the association between physical self-discrepancies and ... · behavior (brunet & sabiston,...

102

Journal of Sport & Exercise Psychology, 2012, 34, 102-123© 2012 Human Kinetics, Inc.

Jennifer Brunet, Catherine Sabiston, Andree Castonguay, and Natalia Bessette are with the Department of Kinesiology and Physical Education, McGill University, Montreal, QC, Canada. Leah Ferguson is with the College of Kinesiology, University of Saskatchewan, Saskatoon, SK, Canada.

The Association Between Physical Self-Discrepancies and Women’s Physical Activity: The Mediating Role of Motivation

Jennifer Brunet,1 Catherine Sabiston,1 Andree Castonguay,1 Leah Ferguson,2 and Natalia Bessette1

1McGill University; 2University of Saskatchewan

The objectives of this study were to test the associations between physical self-discrepancies (actual:ideal and actual:ought) and physical activity behavior, and to examine whether motivational regulations mediate these associations using self-discrepancy (Higgins, 1987) and organismic integration (Deci & Ryan, 1985) theories as guiding frameworks. Young women (N = 205; Mage = 18.87 years, SD = 1.83) completed self-report questionnaires. Main analyses involved path analysis using a polynomial regression approach, estimation of direct and indirect effects, and evaluation of response surface values. Agreement between actual and ideal (or ought) physical self-perceptions was related to physical activity both directly and indirectly as mediated by the motivational regulations (R2 = .24–.30). Specifically, when actual and ideal self-perceptions scores were similar, physical activity levels increased as actual and ideal scores increased. Furthermore, physical activity levels were lower when the discrepancy was such that ideal or ought self were higher than actual self. These findings provide support for integrating self-discrepancy and organismic integration theories to advance research in this area.

Keywords: self-discrepancy theory, self-determination theory, physical self, mediation, response surface

Young women in Western cultures often feel pressured to conform to standards of beauty promoting a slender, fit and well-toned body as the ideal physique (Cash & Pruzinsky, 2004; Pruis & Janowsky, 2010). Consequently, many women report wanting to have a more attractive, leaner, fitter and athletic-looking body than they currently have and this creates an ideal self-perception (Cash & Pruzinsky, 2004). Comparisons between one’s current/actual and ideal physical self result in self-discrepancies (Higgins, 1987) that promote criticism of oneself rather than criticism of the ideal (Thompson & Stice, 2001). These self-criticisms may promote motiva-tion to change oneself to reduce the discrepancy between one’s current physical

Self-Discrepancies and Physical Activity 103

state and ideal physical state, and may give impetus to the use of compensatory behaviors such as physical activity. Self-discrepancies and associated self-criticism may also undermine women’s investment in and enjoyment of physical activity behavior (Brunet & Sabiston, 2009), and reduce the likelihood of participation in activity. Reported statistics show less than 50% of women in Western cultures currently participate in physical activity (Colley et al., 2011). Self-discrepancy perceptions, in particular those focused on the body and physical self, may partially explain these low levels. The purpose of this study was to examine the association between women’s physical self-discrepancies and physical activity behavior while drawing on self-discrepancy (Higgins, 1987) and organismic integration (Deci & Ryan, 1985, 2004) theories. The integration of these two theoretical frameworks helps to recognize underlying mechanisms linking physical self-discrepancies and physical activity behavior (Mathieu & Taylor, 2006), and to identify modifi-able factors (e.g., physical self-discrepancy perceptions) that could be targeted in interventions aimed at increasing physical activity among women (MacKinnon, Fairchild, & Fritz, 2007).

Self-Discrepancy Theory and the Physical Self

Higgins (1987) contends that people hold three types of beliefs about themselves: (i) who they actually are (i.e., actual self), (ii) who they would like to be (i.e., ideal self), and (iii) who they should be (i.e., ought self). The ideal and ought selves correspond to self-evaluative standards or self-guides whereas the actual self is a current self-evaluation. An actual:ideal discrepancy occurs when individuals perceive that their current state is discrepant from their ideal state (e.g., “I am fat but I would ideally like to be thin”). An actual:ought discrepancy occurs when individuals perceive that their current state is discrepant from the state they feel they should be (e.g., “I am fat but I should be thin”). These self-discrepancies result in specific emotional and motivational states that, in turn, may trigger self-regulatory behaviors such as physical activity aimed at decreasing the discrepancy between the self-state and self-guides (Higgins, 1987).

Research on self-discrepancy theory has been limited to global self-assessments rather than more specific discrepancies pertaining to individuals’ perceptions about themselves in specific self domains. Researchers using this global approach have provided evidence to support self-discrepancy theory tenets, but fail to acknowledge that different domains of the self vary on their importance to an individual’s self-concept and related emotions and behaviors. The physical domain is arguably the strongest predictor of the self-system (Fox & Corbin, 1989) and of health behaviors such as physical activity (Fox & Corbin, 1989; Hayes, Crocker, & Kowalski, 1999; Sonstroem, Speliotis, & Fava, 1992). Fox and Corbin (1989) propose that the physi-cal self-concept is defined by four self-perceptions: physical conditioning, sport/physical activity competence, body attractiveness, and strength. Consistent with this multidimensional view of physical self-concept, Marsh and Redmayne (1994) have identified nine self-perceptions (i.e., health, strength, body fat, physical activ-ity, endurance/fitness, sport competence, coordination, appearance, and flexibility) that relate to the physical self. Taken together, assessing physical self-discrepancies related to perceptions of attractiveness, weight/body shape, physical competence, physical strength and fitness/conditioning may provide more knowledge on the

104 Brunet et al.

specific links between the self-system and health-related affective, cognitive, and behavioral outcomes.

Physical Self-Discrepancies and Physical ActivitySelf-discrepancies related to weight and body shape have been studied over the last several decades and used predominantly as indicators of body image. Researchers have often used measures of body figures that incrementally increase in size of fat (e.g., Stunkard, Sorenson, & Schlusinger, 1983) and/or muscle (e.g., Lynch & Zellner, 1999) to garner assessment of discrepancies between one’s actual and ideal figure. Generally, greater weight and body shape discrepancies have been linked to cognitive and affective outcomes such as greater body dissatisfaction and body-related shame (Bessenoff & Snow, 2006; Sabiston, Crocker, & Munroe-Chandler, 2005) and behav-ioral outcomes such as higher eating pathology rates (Anton, Perri, & Riley, 2000; Cafri, van den Berg, & Brannick, 2010; Thompson & Altabe, 1991) and lower levels of physical activity (Anton et al., 2000; Markland, 2009). While informative, this line of inquiry limits physical self-discrepancy perspectives to weight and body shape with no regard to other potentially important physical self-perceptions such as strength, fitness, and attractiveness. Attempts to integrate other physical self-discrepancies are needed to offer a more comprehensive understanding of the relationships between physical self-discrepancies and physical activity behavior. There are also no known accounts of ought self-perspectives in this area of research. Examining both ideal and ought self-discrepancies is important to provide support for the distinctiveness of these constructs and their relationship to physical activity behavior.

Furthermore, the way these limited physical self-discrepancies (i.e., body weight and shape) have been assessed may be wrought with methodological prob-lems based on the reliance on single-item assessments and the use of difference scores between actual and ideal self-perceptions (Cafri et al., 2010). Single-item measures are prone to measurement error, and difference scores are less reliable than the component measures (i.e., actual, ideal, and ought assessments alone) and are ambiguous since they require combining measures of conceptually dis-tinct constructs into a single score (Edwards, 2002). As a way of addressing these limitations, actual and ideal (and ought) self-perceptions should be conceptualized and tested as independent cognitive appraisals that are distinctly associated with various health outcomes, including physical activity, rather than merged within difference scores (Cafri et al., 2010; Edwards & Parry, 1993).

A final limitation of previous research is the possible contradiction between theoretical tenets and empirical evidence. Based on self-discrepancy theory (Hig-gins, 1987), women who perceive actual:ideal and/or actual:ought self-discrepancies may engage in physical activity, a behavior which can improve their physical appearance by increasing muscle tone and reducing excess weight and fat (War-burton, Nicol, & Bredin, 2006). However, researchers have found that greater actual:ideal physical self-discrepancies (focused only on weight and body shape) were associated with lower levels of physical activity behavior in women (Anton et al., 2000; Markland, 2009). It is therefore important to identify mechanisms that may help explain the relationship between self-discrepancies and physical activity. Following Markland’s (2009) study, motivational regulations are proposed to be mechanisms explaining the link between women’s physical self-discrepancies and physical activity behavior.


Organismic Integration Theory and Self-Discrepancies

Researchers examining the association between physical self-discrepancies and health behaviors have tested a direct association between self-discrepancies and behavioral outcomes. However, motivational theories such as self-determination (Deci & Ryan, 1985, 2004) posit that the influence of social and individual fac-tors on behavior is mediated by motivational regulations. According to Deci and Ryan (1985), motivation is a multidimensional construct that consists of several qualitatively distinct motivational regulations that lie on a continuum of self-determination. Self-determined regulations (i.e., intrinsic, integrated, and identified) originate when an individual’s perceived locus of causality is internal and behavioral engagement is done with a sense of enjoyment, volition and choice. In contrast, non-self-determined regulations (i.e., external and introjected) originate when an individual’s perceived locus of causality is external and the behavior is undertaken because she or he feels pressured or compelled to do so, either by herself/himself or by others (Deci & Ryan, 2004). Amotivation, described as lacking the intention to engage in a behavior, lies at one end of the continuum and intrinsic motivation at the other (Deci & Ryan, 2004).

Controlling or pressuring contexts foster non-self-determined motivational regulations and impede self-determined motivational regulations. Given the per-ceived pressure placed on women to obtain and/or maintain an ideal physique, the perception that one falls short of one’s internalized ideal or society’s ideal (i.e., actual:ideal and actual:ought self-discrepancies, respectively) may represent a form of pressure that motivates women to engage in physical activity to reduce any discrepancy between their current state (i.e., actual self) and self-guide states (i.e., ideal and ought selves). In this way, the pressure to attain the ideal and ought selves may hinder women’s behavior by reducing self-determined motivational regulations and increasing non-self-determined motivational regulations. Similarly, women who have high ideal and/or ought self-perceptions coupled with low actual self-perceptions may report low self-determined motivation for physical activity and high non-self-determined motivation. However, it is possible that the two types of self-discrepancies (i.e., actual:ideal and actual:ought) are differentially linked to these motivational regulations. Conceptually, the ideal self represents a woman’s internally referenced conception of the self-guide and reflects her wants and desires. The ought self represents a normatively referenced conception of the self-guide and is based upon what the woman believes society’s standards dictate. Whereas both types of self-discrepancies (i.e., actual:ideal, actual:ought) may be considered controlling factors, it could be argued that the locus of causality is internal for ideal-based self-perceptions and discrepancies, and external for ought-based self-perceptions and discrepancies. Accordingly, the former may have stronger links to introjected regulation where the activity is done to avoid guilt and shame or to attain ego enhancements and feelings of worth. In contrast, ought-based self-perceptions and discrepancies may demonstrate stronger links to external regulation where the activity is done to satisfy an external demand or a socially constructed contingency. Both ideal and ought self-perceptions would be hypothesized to diminish intrinsic motivation, where the activity is done for the inherent feeling of enjoyment and satisfaction since participation in physical activity likely represents a means to an end (i.e., reducing perceived self-discrepancies). Researchers have supported these possible associations indirectly in reporting moderate positive correlations

106 Brunet et al.

between negative body-related affective constructs (e.g., social physique anxiety, body-related shame and guilt) and external/introjected regulations, and negative relationships to identified and intrinsic regulations (Brunet & Sabiston, 2009; Sabiston et al., 2010). There is also evidence of a relationship between body-size self-discrepancies and weight management motives (Ingledew & Sullivan, 2002; Markland & Ingledew, 2007; Sabiston & Chandler, 2009; Sabiston et al., 2005). However, direct testing of the associations between actual:ideal and actual:ought physical self-discrepancies and physical activity behavior, within the combined tenets of self-discrepancy and organismic integration theories, is limited.

Research Linking Self-Discrepancy and Organismic Integration Theories

There is one known study that rests on the combination of self-discrepancy and organismic integration theories. Markland (2009) examined whether the motiva-tional regulations (assessed using the Behavioral Regulation in Exercise Question-naire-2 [BREQ-2]; Markland & Tobin, 2004) mediated the association between body size self-discrepancies (assessed using the Figure Rating Scale; Stunkard et al., 1983) and physical activity behavior (assessed using the Leisure Time in Exer-cise Questionnaire [LTEQ]; Godin & Shephard, 1985) among 102 female hospital workers and church members aged 18–55 years. Bivariate correlations indicated that body-size actual:ideal self-discrepancy was negatively related to intrinsic and identified regulations, and was positively related to external regulation. Based on multiple mediation analysis, self-determined motivation regulations mediated the relationship between self-discrepancy and physical activity behavior (R2 = .33). These findings suggest motivational regulations for physical activity mediate the relationship between actual:ideal self-discrepancy and physical activity behavior. Yet, the relationship between actual:ought self-discrepancy and physical activity behavior was not assessed. One aim of the current study was to investigate the relationships between both actual:ideal and actual:ought physical self-discrepancies and physical activity behavior as mediated by the motivational regulations.

Furthermore, Markland (2009) only examined body size self-discrepancy (a possible proxy for weight or thinness self-perceptions) in spite of the other prominent physical self-perceptions (e.g., attractiveness, physical competence, physical strength, fitness/conditioning) that may predict cognitive and behavioral outcomes. The current research study attempts to fill this gap by assessing physical self-discrepancies on five prominent subdomains of the physical self to develop actual, ideal, and ought physical self-perception factors. The assessment of five physical self-perceptions also helps avoid problems of single-item discrepancy measures found in Markland’s (2009) study and common in body image research (see Cafri et al., 2010). Finally, Markland (2009) used raw discrepancy scores that were obtained by subtracting ideal ratings from actual ratings. Researchers have criticized the use of discrepancy or difference scores stating that they are less reliable and confound the effects of their component measures, which may lead to ambiguous and biased results (see Edwards, 1994, 2002 for discussion). The current study addresses the limitation of biased results by adopting and testing analytical procedures that are proposed to be appropriate for the assessment of self-discrepancies (Edwards, 2002; Edwards & Parry, 1993).


Purpose and HypothesesDrawing from self-discrepancy (Higgins, 1987) and organismic integration (Deci & Ryan, 1985) theories, the purpose of the current investigation was to test the associations between actual, ideal, and ought physical self-perceptions and dis-crepancies and physical activity behavior, as mediated by the motivational regula-tions (see Figure 1). Based on theoretical perspectives (Deci & Ryan, 2004) and empirical findings (Anton et al., 2000; Cafri et al., 2010; Markland, 2009), it was hypothesized that (i) agreement between actual and ideal (or ought) self-perceptions would have positive associations with self-determined regulations and physical activity behavior, and negative ones with non self-determined regulations, and these associations could be both linear and nonlinear; (ii) discrepancies between actual and ideal (or ought) self-perceptions would have negative associations with self-determined regulations and physical activity behavior, and positive associations with non-self-determined regulations, such that self-determined regulations and physical activity levels would be lowest, and non self-determined regulations would be highest, when the discrepancy would be such that ideal or ought self-perceptions were higher than actual self-perceptions; and (iii) the associations between actual and ideal (or ought) self-perceptions agreement and discrepancies and physical activity would be mediated by the motivational regulations.

Methods

Procedures and Participants

Following approval from the University Research Ethics Board, school directors and instructors from Colleges of General and Vocational Education were approached for their support. These colleges are public postsecondary educational institutions in the province of Quebec, Canada, which offer 2-year pre-university and 3-year technical programs. All programs share a common core curriculum that includes French, English, humanities, and physical education courses. Female students from these core classes were briefed on the study during class by a trained research assis-tant and provided with a letter of information for their parents and consent forms. Approximately 1 week later, young women who were interested in participating were asked by the research assistant to sign an informed consent form and complete a self-administered questionnaire package during regular class time, which included measures assessing physical self-discrepancies, motivational regulations, physical activity behavior, and demographics. Young men, and female students who were not interested in participating in this study, were asked to work individually during this time (approximately 20 min). In total, 205 young adult women ranging in age from 16 to 28 years (Mage = 18.87, SD = 1.83; 84.4% < 20 years) and who self-described as predominantly Caucasian (81.4%) took part in this study. Parental consent was obtained for students younger than 18 years of age using a passive consent method approved by the Institutional Review Board because the level of risk involved in this study was minimal. This passive consent approach required parents who did not want their daughter to participate in the research to return the consent form stating a decline. Passive consent procedures result in higher participation rates and a more representative sample compared with active consent procedures (Dent et al., 1993).

108

Fig

ure

1 —

Hyp

othe

size

d m

odel

with

mot

ivat

iona

l reg

ulat

ions

as

med

iato

rs o

f the

rela

tions

hip

betw

een

phys

ical

sel

f-di

scre

panc

ies

and

phys

ical

act

ivity

beh

avio

r.


Measures

Physical Self-Discrepancies. A physical self-discrepancy questionnaire was developed for this study to assess actual, ideal, and ought physical self-perceptions and discrepancies (see Appendix). Each participant rated her actual (physical attributes she currently has), ideal (physical attributes she would ideally like to have), and ought (physical attributes she believes she should have) self-perceptions in relation to strength, attractiveness, thinness, physical ability, and fitness which is consistent with Fox and Corbin’s (1989) subdomains of the physical self and Markland’s (2009) assessment of body-size discrepancy. For each characteristic, women rated the degree to which they believed the attribute described them now, reflected how they would like to be, and reflected how they thought they should be using a 7-point Likert scale anchored by not at all and very much. The measure of self-discrepancies is in line with the conditions set forth by Edwards (2002) such that actual, ideal, and ought ratings are commensurate and the same Likert-type scale is used to assess each component. This resulted in fifteen items assessing actual (five items), ideal (five items), and ought (five items) self-perception latent factors. The scores were reliable based on composite reliability coefficients of .78 (actual), .66 (ideal), and .70 (ought) which were obtained within the framework of confirmatory factor analysis (CFA; Yang & Green, 2010; see Table 1).

Before this study, this measure was validated in a separate sample of young females (N = 184; Mage = 18.46 years, SD = 2.44, 83.1% Caucasian) recruited the same way as the participants in the current study. The scores for actual and ideal self were highly correlated with the respective subdomains of the physical self as assessed using the Physical Self-Description Questionnaire (PSDQ; Marsh, Richards, Johnson, Rocher, & Tremayne, 1994), and with actual reports on the Figure Rating Scale (FRS; Stunkard et al., 1983). In addition, the ideal and ought self-perceptions showed nonsignificant or weak links to PSDQ subscale scores. Furthermore, correlations between the actual and ideal self-perceptions and the FRS discrepancy (a commonly used method of using this scale, in spite of the weaknesses with discrepancy scores) were significant and in the expected direc-tion (r = –.33 and .28, respectively). Finally, similar to previous research (Scalas & Marsh, 2008), actual scores were significantly related to self-esteem assessed on the PSDQ (r = .38), whereas ideal (and ought) self-perceptions were negatively and nonsignificantly related to self-esteem. Details on these analyses can be obtained from the second author.

Motivational Regulations. The BREQ-2 (Markland & Tobin, 2004) was used to assess intrinsic (four items; e.g., “I exercise because it’s fun”), identified (four items; e.g., “I value the benefits of exercise”), introjected (three items; e.g., “I feel guilty when I don’t exercise”), external (four items; e.g., “I exercise because other people say I should”) regulations, and amotivation (four items; e.g., “I don’t see the point in exercising”).1 Responses to each statement were reported on a 5-point scale anchored at the extremes by not true for me to very true for me. Researchers have shown that the BREQ-2 is a reliable (α ≥ .75) measure that assesses the different forms of motivational regulations (Markland & Tobin, 2004; Wilson & Rodgers, 2004) and valid since it differentiates between physically active and inactive women (Landry & Solomon, 2004). The composite reliability coefficients for the subscale scores obtained from the CFA in the current study were above .70 (see Table 1).

110

Tab

le 1

S

core

Ran

ges

, Mea

ns

and

Sta

nd

ard

Dev

iati

on

s (S

D)

for

the

Raw

Sco

res,

an

d R

elia

bili

ty a

nd

C

orr

elat

ion

Co

effic

ien

ts B

etw

een

Phy

sica

l Sel

f-D

iscr

epan

cies

, Mo

tiva

tio

nal

Reg

ula

tio

ns,

an

d P

hysi

cal A

ctiv

ity

Fact

or

Sco

res

Varia

bles

Sco

re

Ran

ges

Mea

n (S

D)

12

34

56

78

1A

ctua

l1–

74.

48 (

.97)

.78

2Id

eal

1–7

6.02

(.6

2).4

1**

.66

3O

ught

1–7

5.71

(.6

9).3

6**

.64*

*.7

0

4E

xter

nal

1–4

1.45

(.5

5)–.

22**

–.15

*–.

19**

.70

5In

troj

ecte

d1–

42.

30 (

.86)

.14*

.15*

.11

.13

.76

6Id

entifi

ed1–

42.

86 (

.76)

.43*

*.2

6**

.17*

–.10

.55*

*.7

6

7In

trin

sic

1–4

2.91

(.8

8).5

3**

.20*

*.1

8**

–.21

**.2

9**

.71*

*.9

2

8LT

EQ

10–

111

44.1

1 (2

2.67

).4

3**

.15*

.11

–.10

.30*

*.4

5**

.46*

*—

Not

e. L

TE

Q1

is th

e ph

ysic

al a

ctiv

ity m

easu

re in

ME

Ts.

Com

posi

te r

elia

bilit

y co

effic

ient

s ar

e on

the

diag

onal

.

*p <

.05.

**p

< .0

01.


Physical Activity. The LTEQ (Godin & Shephard, 1985) was used to assess physical activity behavior. Participants were asked to indicate the number of times they engaged in light, moderate, and vigorous physical activity for more than 15 min during a 7-day period. Scores were converted into metabolic equivalents (METs) by multiplying the weekly frequencies of light, moderate, and vigorous activities by 3, 5, and 9, respectively. METs scores for the different physical activity intensities were summed to create a total physical activity score. The second item on the LTEQ is a 3-point Likert scale frequency score (ranging from often to never) of moderate-to-vigorous activity (i.e., fast heartbeat and sweating) during a typical 7-day period. This item was not used in the current study because it does not assess light intensity physical activity. Researchers have shown that the LTEQ is a reliable (i.e., test–retest) and valid measure of physical activity as evidenced by moderate positive correlations with both objective and subjective measures of physical activity (Jacobs, Ainsworth, Hartman, & Leon, 1993; Miller, Freedson, & Kline, 1994).

Demographics. Self-reported age, household income, ethnicity, height, and weight (used to calculate body mass index) were collected for descriptive purposes.

Data Analysis

Data analysis involved several sequential steps that were performed in PASW 18.0 and LISREL/PRELIS 8.80. First, data were screened for patterns of missing data, outliers, and violations of the assumption of normality (Pedhazur, 1997). Second, descriptive statistics were computed for main study variables. Third, independent maximum likelihood CFA were run to examine the composite reliability coef-ficients of the scores (actual, ideal, and ought physical self-perceptions, and each motivational regulation; Yang & Green, 2010). In this step, estimated factor scores were saved and used in the main analyses (i.e., path analysis) to reduce model complexity (Brown, 2006). Fourth, intercorrelations were examined between these factor scores and the measured physical activity scores.

For the main analyses, path analysis using maximum likelihood estimation was used to investigate the associations between the physical self-perceptions and discrepancies, motivational regulations and physical activity behavior in a direct model (i.e., all factors were estimated to relate to physical activity behavior). Separate models were tested to examine ideal and ought self-perceptions. Physi-cal self-discrepancies were modeled as polynomial regression terms. Specifically, examination of discrepancies included an estimation of the component parts (x1 and x2, where x1 is actual self and x2 is either ideal or ought self), the square of these variables (x1 × x1 and x2 × x2), and the cross-product of these variables (x1 × x2) to assess the linear, nonlinear, and joint relationships between self-perceptions with physical activity, respectively. This approach provides a superior test of conceptual models than approaches that use difference scores between actual and ideal (or ought) measures when the discrepancy between two variables is a central consid-eration (Edwards, 2002). Most notably, it avoids problems with difference scores (i.e., effects of each of the component on the outcome is confounded) because the independent effect of each component is retained. In addition, when used with response surface methodology, it has more explanatory potential than do differ-ence scores since it allowed for the examination of (1) how agreement between the physical self-perceptions related to physical activity, (2) how the degree of

112 Brunet et al.

discrepancy between the physical self-perceptions related to physical activity, and (3) how the direction of the discrepancy between the physical self-perceptions related to physical activity.

The fit of the path model was evaluated using the comparative fit index (CFI), standardized root mean squared residual (SRMR), and the root mean square error of approximation (RMSEA). As recommended by Hu and Bentler (1999), RMSEA values ≤ .06, SRMR values ≤ .08, and CFI values ≥ .95 were used to indicate well-fitting models. However, rather than examining the direct and indirect effect coef-ficients reported in the corresponding matrices in the LISREL output as would be done in a common path analysis, the coefficients in the model were transformed into four surface test values: a1 to a4 (Kazén & Kuhl, 2011). The values of a1 (βx1 + βx2) reflect the linear relationship between the agreement in actual and ideal (or ought) self scores and physical activity motivational regulations and behavior, whereas a2 (βx3 + βx4 + βx5) values reflect the nonlinear relationship between the agreement in actual and ideal (or ought) self scores and physical activity motivational regulations and behavior. The values of a3 (βx1 – βx2) reflect how the direction of the discrepancy between actual and ideal (or ought) self scores is related to physical activity moti-vational regulations and behavior. The values of a4 (βx3 – βx4 + βx5) reflect how the degree of discrepancy between actual and ideal (or ought) self scores relate to physical activity motivational regulations and behavior (Shanock, Baran, Gentry, Pattison, & Heggestad, 2010). Last, the relative contribution of the indirect associations in each model was calculated using the following formula: percentage (%) of indirect effects = (indirect effects / total effects) × 100, where total effects = indirect + direct effects.

ResultsMissing data were minimal (< 2%) and multiple imputation (expectation-maximiza-tion algorithm; Dempster, Laird, & Rubin, 1977) was used to estimate and replace missing observations. The distributional properties of each variable suggested that the normality assumption required for path analysis was met. No outliers were observed for the factor scores and the raw physical activity scores. The scores for the physical self-perceptions and motivational regulations were reliable based on methods of estimating reliability of congeneric measure that have been recommended within the framework of CFA (Yang & Green, 2010; see Table 1). Descriptive statistics for the raw scores and correlations between the factor scores are presented in Table 1. Of note, identified and intrinsic regulation scores were highly correlated (r = .71, p < .001; Table 1). While in line with the simplex pattern of motivational regula-tions outlined in organismic integration theory (Deci & Ryan, 1985), specifying the covariation between intrinsic and identified regulation in the current models led to problems of identification. Given the complexity of the models and the sample size, it was decided to test the main path models with intrinsic (highest form of self-determined motivation), introjected (midpoint on the self-determination continuum), and external regulation (lower end of the self-determination continuum).2

Actual:Ideal Physical Self-Discrepancies

The results of the path analyses are presented in Tables 2 and 3. The fit indices for the model were χ2 = 2.68, df = 1; RMSEA = .09; 90% CI for RMSEA = .00 to .23;


CFI = 1.00; SRMR = .02. In this model, 9%, 7%, 28% and 29% of the variance in external regulation, introjected regulation, intrinsic motivation and physical activity behavior was explained, respectively. Based on the surface value coefficients, a1 was significant and positive for intrinsic motivation and physical activity behav-ior. When actual and ideal self scores were in agreement, intrinsic motivation and physical activity levels increased as actual and ideal scores increased. In addition, a2 was significant and positive suggesting a nonlinear association between the agreement in actual and ideal self scores and physical activity levels. In addition, a3

Table 2 Associations Between the Physical Self-Discrepancies and the Motivational Regulations in the Model Including Ideal Self (First Row) and Ought Self (Second Row)

Direct Associations

External Introjected Intrinsic

Actual Self –.18 (.07)* .08 (.07) .55 (.07)*

–.14 (.07)* .13 (.07) .55 (.07)*

Ideal Self –.04 (.09) .25 (.10)* .01 (.09)

(Ought Self) –.08 (.08) .15 (.09) .01 (.08)

Actual Self Squared .10 (.07) .08 (.07) –.05 (.06)

.11 (.06) .08 (.07) –.04 (.06)

Ideal Self Squared .04 (.05) .11 (.05)* .03 (.05)

(Ought Self Squared) .02 (.06) .07 (.06) .06 (.06)

Actual × Ideal Self .12 (.10) .00 (.10) –.04 (.10)

(Actual × Ought Self) .06 (.09) .09 (.33) –.08 (.08)

Surface Values

a1 –.22 (.16) .33 (.17) .56 (.16)*

–.22 (.15) .28 (.15) .56 (.15)*

a2 .26 (.22) .19 (.22) –.06 (.11)

.19 (.21) .24 (.46) –.06 (.08)

a3 –.14 (.02)* –.17 (.03)* .54 (.02)*

–.06 (.01)* –.02 (.02) .54 (.01)*

a4 .02 (.02) .19 (.02)* .02 (.09)

.08 (.03)* .04 (.20) .10 (.08)

Note. a1 = b1 + b2, where b1 is the beta coefficient of actual self and b2 is the beta coefficient for ideal (or ought) self. a2 = b3 + b4 + b5, where b3 is the beta coefficient of actual self squared, b4 is the beta coefficient for the cross product of actual and ideal (or ought) self, and b5 is the beta coefficient for ideal (or ought) self squared. a3 = b1 – b2, and a4 = b3 – b4 + b5.

*z ≥ 1.96.

114 Brunet et al.

was significant and positive for intrinsic motivation and physical activity levels and negative for external and introjected regulations, indicating that intrinsic motivation physical activity levels were lower, and external and introjected regulations were higher, when the discrepancy was such that ideal self was higher than actual self. Last, a4 was significant and positive for introjected regulation, indicating that a greater degree of discrepancy between actual and ideal self-perceptions was related to higher introjected regulation.

Actual:Ought Physical Self-Discrepancies

The results of the path analyses are presented in Tables 2 and 4. The fit indices for the model were χ2 = 2.57, df = 1; RMSEA = .09; 90% CI for RMSEA = .00 to .23; CFI = 1.00; SRMR = .02. In this model, 9%, 5%, 29% and 30% of the variance in external regulation, introjected regulation, intrinsic motivation and physical activity behavior was explained, respectively. Based on the surface value coefficients, a1 was significant and positive for intrinsic motivation and physical activity levels. In addition, a3 was significant and positive for intrinsic motivation and physical activity levels and negative for external regulation. Last, a4 was significant and positive for external regulation.

Table 3 Associations Between Actual:Ideal Physical Self-Discrepancies and Physical Activity Behavior

IndirectAssociations

DirectAssociations

Indirect effects(%)

Actual Self 3.91 (1.16)* 6.63 (1.94)* 37.10

Ideal Self 1.29 (.94) –.81 (2.21) 61.43

Actual Self Squared .05 (.63) –.42 (1.53) 10.64

Ideal Self Squared .72 (.45) .15 (1.09) 82.76

Actual × Ideal Self –.33 (.89) .56 (2.25) 37.08

External Regulation — –.42 (1.65) —

Introjected Regulation — 4.94 (1.70)* —

Intrinsic Regulation — 6.27 (1.74)* —

Surface Values Coefficients SE Coefficients SE

a1 5.20* 2.10 5.82* .27 47.19

a2 .44* .19 .29 1.81 60.27

a3 2.62* .22 7.44* 4.15 26.04

a4 1.10 1.97 –.83 2.69 56.99

Note. a1 = b1 + b2, where b1 is the beta coefficient of actual self and b2 is the beta coefficient for ideal self. a2 = b3 + b4 + b5, where b3 is the beta coefficient of actual self squared, b4 is the beta coefficient for the cross product of actual and ideal (or ought) self, and b5 is the beta coefficient for ideal self squared. a3 = b1 – b2 and a4 = b3 – b4 + b5.

*z ≥ 1.96.


DiscussionThe purpose of the current study was to examine the relationships between physi-cal self-perceptions and discrepancies and physical activity behavior, as mediated by motivational regulations, in young adult women. Physical activity levels were related directly and indirectly to the agreement between actual and ideal (or ought) self-perceptions, suggesting that the motivational regulations partially mediated the relationship between physical self-discrepancies and physical activity behavior. In addition, physical activity levels were lower when ideal or ought self-perceptions were higher than actual self-perceptions. These findings provide empirical evidence for the links between key constructs embedded in self-discrepancy (Higgins, 1987) and organismic integration (Deci & Ryan, 1985, 2004) theories.

Greater agreement between actual and self-guide states had direct and posi-tive linear links to intrinsic motivation and physical activity behavior, and positive nonlinear links to physical activity behavior. In addition, the degree of discrepancy between actual and ideal (or ought) selves was positively related to introjected and external regulations. This suggests that young adult women’s physical activity moti-vation and behavior is related to the differences between who they perceive they are (actual self) and who they would like to be (ideal self) or who they ought to be

Table 4 Associations Between Actual:Ought Physical Self-Discrepancies and Physical Activity Behavior

IndirectAssociations

DirectAssociations

Indirect Effects(%)

Actual Self 4.24 (1.13)* 6.65 (1.84)* 38.93

Ought Self .83 (.82) –1.77 (1.93) 31.92

Actual Self Squared .07 (.63) –1.39 (1.48) 4.79

Ought Self Squared .75 (.54) –1.18 (1.31) 38.86

Actual × Ought Self –.38 (.81) 3.54 (1.99) 9.69

External Regulation — –.58 (1.63) —

Introjected Regulation — 4.98 (1.67)* —

Intrinsic Regulation — 6.44 (1.73)* —

Surface Values Coefficients SE Coefficients SE

a1 5.07* 1.95 4.88* .09 50.95

a2 .44 .36 2.36 1.80 15.71

a3 3.41* .31 8.42* 3.61 28.83

a4 1.20 1.98 6.11 4.78 16.42

Notes. a1 = b1 + b2, where b1 is the beta coefficient of actual self and b2 is the beta coefficient for ought self. a2 = b3 + b4 + b5, where b3 is the beta coefficient of actual self squared, b4 is the beta coefficient for the cross product of actual and ideal (or ought) self, and b5 is the beta coefficient for ought self squared. a3 = b1 – b2 and a4 = b3 – b4 + b5.

*z ≥ 1.96.

116 Brunet et al.

(ought self). These results are consistent with previous findings that body-size self-discrepancies are positively related to weight management motives among women (Ingledew & Sullivan, 2002; Markland & Ingledew, 2007; Sabiston & Chandler, 2009; Sabiston et al., 2005). These findings are also consistent with Deci and Ryan’s (2004) theoretical proposition in that factors perceived as controlling foster an external locus of causality. From this perspective, it appears that actual:ideal and/or actual:ought self-discrepancies reflect a sense of pressure to conform to perceived physical standards and may relate to increased non-self-determined motivation to participate in physical activity. Thus, encouraging young women to have healthy and realistic ideal and ought self-perceptions to reduce the discrepancies between their actual and self-guide states may help promote self-determined motivational regulations for physical activity in this population.

Contrary to the hypothesis, the degree of discrepancy between actual and ideal (or ought) self-perceptions did not relate to physical activity behavior. These findings do not support Higgins’ (1987) contention that self-discrepancies foster engagement in self-regulatory behaviors used to reduce discrepancies between self-states and self-guides, or at least that physical activity may not be a self-regulatory behavior in this regard. In addition, the current results do not support previous findings linking physical self-discrepancies related to body shape/weight to physical activity (Anton et al., 2000; Markland, 2009). Part of the reason actual:ideal and actual:ought self-discrepancies were not linked to physi-cal activity behavior in this study may be related to a lack of perceived ability to meet one’s ideal or ought selves. While not included in the original theoretical considerations (Higgins, 1987), a can self—which represents individuals’ beliefs regarding the attributes they believe they can possess (Higgins, Tykocinski, & Vookles, 1990)—may help explain these equivocal relationships. That is, the associations between the physical self-discrepancies and physical activity behavior may differ according to the “can” self, such that a positive association may be observed in women who perceive they can achieve their ideal and/or ought selves, whereas a negative association may be observed in women who doubt their abilities. Therefore, the potential moderating role of the “can” self should be considered in future research. Based on previous research (Cumming & Thogersen-Ntoumani, 2011; Fox & Corbin, 1989; Higgins et al., 1990; Martin Ginis, Prapavessis, & Haase, 2008), self-presentational efficacy, social physique anxiety, and the importance of achieving the ideal and/or ought self-states could also be potential moderators. For instance, it is plausible that individuals who experience discrepancies may fear that others negatively evaluate their physique (Woodman & Steer, 2011). In turn, these individuals may be motivated to avoid situations in which aspects of the physique are thought to be evaluated by others and may attempt to avoid behaviors like physical activity in which this is likely to be an outcome. It would be valuable to investigate these factors as moderators to understand when and how discrepancies in self-perceptions relate to physical activity in future investigations.

Furthermore, based on previous studies (Anton et al., 2000; Markland, 2009), it was hypothesized that mediating variables, such as motivational regulations, may help explain the self-discrepancies and physical activity associations. In partial support of this hypothesis, agreement between the physical self-perceptions had indirect and positive associations with physical activity behavior through the moti-


vational regulations. Markland (2009) also found that body-size self-discrepancy related indirectly to physical activity behavior through intrinsic and identified regulations. These findings support the integration of self-discrepancy theory and self-determination theory, and reveal important theoretical insight into how physical self-discrepancies or agreement relate to physical activity behavior.

A key finding in the current study is that ideal and ought scores were highly correlated. This may reflect the internalization or assimilation of individuals’ ideal and ought selves. The ought self represents physical attributes individuals believe they should or ought to possess and are likely shaped by societal standards (Higgins, 1987). Over time, individuals may accept these societal standards as their own with their ideal self reflecting these standards insomuch as their ought selves become their ideal selves. In support of this view, Bessenoff and Snow (2006) reported that personal beauty ideal was highly related to cultural beauty ideal among women. High correlations between ideal and ought selves have been reported in other studies (Ozgul, Heubeck, Ward, & Wilkinson, 2003; Tangney, Niedenthal, Covert, & Barlow, 1998), which raises concerns regarding the discrimi-nant validity of ideal and ought physical self-perceptions. However, the pattern of results observed in this study differed when ideal and ought self-perceptions were included in the path models. For instance, the discrepancy between actual and ideal self-perceptions was not related to external regulation, but the discrepancy between actual and ought self-perceptions was. Thus, there is a need to further develop research questions examining ideal and ought physical self-perceptions and their unique effects on cognition, affect, and behavior. Furthermore, testing specific differences in the associations of actual, ideal, and ought self-perceptions and discrepancies with physical activity motivation regulations and behavior is needed as few studies have assessed ought physical self-perceptions. Moreover, examining the development of these discrepancies and identifying the develop-mental phase(s) in which the ought self is assimilated within the ideal self would be a valuable line of further inquiry.

It is important to acknowledge the limitations of this study. First, the cross-sectional design of this study limits conclusions regarding the causal direction of associations. Although establishing mediation requires that researchers specify a temporal ordering of the independent, mediator, and dependent variables, respec-tively, researchers have stated that the ordering of the variables in cross-sectional studies may be justified on theoretical grounds (Mathieu & Taylor, 2006). In this study, self-discrepancy and organismic integration theories were used to guide model testing (i.e., physical self-discrepancies to motivational regulations to physical activity behavior; Deci & Ryan, 1985, 2004; Higgins, 1987); however, other sequential patterns may also be plausible. For instance, based on Higgins (1987) theory it could be argued that self-regulatory behaviors such as physical activity may reduce physical discrepancies in the long-term. Longitudinal stud-ies would provide additional justification for the sequence tested herein, and experimental studies would strengthen the evidence for the temporal ordering (Lockhart, MacKinnon, & Ohlrich, 2011). Second, physical activity behavior was assessed using a self-report questionnaire which may have limitations (e.g., recall bias, social desirability; Sallis & Saelens, 2000). Third, the complexity of the model (and low degrees of freedom) prevented potential covariates to be included in the path models. If physical self-perceptions, motivational regulations,

118 Brunet et al.

and physical activity levels vary by demographic characteristics, mean-level dif-ferences could emerge. However, based on Deci and Ryan’s (1985) contention that the hypothesized associations embedded in the broader self-determination theory are universal, relationship differences would not be expected. Fourth, the unique relationships between amotivation and integrated regulation and the self-discrepancies were not tested because the BREQ-2 does not include an integrated regulation subscale and the distribution of amotivation scores was non-normal. Although the finding that few women reported being amotivated for physical activity is a common finding in research using the BREQ-2, future studies may want to target inactive populations to better understand this construct. Last, the physical self-discrepancies were measured among young adult women both in the current study and in preliminary analyses on reliability and validity. This pre-cludes the use of this measure with other samples, including youth, older adults, and men. Future research is needed to test the measure with different samples. Along these lines, the reliability of the ideal self-perception scores was low and should be examined in further research.

Despite these limitations, the current study provides important insights into understanding how physical self-perceptions and discrepancies relate to physical activity behavior. Specifically, it provides evidence that integrating self-discrepancy and organismic integration theories holds promise for the study of psychosocial models aimed at explaining physical activity behavior. Since the majority of previ-ous research has focused on a narrow range of discrepancies regarding one’s body (i.e., size, shape, weight), the current study builds on this research by assessing self-perceptions in relation to five salient physical characteristics. Nevertheless, future research strategies may be directed to advance the measurement of actual, ideal, and ought physical self-discrepancies as there may be other physical char-acteristics that could explain additional variance in physical activity behavior. In addition, researchers should consider the possibility that each physical self-perception may differentially relate to physical activity motivation and behavior. Furthermore, although previous studies have demonstrated an association between actual:ideal self-discrepancy and physical activity behavior, few studies have used appropriate methodologies for studying discrepancy scores (Cafri et al., 2010). The current study adds to this literature by including a measure of ought physical self-perceptions and using a sophisticated data analytical approach proposed for analyzing discrepancies in ratings (Cafri et al., 2010; Edwards, 2002; Shanock et al., 2010). This made it possible to examine the extent to which actual, ideal, and ought self-perceptions were associated to physical activity, thus overcom-ing problems associated with ambiguous interpretations and confounded effects (Edwards, 2002). It was also possible to examine the agreement between actual and self-guide states and their association with physical activity, as well as whether physical activity levels were lower when actual scores were lower than self-guide states—information that would have been sacrificed had relative or absolute dif-ference scores been used (Shanock et al., 2010). Based on the current findings, it appears that researchers should attempt to increase women’s actual, ideal, and ought physical self-perceptions and bridge the gap between their self-states and self-guides since agreement between these self-perceptions related positively to physical activity motivation and behavior.


Notes

1. Deci and Ryan (1985) describe integrated regulation as a fourth external regulation, located adjacent to intrinsic motivation on the continuum, and the most self-determined of the external regulations. Integrated regulation occurs when an individual participates in an activity because it is congruent with the personal values, goals, and needs that are part of oneself, but is not inherently enjoyable. This motivational regulation was not discussed in the current study because it is difficult to empirically differentiate between integrated and identified regulation and is not assessed in the BREQ-2 (Markland & Tobin, 2004). In addition, although the BREQ-2 includes an amotivation subscale, amotivation scores were leptokurtotic (kurtosis = 6.30) and positively skewed (skewness = 2.45), and the items were minimally endorsed such that 80.1% of participants reported “not at all” to all four items in this study. Therefore, it was excluded from the analyses and data are not presented for this variable owing to the measurement concerns and lack of reporting of this motivational regulation in the current sample.

2. The results of the path analyses including identified regulation were similar to those including intrinsic motivation. Specifically, the fit indices for the models including ideal and ought self-perceptions and discrepancies, respectively, were as follows: χ2 = 6.67 and 7.72, df = 1; RMSEA = .11 and .12; 90% CI for RMSEA = .02 to .21; CFI = .99; SRMR = .02, and 28% and 19% of the variance in physical activity behavior was explained. Based on the surface value coefficients for the models including actual and ideal self-perceptions, a1 was significant and positive for identified regulation (coefficient = .57, SE = .16) and physical activity behavior (coefficientdirect = 6.20, SE = .38; coefficientindirect = 4.82, SE = 1.87), and negative for external regulation (coefficient = –.14, SE = .02). In addition, a2 was significant and positive for identified regulation (coefficient = .07, SE = .01). In addition, a3 was significant and positive for identified regulation (coefficient = .19, SE = .02) and physical activity behavior (coefficientdirect = 9.30, SE = 4.06; coefficientindirect = .78, SE = .05), and negative for introjected regulation (coefficient = –.17, SE = .03). Last, a4 was significant and positive for introjected regulation (coefficientindirect = .33, SE = .07), identified regulation (coefficientindirect = .61, SE = .26) and physical activity (coefficientdirect = .5.62, SE = .2.65; coefficientindirect = 5.17, SE = 1.02), and negative for external regulation (coefficient = –.26, SE = .08). Based on the surface value coefficients for the models including actual and ought self-perceptions, a1 was significant and positive for identified regu-lation (coefficient = .46, SE = .15) and physical activity behavior (coefficientdirect = 5.93, SE = .17; coefficientindirect = 4.10, SE = 1.73). In addition, a3 was significant and positive for identified regulation (coefficient = .38, SE = .01) and physical activity behavior (coefficientdirect = 9.53, SE = .3.71; coefficientindirect = 2.29, SE = .19) and negative for external (coefficient = –.06, SE = .01). Last, a4 was significant and positive for introjected regulation (coefficientindirect = .25, SE = .07), identified regulation (coefficientindirect = .47, SE = .06) and physical activity (coefficientindirect = 4.01, SE = .97).

Acknowledgments

This research was funded by a Fonds Québécois de la Recherche sur la Société et al Cul-ture new investigator grant awarded to Sabiston. Brunet, Castonguay, and Ferguson were funded by Social Sciences and Humanities Research Council doctoral fellowships during this research project. Brunet and Sabiston made equal contributions to this manuscript.

ReferencesAnton, S.D., Perri, M.G., & Riley, J.R. (2000). Discrepancy between actual and ideal body

images: Impact on eating and exercise behaviors. Eating Behaviors, 1, 153–160.

120 Brunet et al.

Bessenoff, G., & Snow, D. (2006). Absorbing society’s influence: Body image self-discrep-ancy and internalized shame. Sex Roles, 54, 727–731.

Brown, T.A. (2006). Confirmatory factor analysis for applied research: Methodology in the social sciences. New York, NY: The Guilford Press.

Brunet, J., & Sabiston, C.M. (2009). Social physique anxiety and physical activity: A self-determination theory perspective. Psychology of Sport & Exercise, 10, 329–335.

Cafri, G., van den Berg, P., & Brannick, M.T. (2010). What have the difference scores not been telling us? A critique of the use of self-ideal discrepancy in the assessment of body image and evaluation of an alternative data-analytic framework. Assessment, 17, 361–376.

Cash, T.F., & Pruzinsky, T. (2004). Body image: A handbook of theory, research, and clinical practice. New York, NY: Guilford Press.

Colley, R.C., Garriguet, D., Jassen, I., Craig, C.L., Clarke, J., & Tremblay, M.S. (2011). Physical activity of Canadian adults: Accelerometer results from the 2007 to 2009 Canadian Health Measures Survey. Health Reports, 22, 7–14.

Cumming, J., & Thogersen-Ntoumani, C. (2011). Self-presentational cognitions for exercise in female adolescents. Journal of Applied Social Psychology, 41, 429–444.

Deci, E.L., & Ryan, R.M. (1985). Intrinsic motivation and self-determination in human behavior. New York, NY: Plenum Press.

Deci, E.L., & Ryan, R.M. (2004). Handbook on self-determination research. Rochester, NY: The University of Rochester Press.

Dempster, A.P., Laird, N.M., & Rubin, D.B. (1977). Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society. Series A (General), 39, 1–38.

Dent, C.W., Galaif, J., Sussman, S., Stacy, A., Burton, D., & Flay, B.R. (1993). Demographic, psychosocial and behavioural differences in samples of actively and passively consented adolescents. Addictive Behaviors, 18, 51–56.

Edwards, J.R. (1994). The study of congruence in organizational behavior research: Critique and a proposed alternative. Organizational Behavior and Human Decision Processes, 58, 51–100.

Edwards, J.R. (2002). Alternatives to difference scores: Polynomial regression analysis and response surface methodology. In F. Drasgow & N. Schmitt (Eds.), Measuring and analyzing behavior in organizations: Advances in measurement and data analysis (pp. 350–400). San Francisco, CA: Jossey-Bass.

Edwards, J.R., & Parry, M.E. (1993). On the use of polynomial regression equations as an alternative to difference scores in organizational research. Academy of Management Journal, 36, 1577–1613.

Fox, K.R., & Corbin, C.B. (1989). The Physical Self-Perception Profile: Development and preliminary validation. Journal of Sport & Exercise Psychology, 11, 408–430.

Godin, G., & Shephard, R. (1985). A simple method to assess exercise behavior in the com-munity. Canadian Journal of Applied Science, 10, 141–146.

Hayes, S.D., Crocker, P.R.E., & Kowalski, K.C. (1999). Gender differences in physical self-perceptions, global self-esteem and physical activity: Evaluation of the physical self-perception profile model. Journal of Sport & Exercise Psychology, 22, 1–14.

Higgins, E.T. (1987). Self-discrepancy: A theory relating self and affect. Psychological Review, 94, 319–340.

Higgins, E.T., Tykocinski, O., & Vookles, J. (1990). Patterns of self-beliefs: The psychologi-cal significance of relations among the actual, ideal, ought, can and future selves. In J.M. Olson & M.P. Zanna (Eds.), Self-inference processes: The Ontario Symposium, Volume 6. Hillsdale, NY: Lawrence Erlbaum Associates.

Hu, L., & Bentler, P.M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55.

Ingledew, D.K., & Sullivan, G. (2002). Effects of body mass and body image on exercise motives in adolescence. Psychology of Sport and Exercise, 3, 323–338.


Jacobs, D.R., Ainsworth, B.E., Hartman, T.J., & Leon, A.S. (1993). A simultaneous evalu-ation of 10 commonly used physical activity questionnaires. Medicine and Science in Sports and Exercise, 25, 81–91.

Kazén, M., & Kuhl, J. (2011). Directional discrepancy between implicit and explicit power motives is related to well-being among managers. Motivation and Emotion, 35(Suppl.) 317–327.

Landry, J.B., & Solomon, M.A. (2004). African American women’s self-determination across the states of change for exercise. Journal of Sport & Exercise Psychology, 26, 457–469.

Lockhart, G., MacKinnon, D.P., & Ohlrich, V. (2011). Mediation analysis in psychosomatic medicine research. Psychosomatic Medicine, 73, 29–43.

Lynch, S.M., & Zellner, D.A. (1999). Figure preferences in two generations of men: The use of figure drawings illustrating differences in muscle mass. Sex Roles, 40, 833–843.

MacKinnon, D.P., Fairchild, A.J., & Fritz, M.S. (2007). Mediation analysis. Annual Review of Psychology, 58, 593–614.

Markland, D. (2009). The mediating role of behavioural regulations in the relationship between perceived body size discrepancies and physical activity among adult women. Hellenic Journal of Psychology, 6, 169–182.

Markland, D., & Ingledew, D.K. (2007). The relationships between body mass and body image and relative autonomy for exercise among adolescent males and females. Psy-chology of Sport and Exercise, 8, 836–853.

Markland, D., & Tobin, V. (2004). A modification to the Behavioral Regulation in Exercise Questionnaire to include an assessment of amotivation. Journal of Sport & Exercise Psychology, 26, 191–196.

Marsh, H.W., & Redmayne, R.S. (1994). A multidimensional physical self-concept and its relations to multiple components of physical fitness. Journal of Sport & Exercise Psychology, 16, 43–55.

Marsh, H.W., Richards, G.E., Johnson, S., Rocher, L., & Tremayne, P. (1994). Physical Self-Description Questionnaire: Psychometric properties and a multitrait-multimethod analysis of relations to existing instruments. Journal of Sport & Exercise Psychology, 16, 270–305.

Martin Ginis, K.A., Prapavessis, H., & Haase, A.M. (2008). The effects of physique-salient and physique non-salient exercise videos on women’s body image, self-presentational concerns, and exercise motivation. Body Image, 5, 164–172.

Mathieu, J.E., & Taylor, S.R. (2006). Clarifying conditions and decision points for media-tional type inferences in Organizational Behavior. Journal of Organizational Behavior, 27, 1031–1056.

Miller, D.J., Freedson, P.S., & Kline, G.M. (1994). Comparison of activity levels using the Caltrac (R) accelerometer and five questionnaires. Medicine and Science in Sports and Exercise, 26, 376–382.

Ozgul, S., Heubeck, B., Ward, J., & Wilkinson, R. (2003). Self-discrepancies: Measurement and relation to various negative affective states. Australian Journal of Psychology, 55, 56–62.

Pedhazur, E. (1997). Multiple regression in behavioral research: Explanation and prediction. Orlando, FL: Harcourt Brace & College.

Pruis, T.A., & Janowsky, J.S. (2010). Assessment of body image in younger and older women. The Journal of General Psychology, 137, 225–238.

Sabiston, C.M., Brunet, J., Kowalski, K.C., Wilson, P.M., Mack, D.E., & Crocker, P.R.E. (2010). The role of body-related self-conscious emotions in motivating women’s physi-cal activity. Journal of Sport & Exercise Psychology, 32, 417–437.

Sabiston, C.M., & Chandler, K. (2009). Effects of fitness advertising on weight and body shape dissatisfaction, social physique anxiety, and exercise motives in a sample of healthy-weight females. Journal of Applied Biobehavioral Research, 14, 165–180.

Sabiston, C.M., Crocker, P.R.E., & Munroe-Chandler, K.J. (2005). Examining current-ideal discrepancy scores and exercise motivations as predictors of social physique anxiety in exercising females. Journal of Sport Behavior, 28, 668–685.

122 Brunet et al.

Sallis, J.F., & Saelens, B.E. (2000). Assessment of physical activity by self-report: Status, limitations, and future directions. Research Quarterly for Exercise and Sport, 71, 1–14.

Scalas, L.F., & Marsh, H.W. (2008). A stronger latent-variable methodology to actual–ideal discrepancy. European Journal of Personality, 22, 629–654.

Shanock, L.R., Baran, B.E., Gentry, W.A., Pattison, S.C., & Heggestad, E.D. (2010). Poly-nomial regression with response surface analysis: A powerful approach for examining moderation and overcoming limitations of difference scores. Journal of Business and Psychology, 25, 543–554.

Sonstroem, R.J., Speliotis, E.D., & Fava, J.L. (1992). Perceived physical competence in adults: An examination of the physical self-perception profile. Journal of Sport & Exercise Psychology, 14, 207–221.

Stunkard, A., Sorenson, T., & Schlusinger, F. (1983). Use of the Danish adoption register for the study of obesity and thinness. In S. Kety, L.P. Rowland, R.L. Sidman, & S.W. Matthysse (Eds.), The genetics of neurological and psychiatric disorders (pp. 115–120). New York, NY: Raven.

Tangney, J.P., Niedenthal, P.M., Covert, M.V., & Barlow, D.H. (1998). Are shame and guilt related to distinct self-discrepancies? A test of Higgins’s (1987) Hypotheses. Journal of Personality and Social Psychology, 75, 256–286.

Thompson, J.K., & Altabe, M.N. (1991). Psychometric qualities of the Figure Rating Scale. The International Journal of Eating Disorders, 10, 615–619.

Thompson, J.K., & Stice, E. (2001). Thin-ideal internalization: Mounting evidence for a new risk factor for body-image disturbance and eating pathology. Current Directions in Psychological Science, 10, 181–183.

Warburton, D.E.R., Nicol, C.W., & Bredin, S.S.D. (2006). Health benefits of physical activ-ity: The evidence. Canadian Medical Association Journal, 174, 801–809.

Wilson, P.M., & Rodgers, W.M. (2004). The relationship between perceived autonomy support, exercise regulations and behavioral intentions in women. Psychology of Sport and Exercise, 5, 229–242.

Woodman, T., & Steer, R. (2011). Body self-discrepancies and women’s social physique anxiety: The moderating role of the feared body. The British Journal of Psychology, 102, 147–160.

Yang, Y., & Green, S.B. (2010). A note on structural equation modeling estimates of reli-ability. Structural Equation Modeling, 17, 66–81.

Manuscript submitted: March 12, 2011Revision accepted: November 5, 2011


Appendix: The Physical Self-Discrepancy ScaleRead each of the descriptors below and indicate on the rating scale provided how much each word describes you now, what you would like to be, and what you think you should be.

Not at all Somewhat Very Much

Physically Strong

I am 1 2 3 4 5 6 7

I would like to be 1 2 3 4 5 6 7

I should be 1 2 3 4 5 6 7

Attractive

I am 1 2 3 4 5 6 7


I should be 1 2 3 4 5 6 7

Thin

I am 1 2 3 4 5 6 7


I should be 1 2 3 4 5 6 7

Confident in my physical abilities

I am 1 2 3 4 5 6 7


I should be 1 2 3 4 5 6 7

Fit

I am 1 2 3 4 5 6 7


I should be 1 2 3 4 5 6 7

the association between physical self-discrepancies and ... · behavior (brunet & sabiston,...

Documents