, or chance and that there is little they themselves can do to affect it. Individuals with
a powerful others orientation believe that health professionals control health and that
regular contact with them and adherence to their orders is the best way to stay healthy.
An extensive literature has been published on the locus of control construct; much of it
suggests that, other things equal, an internal orientation is preferable (Boyle & Harrison,
1981). Strickland (1978) provides an extensive review of published research on the locus
of control construct related specifically to health behavior.
While the MHLC scales represent an attempt to measure locus of control expectancies in the area of personal health, to date the psychometric characteristic of the scales
have not been examined fully (O'Looney & Barrett, 1983). Wallston et al. (l?78)have
reported preliminary evidence for reliability and predictive and construct validity, which
indicates that the scales are an accurate measure of health-related locus of control.
However, their hypothesized dimensionality has been based entirely upon correlated
evidence that employed total scale scores. Subsequent research has reported contradictory evidence of subscale validity and independence within various populations. For example, Shipley (1981), who studied cigarette smokers {N = 43) in a cessation clinic.
Requests for reprints should be addressed to Richard J. Coelho, Ph.D., Department of Psychiatry, East
Fee Hall, Michigan State University, East Lansing, Michigan 48824-1316.

372

Health Locus of Control

373

found moderate correlations between subscales of .25 to .37, which suggests the independence of subscales. Hartke and Kunce (1982), who used male medical patients (A^
= 86) in a VA hospital, found intercorrelations to be less than .30. In addition, their factor analysis tended to confirm MHLC subscale independence. In contrast. West (1980)
failed to replicate subscale independence for lower-class Black-American and MexicanAmerican men. Intercorrelations were high and ranged from .68 to .80. In a recent investigation of the psychometric properties of the MHLC, O'Looney and Barrett (1983)
found sex differences in the factor structure such that male college students yielded only
two factors, while females yielded the three proposed factors. Their conclusions, along
with those of West (1980), have led to the suggestion that the MHLC instrument be used
with caution.
The MHLC Scale has been utilized in a substantial number of studies that investigated various health conditions and health-related behaviors with a wide range of
populations (Wallston &. Wallston, 1981). However, there has been only one study that
pertained to MHLC beliefs with chronic cigarette smokers who were seeking treatment
(i.e., Shipley, 1981). Cigarette smoking is a major health hazard and has been linked with
the development of various debilitating conditions. The current lack of research in this
area represents a significant gap in the literature. Therefore, it seems important to examine the psychometric properties of the MHLC Scales with cigarette smokers who are
engaged in a preventive health effort.
The present study was designed to provide additional information on health locus of
control beliefs with chronic cigarette smokers and to determine whether a multidimensional orientation (i.e., three domains), as proposed by Wallston et al. (1978), would be
supported.

METHOD

Subjects
Subjects were 146 chronic cigarette smokers (53 males; 93 females) from the adult
community in the Lansing (Michigan) metropolitan area who volunteered to participate
in a smoking cessation study. The mean age was 41 (range = 20-67), mean years smoking was 22 {SD = 10.6), mean baseline daily smoking rate was 30 cigarettes {SD - 12.2),
and 71% had attained some college education. Compared with the normative data
presented by Wallston et al. (1978), which was provided by adult air travellers, mainly
from Tennessee (55%), this sample contained only Michigan residents and had a higher
proportion of females (64%).
Procedure
Responses to items on the MHLC (Form A) provided the data base for the study.
The MHLC questionnaire is an 18-item instrument that uses a 6-point Likert-type format, with self-ratings from "Strongly Disagree" to "Strongly Agree." The three subscales each contain 6 items with a range of potential scores from 6 to 36.
Coefficient alpha (Cronbach, 1970), a measure of internal consistency, first was
measured by each subscale. (See Table 1.) Subscale items were summed to determine
each scale's total score, and these scores then were correlated in order to estimate their
independence. Finally, smokers* responses to the 18 internal, powerful others, and
chance scale items were subjected to principal-component factor analysis with communalities followed by a Varimax rotation. Eigenvalues were obtained, and those factors
whose weights were above 1.0 were subjected to further evaluation. In all, two factors
that accounted for 74% of the variance were identified by this method. (The eigenvalues
for these factors are 2.4 and 1.8.) The factors and factor loadings are summarized in
Table 2.

374

Journal of Clinical Psychology. May 1985. VoL 41, No. 3

RESULTS AND CONCLUSIONS

The means, standard deviations, and alpha reliability coefficients for the MHLC
Scales are presented in Table 1. The highest mean score for this group of cigarette
smokers was on the IHLC Scale; the mean scores on the CHLC and PHLC Scales were
relatively undifferentiated. The mean scores obtained on the IHLC Scale were above the
value that represents neutral beliefs on this dimension (see Wallston & Wallston, 1981),
while the mean scores on the CHLC and PHLC Scales were substantially below their
neutral points. The configuration of scores on the three MHLC dimensions obtained in
the present investigation was generally similar to the pattern for persons engaged in
preventive health behaviors. (See Wallston & Wallston, 1981.)
Alpha reliabilities showed that the instrument had internal consistency for the study
sample. Item total correlations ranged from .35 to .74 for the IHLC; .26 to .58 for
CHLC; and .21 to .66 for PHLC. Overall, these results showed lower internal consistency than Wallston's (1978) normative data, except for the PHLC subscale. The alpha
values indicate that the responses to the scale were consistent. For a 6-item scale, these
values for each subscale were quite good. However, subscale intercorrelation coefficients,
along with several low item total scale correlations, impose a limitation on the value of
the instrument's internal consistency.
Table 1
Means. Standard Deviations, and Alpha Reliability Coefficients for the
MHLC Scales (N = 146)
Scale

Number of items

SD

Alpha

IHLC

27.82

4.51

.70

PHLC

15.89

5.84

.69

CHLC

15.43

5.35

.56

Note.The results reported are for Form A of the MHLC Scales.

The correlational results showed a relationship between subscales that differed

somewhat from the pattern described by Wallston et al. (1978). The IHLC correlated
- . 0 4 (ns) with PHLC, the IHLC correlated - . 3 9 (p <.0I) with CHLC, and the PHLC
correlated .32 {p <.O1) with CHLC. Both the PHLC and CHLC subscales were related
inversely to the IHLC subscale, but this finding is not surprising because both PHLC and
CHLC subscales reflect a belief in an external orientation.
Table 2 shows that two domains characterized the data set. Factor 1 was loaded
greater than .30 by all six of the Internal subscale items and negatively by two of the six
Chance items (Numbers 9, 11). An additional three items from the Chance subscale
(Numbers 2, 4, 15) also loaded negatively on this factor, but below .30. Factor II was
loaded greater than .30 by five of the six Powerful Others subscale items and one item
from the Chance subscale (Number 16). The remaining item from the Powerful Others
subscale (Number 7) loaded below .30 on this factor. These findings do not support the
three independent subscale dimensionality of the MHLC as proposed by Wallston
(1978).
The results indicate that the MHLC instrument measures not a three dimensional
construct, but a bidimensional one with factors composed predominately by items that
measure Internal and Powerful Others control expectancies, respectively. The absence of
a Chance locus of control dimension may be more a result of the type of population
studied, rather than any invalidity in the theoretical assumptions that underlie the health
locus of" control construct. Previous research had indicated that health locus of control
factor structures vary within and between clinical and non-clinical populations (Wallston

Health Locus of Control

375

Table 2

Varimax-rotated Factor Loadings of MHLC Scales for Cigarette Smokers Seeking Treatment
jN = 1461
MHLC Scale/direction
17.

If I take the right actions, I stay healthy. (I)

12.

The main thing which affects my health is what I myself do. (I)

1.

Ifl take care of myself, I can avoid illness. (I)

6. lam in control of my health. (I)

9. Luck plays a big part in determining how soon I will recover from
an illness. (C)
8. When I get sick I am to blame. (I)
11.

My good health is largely a matter of good fortune. (C)

15.

No matter what I do, I'm likely to get sick. (C)

2.

No matter what I do, if I am going to get sick, I will get sick. (C)

4.

Most things that affect my health happen to me by accident. (C)

5. Whenever I don't feel well, I should consult a medically trained

professional. (P)
14. When I recover from an illness, it's usually because other people
(for example, doctors, nurses, family, friends) have been taking
good care of me. (P)
3.

Having regular contact with my physician is the best way for me to

avoid illness. (P)

10.

Health professionals control my health. (P)

18.

Regarding my health, I can only do what my doctor tells me to do. (P)

16.

If it's meant to be, I will stay healthy. (C)

7.

Factor 2

65
64

.08
.00

50
48
44

-.06
.09
.07

40
38
35
27
20
20

.15
.03
.14
.16
.05
.15

.06

.64

-.15

.62

Ifl get sick, it is my own behavior which determines how soon I get
well again. (I)

13.

Factor 1

My family has a lot to do with my becoming sick or staying healthy. (P)

12
09
25
22
10

.61
.54
.52
.32
.25

Note.I = IHLC, C = CHLC; P = PHLC.

& Wallston, 1981). The present investigation, based on chronic cigarette smokers, may
not represent all smokers and is even less likely to represent the general population. The
subjects were self-selected volunteers, and their relationship to the total population of
cigarette smokers or aspiring quitters was undetermined. There is no guarantee that
similar results would be achieved with other self-selected smokers, non-volunteers, or unaided quitters.
Due to social expectations, subjects may have been inclined to endorse statements
that described their ideal or expected beliefs and thus may have described how they actually should be or would like to be. For example, those who indicated an Internal locus
of control orientation may have been influenced by a common social value placed on
quitting smoking (i.e., one should possess the willpower to control one's own habit).
Those who indicated Powerful Others externality knew they were in a research study and
may have endorsed those types of statements simply to make the group facilitator appear
favorable or to assure they would receive an immediate treatment assignment (subjects
knew there would be a delayed treatment condition within the study). It is also
reasonable to assume that the two dimensions found do, in fact, accurately reflect basic
health orientations for these smokers. However, without controlling for social
desirability, these questions remain unanswered. In addition, the subjects' behavior in
volunteering supports the hypothesis that self-selected smokers are motivated to change

376

Journal of Clinical Psychology. May 1985. VoL 41. No. 3

their health behavior beyond that of chance, fate, or luck. Perhaps a Chance locus of control dimension would be more prominent with nonvolunteers who were also aware of the
availability of treatment. Wallston et al. (1978) have suggested, however, that certain
populations and/or health behaviors may only require assessment on one or more of the
MHLC subscales. In the case of self-selected cigarette smokers who volunteer for cessation treatment, it seems beneficial to assess health locus of control orientations from the
Internal and Powerful Others subscales.
The MHLC instrument holds promise for future use. However, additional research
clearly is warranted to clarify further the nature of the multidimensional health locus of
control construct with cigarette smokers, both treatment volunteers and non-volunteers.
Specifically, it might be found that treatments tailored to the smokers* existing beliefs
about personal control over health would facilitate maintenance of treatment gains and
provide for a more cost-effective approach to intervention.

REFERENCES
& HARRISON, B. E. (1981). Factor structure of the health locus of control scale. Journal of
Clinical Psychology. 37. 217-218.
CRONBACH, L. J. (1970). Essentials of psychological testing. New York: Harper & Row.
HARTKE, R. J., & KuNCE, J. T. (1982). Multidimensionality of health-related locus-of-control scale items.
Journal of Consulting and Clinical Psychology. 50. 594-595.
LEVENSON, H . (1975). Multidimensional locus of control in prison inmates. Journal of Applied Social
Psychology. 5. 342-347.
LEWIS, F., MORISKY, D . , & FLYNN, B. (1978). A test of the construct validity of health locus of control:
Effects on self-reported compliance for hypertensive patients. Health Science Monographs. 6. 138-144.
OXooNEY, B. A., & BARRETT, P. T. (1983). A psychometric investigation of the multidimensional health
locus of control questionnaire. British Journal of Clinical Psychology. 22. 217-218.
ROTTER, J. B. (1954). Social learning and clinical psychology. Englewood Cliffs, NJ: Prentice Hall.
ROTTER, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement.
Psychological Monographs. 80. (1, Whole No. 609).
SHIPLEY, R. H . (1981). Maintenance of smoking cessation: Effect of follow-up letters, smoking motivation,
muscle tension, and health locus of control. Journal of Consulting and Clinical Psychology. 49. 982-984.
STRICKLAND, B. (1978). Internal-external expectancies and health-related behaviors. Journal of Consulting
and Clinical Psychology. 46. 1192-1221.
WALLSTON, K. A., & WALLSTON, B. S. (1981). Health locus of control scales. In H. Lefcourt (Ed.), Research
with the locus of control construct (Vol. 1). New York: Academic Press.
WALLSTON, K. A., WALLSTON, B. S., &. DEVELLIS, R. (1978). Development of the multidimensional health
locus of control (MHLC) scales. Health Education Monographs. 6. 161-170.
WALLSTON, B. S., WALLSTON, K. A., KAPLAN, G. D . , & MAIDES, S. A. (1976). Development and validation
of the health locus of control (HLC) scale. Journal of Consulting and Clinical Psychology. 44. 580-585.
WEST, R. B. (1980). The effects of ethnic background and health locus of control on health maintenance practices related to the risk factors of heart disease. Unpublished Master's thesis, Michigan State University,
1980.
BOYLE, E. S.,