Today, one in eight Americans is older than 65. These older adults can be expected to have, on average, an additional 19 years of life ahead of them (Administration on Aging, 2009). The average life expectancy is now approaching 84 years in the United States (Administration on Aging, 2009). In addition to living longer, older Americans are also living healthier than in previous generations (Stewart, Cutler, & Rosen, 2013). However, as the population ages, the rates of chronic illness and disability are expected to increase significantly; thus, demands on the health care system can be expected to increase accordingly.
Exercise participation is recognized as one of the most important health behaviors for the prevention of chronic disease and health promotion among older adults (National Institute on Aging [NIA], 2015). Older adults who exercise have longer life expectancies, increased independence, less frailty, reduced health care costs, as well as reduced risk for chronic disease (NIA, 2015). Physical exercise and the resulting increased energy expenditure are important for health-related outcomes, including morbidity and mortality, as well as reducing the rate of overweight and obesity (Edwards, 2006). Sedentary behavior is linked to health-related problems such as osteoporosis, obesity, type II diabetes mellitus, cardiovascular disease, and various forms of cancer (NIA, 2015). Despite the known benefits of physical activity, approximately 85% of older adults fail to achieve the amount of physical exercise recommended in the Centers for Disease Control and Prevention's (CDC; 2015b) physical activity guidelines for older adults, which is 150 minutes of aerobic activity per week.
The purpose of the current study was to determine whether exercise behaviors are consistent with the constructs of the Theory of Planned Behavior (TPB) in older adults. Another objective was to evaluate whether adding temporal discounting (TD) with TPB constructs increases TPB's ability to predict exercise. These were focused objectives; how exercise behaviors depart from the TPB was not attempted to be determined.
Theory of Planned Behavior
Many factors influence older adults' exercise behavior, such as those identified in the TPB (Figure 1). The TPB (Ajzen & Fishbein, 1980) focuses on individual motivational factors as the determinants of performing a specific behavior, such as exercise. The TPB is an extension of the Theory of Reasoned Action (TRA).
Theory of Planned Behavior (TPB).
Note. The upper dark section shows the Theory of Reasoned Action; the entire figure shows the TPB.
The TPB states that attitude toward behavior, subjective norm, and perceived behavioral control lead to behavioral intention. The TPB asserts that the most important determinant of behavior is an individual's behavioral intention (Glanz, Rimer, & Lewis, 2002).
Attitude toward behavior includes the individual's general feeling or belief as to whether the behavior is favorable. Subjective norm includes normative beliefs and motivation to comply. Normative beliefs are defined as the individual's perception of certain significant others' preferences about whether one should engage in the behavior. Motivation to comply is multifaceted and complex; it is the driving force or reason why an individual would perform a behavior. Perceived behavioral control is concerned with an individual's judgment of how well he or she can perform a behavior under various conditions.
Behavioral intention is defined as the individual's perceived likelihood of performing a behavior and is a direct result of attitude toward a behavior, subjective norm, and perceived behavioral control. Behavioral intention is the cognitive representation of an individual's readiness to perform a given behavior, and it is considered to be the immediate antecedent of behavior. The TPB assumes a causal chain that links attitude toward behavior, subjective norm, and perceived behavioral control to behavioral intention and behavior (Glanz et al., 2002).
To predict individuals' behavioral intentions, awareness of their behavioral beliefs may be as important as awareness of their attitude toward behavior. Both predictors lead to behavioral intention. As a general rule, if an individual's attitude toward behavior and the subjective norm are favorable, then the individual's behavioral intention to perform the behavior in question should be powerful. The addition of perceived behavioral control allows individuals to determine whether they feel they are capable of performing the activity along with attitude toward behavior and subjective norms in determining behavioral intention (Glanz et al., 2002).
The TPB has been used to predict and explain a range of health behaviors and behavioral intentions. According to the literature, these findings have been used in interventions developed to change behavior, such as smoking, alcohol use, contraception use, mammography screening, exercise, seat belt use, substance abuse, and prevention behaviors (Godin & Kok, 1996).
The TPB model is supported as an effective predictor of behavior based on an individual's behavioral intention. Behavioral intention is highly predictive of behavior only when measured within the specific context of the action, such as exercise, and within a specific time frame (Ajzen & Fishbein, 1980). Not all situations fit neatly within the TPB framework. Other constructs or factors may play a role and help explain some of the variance in the TPB in prediction of exercise. Gretebeck et al. (2007) showed that by adding the additional construct of functional ability, 11% of the variance in the TPB could be explained. According to Churchill, Jessop, and Sparks (2008), research has shown that there is significant empirical support for the use of the TPB as a framework to predict behavior. Although the empirical support is impressive, it is still not 100%, especially in situations where individuals may have to make decisions based on what the behavior offers them immediately versus what the behavior may offer them later or over a period of time. The concept of immediate versus delayed benefit is referred to as temporal discounting (TD) (Madden & Bickel, 2010). The current study examines the use of TD in increasing the predictability of the TPB.
TD refers to subjectively devaluing a delayed reinforcement in lieu of a more immediate reinforcement or reward (Bickel & Marsch, 2001). For example, receiving $50 today may be valued more than the promise of receiving $75 in 30 days for some individuals. Typically, as the time an individual has to wait for a reward increases, the perceived value of the future reward decreases (Bickel & Marsch, 2001).
Bickel and Marsch (2001) conceptualized individual differences in TD as an important aspect of behavioral economics. Therefore, how time is discounted or how an individual views the benefits or consequences of a behavior and its immediate and/or delayed results may be one way to account for individual differences in exercise behavior and perceived necessity of effort to receive expected results.
Discounting the future is a behavioral process common to every species (Madden & Bickel, 2010). Studies that measure hypothetical money and health behaviors ask participants to make comparisons between an immediate or near future outcome and a delayed outcome. To determine discount rates, an understanding of the trade-off between costs and benefits must be measured. Discounting rates have been observed as a hyperbolic curve, suggesting that individuals will make far-sighted decisions when planning in advance, but will make short-sighted decisions when costs or benefits are immediate (McAuley, 2013). Typically, the longer the wait time, the more likely the individual will choose a more immediate reward.
Story, Vlaev, Seymour, Darzi, and Dolan (2014) provide a thorough review of TD and unhealthy behaviors. Research shows that discounting future rewards tends to be heavier in those dependent on cigarettes, alcohol, and illicit substances than in individuals free of those dependencies (Story et al., 2014). However, there is mixed evidence of heavy discounting associating with unhealthy eating behaviors and low or no practice of preventive health behaviors. Regarding exercise—a preventive health behavior—two studies were found. Daugherty and Brase (2010) surveyed 467 undergraduate students about health behaviors and assessed TD with the Monetary Choice Questionnaire (MCQ). The MCQ and exercise had a negligible zero-point correlation of −0.03. Bradford (2010) used existing survey data from respondents between ages 24 and 65. He reported associations of high discounters and a panel of health maintenance behaviors, finding that those whose discount rates were in the highest 20% were 15% less likely to engage in weekly vigorous exercise compared to those in the lower 80% (N = 789, p = 0.001) (Bradford, 2010).
With the increasing numbers of older adults and the financial burden of chronic disease, it is crucial that health care programs seek ways to improve exercise participation in this population. Predictor models, such as the TPB, are vital for health care programs to plan and implement successful ways to motivate older adults to exercise and maintain an exercise program. It may prove beneficial for health care providers to be able to predict early on which individuals will participate fully. This knowledge will allow health care providers to determine individuals who may drop out or not reap any benefits due to high rates of TD and to develop better strategies to engage them in exercise.
All constructs of the TPB were not measured. Because attitude, subjective norm, and behavioral control are the three main predictors of behavioral intention, these constructs were chosen by the researchers as the primary measures to examine the predictability of the TPB in the study.
The specific aims of the current study were to (a) examine the TPB in predicting exercise in older adults and (b) examine the TPB in predicting exercise in older adults with the addition of TD. It was proposed that adding the additional variable (i.e., TD) would increase the predictability of the TPB.
A cross-sectional observational study on 137 community-dwelling older adults was performed. Information regarding participants' exercise habits (exerciser or non-exerciser), health status (comorbidities), attitudes toward exercise, subjective norm, perceived behavioral control, behavioral intentions about exercise, TD rates, and demographic information was collected. The study was reviewed and approved by the university institutional review board (IRB). The IRB issued a waiver of consent because no identifying data were being collected.
Participants 60 or older with little to no cognitive impairment who reside in the community were recruited. Based on the Six-item Memory Screener, participants scoring ≥4 were selected to participate (Callahan, Unverzagt, Hui, Perkins, & Hendrie, 2002). Exclusion criteria were individuals with medical conditions or physical limitations that would not allow them to exercise. These conditions were identified using a modified version of the Exercise Assessment and Screening for You (EASY) tool (Resnick et al., 2008). This assessment was adapted/used with permission from the EASY Partners who developed the EASY materials (access http://www.easyforyou.info). Including individuals with such conditions in the current study would not provide an accurate measure of TD, as they had physical limitations and little control of their exercise ability.
One hundred fifty-one community-dwelling adults 60 or older were recruited from 11 churches in rural Arkansas. Church leaders were contacted and gave permission for recruitment to take place during meetings of older adult groups. Handouts were taken and the study was explained on the first visit. On the second visit, data collection began with interested parties. After completion of screening for eligibility, 137 participants met the inclusion and exclusion criteria, ranging in age from 60 to 77 years (mean age = 65.26 years, SD = 4.51 years). The majority of participants were female (61%), Caucasian (56%), and had at least 12 years of education (89%).
Demographic Form. The Behavioral Risk Factor Surveillance System (CDC, 2015a) form was used to gather demographic data regarding age, race, gender, marital status, education, and income status.
Yale Physical Activity Survey (YPAS). The YPAS (De Abajo, Larriba, & Marquez, 2001) was used to collect data on exercise activity. After recruiting a small number of older adults, the YPAS was found to be too time consuming and strenuous along with all other instruments in the study; therefore, the Vigorous Activity Index (VAI) Score in Part 2 of the YPAS questionnaire was the only information gathered from this tool for the remainder of participants. The VAI score was totaled for each participant, with scores ranging from 0 to 60, with 0 being no exercise at all and 60 being >150 minutes per week. This part of the form has participants provide a frequency of how many times during the past month they participated in vigorous activities and for how long. A score of 20 (preselected by the researcher) on the VAI was intended to capture those individuals who were exercising at a minimum of 50 minutes per week. Although this is well below the recommended guidelines for exercise for older adults, they were at least making some attempt to exercise. Solomito Pugh (2006) suggested that moderate construct validity for diverse populations, such as older adults, exists for the YPAS. Reliability was higher for more vigorous activities than for moderate to low intensity activities. Test–retest reliability was reported by De Abajo et al. (2001).
Fitness Attitude Scale. Attitude toward exercise was measured as the total score on the 19-item Fitness Attitude Scale (Kerner & Grossman, 2001). The anchors are −3 (strongly disagree), −2, −1 (disagree), 0, and 1 (agree), 2, and 3 (strongly agree). An example of an item is “I think that the time I spend doing physical fitness activities is not wasted.” The maximum positive and negative scores are 57 and −57, respectively. Reliability of the remaining 19-item scale provided an alpha coefficient of 0.87. Alpha reliability was 0.70 for the 19 items, which demonstrates adequate internal consistency (Kerner & Grossman, 2001).
Expectations of Others Scale. The perception of participants that significant others in their lives believe they should engage in a program of physical activity (subjective norm) was measured as the total score on the 6-item Expectations of Others Scale (Kerner & Grossman, 2001). The anchors are −3 (I should not) to 3 (I should). Participants respond to a prefix in each of the six scale statements: “My ‘significant other’ thinks that (anchor) adhere to a program of physical activity to....” An example of an item is: “…reduce my body weight in the next 12 months.” The maximum positive and negative subjective norm scores are 18 and −18, respectively. Reliability for the 6-item scale revealed an alpha coefficient of 0.84, demonstrating adequate reliability of the scale (Kerner & Grossman, 2001).
Perceived Behavioral Control Scale. Participants' perceived ease or difficulty in performing physical activities was measured as the total score on the 3-item Perceived Control Scale (Kerner & Grossman, 2001). The anchors are 6 (extremely likely), 5, 4 (likely), 3, 2 (unlikely), 1, and 0 (extremely unlikely), and 6 (complete control), 5, 4 (moderate control), 3, 2 (difficult), 1, and 0 (very difficult). An example of an item is: “For me to do physical activities on a regular basis is….” The maximum and minimum scores are 18 and 0, respectively. Alpha coefficient for the 3-item scale was 0.82, demonstrating adequate reliability of the scale (Kerner & Grossman, 2001).
Intention to Exercise Scale. Intention to engage in physical activity was measured as the total score on the 11-item Intention to Engage in Physical Activity Scale (Kerner & Grossman, 2001). The anchors are −3 (very unlikely) to 3 (very likely). An example of an item is: “I intend to adhere to a program of physical activities during the next 3 weeks to get into shape.” The maximum positive and negative scores are 33 and −33, respectively. Reliability for the 11-item scale revealed an alpha coefficient of 0.75, demonstrating adequate reliability of the scale (Kerner & Grossman, 2001).
Monetary Choice Questionnaire (MCQ). The MCQ is a 27-item questionnaire that measures delay discounting rates of an individual. The instrument has been shown to be valid and reliable in test–retest measures (Kirby, 2009). The 5-week test–re-test stability rate was 0.77 (95% confidence interval [0.67, 0.85]), with the 1-year stability rate being 0.71 (Kirby, 2009). Internal reliability was found to be 0.98 (Duckworth & Seligman, 2005). The MCQ is one of the best validated discount rate measures (Duckworth & Seligman, 2005; Kirby, 2009). According to Daugherty and Brase (2010), the MCQ is significantly correlated with exercise, and thus a good choice of measure for the current study. The MCQ administered obtained strictly hypothetical responses as no monetary rewards were issued to participants.
In summary, all instruments had adequate validity and reliability in the general adult population. Therefore, they were chosen as appropriate tools for measurement for the current study.
Path analysis was used to examine the direct and indirect hypothesized relationships between and among attitude, subjective norm, behavioral control, and TD. Path analysis model building involves the generation of hypotheses based on prior research and theoretical assumptions. To construct the path model, two regression models were built. Regression Model 1 examined the predictability of the constructs of attitude (X1), subjective norm (X2), and behavioral control (X3) on behavioral intention (Y1). Attitude was log transformed to meet the normality assumption of the residuals. Model 2 examined the effect of intention (Y1) and TD (Y2) on exercise behavior (Z1). The path map is displayed in Figure 2. Models 1 and 2 are represented by the below equations:
In the equations, i represents the ith individual, є1 and є2 are assumed to be independently and identically distributed according to a normal distribution with mean 0 and variance
. The model coefficients α1, α2, α3, and β1, β2 were estimated by fitting the above two models to the data.
For Model 1, the degree of effects of attitude, subjective norm, and behavioral control on behavioral intention was equal to the sum of the path coefficients. For Model 2, the degree of the effects of behavioral intention and TD was equal to the sum of the path coefficients. All analyses were performed using SPSS version 22.
Mean participant age was 65.26 years (SD = 4.51 years), with females (61%) and Caucasians (56%) comprising the majority of the sample. More than 89% of participants completed high school or higher education, and all income levels were represented. Table 1 provides characteristics and summary statistics of the sample.
Sample Demographics (N = 137)
Regression results indicated that the overall model significantly predicted behavioral intention (F(3, 132) = 41.01, p < 0.001). Figure 2 shows the adjusted standardized coefficients for each path in the path map. This model accounted for 48.5% (47.3% adjusted) of the variance in behavioral intention. All three variables (attitude, subjective norm, and behavioral control) were significant in predicting behavioral intention. Attitude (α1 = 0.443, p < 0.001) and behavioral control (α3 = 0.319, p < 0.001) had the largest effect.
Regression results indicated that the overall model significantly predicted exercise behavior (F(2, 134) = 26.83, p < 0.001). This model accounted for 28.6% (27.5% adjusted) of the variance in exercise behavior. Behavioral intention (standardized β = 0.504, p < 0.001) was the only predictor. TD (standardized β = −0.066, p = 0.413) had no significant effect after adjusting for behavioral intention. The standardized path coefficients for Model 2 are shown in Figure 2.
The proportions of direct and indirect effects of the major variables are shown in Table 2. Attitude, subjective norm, and behavioral control had no direct effect on exercise behavior. Therefore, their indirect effect on exercise behavior was calculated by taking the product of their individual alpha coefficient and beta coefficient of behavioral intention. For example, indirect effect (0.223) of attitude on exercise behavior equals the path coefficient for attitude (α1 = 0.443) multiplied by the path coefficient for behavioral intention (β1 = 0.504). To calculate percentage of a variable's effect on exercise behavior, the sum of the direct and indirect effect of the variable was divided by the sum of the total direct and indirect effects. For example, attitude's total effect (0.223) on exercise behavior was divided by total effects (1.029), equaling 0.217 or 21.7%. TD counts for only 6.4% of total effect on exercise behaviors.
Direct and Indirect Effects of Major Variables on Exercise Behavior
As expected, the data revealed that attitude, subjective norm, and behavioral control collectively had a strong effect on the prediction of behavioral intention (47.3%). Courneya (1995) examined similar constructs (i.e., attitude, subjective norm, and behavioral control) in 288 adults 60 and older and had similar results on the prediction of intention (71%). Blue (1995) reviewed 16 studies using the TPB as a predictor of exercise and examined the relationship among the constructs of the theory and their individual relationships to exercise behavior. Results from the studies reviewed showed that attitude was significantly correlated with behavioral intention (Blue, 1995). These results are consistent with the TPB, which postulates that some behavioral intentions are influenced by attitude toward behavior.
In addition, the data revealed that behavioral intention and TD collectively explained 27.5% of variance in exercise behavior. With no surprise, behavioral intention is the most important variable predicting exercise behaviors, accounting for 49% of the total effect. When directly measuring the constructs of the TPB, behavioral intention is the strongest predictor of behavior (Glanz et al., 2002). Benjamin, Edwards, and Bharti (2005) conducted a study evaluating all constructs of the TPB. With a sample of 109 frail older adults, strong behavioral intention (likely and very likely to exercise) was positively correlated with high levels of exercise (r = 0.683, p = 0.005). Behavioral intention also strongly predicted exercise in a meta-analysis by Downs and Hausenblas (2005) of 111 studies examining the TRA/TPB and its relation to exercise. Behavioral intention (β = 0.42, p < 0.001) emerged as the most significant predictor of exercise.
However, TD was not a significant predictor and only accounted for 6.4% of the total effect. No literature was found that examined discounting rates in relation to exercise behavior. However, based on the results of studies examining risk behaviors, it was expected that in the path model, this variable should have played a bigger role. One potential reason could be overlap of the construct between behavioral intention and TD. A statistically significant correlation exists between behavioral intention and TD (r = −0.410, p < 0.001). This correlation suggests that as intention increases discounting decreases, thus the effect of TD may be diluted by adding the variable behavioral intention. Future studies should explore the theoretical construct of these two concepts and tease out the components that belong to each.
Although the current study showed that TD is not a significant predictor of exercise in this population, the test of predictability of the TPB was supported. Thus, using the variables in the model itself was a decent predictor of exercise behavior.
Limitations of the current study include the purposive sample from rural churches that may not be generalizable to the population as a whole. In addition, issues with the data may lie in the scales used to collect the data. Although the scales have validity and reliability with older adults, they may need to be further adjusted for this specific population (i.e., church goers). Individuals who attend church may have different values and methods of choice than the standard population based on their spiritual beliefs. Lastly, the fact that the data were collected strictly by self-reported measure could also be a limitation. Because all reports were self-written and performed in a church setting, individuals may have not taken time to read the questions thoroughly or may have just picked an answer to move through instead of taking the questions seriously.
Future research in using the concept of delay discounting should focus on better conceptualization and instrumentation to distinguish the possible overlap of delay discounting and behavioral intention. Nevertheless, the results of the current study reinforce previous research in establishing that the TPB is a good predictor of exercise behavior and useful in the older adult population. Being able to predict which individuals will exercise can have a tremendous impact on how health care professionals design programs to get and keep older adults exercising that especially focus on changing, promoting, and/or improving intention to exercise.
- Administration on Aging. (2009). Aging integrated database (AGID). Retrieved from https://agid.acl.gov/DataGlance/ACS
- Ajzen, I. & Fishbein, M. (1980). Understanding attitudes and predicting social behaviors. Englewood Cliffs, NJ: Prentice-Hall.
- Benjamin, K., Edwards, N.C. & Bharti, V.K. (2005). Attitudinal, perceptual, and normative beliefs influencing the exercise decisions of community-dwelling physically frail seniors. Journal of Aging and Physical Activity, 13, 276–293. doi:10.1123/japa.13.3.276 [CrossRef]
- Bickel, W.K. & Marsch, L.A. (2001). Toward a behavioral economic understanding of drug dependence: Delay discounting processes. Addiction, 96, 73–86. doi:10.1046/j.1360-0443.2001.961736.x [CrossRef]
- Blue, C.L. (1995). The predictive capacity of the theory of reasoned action and the theory of planned behavior in exercise research: An integrated literature review. Research in Nursing & Health, 18, 105–121. doi:10.1002/nur.4770180205 [CrossRef]
- Bradford, W.D. (2010). The association between individual time preferences and health maintenance habits. Medical Decision Making, (1), 99–112. doi:10.1177/0272989X09342276 [CrossRef]
- Callahan, C.M., Unverzagt, F.W., Hui, S.L., Perkins, A.J. & Hendrie, H.C. (2002). Six-item screener to identify cognitive impairment among potential subjects for clinical research. Medical Care, 40, 771–781. doi:10.1097/00005650-200209000-00007 [CrossRef]
- Centers for Disease Control and Prevention. (2015a). Behavioral risk factor surveillance system. Retrieved from https://www.cdc.gov/brfss/questionnaires/index/htm
- Centers for Disease Control and Prevention. (2015b). How much physical activity do older adults need? Retrieved from https://www.cdc.gov/physicalactivity/basics/older_adults/index.htm
- Churchill, S., Jessop, D.C. & Sparks, P. (2008). Impulsive and/or planned behaviour: Can impulsivity contribute to the predictive utility of the theory of planned behavior?British Journal of Social Psychology, 47, 631–646. doi:10.1348/014466608X284434 [CrossRef]
- Courneya, K.S. (1995). Understanding readiness for regular physical activity in older individuals: An application of the theory of planned behavior. Health Psychology, 14, 80–87. doi:10.1037/0278-6188.8.131.52 [CrossRef]
- Daugherty, J.R. & Brase, G.L. (2010). Taking time to be healthy: Predicting health behaviors with delay discounting and time perspective. Personality and Individual Differences, 48, 202–207. doi:10.1016/j.paid.2009.10.007 [CrossRef]
- De Abajo, S., Larriba, R. & Marquez, S. (2001). Validity and reliability of the Yale Physical Activity Survey in Spanish elderly. Journal of Sports Medicine and Physical Fitness, 41, 479–485.
- Downs, D.S. & Hausenblas, H. (2005). The theories of reasoned action and planned behavior applied to exercise: A meta-analytic update. Journal of Physical Activity & Health, 2, 76–97. doi:10.1123/jpah.2.1.76 [CrossRef]
- Duckworth, A.L. & Seligman, M.E. (2005). Self-discipline outdoes IQ in predicting academic performance of adolescents. Psychological Science, 16, 939–944. doi:10.1111/j.1467-9280.2005.01641.x [CrossRef]
- Edwards, S. (2006). Physical exercise and psychological well-being. South African Journal of Psychology, 36, 357–373. doi:10.1177/008124630603600209 [CrossRef]
- Glanz, K., Rimer, B.K. & Lewis, F.M. (2002). Health behavior and health education: Theory, research, and practice. San Francisco, CA: Jossey-Bass.
- Godin, G. & Kok, G. (1996). The theory of planned behavior: A review of its applications to health-related behaviors. American Journal of Health Promotion, 11, 87–98. doi:10.4278/0890-1171-11.2.87 [CrossRef]
- Gretebeck, K.A., Black, D.R., Blue, C.L., Glickman, L.T., Huston, S.A. & Gretebeck, R.J. (2007). Physical activity and function in older adults: Theory of planned behavior. American Journal of Health Behavior, 31, 203–214. doi:10.5993/AJHB.31.2.9 [CrossRef]
- Kerner, M.S. & Grossman, A.H. (2001). Scale construction for measuring attitude, beliefs, perception of control, and intention to exercise. Journal of Sports Medicine and Physical Fitness, 41, 124–131.
- Kirby, K.N. (2009). One-year temporal stability of delay-discount rates. Psychonomic Bulletin & Review, 16, 457–462. doi:10.3758/PBR.16.3.457 [CrossRef]
- Madden, G.J. & Bickel, W.K. (2010). Impulsivity: The behavioral and neurological science of discounting. Washington, DC: American Psychological Association.
- McAuley, I. (2013). Behavioural economics and public policy: Some insights. International Journal of Behavioural Accounting and Finance, 4, 18. doi:10.1504/IJBAF.2013.057365 [CrossRef]
- National Institute on Aging. (2015). Real-life benefits of exercise and physical activity. Retrieved from http://nihseniorhealth.gov/exerciseforolderadults/healthbenefits/01.html
- Resnick, B., Ory, M.G., Hora, K., Rogers, M.E., Page, P., Bolin, J.N. & Bazzarre, T.L. (2008). A proposal for a new screening paradigm and tool called Exercise Assessment and Screening for You (EASY). Journal of Aging and Physical Activity, 16, 215–233. doi:10.1123/japa.16.2.215 [CrossRef]
- Solomito Pugh, A.N. (2006). Validity of the Yale Physical Activity Survey for older adults [Master's thesis]. Retrieved from http://digitalcommons.lsu.edu/cgi/viewcontent.cgi?article=2382&context=gradschool_theses
- Stewart, S.T., Cutler, D.M. & Rosen, A.B. (2013). US trends in quality-adjusted life expectancy from 1987 to 2008: Combining national surveys to more broadly track the health of the nation. American Journal of Public Health, 103, e78–e87. doi:10.2105/AJPH.2013.301250 [CrossRef]
- Story, G.W., Vlaev, I., Seymour, B., Darzi, A. & Dolan, R.J. (2014). Does temporal discounting explain unhealthy behavior? A systematic review and reinforcement learning perspective. Frontiers in Behavioral Neuroscience, (8), 76. doi:10.3389/fnbeh.2014.00076 [CrossRef]
Sample Demographics (N = 137)
|Age (years) (mean, SD)||65.26 (4.51)|
| Female||83 (60.6)|
| Male||54 (39.4)|
| Caucasian||77 (56.2)|
| African American||42 (30.7)|
| Hispanic||10 (7.3)|
| Asian||8 (5.8)|
| Grade 1 to 8||6 (4.4)|
| Grade 9 to 11||9 (6.6)|
| High school graduate or GED||56 (40.9)|
| 1 to 3 years of college or tech school||24 (17.5)|
| College graduate||42 (30.6)|
| $0 to $14,999||11 (8.1)|
| $15,000 to $19,999||34 (24.8)|
| $20,000 to $24,999||3 (2.2)|
| $25,000 to $34,999||15 (10.9)|
| $35,000 to $49,999||19 (13.9)|
| $50,000 to $74,999||33 (24.0)|
| ≥$75,000||22 (16.1)|
Direct and Indirect Effects of Major Variables on Exercise Behavior
|Factor||Direct Effect||Indirect Effect|