Motivational Correlates of Exercise Behavior Among College Students

The purpose of this research was to explore the relations between basic psychological needs 49 satisfaction (autonomy, relatedness, perceived competence), intrinsic motivation, attraction 50 toward exercise, and exercise behavior among college students. In this study, 128 participants 51 (including 91 women and 36 men, mean age: 24 ± 7 years) responded to a questionnaire 52 assessing basic psychological needs satisfaction (autonomy, competence, relatedness), intrinsic 53 motivation, attraction (vs. antipathy) toward exercise, and exercise behavior. Frequency of 54 aerobic exercise, frequency of resistance exercise, and total aerobic exercise behavior are 55 positively associated with autonomy, competence, relatedness, intrinsic motivation, and 56 attraction toward exercise. Three exploratory mediation analyses suggest that attraction (vs. 57 antipathy) toward exercise mediates the relation between intrinsic motivation and exercise 58 behavior. Taken together, these data support and extend previous research on the importance of 59 motivationally relevant variables, including autonomy, competence, relatedness, intrinsic 60 motivation, and affective exercise experiences

disease. Yet, data indicate that only one out of every four adults in the US meets combined 78 aerobic and muscle-strengthening activity guidelines, with males achieving the guidelines at 79 higher rates than females (Bennie et al., 2019). 80 When moving from adolescence to early adulthood, students are faced with changes in 81 their physical, social, psychological, and structural domains, which play a role in their perception 82 of barriers and motivational regulation towards physical activity (Diehl et al., 2018). Since 83 physical activity is associated with better health and a decreased risk for chronic illnesses, 84 researchers have recognized the importance of motivating college students to participate and 85 maintain physical activity behavior (Gao et al., 2012). Therefore, there is a need to understand 86 motivational factors that may influence the exercise behavior of college students. To better 87 understand motivational factors, we turn to self-determination theory and affective exercise 88 experiences. 89 affective experiences and with higher forms of self-determined motivation (Ntoumanis, 2005; 118 ) and adherence to physical rehabilitation (Chan et al., 2009). Perceived 119 competence appears to be a meaningful predictor of exercise adherence (Vlachopoulos & 120 Neikou, 2007). In sum, needs satisfaction is a key component in motivation and may affect 121 affective experiences, both of which may influence exercise behavior. 122 Meta-analytic evidence has highlighted the meaningful relation between self-123 determination theory constructs (e.g., satisfaction of basic psychological needs, intrinsic 124 motivation) and exercise behavior (Teixeira et al., 2012). In their review, Rhodes  important predictor of behavior and conceptualized a dual-process framework that highlights the 138 importance of affective, evolutionarily primitive processes as an influence on behavior, in 139 addition to more rational, deliberative processes (e.g., consideration of benefits of an active 140 lifestyle, and consequences of a sedentary lifestyle. The affective-reflective theory of physical 141 inactivity, a dual-process framework, was formalized by Brand and Ekkekakis (2018 reflective, deliberate processes and automatically activated, heuristic processes. Intrinsic 163 motivation for exercise may also be related to attraction toward exercise, core affective exercise 164 experiences, and antecedent cognitive appraisals. People with more intrinsic and self-determined 165 motivations may have more attraction toward exercise, and in turn may engage in more exercise 166 behavior. Examination of items measuring each construct suggests that attraction toward exercise 167 (e.g., "Exercise is high on my priority list...", "Exercise is a tempting activity…", "I would 168 choose exercise over most other activities…") is arguably more likely to be influenced by 169 intrinsic motivation (e.g., exercising for fun, enjoying exercise sessions, getting pleasure and 170 satisfaction from exercise) than the other way around. This highlights the possibility that basic 171 psychological needs for exercise and intrinsic motivation have a strong affective component. 172 Thus, there is a need to study attraction toward exercise and intrinsic motivation together. 173

174
The purpose of this research was to add to the literature by further examining the 175 association between exercise behavior and motivational variables, including autonomy, 176 competence, relatedness, intrinsic motivation, and attraction (vs. antipathy) toward exercise. In 177 addition, several exploratory mediation analyses were conducted. These included analyses to 178 determine whether attraction (vs. antipathy) toward exercise mediated the relation between 179 intrinsic motivation and exercise behavior. 180

181
This study was preregistered (https://aspredicted.org/cw7mu.pdf). Predictor and 182 exploratory variables included autonomy, relatedness, competence, intrinsic motivation, and 183 attraction (vs. antipathy) toward exercise. The dependent variable was student exercise behavior. 184 Exercise behavior was measured in three categories, namely aerobic exercise frequency, 185 resistance exercise frequency, and total aerobic exercise. 186

Measures and Instruments 187
There was one online survey used in this study to collect data using Qualtrics (Provo, 188 UT). The survey included items to assess demographic characteristics (i.e., age, education level, 189 gender, etc.) and four questionnaires. We assessed gender by asking "What is your gender 190 identity?". Response options included "man", "woman", "I prefer not to say", and "I would 191 rather describe". 192

Basic Psychological Needs Satisfaction 193
The fulfillment of students' basic psychological needs (i.e., autonomy, competence, and 194 relatedness) in exercise was measured using the Basic Psychological Needs in Exercise Scale 2004). In this study, internal consistency of the intrinsic regulation subscale was high (α = .926). 214

Affective Experiences 215
To measure affective experiences the Affective Exercise Experiences (AFFEXX) 216 Questionnaire was used to assess affective exercise experiences. This scale includes several 217 statements with bipolar answers on each side that are separated by a 7-point response scale. The 218 primary subscale of interest in this study was the attraction-antipathy subscale, which includes 219 items such as "Exercise is something I look forward to" vs. "Exercise is something I dread". 220 The questionnaire was modified in this study to focus directly on mode (i.e., aerobic or 227 resistance), frequency, and duration of exercise. This section was composed of questions used to 228 assess aerobic and resistance frequency as well as overall, total aerobic exercise (i.e., "How 229 many minutes of planned, purposeful cardiorespiratory or aerobic exercise have you completed 230 in the last 7 days? Do not count activities such as walking to work or completing household 231 chores"). The response format included an open-ended textbox where participants could enter 232 their response. 233 Aerobic exercise frequency was assessed using the question "In the past 7 days, how 234 many days did you engage in planned, purposeful cardiorespiratory or aerobic exercise? Do not 235 count activities such as walking to work or completing household chores." Resistance exercise 236 frequency was assessed with the question "In the past 7 days, how many days did you complete 237 muscle-strengthening activities, such as weightlifting, strength training, or resistance training?" 238 Response options included 0, 1, 2, 3, 4, 5, 6, or 7, corresponding to anywhere from 0 days to 7 239 days of aerobic or resistance exercise. We intended to measure frequency as well as overall 240 behavior (minutes) because motivational variables may differently impact frequency and overall 241 behavior. For example, motivational variables may influence choice of whether to exercise today 242 or not (frequency) more than the decision to continue exercising once the task has been initiated 243 (total behavior).  Participants clicked on a link to access the informed consent form and the survey. Once 254 consent was provided, students were prompted with questions on their demographic 255 characteristics as well as the questions to assess their basic psychological needs, intrinsic 256 motivation, attraction (vs. antipathy) toward exercise, and exercise behavior. Upon completion of 257 the survey, participants were redirected to another survey where they provided their email to be 258 compensated with $5.00 Amazon electronic gift cards. 259 Sample Size Calculation

260
The study was launched with a focus on college students at least 18 years of age. The 261 revised analysis that is presented was not based on that initial power calculation. Based on 262 practical and resource considerations (Lakens, 2022), we aimed to recruit between 120 and 150 263 individuals to balance resource constraints. 264 Preregistration, Data Processing, and Data Availability

265
The preregistration, data, data analysis, and data processing notes are available at 266 https://osf.io/y5guj/. Deviations for preregistration are reported in the discussion. 267 All data were cleaned and analyzed using IBM SPSS version 28 (Armonk, NY), JASP 268 (JASP Team, 2022) and jamovi (The jamovi, 2022). Significance levels were set to p < .05. 269 Before beginning data analysis, the data were cleaned and processed. This included the removal 270 of ten responses from people who never completed the consent form (and never provided any 271 additional data), 22 participants who consented but never provided additional data, and 7 272 participants who did not complete the survey. Further, outliers were identified using Tukey's 273 fences and invalid responses were identified, flagged, and removed from relevant analyses and 274 the questionnaire data was appropriately scored. The full notes on the data processing procedure 275 Hawaiian or Pacific Islander participant, and 3 participants identifying as "Other", which 285 included one "European" and one "Palestinian/Arabic" participant. Finally, most were currently 286 undergraduate students with 91.41% having earned their high school diploma (or equivalent), 287 some college, or Associate degree (2-year degree). 11 participants completed their Bachelor 288 degree or Master degree. Ten participants indicated extreme amounts of exercise behavior (more 289 than 433 minutes per week), which was determined using Tukey's fences (IQR * 1.5). These 290 participants were eliminated from further analysis. 291

Motivational correlates of exercise behavior 293
A series of bivariate correlation analyses were performed to determine the relations between 294 motivational variables (autonomy, competence, relatedness, intrinsic motivation, and attraction 295 [vs. antipathy] toward exercise) and exercise behavior (frequency of aerobic exercise, total 296 aerobic exercise, and frequency of resistance exercise). The assumption of normality was 297 violated in multiple cases and thus Spearman's rho is reported for each correlation. 298

Exploratory mediation analyses 299
A series of three exploratory mediation analyses were completed to test whether attraction (vs. For each mediation analysis, 95% confidence intervals were calculated using 5000 bias-corrected 305 bootstrapped samples. If the 95% bootstrap confidence interval of an indirect effect did not 306 include 0, it was considered statistically significant. 307

309
Results of the correlation analyses used to determine the relations between motivational 310 variables and exercise behavior are presented in Table 1. Notably, motivational constructs were 311 also consistently correlated with each other (  This study was initially conceptualized with a focus on gender differences in motivation 378 and physical activity. Following reviewer and editor comments, we removed discussion and 379 analyses focusing on gender. The analysis was simplified by removing gender to determine the 380 correlations between the basic psychological needs (autonomy, competence, and relatedness), 381 intrinsic motivation, attraction (vs. antipathy) toward exercise, and exercise behavior (frequency 382 of aerobic exercise, total aerobic exercise, and frequency of resistance exercise). Although, it 383 was specified that attraction (vs. antipathy) would be used in mediation analyses, details of the 384 mediation analyses were not included (e.g., predictor variables), making it difficult to label the 385 mediation analyses as confirmatory. 386 Following reviewer comments, we simplified analyses here and removed perceived 387 variety in exercise as a variable. This is partly because of incomplete data and researcher error 388 related to the perceived variety in exercise questionnaire (we did have an error in the survey) 389 (Sylvester et al., 2014a(Sylvester et al., , 2014b. Finally, we did not anticipate all of the data cleaning steps noted 390 in the supplementary material. For example, we did not anticipate that people would respond 391 with a range for self-reported physical activity (e.g., 6-10 hours), which required manual 392 recoding. 393

394
Autonomy, competence, and relatedness were consistently associated with exercise 395 behavior, whether behavior was quantified as frequency of aerobic exercise, frequency of 396 resistance exercise, or total aerobic exercise behavior. This adds to existing literature, which has 397 shown mixed associations between autonomy and exercise; it also contrasts with prior literature,  Table 2 reflective and automatically activated processes. We suggest that the satisfaction of basic needs 426 (i.e., autonomy, competence, and relatedness) may enhance intrinsic motivation, which in turn 427 may enhance attraction toward exercise and ultimately exercise behavior. The data presented 428 here highlight the possibility that attraction-antipathy might be partially influenced by intrinsic 429 motivation regulations, theorized here to be a reflective process (i.e., dependent upon cognitive 430 appraisals). As indicated in our results, intrinsic motivation predicts greater attraction toward 431 exercise, which in turn predicts more frequent aerobic exercise, more total aerobic exercise, and 432 more frequent resistance exercise. This suggests the possibility that satisfying basic 433 psychological needs and enhancing autonomous and self-determined forms of motivation may 434 also increase attraction toward exercise. It should be noted that in this study, intrinsic motivation 435 and attraction toward exercise were highly correlated (Spearman's rho = .802, p < .001), which 436 reinforces attraction toward exercise as a highly motivationally relevant variable and may 437 suggest that intrinsic motivation is driven primarily by affective exercise experiences. Further 438 researchers should investigate this possibility. 439

Limitations 440
This study is not without limitations. We mentioned deviations from the preregistration, 441 above. In addition, this study was based on cross-sectional data. Using cross-sectional data in is not recommended (Thoemmes, 2015), and thus we are not presenting the mediation models 449 testing whether intrinsic motivation mediates the relation between attraction toward exercise and 450 exercise behavior. We theorize that attraction toward exercise is a direct determinant of exercise 451 behavior, and suggest that intrinsic motivation (i.e., exercising for fun, enjoyment, because 452 exercise is pleasurable, and because of pleasure and satisfaction) increases attraction toward 453 exercise. Further, we suggest that satisfying basic psychological needs (i.e., increasing 454 autonomy, competence, and relatedness) may increase intrinsic motivation, and in turn attraction 455 toward exercise. This is supported by further exploratory regression analyses, which indicate that 456 autonomy, competence, and relatedness explain 46.5% of the variance in intrinsic motivation. 457 However, more robust data with more complex research designs (e.g., longitudinal studies, 458 randomized controlled trials) are warranted to test these hypotheses. 459 We also recommend that future researchers more specifically assess self-determined 460 motivations (e.g., autonomy, competence, relatedness, intrinsic motivation) as it relates to 461 aerobic exercise and resistance exercise separately, and we appreciate the reviewer for this 462 suggestion. Finally, we recommend that future researchers explore these questions using device-463 based measures to complement self-reported physical activity. 464

Conclusions 465
Using preregistered measures and methods, further evidence is provided to support the 466 importance of basic psychological needs satisfaction, intrinsic motivation, and attraction (vs. 467 antipathy) toward exercise to understand exercise behavior. We recommend further replication 468 attempts on these results using fully preregistered methods. At present, we recommend that 469 researchers and practitioners continue investigating how to (a) promote psychological needs 470 satisfaction to maximize autonomy, competence, and relatedness, (b) enhance intrinsic