Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 May 1.
Published in final edited form as: Psychol Sport Exerc. 2021 Jan 12;54:101888. doi: 10.1016/j.psychsport.2021.101888

Women’s exercise identity increases after a 16-week exercise RCT and is linked to behavior maintenance at follow-up

Arielle S Gillman 1,1, Courtney J Stevens 1,2, Angela D Bryan 1
PMCID: PMC7901813  NIHMSID: NIHMS1662296  PMID: 33633498

Abstract

Background.

Regular exercise is critical for disease prevention, but adherence to public health guidelines is poor. Exercise identity is purported to be associated with exercise behavior maintenance, but the extant literature is largely cross-sectional and of low/modest quality.

Purpose.

To examine change in exercise identity after completion of a supervised exercise intervention, as well as associations between change in exercise identity and exercise maintenance at 6-months follow-up.

Methods.

N = 276 insufficiently physically active women were randomized to a 16-week, supervised exercise training intervention with 4 conditions fully crossed on intensity (vigorous/moderate) and duration (long/short). Exercise identity was measured pre- and post-intervention and assessments of exercise motivation and behavior frequency were collected at 6-months post-intervention follow-up.

Results.

On average, participants experienced a statistically significant change in exercise identity over the course of the intervention, t(128) = 7.94, p < .001, but identity change scores did not differ across training conditions, p = .91. Identity change was significantly positively related to changes in other theory-informed, motivation-based determinants of exercise, and predicted an additional 16.17 minutes of exercise per week, on average, at follow-up, b = 16.76, t(103) = 2.30, p = .023.

Conclusions.

Participants experienced increased self-identification with exercise after 16-weeks of training, but training volume did not influence the amount of identity change. As expected, greater change in exercise identity was associated with higher levels of exercise behavior at 6-months post-intervention follow-up (ClinicalTrials.gov number NCT02032628).

Keywords: Exercise identity, exercise, physical activity, maintenance, motivation

Lack of sufficient physical activity is a widespread, global problem. Survey data collected from 122 countries has found that 31% of the global adult population is physically inactive (Heyward & Gibson, 2014), and the World Health Organization names physical inactivity as the fourth leading risk factor for global mortality causing 3.2 million deaths annually (WHO, 2017).

A number of interventions have demonstrated success in terms of promoting exercise behavior adoption and participation through the intervention period. However, ongoing maintenance of physical activity/exercise has proven difficult to achieve (Rothman et al., 2015; Timmons et al., 2020). Past research has found that only around 50% of individuals who adopt an exercise routine stick with it for more than six months (Ekkekakis et al., 2011; Marcus et al., 2000, 2006). It has been argued that previously-dominant theoretical models (e.g., the Theory of Planned Behavior [TPB]; Ajzen & Madden, 1986) are limited by lack of guidance regarding underlying mechanisms that govern behavior maintenance and, thus, may be insufficient to explain sustained behavior over time (Rothman, Baldwin, Hertel, & Fuglestad, 2004). This is because a behavior that occurs across time, rather than within a prescribed time frame (i.e., the intervention period), is necessarily subject to fluctuating meta-motivational states, as well as changing contextual factors and barriers that may influence decision making and/or constrain behavioral opportunities (Borland, 2017).

To address these gaps in the literature, more recent reviews and theoretical frameworks argue that the development of a stable exercise identity may be one critical element in the promotion of exercise maintenance (Caldwell et al., 2018; Rhodes, 2017; Strachan & Whaley, 2013). Here, we define one’s exercise identity as the degree to which an individual incorporates one or more exercise-related behaviors into their self-identity or self-schema – an individual with a strong exercise identity would view themselves as “an exerciser” rather than “someone who exercises”. A stable exercise identity can help buffer against environmental and other self-regulation challenges to behavior maintenance (Caldwell et al., 2018). A recent review and meta-analysis of 32 studies found a medium-sized relationship (r = .40, 95% CI [.38-.41]) between exercise identity and frequency of exercise behavior (Rhodes et al., 2016). However, this review also concluded that much of the research conducted to date has been of only “modest” quality, as most studies have been cross-sectional and have predominantly utilized college-based, convenience samples. It is not yet well understood how identifying as an “exerciser” may promote sustained exercise behavior over time, or how such exercise identities may form over time (Rhodes, 2017).

Theoretical basis of identity as a maintenance construct

Research has shown that defining an individual (including the self) in terms of a category label leads to specific expectations about that individuals’ behavior. For example, children given a description of another child using a noun compared to a verbal predicate (i.e., “a carrot eater” versus “she eats carrots whenever she can”) were more likely to infer that the characteristic was a stable trait (Gelman & Heyman, 1999). Similarly, Walton and Banaji (2004) found that adults were more likely to infer stability of an individual’s preferences when that individual was described using a category label, which extended to self-perception. Participants evaluated their own preferences as stronger, more stable, and more resilient when these preferences were described using nouns rather than verbs. That is, labeling (e.g., “I am an exerciser”) may imply that an attribute is stable and central to ones’ identity, whereas simply describing a behavior (e.g., “I am someone who exercises”) does not carry such an implication.

Identity theory posits that if a behavior is thought of as an expression of the self, the individual will not only be likely to change their behavior once, but will be likely to maintain it over time (Stets & Burke, 2003; Strachan, Brawley, & Spink, 2015). That is, self-categorization leads to expectations that one’s behavior should be consistent with one’s identity; to maintain this consistency, individuals may either selectively process information relevant to the identity, or initiate behaviors to reduce perceived dissonance when a behavior is not consistent with that identity (Rhodes, 2017). This phenomenon has been examined in several contexts, including in exercise behavior (Guérin et al., 2019; Ntoumanis et al., 2018; Strachan et al., 2015).

Changes in exercise identity may foster behavior change in concert with other theory-informed, motivation-based determinants of exercise behavior. For example, exercise identity and intrinsic motivation (re: Self-Determination Theory; Ryan & Deci, 2000) have been demonstrated to be reciprocally and longitudinally associated (Strachan et al., 2013), although the directionality of this relationship may depend on the aspect of exercise identity in question (i.e., “role idenity” or “exercise beliefs;” see Ntoumanis et al., 2018), and intrinsic motivation is associated with maintenance of exercise behavior over time (Rodgers et al., 2010; Teixeira et al., 2012). Additionally, meta-analytic work shows that self-identity explains variance in intentions to perform various health behaviors, including exercise, over and above TPB constructs (Rise, Sheeran, & Hukkelberg, 2010). In sum, exercise identity likely covaries with other constructs which may jointly improve exercise behavioral regulation over time (Strachan et al., 2013).

Behavioral engagement and identity formation

It is logical to hypothesize that repeated experience with exercise may be a necessary antecedent to exercise identity formation (Rhodes, 2017). Seminal research from social psychology posits that repetition of certain behaviors may eventually translate into identity formation. For example, research on self-serving bias suggests that when individuals succeed at a socially desirable behavior, such as exercise, they are more likely to attribute that success to stable internal characteristics rather than external influences (Miller & Ross, 1975; Adams et al., 2005; Klesges et al., 2004). Additionally, repetition of behavior may enhance one’s perceived ability to engage in the behavior (Gardiner & Bryan, 2017) or enhance one’s perceived commitment to the behavior, both of which may predict the degree to which one identifies as an exerciser (Kendzierski & Morganstein, 2009). Subsequently, individuals may be motivated to maintain an exercise routine so that their behavior is aligned with their cognitions (regarding identification as an exerciser; Chatzisarantis et al., 2008; Cooper & Feldman, 2020). In this way, a positive feedback loop may be created such that identity supports behavioral maintenance over time, which, in turn, promotes identification as an exerciser.

The current study

The current study utilizes secondary data from a parent randomized controlled trial (RCT) for which the primary aims were to test the influence of increased exercise volume on biological outcomes related to breast cancer. The parent RCT main outcome analyses identified limited effects of increases in physical fitness on healthy methylation patterns of specific genes associated with breast cancer ((Gillman, Helmuth, Koljack, Hutchison, Kohrt, & Bryan, under review)). Due to the focus of the parent RCT on breast cancer prevention, men were not included in the sample. Insufficiently physically active women were recruited to participate in a 16-week supervised exercise intervention and were randomly assigned to one of 4 exercise training conditions: (1) short duration/vigorous intensity, (2) long duration/ vigorous intensity, (3) short duration/moderate intensity, or (4) long duration/moderate intensity. Exercise identity and other theory-informed, motivation-based determinants of exercise behavior were measured at baseline prior to the start of the intervention, immediately after completing 16-weeks of supervised exercise training, and again 6-months later. At 6-month follow-up, participants also completed self-report assessments of their exercise behavior.

Primary objective.

The primary objective of the present study was to examine changes in exercise identity among formerly inactive participants enrolled in the parent trial’s 16-week exercise intervention. Given the link between repeated experiences of a socially desirable behavior and identity formation, we propose that repeatedly engaging in exercise over the course of 16 weeks might lead individuals to adopt an identity as an “exerciser.” Specifically, we hypothesized that exercise identity would increase from baseline to post-intervention across all intervention conditions.

Due to the study design of the parent RCT (i.e., 4 conditions fully crossed on intensity and duration), we also had the opportunity to test exploratory questions regarding whether more intense or more frequent training had any differential impact on identity formation. Engaging in more exercise volume during the intervention may have enhanced exercise identity formation in a dose-response pattern for several reasons. First, participants might have perceived greater exercise training volume to be a closer fit with their schema for what it is to be an “exerciser.” That is, engaging in more vigorous intensity exercise, such as running or jogging, for longer periods of time, may be perceived differently than engaging in shorter bouts of moderate intensity exercise, such as walking. It is also possible that larger increases in cardiorespiratory fitness (measured as changes in VO2max; O’Donovan et al., 2005; Swain & Franklin, 2006) could impact exercise identity (e.g., feeling more physically fit over time may also more closely align with their “exerciser”-related schema). However, due to a lack of a strong theoretical rationale and existing supporting data, these research questions concerning moderation effects were exploratory. We tested the hypothesis that changes in exercise identity would be moderated by exercise volume, such that participants who completed a greater volume of exercise (i.e., assignment to the vigorous intensity and/or longer duration conditions, more total volume completed over the intervention period measured in kcal/kg/min, more exercise sessions attended over the intervention period, or larger pre-post-intervention changes in VO2max) would experience greater change in exercise identity.

Secondary objectives.

Secondary objectives of the present study were: (1) to assess whether changes in exercise identity were subsequently associated with greater exercise maintenance at 6-months follow-up, (2) to determine the extent to which changes in exercise identity may co-vary with changes in other theory-informed, motivation-based determinants of exercise behavior (e.g., intrinsic motivation, TPB constructs), and (3) to determine the stability of the exercise identity construct from post-intervention to 6-months follow-up. First, we hypothesized that change in exercise identity over the course of the intervention would predict more exercise behavior at follow-up. Second, we hypothesized that changes in exercise identity would be positively correlated with changes in other theory-informed, motivation-based determinants of exercise behavior. Additionally, because exercise identity is hypothesized to better predict maintenance than traditional models such as the TPB (Rhodes et al., 2016), we expected that changes in exercise identity would explain significantly more variance in follow-up exercise behavior compared to changes in TPB constructs, in line with previous work in this area (Rise et al., 2010). And finally – third, we hypothesized that exercise identity would decrease, on average, from post-intervention to 6-months follow-up without the structure provided by the supervised exercise training intervention.

Method

Design

The parent RCT utilized a fully crossed design in which participants were randomly assigned to engage in either vigorous or moderate intensity aerobic exercise (i.e., walking or running on a treadmill, elliptical machine, or arc-trainer in the laboratory at 75–85% or 55–65% of VO2max) for either a long or short duration (40 or 20 minutes) four times per week for 16-weeks. Participants responded to survey measures for the current study as part of a larger psychosocial battery administered as part of the parent trial at baseline, post-intervention, and 6-months post-intervention follow-up. The variables of interest for the current study were chosen a priori from the larger survey battery. While the primary trial was preregistered on ClinicalTrials.gov (NCT02032628), the analyses for the current study were not preregistered.

Participants

A total of 276 women were recruited from the Denver, ColoradoMetro Area to participate in the parent trial via flyers, online, and other advertisements in the community. To be eligible, participants had to be insufficiently physically active, defined in this study as < 60 minutes of moderate intensity exercise accumulated per week on average over the past 6-months. Participants also had to be female, between the ages of 30 and 45, menstruating regularly, non-smokers, willing to and physically capable of safely engaging in moderate-vigorous intensity exercise (i.e., without health contraindications as assessed by a study physician), willing to accept random assignment to condition, willing to provide blood and saliva samples for the parent trial’s outcome measures, and planning to remain in the Denver metro area for the 10-months duration of the study. Of the 276 women who enrolled in the parent trial, 136 remained enrolled throughout the 16-week exercise intervention and completed post-intervention measures (including the post-intervention measure of exercise identity), and 116 completed the 6-month follow-up survey. Additional details on participant flow throughout the intervention are provided in Figure 1. The current study includes results from the 136 individuals who completed the intervention, regardless of degree of engagement (i.e., attendance). The likelihood of completing the intervention (and thus, inclusion in analyses) was independent of condition assignment (all condition contrast p’s>.19). Recruitment began on 1/30/2014 and 6-month follow-up procedures were complete as of 8/17/2017; the trial ended once final assessments were complete.

Figure 1.

Figure 1.

Open in a new tab

CONSORT Flow Diagram

On average, participants were 37.54 years old (SD = 4.69) at the time of study enrollment. Sixty percent of the sample identified their race as White, 21.3% identified as Hispanic or Latino, 9.6% identified as Black or African American, 5.9% identified as Asian, 2.2% identified as mixed Race, and .07% identified as Native Hawaiian or Pacific Islander. The majority (40%) of participants reported a bachelor’s degree as their highest degree, followed by master’s degree (22.8%), some college (13.2%), associates or technical school degree (12.5%), doctoral degree (8.1%), and high school diploma or equivalent (4.7%). Most participants were overweight (BMI between 25 and 30; 41.5%) or obese (BMI over 30; 31.9%), and 29.9% were in the “normal” weight category (BMI between 18.5 and 25). See Table 1 for baseline participant descriptive statistics by study condition.

Table 1.

Participant Demographics and Baseline Characteristics

Participant Demographics and Baseline Characteristics
Condition
Characteristic Total N = 136 Mod/Short n = 32 Mod/Long n = 44 Vig/Short n = 32 Vig/Long n = 28
M (SD) Test Statistic p
Age (years) 37.62 (5.07) 37.66 (4.40) 37.56(5.29) 37.21 (4.17) F(3, 132) = 0.06 .98
BMI (kg/m2) 28.12 (5.77) 27.74 (5.54) 28.60 (5.40) 30.03 (5.99) F(3, 131) = 1.00 .40
Race & Ethnicity Χ2(18, N = 136) = 14.64 .68
n Asian 8 2 3 2 1
n Black 13 1 8 1 3
n Hispanic 29 7 8 8 6
n Mixed Race 3 1 1 1 0
n Native Hawaiian/Pacific Islander 1 1 0 0 0
n White 81 19 24 22 18
n Other 1 1 0 0 0
Baseline VO2max (ml/kg/min) 28.68 (6.37) 28.05 (5.97) 27.35 (5.44) 27.61 (5.60) F(3, 130) = 0.31 .82
Baseline minutes of exercise 17.34 (24.4) 16.02 (36.81) 20.31 (38.48) 14.82 (26.37) F(3, 132) = 0.16 .92
Open in a new tab

Note. Mod/Short = moderate intensity-short (20min) duration condition; Mod/Long = moderate intensity-long (40min) duration condition; Vig/Short = vigorous intensity-short duration condition; Vig/Long = vigorous intensity-long duration condition; BMI = body mass index; SD = standard deviation.

Measures

Demographics.

Demographic variables and baseline participant characteristics, including race, age, and education were measured at baseline.

Exercise Identity Scale.

At baseline (α = .87), post-intervention (α = .88), and 6-month follow-up (α = .90) exercise identity was measured with the Exercise Identity Scale (Anderson & Cychosz, 1994), a 9-item measure with a single factor structure examining the degree to which an individual identifies with exercise as an integral part of his or her self-concept. Example items include, “I consider myself an exerciser” and “Physical exercise is a central factor to my self-concept.” Importantly, the Exercise Identity Scale has been associated with exercise maintenance in other samples (Strachan et al., 2005). Responses are measured on a 7-point Likert scale ranging from 1 = Strongly Disagree to 7 = Strongly Agree.

Exercise motivation measures.

Intrinsic motivation for exercise was measured at each time point using the Intrinsic Motivation Inventory (IMI; McAuley, Duncan, & Tammen, 1989), a 21-item measure (e.g., “I think that physical activity is quite enjoyable”) with responses on a 7-point Likert scale ranging from 1 = Strongly Disagree to 7 = Strongly Agree (α = .89). Extrinsic motivation for exercise was measured at the 6-month follow-up using the 13-item Extrinsic Motivation Inventory (EMI; Lee & Diclemente, 2001). Participants rate their agreement with items such as “I would find a lot more time in my schedule to be physically active if someone were paying me for it,” scored on a 5-point Likert scale ranging from 1 = Strongly Disagree to 5 = Strongly Agree (α = .74).

TPB Constructs.

At each time point, participants described their intentions for exercise behavior in the next three months using three items (e.g., “I intend to do aerobic exercise as much as I can in the next month) scored on a 7-point scale (α = .78). Attitudes were measured with a 7-item, 7-point semantic differential scale (e.g, “For me, doing 30 minutes of more of aerobic exercise on at least five days during the week would be: Punishing/rewarding”) (α = .81). Norms were measured using a 6-item (e.g., “Most people who are important to me think I should do aerobic exercise”), 7-point scale (α = .88). Finally, perceived behavioral control was measured with a 9-item (e.g., “I feel confident I could do aerobic exercise even if I was in a bad mood), 7-point scale (α = .92).

Exercise Frequency.

At the 6-month follow-up1, the Godin Leisure-Time Exercise Questionnaire (GLTEQ; Godin & Shephard, 1985) and interviewer administered Stanford 7-Day Physical Activity Recall Assessment (PAR; Blair et al., 1985) were used to measure exercise frequency after the conclusion of the supervised exercise portion of the study. From the GLTEQ, we computed a total exercise score adjusted for exercise intensity based on the number of exercise sessions of mild, moderate, and vigorous intensity exercise participants reported completing in a typical week. Minutes of exercise during a typical week was also assessed with the question, “During a typical 7-Day period (a week), how many TOTAL minutes do you engage in physical activity?” Finally, from the PAR, we summed each participant’s total reported minutes of exercise of at least moderate intensity exercise from the past 7-days.

Attendance, exercise volume, and change in VO2max.

Participant VO2max was assessed at baseline and post-intervention via standard incremental treadmill testing procedures and a VO2max change score was computed for each participant. Participants’ attendance was recorded for each exercise session, and a total exercise volume score for the 16-week intervention (in kcal/kg/week) was computed for each participant based on their individual session attendance, condition assignment, and baseline VO2max scores. Per ACSM Guidelines (ACSM, 2013), the following formula was used to compute total exercise volume: ((absolute intensity-rest)/1000)*5.0*duration*frequency = volume. Thus, the formula used to calculate each individual participant’s total exercise volume was: (([.55 OR .75]*VO2max)-3.5)/1000)*5*[20 OR 40]*[sessions per week].

Procedures

Baseline.

At the baseline visit professional research assistants obtained participants’ written informed consent. Participants then completed an eligibility medical screening by a study physician. Eligible participants then completed baseline assessments via an online survey administered via REDCap (Harris et al., 2009). Participants were compensated $20 for this study appointment. At a second baseline visit, participants completed the graded exercise test to determine VO2max. The principal investigator/statistician used an online random number generator to generate the random allocation sequence. No blocking strategies were used to generate randomization; professional research assistants assigned participants to the next condition noted in the randomly generated list of conditions.

Supervised Exercise Visits.

Participants were expected to complete 4 supervised exercise sessions per week over the course of 16 weeks at the study exercise lab facility. Participants were compensated $160 for completing all 64 sessions, or $2.50 per session completed. At each visit, participants were fitted with a digital heart rate monitor. Trained staff were available at all times during these exercise sessions and helped to insure participants were exercising within their prescribed heart rate range. Participants exercised for either 20 minutes or 40 minutes depending on their duration condition assignment, and at either 55–65% or 75–85% of VO2max depending on their intensity condition assignment. Participants exercised either on stationary bikes, treadmills, or elliptical machines.

Post-intervention.

At the conclusion of the 16-week exercise intervention, participants completed post-intervention measures of exercise identity and other theory-based constructs as well as a second VO2max test.

Follow-up.

The focus of the final 6 months of the study was to naturalistically track the maintenance (or not) of exercise behavior following the supervised exercise component of the study. Participants were not given any specific exercise instructions following the conclusion of the supervised exercise visits. At 6 months post-intervention, participants returned to the lab and completed self-report measures of exercise behavior and theory-based constructs. Participants were compensated $30 for completing this study appointment.

Analysis Plan

The originally recruited sample size was selected to permit detection of analyses for primary outcome measures of the RCT that were not relevant to the current paper. Post-hoc power calculations were conducted with G*Power, based on a within-between repeated measures analysis comparing group means on repeated measures of identity pre-to-post intervention. Power was estimated to detect a small to moderate effect (f=.20) with alpha of .05 and power of .90 for the difference between four groups using two repeated measurements of identity and estimating a correlation of .5 between repeated measures. The sample size needed to achieve power of .90 was 96, less than the current analytic sample of 136 participants.

To examine the hypothesis that exercise identity would increase significantly across all groups, and the exploratory hypothesis that this change might be greater for participants in higher-volume conditions (e.g., vigorous intensity and long duration conditions), we conducted a repeated measures analysis of variance (ANOVA) with identity scores as the dependent variable, time (baseline vs. post-intervention) as a within-subjects factor, and intensity, duration, and their interaction as between-subjects factors. The main effect for time addresses whether change in identity from baseline to follow-up was different from zero, and the time by intensity, duration, and intensity*duration interactions address the exploratory question of moderation of identity change by the volume of exercise completed. We conducted additional repeated measures ANOVAs using the different assessments of participation in exercise behavior as independent variables—number of sessions attended, exercise volume completed, and change in VO2max over the intervention period, to further explore moderation of identity change by exercise behavior and change in fitness.

To examine associations between identity change and change in other psychosocial constructs, we first computed change scores for each variable (e.g., post-intervention identity – baseline identity), and then computed Pearson correlations between these change scores.

To examine the hypotheses that larger increases in identity would be associated with higher levels of follow-up exercise, we regressed the outcome variables of interest on the identity change score, controlling for intensity, duration, and their interaction, to examine whether larger changes in identity scores were associated with higher levels of exercise behavior and motivation at 6-months follow-up. Further, to examine whether change in identity explained variance in follow-up exercise over and above change in other TPB constructs, we first ran models predicting follow-up exercise behavior outcomes by change in TPB constructs (controlling for intensity, duration, and their interaction), and then examined the change in R2 when change in identity was added as a predictor.

Finally, to address the hypothesis that exercise identity would decline after the intervention, we conducted a regression with change in exercise identity from post-intervention to 6-month follow-up as the dependent variable and intensity, duration, and their interaction (contrast-coded) as independent variables. The intercept from this model allowed us to examine whether changes in exercise identity from post-intervention to follow-up were significantly different from zero, holding condition constant. To examine if follow-up identity remained higher than baseline-identity, we conducted a similar analysis, instead using change in identity from baseline to 6-month follow-up as the dependent variable.

Results

Descriptive Statistics

We present baseline descriptive statistics among the 136 participants who completed the 16-week exercise intervention and are therefore included in the current analyses examining pre-post intervention identity change. At baseline, participants reported exercising an average of 17.1 minutes per week during a typical week (SD = 32.42). On average, VO2max at baseline was 27.94 ml/kg/minute (SD = 5.83), indicating poor to fair baseline cardiorespiratory fitness for women in this age group (ACSM, 2013). At baseline, the mean exercise identity score was 3.27 (SD = 1.12). A score of 4 on the identity scale represented “neither agree nor disagree” to the identity-related statements; thus, on average, participants did not identify strongly as exercisers before the study, which would be expected, as participants were recruited on the basis that they were insufficiently physically active.

While the study procedures instructed participants to come in to the lab 4 times per week, not all participants had perfect attendance; on average, participants attended 51.04 (SD = 10.95) sessions out of 64, or 80% of sessions. For this reason, we included attendance as another relevant predictor/covariate to examine the effects of participant engagement in the intervention on identity change. There was no omnibus effect of condition on the number of sessions attended, F3,132=1.45, p=.23, change in VO2max, F3,132=0.24, p=.87, or self-reported minutes of exercise reported at follow-up, F3,105=2.17, p=.10.

Baseline identity for consented participants did not predict study completion (i.e., inclusion in current manuscript), z=−1.02, p=.30. Baseline identity was marginally positively associated with the number of sessions attended, b=1.54, p=.07. Baseline identity was not associated with change in VO2max, b=.12, p=.62. Further, these associations between baseline identity and these factors were not moderated by intensity, duration, or their interaction (p’s>.14)

Change in Identity Over Time and by Exercise Intensity and Duration

Identity means at baseline, post-intervention, and follow-up, overall and by intensity and duration, are presented in Table 2. The first hypothesis was that exercise identity would increase over the course of the intervention. The time effect in the repeated measures ANOVA indicated that on average, participants’ exercise identity increased by .78 points (95% CI [.56, .98]) after the intervention; this increase was significantly different from zero, t(128) = 9.94, p < .001.

Table 2.

Baseline, Post-Intervention, and 6-Month Follow-Up Descriptive Statistics Across Conditions

Baseline M (SD) Post-intervention M (SD) Follow-up M (SD)
Identity Scores
 Overall (N=136) 3.27 (1.12) 4.07 (1.21) 3.82 (1.37)
 Moderate intensity/20 min. (n=32) 3.30 (0.99) 4.00 (1.42) 3.78 (1.57)
 Moderate intensity/ 40 min. (n=44) 3.20 (1.15) 4.04 (1.25) 3.96 (1.32)
 Vigorous intensity/20 min. (n=32) 3.14 (1.09) 3.95 (1.02) 3.49 (1.20)
 Vigorous intensity/ 40 min. (n=28) 3.48 (0.00) 4.32 (1.10) 4.06 (1.40)
Intrinsic Motivation
 Overall 4.13 (0.89) 4.40 (0.91) 4.42 (0.97)
Extrinsic Motivation
 Overall 2.64 (0.59) 2.73 (0.67) 2.73 (0.66)
Open in a new tab

Note. Post-intervention refers to immediately after the 16-week exercise intervention; Follow-up refers to 6-month follow-up. Identity scores refer to the mean on the 9-item identity scale (Anderson & Cychosz, 1994), where responses are measured on a 7-point Likert scale ranging from 1 = Strongly Disagree to 7 = Strongly Agree. Intrinsic motivation scores refer to the mean on the 21-item Intrinsic Motivation Inventory (McAuley et al., 1989), where responses are measured on a 7-point Likert scale ranging from 1 = Strongly Disagree to 7 = Strongly Agree. Extrinsic scores refer to the mean on the 13-item Extrinsic Motivation Inventory where responses are measured on a 5-point Likert scale ranging from 1 = Strongly Disagree to 5 = Strongly Agree (Lee & Diclemente, 2001).

Our exploratory question asked whether changes in exercise identity would be moderated by intensity and/or duration such that those who completed the greatest volume of exercise over the course of the intervention (i.e., vigorous intensity or duration of exercise, or both) might experience the most change in identity. The repeated measures ANOVA showed no omnibus condition by time interaction, F(3, 128) = .18, p = .91, R2 = .004, and the individual one-degree of freedom tests of moderation by moderate versus vigorous intensity, long versus short duration, or their interaction indicated no significant differences (all p’s >.53).

For the analyses examining additional between-subjects measures of exercise completed, we found that change in identity was not moderated by the number of sessions attended, b=0.009, t(130) = 0.96, p = .33, R2 = .007. Additionally, change in identity was not associated with total exercise volume completed during the supervised exercise intervention in kcal/kg/week, b=−0.014, t(128) = −0.53, p = .59, R2 = .002. Finally, change in identity was not associated with changes in VO2max over time, b=0.030, t(119) = 0.084, p = .40, R2 = .006.

Associations Between Change in Identity, Change in Motivation and TPB Constructs, and Exercise Behavior Maintenance

Change in identity and change in theory-informed, motivation-based determinants of exercise.

Correlations between identity change, change in psychosocial constructs, and follow-up exercise behavior are presented in Table 3. Change in identity from baseline to post-intervention was positively correlated with change in each of the theory-informed, motivation-based constructs of interest: intentions, attitudes, perceived behavioral control, norms, extrinsic motivation, and intrinsic motivation.

Table 3.

Pearson correlations between identity change, change in motivation and Theory of Planned Behavior (TPB) constructs, and follow-up exercise behavior

2 3 4 5 6 7 8 9 10
1. Identity Δ
2. Intentions Δ 0.22*
3. Attitudes Δ 0 32*** 0.38***
4. PBC Δ 0.31*** 0.42*** 0.16
5. Norms Δ 0.19* 0.22* 0.21* 0.15
6. Extrinsic Δ 0.31*** 0.08 0.19* 0.01 0.03
7. Intrinsic Δ 0 47*** 0.18* 0.25** 0.37*** −0.1 0.22*
8. Follow-up minutes 0.22* 0.21* 0.06 0.08 0.16 −0.06 0.09
9. Follow-up PAR 0.08 0.12 0.1 −0.04 −0.06 −0.01 −0.23* 0 37***
10. Follow-up GLTEQ 0.21* 0.18 0.06 −0.02 0.19 −0.06 0.02 0.52*** 0.31**
Open in a new tab

Note. A=change (posttest score - pretest score). Follow-up minutes = self-reported minutes of exercise in a typical week, follow-up PAR = 7-day Physical Activity Recall moderate/vigorous intensity exercise minutes, follow-up GLTEQ = Godin Leisure-Time Exercise Questionnaire total intensity-adjusted exercise score.

***

p<.001

**

p<.01

*

p<.05

Exercise Behavior Maintenance.

Additionally, we examined the hypothesis that larger changes in identity would be associated with greater maintenance of exercise behavior at follow-up. The three measures of exercise behavior maintenance at follow-up (self-reported number of minutes exercised during a typical week, GLTEQ intensity-adjusted total exercise score, and PAR moderate intensity minutes) were moderately correlated with one another (see Table 3). At the 6-month follow-up, participants reported exercising an average of 78.54 minutes during a typical week (SD = 85.45). Identity change was significantly associated with self-reported minutes of exercise during a typical week at follow-up, controlling for condition. Specifically, a one-point increase on the identity measure after the intervention was associated with 15.88 more minutes of exercise per week on average at follow-up, b=15.88, t(100)=2.20, p=.030, R2 =.046. Additionally, identity change was significantly related to the intensity-adjusted total exercise score from the GLTEQ, b=3.84, t(94)=2.11, p=.0378, R2 = .046. However, identity change was not significantly associated with minutes of moderate exercise in the past seven days as measured by the PAR, b=10.09, t(95) =.786, p=.434, R2 = .006.

Further, we examined whether change in identity predicted variance in follow-up exercise behavior, over and above change in TPB constructs. In total, over and above condition, change in TPB constructs (attitudes, norms, perceived behavioral control, and intentions) explained an additional 9.7 percent of the variance in follow-up exercise minutes, F(4, 88)=2.35, p=.06. Adding identity change to the TPB model explained an additional 3.9% of the variance, F(1, 87)=3.5, p=.06. Change in TPB constructs explained 7.9 percent of the variance in the intensity-adjusted exercise score over and above condition, F(4, 82)=1.77, p=.14; adding identity change to the TPB model explained an additional 5.3% of the variance, F(1, 81)=4.58, p=.035. Lastly, change in TPB constructs explained 3.4 percent of the variance in PAR minutes over and above condition, F(4, 84)=0.86, p=.49. Adding identity change to the TPB model explained an additional 0.2% of the variance, F(1, 83)=0.18, p=.66.

Maintenance of exercise identity at follow-up.

Finally, we examined whether participants “maintained” their exercise identity at the 6-month follow-up, and whether condition moderated identity maintenance. Controlling for condition, identity at follow-up was significantly lower than identity at post-intervention, adjusted M difference = −0.29, 95% CI [0.09, 0.49], t(104) = 2.90, p = .004, and condition did not moderate changes in identity (all contrast code p’s >.31). However, regardless of condition, identity at 6-months follow-up was still significantly higher than identity at baseline, M difference = 0.55, 95% CI [0.78, 0.33], t(107) = 4.88, p < .001, and condition did not moderate baseline to follow-up change in identity (all contrast code p’s>.40) (See Table 2).

Discussion

The purpose of this study was to examine whether repetition of exercise behavior through participation in a 16-week exercise training intervention might influence exercise identity, and whether changes in exercise identity may in turn impact exercise motivation and longer-term exercise maintenance. On average, exercise identity was significantly increased post-intervention. Interestingly, the degree to which exercise identity increased did not depend on the volume of exercise completed as measured by intensity or duration condition, total number of exercise sessions completed during the 16-week intervention, or the total volume of exercise completed during the 16-week intervention (in kcal/kg/week). Change in exercise identity was also not associated with changes in VO2max over time. This suggests that engaging in regular, structured exercise for a prolonged period of time (in this case, 16 weeks) can significantly, positively influence exercise identity. Further, changes in exercise identity may occur even if exercise intensity is low and/or if exercise bouts are short (i.e., 20 minutes). Additionally, changes in exercise identity may occur independent of or preceding measurable physiological changes in physical fitness (i.e., VO2max). The findings suggest that repeatedly engaging in even modest amounts of exercise for a prolonged time period may promote increased exercise identity.

These findings are notable given that neither formation of exercise identity nor enhancement of exercise motivation were expressly targeted by the parent trial’s exercise training intervention. This suggests individuals might spontaneously incorporate health behavior engagement into their self-concept after regularly engaging in that behavior for some time. Few studies have examined how to change exercise identity over time (for exceptions, see Cooke et al., 2019; Husband et al., 2019); the present study is the first to our knowledge to examine whether exercise engagement alone (in a structured setting) is associated with a change in exercise identity.

We also found that pre/post-intervention changes in exercise identity were positively associated with self-report measures of exercise behavior at 6-months follow-up including minutes of exercise during a typical week and intensity adjusted total exercise score from the GLTEQ. However, change in exercise identity was not significantly associated with minutes of exercise as measured by the PAR. This is not entirely surprising, given the lower correlation between this measure and the other two self-report measures of exercise, and because the PAR only examines past-week exercise behavior rather “average” exercise in the past six months. Additionally, lack of consistency between findings using both the PAR and other self-report measures of exercise frequency has been observed in past research (Bryan et al., 2013).

The observed relationships between exercise identity and follow-up exercise behavior are consistent with results from correlational studies; our study extends prior work by using a longitudinal design (Rhodes et al., 2016). Additionally, while not entirely consistent across measures of exercise behavior, we found some evidence that exercise identity explained variance in behavior and intentions over and above TPB constructs. These findings are consistent with suggestions by Rhodes et al. (2016) that identity is more suited as a maintenance construct compared to traditional models of behavior adoption like the TPB, and support previous meta-analytic findings by Rise and colleagues (2010) that identity can improve the explanatory power of the TPB.

Change in exercise identity was also significantly, positively associated with change in determinants of exercise behavior, including intrinsic and extrinsic motivation, intentions, attitudes, perceived behavioral control (PBC), and norms. These findings are consistent with several prior studies that found significant associations between exercise identity and intrinsic motivation (e.g., Ntoumanis et al., 2018; Reifsteck et al., 2016; Strachan et al., 2013; Vlachopoulos et al., 2011), as well as more extrinsic forms of motivation (Strachan et al., 2013). As discussed, meta-analytic work has also found significant associations between identity and intentions, norms, attitudes, and PBC (Rise et al, 2010). Future work is needed to parse the directionality of these associations; as identity is hypothesized to be a maintenance construct (Rhodes, 2017), it may take longer to develop than other motivation constructs. We attempted to examine directionality of the relationships between identity and intrinsic/extrinsic motivation via a cross lagged panel model with our data; however, in that model, the stability coefficients were too high such that was little variance left to test crossed lagged paths; therefore, we did not feel that this was a particularly informative analysis to report in the current paper.

Our findings are consistent with theoretical work on identity and behavior. Social psychological theories suggest that engaging in a desired behavior with success may lead one to attribute that behavior to stable internal characteristics (Miller & Ross, 1975; Adams et al., 2005; Klesges et al., 2004), which is supported by the findings regarding change in exercise identity observed in the present study. In turn, identity theory would posit that once an individual defines themselves in terms of a category label (e.g., demonstrates an increase in exercise identity), that individual should be motivated to engage in behavior that is consistent with that identity (Stets & Burke, 2003; Strachan, Brawley, & Spink, 2015).

This study has some important limitations. First, the 6-month follow-up relied on self-report measurement of exercise, which may overestimate actual exercise engagement (Downs et al., 2014; Prince et al., 2008; Troiano et al., 2008). Future research examining the role of exercise identity in exercise maintenance might consider measuring exercise behavior more intensively and objectively, for example, through the use of a wearable fitness tracker or ecological momentary assessment (EMA; Dunton, Rothman, Leventhal, & Intille, 2019). Additionally, rather than measuring the full spectrum of autonomous motivation as outlined by SDT (Ryan & Deci, 2000), we only measured extrinsic and intrinsic motivation for exercise, limiting our ability to fully examine the relationships between identity and other types of motivation (e.g., identified and integrated motivation). Due to the nature of the parent trial, our sample was comprised entirely of women, so it is unclear whether results would generalize to men. This study was also conducted in the context of a supervised exercise intervention; therefore, our findings may not generalize to a more ecologically valid context. The changes in exercise identity we observed occurred in a sample of participants who voluntarily signed up to participate in a research study meant to increase exercise behavior. Our study did not include a no-exercise control group; therefore, the study design does not rule out that it is possible (though, we believe unlikely) that changes in exercise identity may have occurred over time even without any changes in exercise behavior. Finally, analyses for the current study were not preregistered. It is possible that a subset of these results are reflective of false positives, though the robust associations between change in identity and change in other theory-informed, motivation-based determinants of exercise provides some confidence that the observed associations may not be false positives. Importantly, however, if we were to adjust our p-values for the analyses examining associations between change in identity and follow-up exercise behavior accounting for the fact that we used three different measures of exercise behavior at follow-up (e.g., lower alpha level to .017), our observed associations for two out of the three measures would no longer be considered significant.

Despite these limitations, our study has several notable strengths. First, because this study was conducted in the context of a supervised exercise intervention, we were able to objectively verify that participants regularly engaged in exercise, at a pre-specified intensity and duration, during the intervention portion of the study. Second, our study’s longitudinal design addresses a crucial gap in the literature as the link between identity and behavior has most often been examined cross-sectionally (Rhodes et al., 2016). Thus, we have provided evidence that formation of exercise identity over time is linked to behavior maintenance at a later time point. Further, our study utilized a relatively diverse sample of women recruited from the community, which is a strength as previous studies on exercise identity have primarily been conducted in college-based convenience samples (Rhodes et al., 2016).

Overall, this study provides evidence that repetition of exercise behavior can lead formerly insufficiently physically active women to begin to identify as an “exerciser.” Identifying demographic or situational factors that might predict the degree to which individuals are willing and able to incorporate exercise behavior into their identity is an important area for future research. Likewise, future studies might examine the “threshold” of exercise behavior necessary for the behavior to become part of one’s identity.

Supplementary Material

1

Highlights.

  • Previously-inactive women who completed a 16-week exercise intervention significantly increased their “exercise identity” over time

  • Increases in exercise identity weren’t related to the amount or intensity of exercise completed

  • Women whose exercise identity increased more were more likely to report exercising more 6 months after the intervention

Acknowledgements

Preparation of this manuscript was supported in part by the National Cancer Institute (R01CA179963, PI: Bryan) and the National Institute of Mental Health (T32 MH073553, PI: Bruce, Fellow: Stevens). The content is solely the responsibility of the authors and does not represent the official views of the National Institutes of Health.

Footnotes

Conflict of interest: The authors declare that they have no conflict of interest.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

1

Physical activity was also measured at baseline. This baseline measurement was used to confirm participants’ status as insufficiently active at the beginning of the intervention. As expected, participants reported engaging in low levels of exercise at baseline. Given this, we chose not to examine change in exercise as a dependent variable, since these values were essentially equivalent to follow-up exercise values; correlations between change in exercise and the raw follow-up exercise value were r=.922 (self-reported minutes), r=.920 (GLTEQ intensity-adjusted exercise score), and r=.947 (PAR) respectively.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Ajzen I, & Madden T (1986). Prediction of goal-directed behavior: Attitudes, intentions, and perceived behavioral control. Journal of Experimental Social Psychology. /citations?view_op=view_citation&continue=/scholar%3Fhl%3Den%26start%3D40%26as_sdt%3D0,6%26scilib%3D1&citilm=1&citation_for_view=IGoukmgAAAAJ:M05iB0D1s5 AC&hl=en&oi=p [Google Scholar]
  2. Anderson DF, & Cychosz CM (1994). Development of An Exercise Identity Scale. Perceptual and Motor Skills, 78(3), 747–751. 10.1177/003151259407800313 [DOI] [PubMed] [Google Scholar]
  3. Blair SN, Haskell WL, Ho P, Paffenbarger RS, Vranizan KM, Farquhar JW, & Wood PD (1985). Assessment of habitual physical activity by a sevenday recall in a community survey and controlled experiments. American Journal of Epidemiology, 122(5), 794–804. 10.1093/oxfordjournals.aje.a114163 [DOI] [PubMed] [Google Scholar]
  4. Borland R (2017). CEOS Theory: A Comprehensive Approach to Understanding Hard to Maintain Behaviour Change. Applied Psychology: Health and Well-Being, 9(1), 3–35. 10.1111/aphw.12083 [DOI] [PubMed] [Google Scholar]
  5. Bryan AD, Magnan RE, Hooper AEC, Ciccolo JT, Marcus B, & Hutchison KE (2013). Colorado stride (COSTRIDE): Testing genetic and physiological moderators of response to an intervention to increase physical activity. International Journal of Behavioral Nutrition and Physical Activity, 10(1), 139 10.1186/1479-5868-10-139 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Caldwell AE, Masters KS, Peters JC, Bryan AD, Grigsby J, Hooker SA, Wyatt HR, & Hill JO (2018). Harnessing centred identity transformation to reduce executive function burden for maintenance of health behaviour change: the Maintain IT model. Health Psychology Review, 12(3), 231–253. 10.1080/17437199.2018.1437551 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chatzisarantis NLD, Hagger MS, & Wang JCK (2008). An Experimental Test of Cognitive Dissonance Theory in the Domain of Physical Exercise. Journal of Applied Sport Psychology, 20(1), 97–115. 10.1080/10413200701601482 [DOI] [Google Scholar]
  8. Cooke LM, Duncan LR, Deck SJ, Hall CR, & Rodgers WM (2019). An examination of changes in exercise identity during a mental imagery intervention for female exercise initiates. International Journal of Sport and Exercise Psychology. 10.1080/1612197X.2019.1593216 [DOI] [Google Scholar]
  9. Cooper J, & Feldman LA (2020). Helping the “couch potato”: A cognitive dissonance approach to increasing exercise in the elderly. Journal of Applied Social Psychology, 50(1), 33–40. [Google Scholar]
  10. Downs A, Van Hoomissen J, Lafrenz A, & Julka DL (2014). Accelerometer-Measured Versus Self-reported Physical Activity in College Students: Implications for Research and Practice. Journal of American College Health, 62(3), 204–212. 10.1080/07448481.2013.877018 [DOI] [PubMed] [Google Scholar]
  11. Dunton GF, Rothman AJ, Leventhal AM, & Intille SS (2019). How intensive longitudinal data can stimulate advances in health behavior maintenance theories and interventions. Translational Behavioral Medicine. 10.1093/tbm/ibz165 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Ekkekakis P, Parfitt G, & Petruzzello SJ (2011). The pleasure and displeasure people feel when they exercise at different intensities. Sports Medicine, 41(8), 641–671. [DOI] [PubMed] [Google Scholar]
  13. Gardiner CK, & Bryan AD (2017). Monetary Incentive Interventions Can Enhance Psychological Factors Related to Fruit and Vegetable Consumption. Annals of Behavioral Medicine, 51(4), 599–609. 10.1007/s12160-017-9882-4 [DOI] [PubMed] [Google Scholar]
  14. Gelman S, & Heyman G (1999). Carrot-eaters and creature-believers: The effects of lexicalization on children’s inferences about social categories. Psychological Science. /citations?view_op=view_citation&continue=/scholar%3Fhl%3Den%26start%3D80%26as_sdt%3D0,6%26scilib%3D1&citilm=1&citation_for_view=IGoukmgAAAAJ:L8Ckcad2t8M C&hl=en&oi=p [Google Scholar]
  15. Godin G, & Shephard R (1985). A simple method to assess exercise behavior in the community. Canadian Journal of Applied Sport Sciences. Journal …. /citations?view_op=view_citation&continue=/scholar%3Fhl%3Den%26start%3D40%26as_sdt%3D0,6%26scilib%3D1&citilm=1&citation_for_view=IGoukmgAAAAJ:pqnbT2bcN3 wC&hl=en&oi=p [PubMed] [Google Scholar]
  16. Guérin E, Strachan S, & Fortier M (2019). Exercise and well-being: Relationships with perceptions of exercise identity-behaviour consistency, affective reactions to exercise and passion. International Journal of Sport and Exercise Psychology, 17(5), 445–458. 10.1080/1612197X.2017.1421681 [DOI] [Google Scholar]
  17. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, & Conde JG (2009). Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. Journal of Biomedical Informatics, 42(2), 377–381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Heyward VH, & Gibson A (2014). Advanced fitness assessment and exercise prescription 7th edition. Human kinetics. [Google Scholar]
  19. Husband CJ, Wharf-Higgins J, & Rhodes RE (2019). A feasibility randomized trial of an identity-based physical activity intervention among university students. Health Psychology and Behavioral Medicine, 7(1), 128–146. 10.1080/21642850.2019.1600407 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kendzierski D, & Morganstein MS (2009). Test, revision, and cross-validation of the physical activity self-definition model. Journal of Sport and Exercise Psychology, 31(4), 484–504. 10.1123/jsep.31.4.484 [DOI] [PubMed] [Google Scholar]
  21. Lee RE, & Diclemente CC (2001). Extrinsic And Intrinsic Motivation: Which Is Important For Exercise? Medicine &amp, 33(5), S112 https://insights.ovid.com/medicine-science-sports-exercise/mespex/2001/05/001/extrinsic-intrinsic-motivation/633/00005768 [Google Scholar]
  22. Marcus BH, Forsyth LH, Stone EJ, Dubbert PM, McKenzie TL, Dunn AL, & Blair SN (2000). Physical activity behavior change: issues in adoption and maintenance. Health Psychology, 19(1S), 32. [DOI] [PubMed] [Google Scholar]
  23. Marcus BH, Williams DM, Dubbert PM, Sallis JF, King AC, Yancey AK, Franklin BA, Buchner D, Daniels SR, Claytor RP, American Heart Association Council on Nutrition, Physical Activity, and Metabolism (Subcommittee on Physical Activity), American Heart Association Council on Cardiovascular Disease in the Young, & Interdisciplinary Working Group on Quality of Care and Outcomes Research. (2006). Physical activity intervention studies: what we know and what we need to know: a scientific statement from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism (Subcommittee on Physical Activity); Council on Cardiovascular Disease in the Young; and the Interdisciplinary Working Group on Quality of Care and Outcomes Research. Circulation, 114(24), 2739–2752. 10.1161/CIRCULATIONAHA.106.179683 [DOI] [PubMed] [Google Scholar]
  24. McAuley E, Duncan T, & Tammen VV (1989). Psychometric Properties of the Intrinsic Motivation Inventory in a Competitive Sport Setting: A Confirmatory Factor Analysis. Research Quarterly for Exercise and Sport, 60(1), 48–58. 10.1080/02701367.1989.10607413 [DOI] [PubMed] [Google Scholar]
  25. Medicine AC of S. (2013). ACSM’s guidelines for exercise testing and prescription. Lippincott Williams & Wilkins. [Google Scholar]
  26. Miller DT, & Ross M (1975). Self-serving biases in the attribution of causality: Fact or fiction? Psychological Bulletin, 82(2), 213–225. 10.1037/h0076486 [DOI] [Google Scholar]
  27. Ntoumanis N, Stenling A, Thøgersen-Ntoumani C, Vlachopoulos S, Lindwall M, Gucciardi DF, & Tsakonitis C (2018). Longitudinal associations between exercise identity and exercise motivation: A multilevel growth curve model approach. Scandinavian Journal of Medicine & Science in Sports, 28(2), 746–753. 10.1111/sms.12951 [DOI] [PubMed] [Google Scholar]
  28. O’Donovan G, Owen A, Bird SR, Kearney EM, Nevill AM, Jones DW, & Woolf-May K (2005). Changes in cardiorespiratory fitness and coronary heart disease risk factors following 24 wk of moderate- or high-intensity exercise of equal energy cost. Journal of Applied Physiology (Bethesda, Md. : 1985), 98(5), 1619–1625. 10.1152/japplphysiol.01310.2004 [DOI] [PubMed] [Google Scholar]
  29. Physical activity fact sheet. (2017). In World Health Organization. World Health Organization. [Google Scholar]
  30. Prince SA, Adamo KB, Hamel M, Hardt J, Connor Gorber S, & Tremblay M (2008). A comparison of direct versus self-report measures for assessing physical activity in adults: a systematic review. International Journal of Behavioral Nutrition and Physical Activity, 5(1), 56 10.1186/1479-5868-5-56 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Reifsteck EJ, Gill DL, & Labban JD (2016). “Athletes” and “exercisers”: Understanding identity, motivation, and physical activity participation in former college athletes. Sport, Exercise, and Performance Psychology, 5(1), 25–38. 10.1037/spy0000046 [DOI] [Google Scholar]
  32. Rhodes RE (2017). The Evolving Understanding of Physical Activity Behavior (pp. 171–205). 10.1016/bs.adms.2016.11.001 [DOI] [Google Scholar]
  33. Rhodes RE, Kaushal N, & Quinlan A (2016). Is physical activity a part of who I am? A review and meta-analysis of identity, schema and physical activity. Health Psychology Review, 10(2), 204–225. 10.1080/17437199.2016.1143334 [DOI] [PubMed] [Google Scholar]
  34. Rise J, Sheeran P, & Hukkelberg S (2010). The role of self-identity in the theory of planned behavior: A meta-analysis. Journal of Applied Social Psychology, 40(5), 1085–1105. 10.1111/j.1559-1816.2010.00611.x [DOI] [Google Scholar]
  35. Rodgers WM, Hall CR, Duncan LR, Pearson E, & Milne MI (2010). Becoming a regular exerciser: Examining change in behavioural regulations among exercise initiates. Psychology of Sport and Exercise, 11(5), 378–386. 10.1016/j.psychsport.2010.04.007 [DOI] [Google Scholar]
  36. Rothman AJ, Baldwin AS, Hertel AW, & Fuglestad PT (2004). Disentangling behavioral initiation and behavioral maintenance In Handbook of Self-Regulation. Research, Theory, and Applications; Guilford Press: New York, NY, USA: (pp. 130–148). [Google Scholar]
  37. Rothman AJ, Gollwitzer PM, Grant AM, Neal DT, Sheeran P, & Wood W (2015). Hale and Hearty Policies: How Psychological Science Can Create and Maintain Healthy Habits. Perspectives on Psychological Science, 10(6), 701–705. 10.1177/1745691615598515 [DOI] [PubMed] [Google Scholar]
  38. Stets JE, & Burke PJ (2003). A sociological approach to self and identity. Handbook of Self and Identity, 128–152. http://wat2146.ucr.edu/papers/02a.pdf [Google Scholar]
  39. Strachan S, Brawley L, & Spink K (2015). Self-regulatory efficacy’s role in the relationship between exercise identity and perceptions of and actual exercise behaviour. Psychology of Sport and … /citations?view_op=view_citation&continue=/scholar%3Fhl%3Den%26start%3D20%26as_sdt%3D0,6%26scilib%3D1&citilm=1&citation_for_view=IGoukmgAAAAJ:bFI3QPDXJZ MC&hl=en&oi=p [Google Scholar]
  40. Strachan SM, Fortier MS, Perras MGM, & Lugg C (2013). Understanding variations in exercise-identity strength through identity theory and self-determination theory. International Journal of Sport and Exercise Psychology, 11(3), 273–285. 10.1080/1612197X.2013.749005 [DOI] [Google Scholar]
  41. Strachan SM, & Whaley DE (2013). How aspects of the self influence exercise behavior. Routledge Handbook of Physical Activity and Mental Health, 212. [Google Scholar]
  42. Strachan S, Woodgate J, Brawley L, & Tse A (2005). The Relationship of Self‐Efficacy and Self‐Identity to Long‐Term Maintenance of Vigorous Physical Activity. Journal of Applied Biobehavioral Research, 10(2), 98–112. [Google Scholar]
  43. Swain DP, & Franklin B. a. (2006). Comparison of cardioprotective benefits of vigorous versus moderate intensity aerobic exercise. The American Journal of Cardiology, 97(1), 141–147. 10.1016/j.amjcard.2005.07.130 [DOI] [PubMed] [Google Scholar]
  44. Teixeira PJ, Carraça EV, Markland D, Silva MN, & Ryan RM (2012). Exercise, physical activity, and self-determination theory: A systematic review. International Journal of Behavioral Nutrition and Physical Activity, 9(1), 78 10.1186/1479-5868-9-78 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Timmons JF, Griffin C, Cogan KE, Matthews J, & Egan B (2020). Exercise Maintenance in Older Adults 1 Year After Completion of a Supervised Training Intervention. Journal of the American Geriatrics Society, 68(1), 163–169. 10.1111/jgs.16209 [DOI] [PubMed] [Google Scholar]
  46. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, & McDowell M (2008). Physical Activity in the United States Measured by Accelerometer. Medicine & Science in Sports & Exercise, 40(1), 181–188. 10.1249/mss.0b013e31815a51b3 [DOI] [PubMed] [Google Scholar]
  47. Vlachopoulos SP, Kaperoni M, & Moustaka FC (2011). The relationship of self-determination theory variables to exercise identity. Psychology of Sport and Exercise, 12(3), 265–272. 10.1016/j.psychsport.2010.11.006 [DOI] [Google Scholar]
  48. Walton GM, & Banaji MR (2004). Being What You Say: The Effect of Essentialist Linguistic Labels on Preferences. Social Cognition, 22(2), 193–213. 10.1521/soco.22.2.193.35463 [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1662296-supplement-1.docx (14.2KB, docx)

RESOURCES