INT’L. J. AGING AND HUMAN DEVELOPMENT, Vol.

67(1) 25-42, 2008

AGE, EDUCATION, AND THE GENDER GAP

IN THE SENSE OF CONTROL

BRITT SLAGSVOLD
Norwegian Social Research – NOVA
ANNEMETTE SØRENSEN
Norwegian Social Research – NOVA
and Stanford University, California

ABSTRACT

High sense of control is related to benefits in many aspects of life, and

education is known to be strongly related to sense of control. In this article we
explore why women tend to feel a lower sense of control than men, and why
the sense of control tends to be lower among the elderly than among younger
people. In particular we explore the role played by education in explaining
age- and gender differences in sense of control. The analysis is based on data
from the first wave of the Norwegian NorLAG study, with a representative
sample of adults aged 40-79 in 30 municipalities. We find that education
accounts for some of the age and gender differences in sense of control, but
the mediating effects of education are rather modest. We find an increasing
gender gap in sense of control with age, and this increasing gap is completely
explained by differences in education. Gender differences in sense of control
is explained completely by four factors, which are related to resources
and power; physical health, education, living with a partner, and leadership
experience. Age differences in sense of control are only partially explained.
Education, physical health and employment status cuts the age effect on
sense of control to half. The effect of education on sense of control is partly
mediated through what we suggest are tangible benefits of education, namely
health, employment, and leadership experience. Education also influences
individuals through socialization mechanisms. We view agentive orientation
as a psychological benefit of education, and measure this characteristic with
Bem’s (1981) sex-role scale on masculinity. Agentive orientation completely
explains the remaining effect of education on sense of control.

25

Ó 2008, Baywood Publishing Co., Inc.

doi: 10.2190/AG.67.1.b
http://baywood.com
26 / SLAGSVOLD AND SØRENSEN

INTRODUCTION

In this article we explore why women tend to feel a lower sense of control than
men, and why the sense of control tends to be lower among the elderly than among
younger people. We are in particular interested in the role played by education in
the development of these differences by age and gender.
Educational opportunities expanded rapidly after the Second World War. Both
men and women took advantage of these new opportunities, but women more
than men, so that the gender gap in educational attainment narrowed rapidly.
Women’s increased education in turn opened opportunities in the labor market
that were closed to older cohorts of women. This means that women have become
less economically dependent on men, have had greater opportunities for being
in positions of some power, and have been in situations where they had to make
decisions on their own.
While education clearly provides people with skills and resources that can be
used in the labor market, education also have psychological benefits by increasing
individual capabilities such as independence, forcefulness, agency, and the ability
to continue to build on the skills gained through the educational process. In other
words, even if a well educated individual does not make use of his or her
educational skills to gain income or influence in the public domain, they are still
likely to have been molded by education, and have gained insights and learned
practices that can be quite useful in terms of navigating daily life.
This dual effect of education, providing tangible opportunities and psycho-
logical benefits, should lead us to expect that well educated individuals are more
likely to feel that they have a great degree of control over their lives, both because
they are more likely to have experienced success in exerting control, and because
they may have more trust in their ability to overcome obstacles, to make realistic
plans, and to believe that personal agency can be effective. A number of studies
have shown that education is positively associated with sense of control or mastery
(Argyle, 1994; Mirowsky & Ross, 2002; Pearlin, Menaghan, Lieberman, &
Mullan, 1981; Ross & Sastry, 1999; Ross & Willigen, 1997; Schieman, 2001).
Research also tends to show that men feel more in control than women do and
that older people score lower on sense of control than younger people (e.g., Ross
& Mirowsky, 2002). More recently, Ross and Mirowsky (2002) have also found
that the gender gap in feelings of control increases with age, and they find that
education is important in accounting for this gender gap.
Apart from Ross and Mirowsky’s innovative and important research
(Mirowsky, 1995; Mirowsky & Ross, 1998, 2003, 2005; Mirowsky, Ross, &
Reynolds, 2000; Ross & Mirowsky, 2002, 2006; Ross & Willigen, 1997), there are
few studies on age- and gender-differentiated effects of education on various
aspects of “psychological capital.” Considering the dramatic increase in educa-
tional level among the future elderly in most populations, especially among
women, possible psychological consequences of education may be considerable,
 SENSE OF CONTROL / 27

and such changes may be relevant for a wide range of social-political challenges in
an aging society. Among the possible psychological consequences of education,
sense of control is especially relevant because control beliefs have impact on
behavior and well-being in most life domains.
In this article we ask whether differences in educational attainment can explain
the gender and age differences in the sense of control for a Norwegian sample of
40 to 79 year olds, and we explore the mechanisms through which education
may bring such differences about. With respect to this last question, we are
especially interested in seeing whether it is the tangible benefits of education
that increase a sense of mastery, or whether more psychological benefits of edu-
cation also play a role.

Sense of Control

The sense of personal control has been defined, described, and measured in
different ways in the literature, but the different terms are highly related and
overlaps to a large extent (Pearlin & Pioli, 2003; Ross & Mirowsky, 2002;
Schieman, 2001; Skinner, 1996). Examples of related concepts are mastery
(Pearlin et al., 1981), personal autonomy (Seeman, 1983), locus of control (Rotter,
1966), self-efficacy (Bandura, 1986), and learned helplessness (Seligman, 1975).
The different concepts have been developed within different theoretical models,
studying a variety of different events and experiences. They therefore vary
somewhat in connotations and focus. When studying the effect of control beliefs
on the same types of individual behavior, the nuances between the concepts does,
however, make little difference. The different terms cover much of the same
phenomenon (Mirowsky & Ross, 2002). The sense of personal control refers to
persons’ sense or belief regarding the extent to which they can control or influence
outcomes (Mirowsky & Ross, 1998). Individuals with a strong sense of control
believe and expect that their actions can affect their world.
The concept of sense of control can be separated in two components: assump-
tions about the controllability of the environment and beliefs about oneself.
Individuals with a high sense of control assume first that it is possible to influence
one’s environment and own life, and second that they have the ability and capacity
to do so. Bandura (1977) calls these two components for outcome expectancy and
self-efficacy expectancy. Others use concepts as contingency and competency
beliefs (Skinner, 1996) or task-beliefs and self-beliefs (Abeles, 1991; Fung,
Abeles, & Carstensen, 1999). We shall here use Bandura’s terms for the two
aspects of sense of control.
The distinction between the two aspects of the control-concept is however
often not clear in empirical studies of sense of control, and most measures of
sense of control merge the two aspects (Ajzen, 2002; Skinner, 1996). For studies
that focus on consequences of sense of control this merge makes little differ-
ence, because for positive benefits of sense of control both outcome beliefs and
28 / SLAGSVOLD AND SØRENSEN

self-efficacy beliefs are necessary, i.e., a belief that the environment is responsive
and that the person is able to use this responsiveness (Skinner, 1996). The two
components of sense of control also influence each other reciprocally through
behavior and attributions of success and failure: Experiences of a responsive
environment will reinforce self-efficacy, and vice versa; self-efficacy beliefs
will influence attempts to control the environment and consequently affect the
chance of experiencing success in controlling the environment. So even though
the two aspects of control beliefs are conceptually and empirically distinct, they
are also highly intercorrelated (Ajzen, 2002; Pearlin & Pioli, 2003). For the
purpose of studying behavior modification or explaining behavior it is often
sufficient to merge the two factors in one measure.
If we want to understand how education, age, and gender affect sense of control,
the distinction between the two may, however, be important because the two
aspects of sense of control may be influenced through different mechanisms
and experiences. Age-related changes, such as losing a spouse or involuntary
retirement, may for example lead to doubt about the possibility of exerting
control over one’s life, i.e., affect outcome beliefs. Similarly, to the extent that
gender differences in sense of control can be attributed to differences in the
experience of power, control, and opportunity, we would see this primarily as a
reflection of an effect on outcome beliefs. Education, however, may be related to
the sense of control both because higher education leads to stronger beliefs in
environmental responsiveness as well as increased self-efficacy beliefs.
Self-efficacy beliefs may more specifically be reflected in self-image and
self-descriptions. Self-efficacy beliefs, as perceived ability to cope with life
challenges, are usually measured by descriptions of how the person approaches
goals or handles problems, or by confidence in personal ability to reach goals and
solve problems (e.g., Bosscher & Smith, 1998; Schwarzer, 1993; Sherer et al.,
1982). As an aspect of self-image, self-efficacy may however be reflected in more
general self-descriptions of agency and instrumentality such as having go-ahead
spirit, and being forceful, assertive, aggressive, self-confident, and determined.
When we explore the relationship between education and sense of control, we will
tentatively see such self-descriptions, which we will call “agentive orientation,” as
having an effect on the sense of self-efficacy component of the control-concept.
The literature generally finds that a high sense of control has clear benefits.
Sense of control is for example positively associated with physical and mental
health (Jang, Kim, & Chiriboga, 2006; Marmot et al., 1998; Rodin, 1986;
Rosenfield, 1999), coronary heart disease (Marmot, Bosma, Hemingway, Brunner
& Stansfeld, 1997), good immunological functioning (Rodin & Timko, 1992),
and longevity (Langer & Rodin, 1976; Seligman, 1975). Control beliefs have
a mediating or moderating effect on a diversity of outcomes; sense of control
mediates the effect of stressful life events on depression (Chou & Chi, 2001),
and high sense of control moderates the negative effects of economic hardships
(Krause, 1987; Pudrovska, Schieman, Pearlin, & Nguyen, 2005), the negative
 SENSE OF CONTROL / 29

effects of low income on health and well-being (Lachman & Weaver, 1998),
and buffers the adverse effects of disability on depression (Jang, Haley, Small,
& Mortimer, 2002). Control beliefs also are associated with good academic
performance, success at work, and good relations to other people (George, 2003).
A high sense of control has broad positive consequences, whether it is possible
to influence outcomes or not (Mirowski & Ross, 2002; Skinner, 1996; Smith
et al., 2000). There are thus good reasons to try to understand how a high sense
of control may come about.

Age, Gender, and Sense of Control

Several studies find a gender gap in sense of control (Pearlin et al., 1981;
Rosenfield, 1999; Ross & Mirowsky, 2002; Rotter, 1966). This is thought to
have its roots in the fact that women are more likely to find themselves in
disadvantaged situations with little power to exert influence over their own lives.
To be in a disadvantaged situation can affect sense of control both directly
and indirectly. Control beliefs are affected directly through experiences of help-
lessness, which again is generalized to other aspects of life. Indirectly, control
beliefs can be affected through accentuation processes (Dannefer, 2003) whereby
initial disadvantage, social or personal, increases the chances of other disadvan-
tages, such as poor health, problems in the labor market, and poor financial
circumstances. In these processes of “cumulative disadvantages” (Dannefer,
1987; O’Rand, 2006), sense of control may be an essential causal component:
a disadvantaged situation, and especially low education, reinforces low control
beliefs, which leads to more helpless behavior and fewer whole-hearted attempts
to change the situation, which again may make the disadvantages increase (Ross
& Mirowsky, 2002; Schieman, 2001). Processes of such mutually reinforcing
consequences which reduce sense of control are assumed to contribute to explain
why women more often than men have poor health and mental health problems
such as depression and anxiety (Mirowsky & Ross, 2005; Rosenfield, 1999;
Ross & Mirowsky, 2002). As gender inequality in education and life chances
decline, we should expect gender differences in sense of control to decline as
well, because women’s and men’s life courses are converging.
Ross and Mirowsky (2002) also found that the gender gap in sense of control
is greater among the old than among the young. If it is gender inequality in
education and employment that is at the root of gender differences in sense of
control, then it seems likely that the increase in the gender gap with age is a
cohort rather than an aging phenomenon. Older women feel less in control of
their lives than men, because their lower education, traditional sex roles, and
limited opportunities prevented them from occupying positions in society that
encouraged the development of a sense of control. Put differently, the increasing
gender gap with age may be accounted for by the effects that education has on
people’s sense of control.
30 / SLAGSVOLD AND SØRENSEN

Previous research also finds a decline in belief of control with age from midlife
to old age (Carmel & Bernstein, 2003; Mirowsky, 1995; Ross & Mirowsky,
2002; Schieman & Campbell, 2001; Smith et al., 2000; Wolinsky, Wyrwich,
Babu, Kroenke, & Tierney, 2003). This decline may be due to aging processes
resulting in an increase in health problems and reduced physical functioning,
or it may reflect cohort differences in education and life conditions (Mirowsky,
1995; Rodin, 1986).

Education as a Mediating Factor

How can education contribute to a sense of control? Education opens wide

opportunities and is a personal resource in most areas of life. Clearly education
improves an individual’s chances of being in positions that provide oppor-
tunities for having influence and exerting control over one’s environment.
Higher education provides better opportunities in labor markets, gives access
to more independent jobs, higher earnings, and higher social status. Being in a
position with real possibilities of exerting control or having influence on the
external world provides the lesson that it is possible to exert control, i.e., it
affects outcome beliefs. The relationship between actual opportunity for control
and control beliefs is well established (e.g., Heckhausen & Schulz, 1995;
Mirowsky & Ross, 1998; 2002; Pearlin et al., 1981), and has been demon-
strated both in experimental studies (Bandura & Wood, 1989; Langer & Rodin,
1976; Reich & Zautra, 1990; Slivinske & Fitch, 1987), and in studies of
related constructs (Kohn & Schooler, 1983). No doubt, one’s sense of control
will also be strengthened by the experience of having actually exerted control
or had influence.
Education can also have an influence on sense of control that is not channeled
through actual experience, but rather through the educational process. Through
education one acquires attitudes and beliefs which increase the chance of
becoming an active and effective agent in one’s own life. You learn to solve
problems, develop skills and ideas, learn how to find and use new information, to
think logically and rationally, and to make and execute plans (Mirowsky & Ross,
1998, 2005; Ross & Sastry, 1999). Education can be viewed as a resource that
is internalized and becomes part of the person (Ross & Mirowsky, 2006), and
can thus alter the very nature of the person (Becker, 1975) by influencing the
way of approaching life, self-image, ambitions, basic assumptions, even values.
Hyman and Wright (1979) summed it up well: “education is an important force . . .
in molding character as well as intellect” (p. 61).
We see then that education can influence both aspects of sense of control,
outcome, and self-efficacy beliefs. In the analyses below, we shall make use of
this insight when we examine the mechanisms through which education affects
sense of control.
 SENSE OF CONTROL / 31

DATA AND MEASURES

The analyses are based on data from the first wave of the Norwegian NorLAG
study (The Norwegian study of Life course, Aging, and Generation). A repre-
sentative random sample stratified by age and sex was selected among adults
aged 40-79 in 30 municipalities in Norway representing different geographic
regions and contexts. The data were collected between 2002 and 2003. Respon-
dents were initially interviewed over the phone; after the interview a mailed
questionnaire with supplementary questions were sent to the study participants.
Data from public registers were added with the respondents’ informed consent.
The response rate of the telephone interview was 67% (n = 5589). The postal
questionnaire was returned by 75% (n = 4169) of these, resulting in a combined
response rate of 50%. In this article, only respondents with non-missing data
on all the variables used in the analysis are included. The size of the analysis
sample is 3861; of these 1838 are men and 2023 women.
The dependent variable in the analyses, sense of control, is measured by The
Personal Mastery Scale (Pearlin & Schooler, 1978). It’s comprised of seven
items that use a 5-point Likert scale ranging from strongly agree to strongly
disagree. This is a short and robust scale focusing on control of aspects of life
that people consider personally important (Pearlin & Pioli, 2003). The original
version uses a 4-point scale. The seven items are:

1. I have little control over the things that happen to me;

2. what happens to me in the future mostly depends on me;
3. there is really no way I can solve some of the problems I have;
4. there is little I can do to change many of the important things in my life;
5. I often feel helpless in dealing with the problems of life;
6. sometimes I feel I’m being pushed around in life; and
7. I can do just about anything I really set my mind to do.

The Personal Mastery Scale is one among many measures of sense of control
which combine the outcome aspect and self-efficacy aspect of sense of control
in the same items. The first item, for example, “I have little control over the
things that happen to me,” can be interpreted as referring to self-efficacy beliefs
“I have little control . . . ,” or to outcome beliefs “. . . things that happen to me.”
The index on sense of control averages the seven response scores. In this sample
it ranges from 9 to 35 (a = .72) with a high score indicating high levels of mastery.
The mean score on mastery is a little lower for women than men with women
scoring on average 25.2 and men 25.8 (see Table 1).
The mean and standard deviation of the variables in the multivariate analyses
are shown in Table 1. Gender is coded 1 for women, 0 for men. The respondent’s
age is measured in years, ranging from 40 to 79. Educational attainment is
measured in five categories corresponding to 8-10 years of schooling, 11-12,
13-14, 14-17, and 18 or more years. The men in the sample are slightly older than
32 / SLAGSVOLD AND SØRENSEN

Table 1. Means and Standard Deviations of Variables

Used in the Analysis; by Sex

Variable Men Women

Sense of control 25.8 (4.6) 25.2 (5.0)

Age 57.8 (10.8) 56.5 (10.6)
Education 2.7 (1.2) 2.6 (1.2)
Health 48.8 (10.1) 46.7 (11.8)
Not employed 0.4 (0.5) 0.4 (0.5)
With partner 0.8 (0.4) 0.7 (0.5)
No leadership 0.5 (0.5) 0.7 (0.4)
Agentive orientation 19.1 (3.4) 18.3 (3.7)

the women and have slightly higher educational attainment. Physical health is
measured by SF-12, which is a 12-item version of Short form health survey (based
on a 36-item version, SF-36) (Gandek et al., 1998). The scale ranges from 10 to
65 in this sample. Women score a little lower than men. Partnership status is
measured by a dummy variable coded 1 if the respondent lives with a partner,
0 otherwise. Physical health and partnership status vary with age, gender, and
education, and both are known to be related to sense of control. In the analyses
below, we regard physical health and partnership status as mediating variables
between age, gender, and education and sense of control.
Employment status is measured by a dummy variable coded 1 if the respondent
is not employed, 0 otherwise. A measure of success in the labor market is captured
by a dummy variable that is coded 1 if the respondent never has been in a
leadership position in the labor market, 0 otherwise. Not surprisingly, men are
much more likely to have been in a leadership position than women. It would
have been desirable to have a better measure on success in the labor market, but
the data do not contain detailed employment histories for the older members of
the sample. Current employment and leadership experience are influenced by
education, and can mediate the effect of education on sense of control.
We do not use income as a variable in the analyses because we only have
information on personal income, and not on household income. Personal income
is a poor indicator of financial circumstances for women.
Six items from Bem’s Sex Role Inventory (Bem, 1981) are used to construct a
scale of what we call agentive orientation, which we assume reflects formative
or molding aspects of education. The items consist of self-descriptions, where
the persons characterize themselves on a 5-point scale as being aggressive,
having leadership abilities, defending own beliefs, being willing to take risks,
 SENSE OF CONTROL / 33

being assertive, and forceful. Bem (1981) originally called these qualities mascu-
line; others treat them, as we do here, in gender neutral terms as “competency”
(Olds & Shaver, 1980) or “instrumentality” (Hermann & Betz, 2004). We prefer
the label “agentive orientation” because the six items all reflect aspects of the
individual’s description of him- or herself as behaving or not behaving as an agent
in one’s own life. The index ranges from 6 to 30, with a high score indicating a
high level of agentive orientation (a = .72). As shown in Table 1, men score higher
on this scale than women do.

RESULTS
In the first part of the analysis, we examine the mechanisms through which
age and gender are related to sense of control. We view this as a partial replication
of previous research done in the United States by Ross and Mirowsky (2002).
It is a partial replication because we do not have measures identical to theirs,
but the hypotheses we test are similar. If our study confirms the findings of
Ross and Mirowsky (2002), it may be said to strengthen the validity of their
findings substantially. In the second part of the analysis we go a step further and
explore the relationship between education and sense of control taking both
tangible and psychological benefits of education into consideration.

Sense of Control

Ross and Mirowsky (2002) test three hypotheses in a national probability

sample of United States households; we test the same hypotheses in a Norwegian
sample, building on the same theoretical assumptions:

1. Men have a higher sense of control than women and sense of control
declines with age. This hypothesis is based on the expectation that
inequality in resources and power between men and women, and between
age cohorts, will affect the sense of control.
2. The gender gap in sense of control is larger at older ages. This hypothesis
is based on the expectation that women’s disadvantage accumulates over
the life course.
3. Education, employment, physical health, and partnership status help
account for the effects of gender and age. This hypothesis is based on the
expectation that important resources which impact the sense of control
vary with age and gender.
The multivariate analyses used to examine these three hypotheses are presented
in Table 2, Model 1 through 9. There is as expected a negative effect of gender as
well as of age on sense of control (Model 1). Women score on average .7 points
lower than men, and there is a decline by about 1.3 points for each 10 years
increase in age. It is also the case that the gender difference increases with age
 Table 2. Regression of Sense of Control on Age, Gender, Education, Physical Functioning, Work, Partner,
Leadership Experience, and Agentive Orientation. Unstandardized Regression Coefficients.

Model Model Model Model Model Model Model Model Model

I II III IV V VI VII VIII IX

Age –.13*** –.11*** –.10*** –.11*** –.09*** –.06*** –.06*** –.06*** –.05***
Female –.72*** 1.00 .65 –.61*** –.33* –.32* –.27* –.12 –.13
Age*Sex –.03* –.02
34 / SLAGSVOLD AND SØRENSEN

Education .56*** .57*** .31*** .27*** .27*** .22*** .04

Physical healtha .14*** .12*** .12*** .12*** .12***
Not employed –1.1*** –1.1*** –.99*** –.82***
Has a partner .47** .46** .52**
Not a leader –.82*** –.33*
Agentive orientationa .42***

Constant 33.68 31.06 29.42 31.31 23.68 24.32 23.79 25.26 15.93
Radj2 .079 .080 .098 .098 .169 .176 .178 .183 .244
aThere is measurement error in the physical health scale (a of .81) and in the measure of agentive orientation (a = .72). This is corrected through
the procedure eivreg in STATA.
*p < .05, **p < .01, ***p < .001.
 SENSE OF CONTROL / 35

(Model 2). Among persons aged 40-49 the difference in sense of control between
men and women is only .4; among those aged 70-79 it is 1.1.
The interaction between gender and age is due to differences in education
(Model 3). Among the older age groups, gender differences in education are
greater than among the younger. If older women had as much education as older
men, the gender difference in sense of control would not increase with age.
Since the interaction effect disappears once education is in the model, we
re-estimate the model with just age, gender, and education as independent vari-
ables (Model 4). In this model it is clear that education has a positive effect on
sense of control. It is also clear that it can account for only a small part of the
effects of age and gender.
What accounts for the remaining effect of gender? An examination of the
results reported in Model 5 through 8 shows first that women’s poorer physical
health is an important contributor to their lower sense of control. If women and
men did not differ on physical health, the difference between them in sense of
control would be cut in half. Not being currently employed has a negative effect
on sense of control, but it is not a mediator of the gender effect, since men’s and
women’s employment rates are similar (Model 6). Living with a partner has a
positive effect on sense of control (Model 7). After controlling for partnership
status, the effect of gender is only marginally significant. The remaining effect of
gender is accounted for by the variable “never in a leadership position” (Model 8).
Women are significantly less likely to have that experience, and this is one
reason they have a lower sense of control.
We have thus been successful in accounting for the effect of gender on sense
of control. In addition to education, there are three important mediators of the
effect of gender on sense of control, namely physical health, living with a partner,
and leadership experience. One could say that if women were in as good physical
health as men, if they lived with a partner as often as men do in the second half
of life, if their leadership experience were on par with men, and if there were
no educational differences, then there would be no gender difference in sense
of control. This is of course to simplify a pattern of complex and reciprocal
relationships, because these mediators do not only influence, but can also be
influenced by control beliefs. In fact, much of the literature on sense of control
has focused on the effect of sense of control on health and health behavior
(Mirowsky & Ross, 2005; Ross & Wu, 1995).
Models 1 through 8 are also useful for shedding light on why sense of control
declines with age, although we do not succeed in accounting for more than
half of the age effect. The key question is whether we should see the age effect as
reflecting an aging or cohort effect. Education is the main mediator of a cohort
effect, since there are clear educational differences by birth year. It is clear
from Model 4 that education only contributes marginally to a reduction in the age
effect. This suggests that the effect of age most likely should be seen as the result of
the aging process.
36 / SLAGSVOLD AND SØRENSEN

As people get older, their health declines, they are less likely to be employed,
and less likely to share their daily life with a partner. These three age-related
factors all are potential mediators of the effect of age on sense of control.
Controlling for health reduces the age effect by about 20% (Model 5). Employ-
ment status leads to a further reduction, while partnership status plays only a
marginal role in explaining the age effect (Model 6 and 7). In light of the
modest role played by education, we conclude that aging rather than cohort
differences most likely are behind the 50% of the age effect that we are able to
account for.
As part of the process of aging we have already taken the most obvious
factors into consideration; physical functioning, employment, and living with
a partner. But aging may imply other changes, both negative and positive, which
can affect sense of control. An increased tendency to have minor depressive
symptoms with age, which we find in this sample (Clausen & Slagsvold, 2005),
may for example affect sense of control negatively. This is indeed the case, but
depressive symptoms do not contribute to a reduction in the age effect (results
not shown). Other reasons for the decline in sense of control with age that we do
not have measures of include increased wisdom and insight regarding human
vulnerability and acceptance of uncertainty and unpredictability in life.
These results confirm the three hypotheses we set out to test, and they mostly
confirm what Ross and Mirowsky (2002) found in their U.S. sample. Age and
gender (being woman) are negatively related to sense of control and the gender gap
increases with age. For the Norwegian sample, this interaction effect was com-
pletely due to the increase with age in the education gender gap. Thus processes of
cumulative disadvantages over the years, as Ross and Mirowsky suggest, do not
contribute to an increased gender gap in sense of control in this Norwegian
sample. In the U.S. sample, the interaction effect was reduced, but remained sig-
nificant after education and variables capturing aspects of the employment experi-
ence, economic circumstances, and partnership status were added to the model.
It is difficult to ascertain why this should be the case, but the interaction effect
in the U.S. data appears to be somewhat larger than is the case in the Norwegian
data, and the U.S. sample includes younger people between age 30 and 39.
We also find similar effects of health, partnership status, and characteristics
of the employment experience as Ross and Mirowsky (2002), and we are able to
account completely for the gender effect by these three factors and to a lesser
degree by education. The model specification of the U.S. study does not permit
a direct comparison, because the age-gender interaction remains in the model,
but once health measures are added, this interaction is only significant at the .10
level, and it is unclear whether a main effect of gender would emerge if the
interaction term was dropped.
Overall, it seems that the pattern of effects in two rather different Western
societies, Norway and the United States, seems quite similar with respect to the
factors that are associated with variations in men’s and women’s sense of control.
 SENSE OF CONTROL / 37

Education and Sense of Control

It is clear from the analyses so far that education has a positive effect on sense
of control, and that it plays a minor role in terms of mediating the effects of age
and gender. In the last part of the analyses, we will focus on the mechanisms
through which education affects sense of control. It is here of special interest
to ascertain whether it is the tangible benefits of education that matter or whether
the more psychological benefits also add to sense of control. In the analyses
presented above, good physical health, employment, living with a partner, and
leadership experience all are related to education. Well educated men and
women report better physical health, they are more likely to be employed, more
likely to live with a partner, and more likely to have had leadership experience
in employment. We suggest that each of these four factors can be viewed as
“tangible” benefits of education. They result, at least partly, from status, skills, and
knowledge gained through education, and these factors generate economic and
social resources and experiences that in turn affect sense of control. The results
presented in Model 4 through 8 suggest that three of these factors (the exception
being partnership status) mediate the effect of education on sense of control.
Controlling for physical health, employment status and leadership experience
reduces the effect of education from .57 to .22, or by two-thirds. Education retains
a significant effect, however. Is it possible that this is due to the more personal
or psychological benefits that education provides people?
We suggested earlier that through education one acquires attitudes and beliefs
that increase the chance of becoming an active and effective agent in one’s own
life. Education may have a socializing effect that alters the person’s basic assump-
tions about oneself and the environment. This formative influence of education
may also alter the ways of approaching situations in life. We suggest that agentive
orientation may be viewed as one such psychological consequence of educa-
tion; being aggressive, having leadership abilities, defending own beliefs, being
willing to take risks, being assertive and forceful.
Can agentive orientation account for the remaining education effect on sense
of control? This is indeed the case. Agentive orientation has a positive effect on
sense of control (Model 9) and the effect of education becomes insignificant.
We find then a rather strong relationship between education and sense of
control. This relationship is partly explained by tangible benefits, which provide
opportunities and experiences of control, which over time are assumed to become
deep-seated assumptions of control in the individual. These findings are well
established in the literature. However, we also find that education has an effect
on sense of control through agentive orientation. We suggest that agentive orienta-
tion is a result of the psychological benefits of education. Through education indi-
viduals are stimulated, encouraged, reinforced, and socialized to behave as persons
who take charge and who trust their abilities and skills. A self-image of agency
is nurtured through education. This in turns results in a belief in self-efficacy.
38 / SLAGSVOLD AND SØRENSEN

DISCUSSION AND CONCLUSIONS

We find that education accounts for some of the age- and gender differences
in sense of control, but the mediating effect is rather modest. In contrast, education
accounts completely for the increasing gender gap in sense of control with age.
Gender differences in sense of control is explained completely by four factors,
which are related to resources; physical health, education, living with a partner,
and leadership experience. Age differences in sense of control are only partially
explained. Education, physical health, and employment status cut in half the
age effect on sense of control. These results are consistent with research done
in the United States by Ross and Mirowsky (2002) and support the validity and
robustness of their findings. The two studies find the same pattern of association
between age, gender, education, and sense of control, even though they are done
in two societies that are very different regarding the degree to which there is
equal opportunity for access to education. The two societies also differ in the
extent to which the welfare state protects individuals and families against the
economic and social consequences of having little education.
We find that the decline in sense of control with age largely should be seen as
an aging effect due to factors such as declining health and retirement from work.
Education accounted for a small part of the age effect, thus suggesting that cohort
effects are not very important. Since we cannot account for all of the age effect,
other unmeasured factors are clearly at work. These include unmeasured cohort
differences in life experience. The older cohorts in this study did for example grow
up when Norway was among the poorest countries in Europe (Organisation for
Economic Cooperation and Development [OECD], 2006), and they experienced
the German occupation of WWII during childhood and youth. In contrast, the
younger cohorts grew up in an increasingly wealthy society where individualistic
ideas gained more prominence. It is also possible that a lower sense of control
develops as one ages because experience shows that life can change in a second
due to forces outside human control.
We suggested that age- and gender-related effects on sense of control largely
stem from their effects on outcome beliefs, while education has an effect both on
outcome beliefs and on self-efficacy beliefs. We are not able to test this directly,
since our measure of sense of control does not distinguish between outcome and
self-efficacy beliefs. Our findings do however indirectly support this idea: The
factors that mediate the effect of age and gender on control are tangible resources,
related to objective possibilities of control, i.e., to outcome beliefs. Agentive
orientation, which we assume reflects self-efficacy beliefs, contributes signifi-
cantly to explain the education-control relationship, but does not explain gender-
control or age-control relationships.
The measure of agentive orientation consists of six items on masculinity from
Bem’s (1981) Sex Role Inventory. Our findings suggest that Bem’s scale on
masculinity is meaningfully interpreted in gender neutral terms. Both men and
 SENSE OF CONTROL / 39

women become increasingly more “masculine” with increasing education, but

women more than men. In fact, the gender gap in agentive orientation disappears
among persons with high education; well-educated women are as agentive
oriented as men at the same high educational level (results not shown). This
pattern, that women benefit more from education than men, is parallel to Ross
and Mirowsky’s (2006) finding in a recent study on depression, where they find
that education has a larger impact on depression for women than for men.
The topic of sense of control and it’s antecedents is complex because many
relevant variables may influence each other reciprocally. We have assumed for
example that education helps develop an agentive orientation. It is however,
also possible that a high sense of control may contribute to the development
and support of an agentive orientation. It is likely that there are feedback loops
between sense of control and agentive orientation. Longitudinal data are needed
to examine such interrelationships over time, while data we have used here are
cross-sectional. Even so, we can safely anticipate that increased education will
have far-reaching consequences.
The future population of the elderly, especially women, will have much more
education than is the case for the elderly today. This will have implications for many
different aspects of the quality of life for the elderly, including health, mortality, and
sense of control. The mechanisms through which education has these effects are not
well understood. Our finding that education affects sense of control both through
tangible effects and through agentive orientation should be seen as a step toward a
better understanding of the role played by education in individual lives.

REFERENCES
Abeles, R. P. (1991). Sense of control, quality of life, and frail older people. In J. E. Birren,
J. C. Rowe, J. E. Lubben, & D. E. Deutchman (Eds.), The concept and measurement of
quality of life in the frail elderly (pp. 297-315). San Diego, CA: Academic Press.
Ajzen, I. (2002). Perceived behavioral control, self-efficacy, locus of control, and the
theory of planned behavior. Journal of Applied Social Psychology, 32(4), 665-683.
Argyle, M. (1994). The psychology of social class. London, UK: Routledge.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.
Psychological Review, 84, 191-215.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory.
Englewood Cliffs, NJ: Prentice-Hall.
Bandura, A., & Wood, R. (1989). Effect of perceived controllability and performance
standards on self-regulation of complex decision-making. Journal of Personality and
Social Psychology, 56, 805-814.
Becker, G. S. (1975). Human capital: A theoretical and empirical analysis, with special
reference to education (2nd ed.). New York: Columbia University Press.
Bem, S. L. (1981). Bem sex-role inventory. Professional manual. Palo Alto, CA:
Consulting Psychologists Press.
Bosscher, R. J., & Smith, J. H. (1998). Confirmatory factor analysis of the general
self-efficacy scale. Behaviour Research and Therapy, 36, 339-343.
40 / SLAGSVOLD AND SØRENSEN

Carmel, S., & Bernstein, J. H. (2003). Gender differences in physical health and psycho-
social well being among four age-groups of elderly people in Israel. International
Journal of Aging & Human Development, 56, 113-131.
Chou, K. L., & Chi, I. (2001). Stressful life events and depressive symptoms: Social support
and sense of control as mediators or moderators? International Journal of Aging &
Human Development, 52, 155-171.
Clausen, S. E., & Slagsvold, B. (2005). Ageing and depression: An epidemiologic study.
Tidsskrift for Norsk Psykologforening, 42, 779-784.
Dannefer, D. (1987). Aging as intra-cohort differentiation: Accentuation, the Matthew
effect, and the life course. Sociological Forum, 2, 211-236.
Dannefer, D. (2003). Cumulative advantage/disadvantage and the life course: Cross-
fertilizing age and social science theory. Journals of Gerontology Series B-Psychological
Sciences and Social Sciences, 58(6), S327-S337.
Fung, H. H., Abeles, R. P., & Carstensen, L. L. (1999). Psychological control in later
life. Implications for life-span development. In J. Brandstädter & R. M. Lerner (Eds.),
Action & self-development. Theory and research through the life span (pp. 345-372).
Thousand Oaks, CA: Sage.
Gandek, B., Ware, J. E., Aaronson, N. K., Apolone, G., Bjorner, J. B., & Brazier, J. E.
(1998). Cross-validation of item selection and scoring for the SF-12 Health Survey
in nine countries: Results from the IQOLA Project. International Quality of Life
Assessment. Journal of Clinical Epidemiology, 51, 1171-1178.
George, L. K. (2003). Commentary: Embedding control beliefs in social and cultural
context. In S. H. Zarit, K. W. Schaie, & L. I. Pearlin (Eds.), Personal control in social
and life course contexts. New York: Springer.
Heckhausen, J., & Schulz, R. (1995). A life-span theory of control. Psychological Review,
102, 284-304.
Hermann, K. S., & Betz, N. E. (2004). Path models of the relationships of instrumentality
and expressiveness to social self-efficacy, shyness, and depressive symptoms. Sex
Roles, 51, 55-66.
Hyman, H. H., & Wright, C. R. (1979). Education’s lasting influence on values. Chicago,
IL: University of Chicago Press.
Jang, Y., Haley, W. E., Small, B. J., & Mortimer, J. A. (2002). The role of mastery and
social resources in the associations between disability and depression in later life.
Gerontologist, 42, 807-813.
Jang, Y., Kim, G., & Chiriboga, D. A. (2006). Correlates of sense of control among older
Korean-American immigrants: Financial status, physical health constraints, and environ-
mental challenges. International Journal of Aging & Human Development, 63, 173-186.
Kohn, M. L., & Schooler, C. (1983). Work and personality: An inquiry into the impact
of social stratification. Norwood, NJ: Ablex.
Krause, N. (1987). Chronic strain, locus of control, and distress in older adults. Psychology
and Aging, 2, 375-382.
Lachman, M. E., & Weaver, S. L. (1998). The sense of control as a moderator of social
class differences in health and well-being. Journal of Personality and Social
Psychology, 74, 763-773.
Langer, E. J., & Rodin, J. (1976). The effects of choice and enhanced personal
responsibility for the aged: A field experiment in an institutional setting. Journal of
Personality and Social Psychology, 34, 191-198.
 SENSE OF CONTROL / 41

Marmot, M. G., Bosma, H., Hemingway, H., Brunner, E., & Stansfeld, S. (1997). Contri-
bution of job control and other risk factors to social variations in coronary heart
disease incidence. Lancet, 350, 235-239.
Marmot, M. G., Fuhrer, R., Ettner, S. L., Marks, N. F., Bumpass, L. L., & Ryff, C. D.
(1998). Contribution of psychosocial factors to socioeconomic differences in health.
Milbank Quarterly, 76, 403-448.
Mirowsky, J. (1995). Age and the sense of control. Social Psychology Quarterly, 58, 31-43.
Mirowsky, J., & Ross, C. E. (1998). Education, personal control, lifestyle and health—
A human capital hypothesis. Research on Aging, 20, 415-449.
Mirowsky, J., & Ross, C. E. (2002). Social causes of psychological distress (2nd ed.).
New York: Aldine de Gruyter.
Mirowsky, J., & Ross, C. E. (2003). Education, social status and health. New York:
Aldine de Gruyter.
Mirowsky, J., & Ross, C. E. (2005). Education, learned effectiveness and health. London
Review of Education 3, 205-220.
Mirowsky, J., Ross, C. E., & Reynolds, J. (2000). Links between social status and health
status. In E. B. Chloe, P. Conrad, & A. M. Fremont (Eds.), Handbook of medical
sociology (5th ed., pp. 47-67). New York: Kluwer.
Olds, D. E., & Shaver, P. (1980). Masculinity, femininity, academic performance, and
health: Further evidence concerning the androgyny controversy. Journal of Per-
sonality, 48, 323-341.
O’Rand, A. (2006). Stratification and the life course. Life course capital, life course risks,
and social inequality. In R. H. Binstock & L. K. George (Eds.), Handbook of aging and
the social sciences (6th ed., pp. 145-162). Amsterdam: Elsevier.
Organisations for Economic Cooperation and Development [OECD]. (2006). The world
economy (Vol. 2, Historical statistics). Paris: OECD.
Pearlin, L., Menaghan, E., Lieberman, M., & Mullan, J. (1981). The stress process.
Journal of Health and Social Behavior, 22, 337-356.
Pearlin, L. I., & Pioli, M. F. (2003). Personal control: Some conceptual turf and future
directions. In S. H. Zarit, L. I. Pearlin, & K. W. Schaie (Eds.), Personal control in
social and life course contexts. New York: Springer.
Pearlin, L. J., & Schooler, C. (1978). The structure of coping. Journal of Health and
Social Behavior, 19, 2-21.
Pudrovska, T., Schieman, S., Pearlin, L. I., & Nguyen, K. (2005). The sense of mastery
as a mediator and moderator in the association between economic hardship and
health in late life. Journal of Aging and Health, 17, 634-660.
Reich, J. W., & Zautra, A. J. (1990). Dispositional control beliefs and the consequences
of a control-enhancing intervention. Journal of Gerontology, 45, P46-P51.
Rodin, J. (1986). Aging and health: Effects of the sense of control. Science, 233,
1271-1276.
Rodin, J., & Timko, C. (1992). Sense of control, aging, and health. In M. G. Ory, R. P.
Abeles, & P. D. Lipman (Eds.), Aging, health, and behavior. Newbury Park, CA: Sage.
Rosenfield, S. (1999). Splitting the difference. Gender, the self, and mental health. In
C. S. Aneshensel & J. C. Phelan (Eds.), Handbook of the sociology of mental health
(pp. 209-224). New York: Kluwer Academic/Plenum Publishers.
Ross, C. E., & Mirowsky, J. (2002). Age and the gender gap in the sense of personal
control. Social Psychology Quarterly, 65, 125-145.
42 / SLAGSVOLD AND SØRENSEN

Ross, C. E., & Mirowsky, J. (2006). Sex differences in the effect of education on
depression: Resource multiplication or resource substitution? Social Science &
Medicine, 63, 1400-1413.
Ross, C. E., & Sastry, J. (1999). The sense of personal control. Social structural causes and
emotional consequences. In C. S. Aneshensel & J. C. Phelan (Eds.), Handbook of the
sociology of mental health (pp. 369-394). New York: Kluwer Academic/Plenum Publishers.
Ross, C. E., & Willigen, M. V. (1997). Education and the subjective quality of life.
Journal of Health and Social Behavior, 38, 275-297.
Ross, C. E., & Wu, C. L. (1995). The links between education and health. American
Sociological Review, 60, 719-745.
Rotter, J. B. (1966). Generalized expectancies for internal vs. external control of reinforce-
ment. Psychological Monographs, 80, 1-21.
Schieman, S. (2001). Age, education, and the sense of control: A test of the cumulative
advantage hypothesis. Research on Aging, 23, 153-178.
Schieman, S., & Campbell, J. E. (2001). Age variations in personal agency and self-
esteem: The context of physical disability. Journal of Aging and Health, 13, 155-185.
Schwarzer, R. (1993). Measurement of perceived self-efficacy: Psychometric scales for
cross cultural research. Berlin: Freie Universität Berlin, Institut für Psychologie.
Seeman, M. (1983). Alienation motifs in contemporary theorizing: The hidden continuity
of classic themes. Social Psychology Quarterly, 46, 171-184.
Seligman, M. E. P. (1975). Helplessness: On depression, development, and death. San
Francisco, CA: Freeman.
Sherer, M., Maddux, J. E., Mercandante, B., Prentice-Duun, S., Jacobs, B., & Rogers,
R. W. (1982). The self-efficacy scale: Construction and validation. Psychological
Report, 51, 663-671.
Skinner, E. A. (1996). A guide to constructs of control. Journal of Personality and Social
Psychology, 71, 549-570.
Slivinske, L. R., & Fitch, V. L. (1987). The effect of control enhancing interventions on
the well-being of elderly individuals living in retirement communities. Gerontologist,
27, 176-181.
Smith, G. C., Kohn, S. J., Savage-Stevens, S. E., Finch, J. J., Ingate, R., & Lim, Y. O.
(2000). The effects of interpersonal and personal agency on perceived control and
psychological well-being in adulthood. Gerontologist, 40, 458-468.
Wolinsky, F. D., Wyrwich, K. W., Babu, A. N., Kroenke, K., & Tierney, W. M. (2003).
Age, aging, and the sense of control among older adults: A longitudinal reconsider-
ation. Journals of Gerontology Series B-Psychological Sciences and Social Sciences,
58, S212-S220.

Direct reprint requests to:

Britt Slagsvold
Norwegian Social Research – NOVA
P.O. Box 3223 Elisenberg
0228 Oslo, Norway
e-mail: bsl@nova.no