Psychosocial Predictors and Exercise Intentions and Behavior Among Individuals With Spinal Cord Injury

Transcription

1 ADAPTED PHYSICAL ACTIVITY QUARTERLY, 2004, 21, Human Kinetics Publishers, Inc. Psychosocial Predictors and Exercise Intentions and Behavior Among Individuals With Spinal Cord Injury Amy E. Latimer and Kathleen A. Martin Ginis McMaster University B. Catherine Craven Toronto Rehabilitation Institute Using the theory of planned behavior as a theoretical framework, the present study examined psychosocial predictors of exercise intentions and behavior among 124 men and women with spinal cord injury. Theory of planned behavior constructs were measured using an exercise specific questionnaire for individuals with spinal cord injury. Exercise behavior was assessed using an adapted version of the Godin Leisure Time Exercise Questionnaire. Regression analyses indicated that the theory of planned behavior had limited utility in this population. Among individuals with tetrapelgia, perceived behavioral control was the only determinant of exercise intentions and behavior. Among people with paraplegia, none of the theory of planned behavior constructs predicted exercise intentions or behavior. These results have methodological and practical implications for future research and exercise interventions, respectively. There is accumulating evidence of the benefits of regular exercise for individuals with spinal cord injury (SCI). For instance, regular exercise training has been shown to produce significant improvements in both cardiovascular and muscular fitness (Duran, Lugo, Ramirez, & Eusse, 2001; Jacobs, Nash, & Rusinowski, 2001). In addition to these physiological benefits, a randomized controlled trial indicated that people with SCI who engaged in a twice-weekly center-based exercise program also accrued psychological benefits from exercise participation (Hicks et al., 2003). Specifically, compared to a control group, exercisers reported less pain and stress, fewer depressive symptoms, and greater health-related quality of life. Despite these encouraging findings, however, several studies indicate that few individuals with SCI are sufficiently active to achieve these benefits (Buchholz, McGillvray, & Pencharz, 2003; Janssen, van Oers, van der Woude, Hollander, 1994; Pentland, Harvey, Smith, & Walker, 1999). This is a very serious health concern for people with SCI, as an inactive lifestyle has many health liabilities (e.g., increased risk Amy Latimer and Kathleen Martin Ginis are with McMaster University, Department of Kinesiology, Hamilton, ON, L8S 4K1 Canada. E mail: martink@mcmaster.ca. B. Catherine Craven is with Toronto Rehabilitation Institute, Toronto, Ontario, Canada. 71

2 72 Latimer, Martin Ginis, and Craven of heart disease, hypertension, diabetes mellitus; Heath & Fentem, 1997). Hence, there is a pressing need to identify determinants of exercise participation among people with SCI. By identifying determinants, researchers and practitioners can then begin to develop strategies to increase physical activity in this population (Rimmer, Braddock, & Pitetti, 1996). Some factors affecting exercise participation in persons with SCI, such as time limitations (Martin et al., 2002; Shifflet, Cator, & Megginson, 1994), peer influence (Wu & Williams, 2001), and severity of injury (Godin, Shepard, Davis, & Simard, 1989; Hedrick & Broadbent, 1996; Wu & Williams, 2001), have already been identified through qualitative and descriptive studies. Although these studies have begun to advance our understanding of exercise behavior among individuals with SCI, unfortunately, they have been largely atheoretical. There is great benefit in using a theoretical framework to examine exercise determinants. Theory-driven research helps to direct the researcher s inquiries and provides a more comprehensive picture of exercise, a definite need in this scarcely studied area (Rejeski, 1995). Of further benefit, using a theory to understand the behavior of a special population may help to expand or limit the theory s tenets, a major contribution to the scientific field and to theory development (Crocker, 1993). With these advantages in mind, the theory of planned behavior (TPB; Ajzen, 1985) was used for the current investigation of exercise behavior among individuals with SCI. The Theory of Planned Behavior: A Theoretical Framework The TPB, an extension of the theory of reasoned action (TRA; Fishbein & Ajzen, 1975) was developed by Ajzen (1985) to account for goal-directed behaviors over which the individual does not have complete volitional control. The TPB posits that behavior is determined by an individual s intentions to perform the behavior. In turn, intentions are determined by attitude, subjective norms, and perceived behavioral control. Attitude (i.e., one s positive or negative evaluation of the behavior) is a function of behavioral beliefs, which refer to the perceived advantages and disadvantages of performing the behavior. Subjective norms (i.e., the perceived social pressure to perform the behavior) are a function of normative beliefs, which refer to the perceived pressure from important others to perform the behavior. Perceived behavioral control (i.e., the perceived ease or difficulty of performing the behavior) is a function of control beliefs, which refer to the beliefs about the presence of factors that may facilitate or impede performance and the perceived impact of these factors on the individual. It is this latter construct that takes into account individuals lack of volitional control over certain behaviors. The TPB has been used extensively to guide the investigation of exercise behavior among a variety of clinical and nonclinical populations (e.g., Courneya & Friedenreich, 1997; Hausenblas, Carron, & Mack, 1997). The wide use of the TPB compared to other psychosocial theories (e.g., health belief model, Rosenstock, 1974; protection motivation theory, Rogers, 1975; TRA, Fishbein & Ajzen, 1975) is largely due to (a) its ability to account for complex behaviors lacking complete volitional control (e.g., exercise) and (b) the substantial body of evidence to support the model (Maddux, 1993). For instance, three separate reviews of the TPB and exercise behavior literature have confirmed the usefulness of the TPB s theoretical constructs for predicting exercise behavior (Godin & Kok, 1996; Hagger,

3 Exercise Determinants and Spinal Cord Injury 73 Chatzisarantis, & Biddle, 2002; Hausenblas et al., 1997). It has been concluded from these reviews that perceived behavioral control and intentions predict a moderate amount of variance in exercise behavior (i.e., 35-49%). Although all three antecedents of intentions have been shown to predict exercise intentions, perceived behavioral control and attitude have consistently emerged as the strongest predictors. For individuals with SCI, the TPB constructs should be particularly relevant in predicting exercise behavior (Crocker, 1993). Indeed, anecdotal reports from people with SCI indicate that many of the reasons for an inactive lifestyle are encompassed by the TPB constructs (Hicks et al., 2003; Martin et al., 2002; Rimmer et al., 1996). These reasons include a lack of knowledge concerning the importance and benefits of exercise (i.e., aspects of attitude), limited access to transportation to and from the exercise site, inaccessible facilities and equipment, perceptions of an inability to exercise as a result of a disability (i.e., factors that affect perceived behavioral control), and a lack of enthusiasm from significant others to exercise (i.e., factors related to subjective norms; Hicks et al., 2003; Martin et al., 2002; Rimmer et al., 1996). Given preliminary anecdotal support for using the TPB constructs to explain exercise in the SCI population, the purpose of this study was to examine predictors of exercise behavior among individuals with SCI using the TPB as a guiding theoretical framework. The findings will extend our understanding of the theoretical tenets of the TPB and will help to direct the development of exercise-enhancing interventions in persons with SCI. Our research hypotheses stemmed from the TPB and were based on findings from earlier TPB and exercise literature reviews. Specifically, it was hypothesized that (a) exercise intentions would be predicted by attitudes, subjective norms, and perceived behavioral control with attitude and perceived behavioral control each accounting for more variance than subjective norms; and (b) exercise behavior would be predicted by intentions and perceived behavioral control. Participants Method A sample of 124 individuals with SCI (70 with tetrapelgia, 54 with paraplegia) participated in the study. According to Green s (1991) guidelines for determining regression analysis sample size, 124 was the necessary number to detect a mediumsized effect for the overall model. Participants demographic and medical profiles are presented in Table 1. Using convenient sampling methods, 25% of participants were recruited through an information network (Canadian Paraplegic Association). The remaining 75% of participants were recruited from service programs for individuals with SCI at three sites in southern Ontario. The service programs included medical and wheelchair fitting clinics, exercise programs, and a wheelchair basketball league. Chi square tests and a multivariate analysis of variance (MANOVA) revealed no differences in either demographic characteristics or response variables between individuals recruited from the information network and the service programs (all ps >.05). Accordingly, data were collapsed across the two recruitment sources to maximize the variance of the TPB measures. Greater variance in the predictors improves the precision of regression estimates (Wonnacott & Wonnacott, 1987).

4 74 Latimer, Martin Ginis, and Craven Table 1 Participant Demographic Data Nature of Years of Gender Years Completeness injury age (%) post injury of injury (%) Tetraplegia ± Female: 9.10±10.04 Incomplete: (n = 70) n = 22 (31) n = 45 (61) Male: Complete: n = 48 (69) n = 22 (31) Unknown: n = 3 (4) Paraplegia ± Female: 8.10 ± 7.95 Incomplete: (n = 54) n = 16 (30) n = 29 (54) Male: Complete: n = 38 (70) n = 25 (46) Separate ANOVAs revealed no between group differences in either age or years post injury. Chi Square analyses revealed no between group differences in either gender or completeness of injury. There were fewer women in the study than men. This is typical of the SCI population, of which 80% is male (Canadian Paraplegic Association, 2000). A MANOVA indicated no gender differences in the TPB variables and exercise frequencies (p >.05). Consequently, to reduce the likelihood of a type II error resulting from the small sample of women, data were collapsed across genders rather than analyzed separately. Procedure In accordance with the cross-sectional correlation design of the study, participants were asked to complete and return a brief questionnaire. Our initial approach to questionnaire distribution was through the mail. Specifically, 140 members of the information network were mailed a copy of the survey. The instructions provided participants the option of completing the questionnaire independently or by telephone with the assistance of a researcher. The mailed questionnaire return rate was 22% (n = 31). Having not met the sample size requirements stipulated by Green (1991), additional participants were recruited from service programs offered in the community. A research assistant approached potential participants asking them to complete a brief questionnaire. Participants were given the option of completing the questionnaire independently or in a face-to-face interview with a research assistant. Measures A questionnaire tailored for individuals with SCI was developed according to Ajzen s (2003a) recommendations to assess each TPB construct in an exercise

5 Exercise Determinants and Spinal Cord Injury 75 context. Participants were instructed that exercise included all strength training and aerobic training activities performed 3 times per week. To ensure that salient beliefs for the participant group were incorporated into the questionnaire, 3 focus groups including rehabilitation experts, family members, and individuals with SCI were conducted to assess beliefs concerning SCI and exercise participation (Martin et al., 2002). Using the information gathered from the focus groups, a questionnaire was designed to assess individuals attitudes, subjective norms, perceived behavioral control, and intentions to exercise regularly. Individuals with SCI reviewed a pilot version of the questionnaire to confirm its face validity before it was administered. In the pilot questionnaire, attitude, subjective norms, and perceived behavioral control were assessed both directly and indirectly. The direct measures required participants to judge each construct on a set of scales (e.g., perceived behavioral control item: I feel like I have control over whether I do or do not exercise 3 times per week ). The indirect measures 1 required participants to rate their corresponding beliefs (e.g., perceived behavioral control item: Transportation difficulties would make it difficult/easy to exercise 3 times per week ). Pilot study results indicated that the indirect measures of attitude and subjective norms were better predictors of intentions than their corresponding direct measures were. In contrast, the direct measure of perceived behavioral control was a better predictor of intentions than the indirect measure was. Consequently, indirect measures of attitude and subjective norms were used and a direct measure of perceived behavioral control was used. The items included in the final questionnaire are described below. Attitudes. Behavioral beliefs were assessed as an indicator of attitudes. Participants rated their behavioral belief strength for seven salient exercise participation outcomes: (a) increased mobility, (b) increased ease of performing activities of daily living, (c) increased energy level, (d) increased confidence to do the things you want or need to do, (e) prevention of secondary health conditions, (f) increased self-esteem, and (g) increased confidence in social situations. Items were preceded by the question, To what extent do you think exercising 3 times a week will (insert outcome)...? and rated on a scale ranging from 1 (not at all) to 7 (very much). The total attitude score was calculated. According to Nunnally s (1978) criterion, the scale demonstrated adequate internal consistency (α =.91). Subjective Norms. The subjective norms measure assessed individuals normative beliefs. Participants indicated the perceived social pressure to exercise from three social influences: (a) family, (b) friends, and (c) doctor. Each item was preceded by the statement, My thinks that I should engage in regular physical activity and rated on a scale ranging from 1 (disagree) to 7 (agree). In the case that a certain social influence was not relevant for an individual (e.g., individuals with no family), there was an option of responding do not have. The subjective norms scale was calculated as the mean score of items completed (i.e., for participants who completed all three items, the mean was calculated for the 3 items; for participants who completed only 2 items, the mean was calculated for these 2 items). Internal consistency for this scale was adequate (α = 90). Perceived Behavioral Control. The perceived behavioral control measure was a 2-item scale: (a) I feel that I have control over whether I do or do not exercise 3 times per week, and (b) I have confidence in my ability to exercise 3 times per week. Responses were rated on a 7-point scale ranging from 1 (disagree) to 7

6 76 Latimer, Martin Ginis, and Craven (agree) and summed to produce a total score. As outlined by Ajzen ( 2003a), this is a standard measure of perceived behavioral control including both a controllability item (item 1) and a self-efficacy item (item 2). Intentions. Intentions were measured using a single item: I intend to exercise 3 times a week for at least 30 minutes each time, in the forthcoming month. Responses were rated on a 7-point scale ranging from 1 (extremely unlikely) to 7 (extremely likely). A single item to assess intentions is common in the exercise domain and is consistent with TPB guidelines (Ajzen, 1991; Courneya & McAuley, 1993). Exercise Behavior. Exercise behavior was assessed using a modified version of the validated Godin Leisure Time Exercise Questionnaire (GLTEQ; Godin & Shephard, 1985). Because the GLTEQ has been shown to be sensitive to activity level differences among individuals with tetraplegia versus paraplegia (Noreau, Shephard, Simard, Pare, & Pomerleau, 1993), it was considered appropriate for this study. Using a separate item for each intensity level, the questionnaire assessed the frequency of mild, moderate, and strenuous exercise completed during free time for at least 15 min in a typical week. The traditional examples of activity used to describe each intensity of exercise (e.g., running, walking) were modified for individuals with SCI (see Appendix A). The activities listed were drawn from focus group discussions (Martin et al., 2002). In accordance with the recommendation of Streiner and Norman (1999), one individual with SCI confirmed the face validity of the modifications. The fourth GLTEQ item, How often in a 7-day period do you engage in physical activity long enough to work up a sweat, was not included in the modified scale because many individuals with SCI are unable to sweat as a result of their injury. The GLTEQ is typically scored by weighting each exercise frequency by its estimated metabolic equivalent (METs) and summing these values for a total score. However, due to changes in metabolic functioning associated with spinal cord injury, these standard MET values cannot be applied to individuals with SCI (Mollinger et al., 1985; Monroe et al., 1998). With no MET values available, a total energy expenditure value could not be calculated. Consequently, our measure of exercise was simply the number of weekly bouts of activity performed at each of these three intensity levels. Analyses. Correlational analyses were first conducted to examine the relationship among the TPB constructs. Second, to test the hypotheses, hierarchical regression analyses were conducted. Separate analyses were conducted for individuals with tetraplegia versus paraplegia because previous research has identified spinal cord lesion level as a moderator of exercise behavior (Hedrick & Broadbent, 1996; Noreau et al., 1993). Correlational Analyses Results Consistent with the tenets of the TPB, for individuals with tetraplegia, intentions to exercise were significantly correlated with subjective norms (r =.36, p =.002), attitudes (r =.34, p =.004) and perceived behavioral control (r =.66, p <.01). Furthermore, strenuous and moderate intensity exercise behavior were each

7 Exercise Determinants and Spinal Cord Injury 77 Table 2 Descriptive Statistics and Bivariate Correlations Among the Theory of Planned Behavior Variables and Years Post Injury Mean SD Tetraplegia (n = 70) 1. YPI 9.09 ± Attitude ± SN.13.54** 6.40 ± PBC.02.37**.30* ± Intentions.23.34**.36**.66** 5.49 ± Mild Ex.27* * ± Moderate Ex.14.46**.21.45**.42** ± Strenuous Ex **.33**.10.41** 1.35 ± 2.08 Paraplegia (n = 54) 1. YPI 8.10 ± Attitude ± SN.27*.55** 6.22 ± PBC.10.29* ± Intentions.30* * 5.92 ± Mild Ex ± Moderate Ex ± Strenuous Ex ± 2.21 Note. *p <.05, **p <.01, between group difference as determined by a MANOVA of exercise behavior. YPI = years post injury, PBC = perceived behavioral control, SN = subjective norms, Ex = exercise.

8 78 Latimer, Martin Ginis, and Craven correlated with intentions (rs >.33, ps.005) and perceived behavioral control (rs >.40, ps <.001). Mild intensity exercise was correlated with perceived behavioral control only (r =.24, p =.05). These results are presented in Table 2. Contrary to the predictions of the TPB, for individuals with paraplegia, attitudes and subjective norms did not correlate with intentions. Perceived behavioral control was the only TPB variable that was significantly correlated with intentions, r =.30, p =.031. Strenuous intensity exercise correlated with perceived behavioral control (r =.26, p =.05) but not intentions. Mild and moderate intensity exercise behaviors were not correlated with either intentions or perceived behavioral control. Regression Analyses Prior to conducting these analyses, regression assumptions were tested. Inspection of the residuals on P-P plots indicated that data were slightly skewed and heteroscedastic according to criteria outlined by Tabachnick and Fidell (2001). Because regression is robust with respect to deviations from normality, slightly skewed data still provide a reliable and accurate estimate of regression weights (Kleinbaum, Kupper, Muller, & Nizam, 1998). Thus, no steps were taken to correct for skewness. Heteroscedasticity can result from an association between the outcome variable and another uncontrolled variable. In accord with Tabachnick and Fidell s suggestions for correcting hetroscedasticity, a potentially related variable, number of years post injury (Janssen et al., 1996), was controlled for in each regression. This reduced the heteroscedasiticity of the residuals, and thus number of years post injury was retained in the following analyses. In the first regression model, exercise intentions were regressed on number of years post injury (Step 1), attitude and subjective norms (Step 2), and perceived behavioral control (Step 3). The order of variable entry was based on the tenets of the TPB (Ajzen, 1985) and previous research examining the predictive capabilities of the TPB in the exercise domain (Courneya, 1995; Wankel, Mummery, Stephens, & Craig, 1994). In the second model, exercise behavior was regressed on number of years post injury (Step 1), intentions (Step 2), and perceived behavioral control (Step 3). Separate models were used to examine the predictors of each level of exercise intensity. Hypothesis 1: Predicting Exercise Intentions. Hierarchical regression analyses (Table 3) indicated that the overall models were significant for both individuals with tetraplegia, R 2 adjusted =.44, F(4, 61) = 13.89, p <.01 and paraplegia, R 2 adjusted =.13, F(4, 46) = 2.93, p =.03. Inspection of the beta weights indicated that our first hypothesis was not supported. That is, for individuals with tetraplegia, perceived behavioral control was the only TPB construct that predicted intentions (β =.59, p <.01). For individuals with paraplegia, none of the TPB constructs predicted intentions. Hypothesis 2: Predicting Exercise Behavior. Hierarchical regression analyses revealed that our second hypothesis was not supported (see Table 4). For individuals with tetraplegia, intentions were not a significant predictor of exercise behavior at any intensity. The contribution of perceived behavioral control was significant only in the moderate intensity exercise model (ΔR 2 =.06, p =.03, β =.33, p =.31). For individuals with paraplegia, the TPB constructs did not predict exercise behavior at any intensity.

9 Exercise Determinants and Spinal Cord Injury 79 Table 3 Hierarchical Linear Regression Models Predicting Intentions to Exercise for Individuals With Tetraplegia and Paraplegia adjr 2 ΔR 2 β Tetraplegia (n = 70) Step 1 YPI.05.06*.22* Step 2 Attitude.14.11*.05 SN.18 Step 3 PBC.44.30**.59** Paraplegia (n = 54) Step 1 YPI *.28* Step 2 Attitude SN.01 Step 3 PBC Note. *p <.05, **p <.01, YPI = years post injury, PBC = perceived behavioral control, SN = subjective norms. Discussion Using the TPB as a theoretical framework, this study examined predictors of exercise intentions and behavior among individuals with SCI. The regression analyses indicated that the TPB had limited utility in this population. Among individuals with tetrapelgia, perceived behavioral control was the only determinant of exercise intentions and behavior. Among people with paraplegia, none of the TPB constructs predicted exercise intentions or behavior. These results are in sharp contrast to those seen in studies of other populations where the TPB constructs have been identified as significant predictors of exercise intentions and behavior. Perceived behavioral control was the only TPB variable to emerge as a significant predictor of exercise intentions and behavior, but it was only significant for the sample of individuals with tetraplegia. This finding suggests that lesion level moderates the effects of perceived behavioral control. Given the difference in functional limitations between people with paraplegia versus tetraplegia (American Spinal Injury Association, 1992), the magnitude and number of barriers impeding exercise participation tend to be greater for the latter than the former (Tasiemski, Bergstrom, Savic, & Gardner, 2000). With greater barriers to overcome, and the fact that perceived behavioral control reflects confidence in one s ability to overcome these barriers, it is perhaps not surprising that perceived behavioral control was a more important determinant of intentions and behavior for individuals with tetraplegia. Hence, interventions that target underlying control beliefs (e.g., interventions that teach strategies for overcoming exercise barriers) could have

10 80 Latimer, Martin Ginis, and Craven Table 4 Hierarchical Regression Models Predicting 3 Intensities of Exercise for Individuals With Tetraplegia and Paraplegia Mild Moderate Strenuous intensity intensity intensity adjr 2 ΔR 2 β adjr 2 ΔR 2 β adjr 2 ΔR 2 β Tetraplegia (n = 70) YPI.06.07* Intentions ** **.17 PBC *.33* Paraplegia (n = 54) YPI < Intentions PBC.04 < Note. *p <.05, **p <.01, YPI = years post injury, PBC = perceived behavioral control.

11 Exercise Determinants and Spinal Cord Injury 81 positive effects on exercise intentions and behavior vis a vis enhanced perceived behavioral control (Ajzen, 2003b). There are a couple of possible explanations for the TPB s predictive weakness. The first is a measurement issue. Although Azjen (2003a, 2003b) suggests that indirect and direct measures examine the same constructs, the relationships between intentions and its antecedents tend to be stronger when direct measures are used (Ajzen, 1991). Our use of indirect measures of attitude and subjective norms may have attenuated relationships among the TPB constructs. If this study was replicated using direct measures, stronger relationships among the TPB variables might emerge. A second possible explanation is reflected in the unique contributions of years post injury to the prediction of exercise intentions (i.e., as years post injury increased, intentions decreased). As time post SCI increases, there is often a decrease in physical capacity and an increase in bodily pain due to the prolonged amount of time spent seated in a wheelchair (Janssen et al., 1996). Combined, these conditions can make it difficult for a person with a SCI to perform physical activities. Perhaps the physiological stress of chronic SCI is so wearing that over time, regardless of knowing and being told that exercise is beneficial, individuals simply lose incentive to be active. If this is the case, efforts must be directed toward the promotion of exercise during the acute phase of SCI and the maintenance of an active lifestyle in the chronic phase. Maintaining an active lifestyle will help to prevent decrement in physical capacity and the exacerbation of pain symptoms (Hicks et al., 2003). Another possible reason for the TPB s failure to better predict exercise behavior may reflect limitations of our exercise behavior measure and its ability to accurately capture exercise performed by people with SCI. Although we modified the GLTEQ to reflect a variety of exercise activities that people with SCI might perform, the list was not exhaustive, and some common activities might have been overlooked. It is also possible that despite the questionnaire s instructions, some participants incorrectly included physical activities of daily living (e.g., wheeling for transportation, transferring) when responding to the exercise measure. This would result in a lack of correspondence between the TPB questionnaire items (which targeted thoughts about structured exercise bouts) and the exercise questionnaire items and would diminish the correlations between these variables. The development of valid and reliable measures of structured exercise and physical activities of daily living for people with SCI would alleviate these problems. Our cross-sectional sample is a weakness of this study. In assessing determinants of a behavior, a prospective design is preferred such that the measurement of antecedents precedes the measurement of the actual behavior (e.g., behavior is assessed a week after antecedents). With the majority of our participants drawn from drop-in medical clinics, it could not be assured that participants would be in attendance the following week to complete a follow-up assessment. We recognize, however, the need for prospective studies of the TPB in this population. We also acknowledge the limited generalizability of our findings primarily to individuals who utilize community service programs for people with SCI. It is possible that the people who use these programs are in the community more often than nonusers and are thus more aware of the benefits of exercising and staying physically fit. Although a random sample of people with SCI would provide greater

12 82 Latimer, Martin Ginis, and Craven generalizability, there are, of course, great difficulties in accessing this type of sample. The 22% return rate for our mailed questionnaire highlights the challenge of reaching sufficiently large numbers of people with SCI through population survey techniques. Note, however, we did not employ any techniques that would have increased return rate (e.g., advance notice of survey arrival, post card reminders; Poretta, Kozub, & Lisboa, 2000). Implementation of such strategies is strongly recommended for future survey-based investigations among individuals with SCI. In summary, the findings from this study indicate limited utility of the TPB to predict exercise behavior among individuals with SCI. The lack of support for the tenets of the TPB suggests that an alternative framework that captures a broader range of cognitions (e.g., social cognitive theory; Bandura, 1997) might be more suitable for understanding exercise behavior among individuals with SCI. Yet, despite this limited prediction of exercise behavior, this study does begin to provide some direction for facilitating exercise participation among individuals with SCI and tetraplegia, in particular. With perceived behavioral control emerging as a significant predictor of exercise intentions, there is a definite need to promote a type of exercise activity that reduces the number of difficult-to-control barriers faced by individuals with tetraplegia. One possibility might be to promote a regime of physically active activities of daily living (e.g., lifting groceries, wheeling up ramps rather than taking the elevator). The efficacy of lifestyle physical activity has begun to be established in populations without a physical disability (Anderson et al., 1999; Dunn et al., 1999) but is an avenue that has not yet been explored among those with disability resulting from an SCI. References Ajzen, I. (1985). From intentions to action a theory of planned behavior. In J. Kuhl & J. Beckmann (Eds.), Action control: From cognition to behavior (pp ). Heildelberg: Springer-Verlag. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, Ajzen, I. (2003a). Construction of a Standard Questionnaire for the Theory of Planned Behavior. Retrieved May 5, 2003 from www. unix.oit.umass.edu/~aizen/tpb Ajzen, I. (2003b). Behavioral Interventions Based on the Theory of Planned Behavior. Retrieved May 5, 2003 from unix.oit.umass.edu/~aizen/tpb Ajzen, I. & Driver, B.L. (1992). Application of the theory of planned behavior to leisure choice. Journal of Leisure Research, 24, American Spinal Injury Association. (1992). International medical society of paraplegia standards for neurologic and functional classification of spinal cord injury. Chicago: ASIA/IMOSP. Anderson, R.E., Wadden, T.A., Bartlett, S.J., Zemel, B., Verde, T.J., & Franckowiak, S.C. (1999). Effects of lifestyle activity vs. structured aerobic exercise in obese women: A randomized trial. Journal of American Medical Association, 281, Bandura, A. (1997). Self-Efficacy: The exercise of control. New York: Freeman. Buchholz, A.C., McGillvray, C.F., & Pencharz, P.B. (2003). Physical activity levels are low in free-living adults with chronic paraplegia. Obesity Research, 11, Canadian Paraplegic Association. (2000, July 27). The workforce participation survey of Canadians with spinal cord injury. Retrieved January 23, 2003 from http: //

13 Exercise Determinants and Spinal Cord Injury 83 Courneya, K.S. (1995). Understanding readiness for regular physical activity in older individuals: An application of the theory of planned behavior. Health Psychology, 14, Courneya, K.S., & Friedenreich C.M. (1997). Determinants of exercise during colorectal cancer treatment: An application of the theory of planned behavior. Oncology Nursing Forum, 24, Courneya, K.S., & McAuley, E. (1993). Predicting physical activity from intentions: Conceptual and methodological issues. Journal of Sport & Exercise Psychology, 5, Crocker, P.R.E. (1993). Sport and exercise psychology and research with individuals with physical disabilities: Using theory to advance knowledge. Adapted Physical Activity Quarterly, 10, Dunn, A.L., Marcus, B.H., Kampert, J.B., Garcia, M.E., Kohl, H.W., & Blair, S.N. (1999). Comparison of lifestyle and structured interventions to increase physical activity and cardiorespiratory fitness: A randomized trial. Journal of American Medical Association, 281, Duran, F.S., Lugo, L., Ramirez, L., & Eusse, E. (2001). Effects of an exercise program on the rehabilitation of patients with spinal cord injury. Archives of Physical Medicine and Rehabilitation, 82, Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intentions, and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley. Gagne, C., & Godin, G. (2000). The theory of planned behavior: Some measurement issues concerning belief-based variables. Journal of Applied Social Psychology, 30, Godin, G., & Kok, G. (1996). The theory of planned behavior: A review of its applications to health related behaviors. Science and Health Promotion, 11, Godin, G., Shepard, R.J., Davis, G.M., & Simard, C. (1989). Prediction of exercise in lower limb disabled adults: The influence of cause of disability (traumatic or atraumatic). Journal of Social Behavior and Personality, 4, Godin, G., & Shephard, R.J. (1985). A simple method to assess exercise behavior in the community. Canadian Journal of Applied Sports Science, 10, Green, S.B. (1991). How many subjects does it take to do a regression analysis? Multivariate Behavioral Research, 20, Hagger, M.S., Chatzisarantis, N.L.D., & Biddle, S.J.H. (2002). A meta-analytic review of the theories of reasoned action and planned behavior in physical activity: Predictive validity and the contribution of additional variables. Journal of Sport and Exercise Psychology, 24, Hankins, M., French, D. & Horne, R. (2000). Statistical guidelines for studies of the theory of reasoned action and the theory of planned behavior. Psychology and Health, 15, Hausenblas, H.A., Carron, A.V., & Mack, D.E. (1997). Application of the theories of research action and planned behavior to exercise behavior: A meta analysis. Journal of Sport and Exercise Psychology, 19, Heath, G.W., & Fentem, P.H. (1997). Physical activity among persons with disabilities: A public health perspective. Exercise and Sport Science Reviews, 25, Hedrick, B.N. & Broadbent, E. (1996). Predictors of physical activity among university graduates with physical disabilities. Therapeutic Recreation Journal, Hicks, A.L., Martin, K.A., Ditor, D.S., Latimer, A.E., Craven, C., Bugaresti, J. et al. (2003). Long-term exercise training in persons with spinal cord injury: Effects on strength, arm ergometry performance and psychological well-being. Spinal Cord, 41,

14 84 Latimer, Martin Ginis, and Craven Jacobs, P.L., Nash, M.S., & Rusinowski, J.W. (2001). Circuit training provides cardiorespiratory and strength benefits in persons with paraplegia. Medicine and Science in Sports and Exercise, 33, Janssen, T.W.J., van Oers, C.A.J.M., Rozendall, E.P., Willemsen, E.M., Hollander, A.P. & van der Woude, L.H.V. (1996). Change in physical strain and physical capacity in men with spinal cord injuries. Medicine and Science in Sports and Exercise, 28, Janssen, T.W.J., van Oers, C.A.J.M., van der Woude, L.H.V., & Hollander, A.P. (1994). Physical strain in daily life of wheelchair users with spinal cord injuries. Medicine and Science in Sports and Exercise, 26, Kleinbaum, D.G., Kupper, L.L., Muller K.E. & Nizam, A. (1998). Applied regression analysis and other multivariate methods (3 rd ed.). Boston, MA: Duxbury Press. Maddux, J.E. (1993). Social cognitive models of health and exercise behavior: An introduction and review of conceptual issues. Journal of Applied Sport Psychology, 5, Martin, K.A., Latimer, A.E., Francoeur, C., Hanley, H., Watson, K., Hicks, A.L. et al. (2002). Sustaining exercise motivation and participation among people with spinal cord injury: Lessons learned from a 9-month intervention. Palaestra, 18, Mollinger, L.A., Spurr, G.B., el Ghatit, A.Z., Barboriak, J.J., Rooney, C.B., Davidoff, D.D. et al. (1985). Daily energy expenditure and basal metabolic rates of patients with spinal cord injury. Archives of Physical Medicine and Rehabilitation, 66, Monroe, M.B., Tataranni, P.A., Pratley, R., Manore, M.M., Skinner, J.S., & Ravussin, E. (1998). Lower daily energy expenditure as measured by a respiratory chamber in subjects with spinal cord injury compared with control subjects. American Journal of Clinical Nutrition, 68, Noreau, L., Shephard, R.J., Simard, C., Pare, G., & Pomerleau, P. (1993). Relationship of impairment and functional ability to habitual activity and fitness following spinal cord injury. International Journal of Rehabilitation Research, 16, Nunnally, J.C. (1978). Psychometric theory. (2 nd ed.) New York: McGraw-Hill. Pentland, W., Harvey, A.S., Smith, T., & Walker, J. (1999). The impact of spinal cord injury on men s time use. Spinal Cord, 37, Porretta, D.L., Kozub, F.M., & Lisboa, F.L. (2000). Documentary analysis of survey research in adapted physical activity: Adapted Physical Activity Quarterly, 17, Rejeski, W.J. (1995). Motivation for exercise behavior: A critique of theoretical directions. In G.C. Roberts (Ed.), Motivation in sport and exercise (pp ). Champaign, IL: Human Kinetics. Rimmer, J.H., Braddock, D., & Pitetti, K.H. (1996). Research on physical activity and disability: An emerging national priority. Medicine and Science in Sports and Exercise, 28, Rogers, R.W. (1975). A protection motivation theory of fear appeals and attitude change. Journal of Applied Social Psychology, 91, Rosenstock, I.M. (1974). The health belief model and preventive health behavior. Health Education Monographs, 2, Shifflet, B., Cator, C., & Megginson, N. (1994). Active lifestyle adherence among individuals with and without disability. Adapted Physical Activity Quarterly, 11, Streiner, D.L., & Norman, G.R. (1999). Health measurement scales: A practical guide to their development and use (2 nd ed.). New York: Oxford University Press. Tabachnick, B.G., & Fidell, L.S. (2001). Using multivariate statistics. (4th ed.) Boston: Allyn and Bacon. Tasiemski, T., Bergstrom, E., Savic, G., & Gardner, B.P. (2000). Sports, recreation and employment following spinal cord injury a pilot study. Spinal Cord, 38,

15 Exercise Determinants and Spinal Cord Injury 85 Wankel, L.M., Mummery, W.K., Stephens, T., & Craig, C.L. (1994). Prediction of physical activity intentions from social psychological variables: Results from the Campbell s Survey of Well-Being. Journal of Sport & Exercise Psychology, 16, Wonnacott, T., & Wonnacott, R. (1987). Regression: A second course in statistics. Malabar, FL: Robert E. Krieger. Wu, S.K. & Williams, T. (2001). Factors influencing sport participation among athletes with spinal cord injury. Medicine and Science in Sports and Exercise, 33, Authors Note 1 Indirect measures of the TPB constructs are based on the expectancy-value model where salient beliefs are multiplied by the value an individual places on each belief. A number of studies have reported that measuring the value component does not substantially improve predictions of intentions and, in fact, may attenuate them (Ajzen, 1991; Ajzen & Driver, 1992; Gagne & Godin, 2000; Hankins et al., 2000). Consequently, the indirect measures of the TPB constructs were based on expectancy measures only (i.e., behavioral, normative, and control beliefs) and did not incorporate the corresponding value measures. Acknowledgment Amy Latimer s research was supported by a Fellowship from the Social Sciences and Humanities Research Council of Canada. Appendix A Modified Godin Leisure Time Exercise Questionnaire Items Exercise intensity Descriptor Activity example Strenuous Exhausting Vigorous wheeling, sledge hockey,basketball, wheelchair rugby, vigorous swimming Moderate Not exhausting Moderate wheeling, tennis, easy hand cycling (stationary or hand cycle), volleyball, badminton easy swimming, dancing Mild Minimal effort Yoga, archery, fishing from river bank, bowling, horseshoes, golf, snow-mobiling, easy wheeling, car maintenance/washing playing with pets, housework (i.e., cleaning, vacuuming, repairs) Note. The scale instructions are as follows: Considering a 7-day period (one week), how many times on the average do you do the following kinds of exercise for more than 15 minutes during your free time (write appropriate number on each line).