Motivational characteristics and resistance training in older adults: A randomized controlled trial and 1‐year follow‐up

The aim of this study was to investigate the effects of a 9‐month supervised resistance training intervention on motivational and volitional characteristics related to exercise, and whether the absolute level and/or intervention‐induced change in these characteristics predict self‐directed continuation of resistance training 1 year after the intervention. Community dwelling older adults aged 65‐75, who did not fulfill physical activity recommendations, were randomized into resistance training intervention groups: training once‐ (n = 26), twice‐ (n = 27), three‐times‐a‐week (n = 28) or non‐training control group (n = 25). Training groups participated in supervised resistance training for 9 months: during months 1‐3 all groups trained twice‐a‐week and then with allocated frequencies during months 4‐9. Exercise‐related motivation, self‐efficacy, and planning were measured with questionnaires at baseline, month‐3 and month‐9. The continuance of resistance training was determined by interviews 6 and 12 months after the end of the intervention. The intervention improved action and coping planning as well as intrinsic motivation (group × time P < .05). During 1‐year follow‐up, 54% of participants did not continue self‐directed regular resistance training, 22% continued regular resistance training once‐a‐week, and 24% twice‐a‐week. Increases in exercise self‐efficacy and intrinsic motivation related to training during the intervention predicted continuation of resistance training twice‐a‐week. Resistance training improved exercise‐related motivational and volitional characteristics in older adults. These improvements were linked to continuing resistance training 1 year after the supervised intervention. The role of these characteristics should be taken into account when promoting long‐term resistance training participation among older adults.


| INTRODUCTION
Regular and progressive resistance training (RT) offers major benefits to older adults' physical functioning and health, such as maintaining and increasing muscle mass, strength and power. [1][2][3] According to both the American College of Sport Medicine 4 and Finnish national physical activity recommendations, 5 older adults should participate in muscle-strengthening activities at least twice-a-week. However, in Finland, only 11% of older adults aged 65-74 report meeting these recommendations. 6 Hence, understanding the possible reasons for non-participation and participation in RT are important.
There are many theories explaining the potential psychological reasons for participating in exercise, and motivation is one of the key characteristics behind a behavior. 7,8 According to self-determination theory by Ryan and Deci,8 motivation can arise from the satisfaction and pleasure of the activity itself (intrinsic motivation) or from extrinsic outcomes achieved via the activity (extrinsic motivation). Motivation is seen as a continuum from more extrinsic motivation to more intrinsic and autonomous motivation. 8 In the Health Action Process Approach (HAPA) by Schwarzer, motivation is distinguished from volition: the motivational phase leads to intention and secondly, the volitional phase leads to actual behavior. 9 In the volitional phase, implementing intentions is closely related to action planning, which includes planning when, where and how to exercise, as well as coping planning, which describes how difficult situations will be confronted. 9,10 Both of these have been shown to be predictors of exercise adherence. 11 In addition, self-efficacy is needed in both motivational and volitional phases. 9,12 Based on Bandura's social cognitive theory, self-efficacy describes how confident an individual is about their abilities to succeed in different situations or tasks. 13 High self-efficacy is linked to commitment and intrinsic motivation. 13 Related to health-enhancing behavior, self-efficacy reflects, for instance, capabilities to perform an action and to overcome different barriers (coping self-efficacy). 9,12 Although exercise-related motivational and volitional characteristics have been shown to be determinants of exercise participation, 11,14,15 much less is known specifically regarding RT. Winett et al. 16,17 suggest that RT may require a specific set of beliefs compared to other physical activities because of its nature. For example, confidence to lift external loads and increase the training load regularly is a specific requirement in RT. Nevertheless, there is some evidence that exercise selfefficacy is also closely related to RT; exercise self-efficacy beliefs have been associated with the continuance of RT after intervention in older adults, 18 but not in all studies. 19 Interestingly, there is some evidence that the association between motivational and volitional characteristics and RT may be bidirectional; these characteristics are related to adoption and continuance of physical activity, but also RT in itself may be able to improve these characteristics. For example, RT has had a positive effect on confidence to lift external loads in older adults 20 and on exercise self-efficacy in younger adults. 21 In addition, there is evidence that RT interventions lasting 12-week (three times-a-week) can improve introjected, identified, and intrinsic motivation for exercise. 22,23 However, it is currently unclear whether RT training frequency influences these findings in older adults.
Therefore, the purpose of this study was to; (i) investigate the effects of a 9-month RT intervention with different training frequencies on exercise-related motivational and volitional characteristics, assessed here by motivation, self-efficacy, action, and coping planning, and (ii) investigate whether their absolute level and/or intervention-induced change predicted self-directed RT continuance 1 year after the intervention. Based on the review of the literature, it was hypothesized that RT will improve motivational and volitional characteristics in older adults and that higher levels of exercise motivation and self-efficacy predict RT continuance after the intervention.
The present study was a secondary analysis of a randomized controlled trial "Get in Shape in the Team Research" (Clinical trials (clinicaltrials.gov) register number NCT02413112). Previous studies that have published the physical outcomes of the trial showed that maximum strength 24 and cardiorespiratory fitness 25 improved after 3 months of training, and a higher training frequency provided greater benefit for maximum dynamic strength, but not for functional capacity, over the 9-month intervention. 26 2 | METHODS

| Study design and participants
This study was a secondary analysis of a parallel-group randomized controlled trial. Ethical approval for the study was obtained from the Ethical Committee of the University of Jyväskylä and all participants signed written informed consent prior to participation. The study design, participants, and intervention have been described in detail in previous studies. 24,26 Participants were randomly assigned to 4 groups: RT once-a-week (RT1 n = 26), RT twice-a-week (RT2 (n = 27), RT 3 times-a-week (RT3 n = 28), and non-training control group (CG, n = 25). Pre-trial power analysis for primary outcomes of the study (maximum strength and functional capacity) was based on the effect sizes reported in a meta-analysis by Liu and Latham. 1 With a 75:25 intervention-to-control ratio, a sample size of 60 for strength and 88 for functional capacity was sufficient to reach 80% probability to observe treatment difference with a 5% level of significance.
The flow diagram of the study is shown in Figure 1. Participants were selected by random sampling from a population register of community-dwelling older adults aged 65-75 living in the Jyväskylä area. Recruitment occurred in October-December 2014. Inclusion criteria were; (a) leisure-time aerobic exercise less than 3 h/wk, (b) no previous regular RT experience, (c) BMI<37, (d) no previous testosterone-altering treatment, (e) no serious cardiovascular disease that may affect participating in RT, (f) no medication related to the neuromuscular or endocrine systems, (g) capability to walk without walking aids, and (h) no smoking. Two thousand invitation letters were sent with a response rate of 23% (n = 454). Those who met the inclusion criteria (n = 148) were invited to an information session and 116 participants attended a physician's examination: 8 persons were excluded because of medical reasons and 2 persons were not interested in participating. Hence, 106 participants remained for randomization. Randomization and allocation were made by the principal investigator of the study who was not involved in data collection. Randomization was performed by an online random number generator in a block of 100 participants for 4 groups, 25 participants for each. Because it was assumed that higher training frequency is likely to increase non-compliance and dropout rates, 3 participants were randomized to RT3, 2 to RT2, and 1 to RT1 from the remaining 6 participants. Two participants dropped out from CG because of the randomization result.

| Intervention
Training groups (RT1, RT2, and RT3) participated in a 9-month supervised whole-body RT intervention at the Faculty of Sport and Health Sciences gym, University of Jyväskylä. The exact training program has been previously reported, 24,26 but briefly, each 1-hour training session included a 10-minute warm-up and 8-9 exercises for different muscle groups (eg leg press, knee extension and flexion, chest press, pulldown, pushdown, pec deck, ab crunch, and back extension) and was supervised by trained personnel. All training groups followed an alternating 2-session training program throughout the intervention. Months 1-3: To become familiar with RT and to build capacity for high-load RT, all training groups (RT1, RT2, and RT3) participated in supervised RT twice-a-week for the first 3 months. The focus of this initial period was on muscular endurance using low F I G U R E 1 Flowchart of the study loads, with participants performing 2-3 sets of 14-20 repetitions with 0-1-minute rest between sets for each exercise. Months 4-9: The training groups started to exercise with different frequencies: completing the 2-session program took 2 weeks from RT1, 1 week from RT2 and RT3 completed 3 cycles in 2 weeks. The frequency-specific 6-month training period focused on developing muscle hypertrophy and maximum strength. The participants performed 4-5 sets of 4-12 repetitions with 1-3-minute rest between sets. CG was instructed not to change their lifestyle during the intervention and after month-9 measurements they had an opportunity to participate in supervised RT twice-a-week for 6 months.
After the 9-month intervention, participants of the training groups could no longer train in the same gym as during the intervention and had to arrange possible continuance of training by themselves. To lower the threshold to continue training independently, participants were given a list of possible low-cost gyms in the city area from the principal investigator upon completion of the month-9 tests.

| Measurements
Motivational and volitional characteristics were measured with computer-based questionnaires related to exercise or physical activity in general (not specifically RT).
Exercise self-efficacy was measured by 10 questions developed by Schwarzer: 5 questions evaluated action and coping self-efficacy 12 and 5 questions barrier self-efficacy. 27 These scales have satisfactory psychometric properties 12,27 and have been used among Finnish adults and older adults. 28,29 Participants were asked to evaluate how confident they were to maintain exercise in different situations related to adoption (eg "even if I have to make a detailed plan to exercise") and to overcome barriers (eg "even when I am busy"). The answers were given on a scale from 1 = very certain I cannot to 4 = very certain I can, and aggregated. Cronbach's alphas were 0.91, 0.92, and 0.92 at baseline, month-3, and month-9, respectively.
Action planning and coping planning to exercise were both assessed with 4 questions. 10 Participants were asked whether they had a detailed plan when, where, how, and how often to exercise (action planning) and how to combat the different barriers and obstacles to exercise (coping planning). The scale was from 1 = not at all true to 4 = very true. The scale has good psychometric properties in a wide population age range. 10 Cronbach's alphas were 0.89, 0.90, and 0.89 for action planning and 0.90, 0.91, and 0.86 for coping planning at baseline, month-3, and month-9, respectively.
Motivation for physical activity and training was measured with the Exercise Self-Regulation Questionnaire (SRQ-E). 30 The questionnaire has shown adequate psychometric properties among different populations. [31][32][33] The questionnaire has two parts: 16 statements measure motivation regarding physical activity (eg "I try, or would like to try, to be physically active regularly because I enjoy physical activities") and 12 statements regarding training (eg "I exercise/work out (or would like to work out) because I feel pressure to work out"). The scale was from 1 = not at all true to 7 = very true. The mean scores for each motivational regulation (external, introjected, identified, and intrinsic) were calculated separately for physical activity and training. Cronbach's alphas showed an adequate fit 34 (range 0.70-0.92) to all 4 regulation styles for physical activity and to 3 regulation styles for training. The exception was introjected regulation for training: Cronbach's alphas were 0.47 at baseline, 0.57 at month-3, and 0.39 at month-9. The 3 items measuring introjected regulation for training were "I exercise/ work out because I would feel bad about myself if I didn't do it",…"I'd be afraid of falling too far out of shape if I didn't" and …"I feel pressured to work out." Deleting any of these items would not have improved alpha levels.
Leisure-time aerobic exercise was estimated at baseline by a single question specifying the average weekly minutes participants engaged in leisure-time aerobic exercise. During the intervention, diaries tracked daily activities, as well as the duration and intensity of the activities recorded. 35 The average weekly minutes for leisure-time aerobic exercise for months 1-3 and 4-9 were calculated from the diaries.
Continuance of RT: In interviews 6 months after the intervention ("follow-up 1″) and 12 months after the intervention ("follow-up 2"), participants were asked whether they had continued RT. At follow-up 1, the interviews were conducted face-to-face to all training group participants who participated in follow-up strength and functional capacity measurements (n = 66). At follow-up 2, interviews were conducted by telephone from all training group participants except the 2 dropouts and 1 participant (from RT3) who was not reachable (n = 78). From those who had continued RT, the regularity and average number of RT sessions per ordinary week was asked (common short breaks, eg because of flu or travel should not be taken into account) in both interviews. Training either individually or in a supervised group was accepted. The average weekly frequency to participate in RT during the year was calculated from values reported in follow-up 1 and follow-up 2. Those who reported that they had continued RT regularly during the year and at least 6 months twice-a-week and during the other 6 months once to twice-a-week (average for the year 1.75-2.5 times-a-week) were categorized as twice-a-week continuers. The cutoff value of 1.75 was used because these participants intended to participate in RT twice-a-week (and mostly achieved their intention) but there were some weeks when they trained only once-a-week due to holiday/illness or some other reason. Those who reported that they had continued RT regularly during the year at least once-a-week but not twicea-week throughout the year (average for the year 1-1.5 timesa-week) were classified as once-a-week continuers. All others, including both participants who had not continued RT at all and participants who had continued less than once-a-week on average (eg some participants reported training approx. oncea-month), were classified as non-continuers.

| Statistical analysis
Analyses were performed using IBM SPSS Statistics 24.0. A significance level of P < .05 was used in all analyses. The first study aim, to investigate the effects of intervention on motivational and volitional characteristics was analyzed by an intention-to-treat principle, with the exception of the 2 participants that dropped out immediately after the randomization before baseline measurements for motivational and volitional characteristics. The differences between groups at baseline were analyzed by one-way ANOVA for continuous variables and Chi-Square test for categorical variables. The generalized estimation equation (GEE) method with an unstructured working correlation matrix was used to analyze group × time differences between groups. Based on the structure of the intervention, 3 sets of GEE analyses were performed; (a) baseline -month-3 between CG and the training group (RT1, RT2, and RT3), (b) month-3 -month-9, and (c) baseline -month-9 between CG, RT1, RT2, and RT3. Because the amount of leisure-time aerobic exercise (min/wk) was not stable in all groups during the intervention (assessed by physical activity diaries), GEE analyses were repeated and adjusted for the amount of leisure-time aerobic exercise. The diaries were missing from 7 participants (1 from each RT group and 4 from CG), hence those participants were excluded from the adjusted analyses. The results of analyses with and without aerobic exercise adjustment were similar, therefore only the adjusted GEE results are shown in the Results section. In addition to GEE analyses, the standardized effect sizes for differences in change-scores between groups were calculated with Cohen's d formula. 36 The changes in motivational and volitional characteristics during the intervention were calculated by subtracting the baseline value from the intervention completion (month-9) value. Within-group differences were analyzed by paired sample t-tests.
The second study aim was to investigate predictors of continuing RT after the intervention. Both the absolute level at post-intervention (month-9) and change during the intervention (month-9 value -baseline value) of motivational and volitional characteristics were analyzed as possible predictors. Differences in possible predictors between non-continuers, once-a-week continuers, and twice-a-week continuers were analyzed by one-way ANOVA and Chi-Square tests.

| Descriptive statistics
The flow of the participants throughout the study is shown in Figure 1. Three participants dropped out from CG, 1 from RT1, and 1 from RT2 (Figure 1).The dropout from RT1 was due to back pain induced by the strength testing in month-6, and other dropouts occurred due to health (acid reflux events, stress-related high blood pressure, cancer recurrence) or personal issues unrelated to the study.
There were some adverse events, as judged by the investigators, in the study. Adverse effects were collected by selfreports: participants reported more serious ones directly to the principal investigator and documented minor ones in their training diaries. The most serious adverse event was the back pain induced by the strength testing. One participant strained a hamstring muscle walking home after strength measurements at month-3 and could continue the training program but no longer participate in maximum muscle strength measurements. Based on information collected from training diaries, about one-third of the participants in training groups had some temporary pain or soreness, for example, in the knee joint, shoulder or back. For these participants, the training program was adjusted (eg use of lighter loads, modified exercise technique) for a maximum of 2-3 weeks. These discomforts were judged to be part of the adaptation processes, and indeed these feelings were short-term and did not adversely affect participation rates during the 9-month intervention period.
Participants' demographic characteristics and baseline values in motivational and volitional characteristics are presented in Table 1. The only significant difference between groups at baseline was in coping planning (ANOVA F(df) = 3.30(3), P = .024), which had a higher value in RT1 compared to RT3.

| Changes in exercise-related motivational and volitional characteristics during the intervention
The effects of the intervention on motivational and volitional characteristics are presented in Table 2. The mean values at each measurement point and effect sizes for mean change between the groups can be found from supporting information. After 3 months of intervention, there were significant group × time effects for exercise self-efficacy, action and coping planning, and intrinsic motivation related to training. From month-3 to month-9, there were no statistically significant between-group differences. Throughout the 9-month intervention, changes occurred in several measured motivational and volitional characteristics. Action planning improved in all training groups compared to CG. Coping planning and intrinsic T A B L E 1 Participants' demographic, motivational, and volitional characteristics at baseline. Mean values with standard deviations in parentheses or frequencies presented

T A B L E 2 Effect of the intervention on motivational and volitional characteristics, analyzed by generalized estimated equations (GEE),
p-values presented. Models adjusted by mean aerobic exercise (min/wk) from a months 1-3, b months 3-9, and c months 1-9. PA, motivational regulation related to physical activity, TR, motivational regulation related to training. Significant differences between d CG vs RT1 P = .019, CG vs RT2 P < .001, CG vs RT3 P < .001; e CG vs RT2 P = .010, CG vs RT3 P = .001; f CG vs RT2 P = .026; RT1 vs RT2 P = .004; g CG vs RT2 P = .013, CG vs RT3 P = .009; h CG vs RT2 P = .008, CG vs RT3 P < .001, RT1 vs RT2 P = .047, RT1 vs RT3 P = .003. motivation related to physical activity improved in RT2 and RT3 compared to CG. Both RT2 and RT3 increased their intrinsic motivation related to training and RT2 also introjected regulation related to physical activity compared to CG and RT1. The changes in these variables are shown in Figure 2.

| Predictors of RT continuance
Out of 78 participants, 31 (40%) reported that they did not continue RT at all after the intervention. Five persons (6%) continued RT immediately after the intervention but quit prior to follow-up 1 (6 months after the intervention) and 6 persons continued RT occasionally (less than once-a-week on average) throughout follow-up (12 months). Therefore, 42 persons (54%) were counted as non-continuers. Seventeen persons (22%) reported to have trained regularly once-a-week on average during the whole year and 19 persons (24%) twice-a-week on average. The average weekly frequency to participate in RT varied between 1.0 and 2.5 (mean 1.58, SD 0.48). There were no significant differences between intervention groups in continuance rates ( Table 3). The level of motivational and volitional characteristics at post-intervention (month-9) and their change during the intervention (baseline to month-9) were analyzed as possible predictors for continuance of RT (Table 3). Twicea-week continuers had a greater increase in exercise self-efficacy during the intervention than non-continuers and a greater increase in intrinsic motivation related to training compared to both non-continuers and once-aweek continuers.

| DISCUSSION
The purpose of this study was to (a) investigate the effects of an RT intervention with the different training frequencies on exercise-related motivational and volitional characteristics, and (b) whether the absolute level and/or change in these characteristics predict self-directed continuance of RT 1 year after the intervention. RT led to positive changes in exercise self-efficacy, planning, and autonomous motivation. Twentytwo percent of participants continued regular RT once-aweek and 24% twice-a-week during the next year after the intervention. A greater increase in exercise self-efficacy and intrinsic motivation related to training during the intervention predicted continuation of RT twice-a-week during follow-up.
The effect of RT on exercise self-efficacy and motivation is in-line with previous studies, [20][21][22][23]37 indicating that participation in RT intervention can increase self-confidence to maintain exercise behavior and increase intrinsic motivation to exercise in addition to physical outcomes. However, the improvements in exercise self-efficacy diminished after 3 months in the present study. It is plausible that at month-3 measurements participants were more confident because the intervention was going to continue but at post-measurements they had started to think about training independently. 38 Action and coping planning in relation to RT have not been studied previously, but the results of the present study suggest that participating in an RT intervention can also increase exercise planning. In relation to motivation, a limitation of this study was that the scale for introjected regulation for training had low internal consistency (alpha values 0.39-0.57). Low alpha reliability implies that the items intended to measure introjected regulation do not measure it in a consistent way in this study sample, even though the scale has had good psychometric properties in previous studies. [31][32][33] Therefore, the results for introjected regulation for training should be treated with extreme caution.
A possible source for exercise motivation could be the improvements in strength and functional capacity during the intervention. [24][25][26] However, in the present study, improvements in strength and functional capacity did not influence changes in motivational and volitional characteristics (tested in GEE analyses adjusted by change in strength and functional capacity, the results not included in the paper). In addition, other characteristics of the intervention may have affected the results. For instance, training in small groups (8-10 persons) and having support/encouragement from the research staff may have affected motivational and volitional characteristics. Hence, even though RT interventions seem to be an effective way to increase intrinsic motivation and planning, it is not clear whether performing RT itself (ie without these potential confounding factors) led to these findings. To examine this, the relationship between RT and these characteristics should be studied external to a structured scientific, group-based intervention.
When comparing the results between different training frequencies, there were no differences between groups from month-3 to month-9. It should be noticed, however, that when comparing month-9 results to baseline, RT1 only improved their action planning compared to CG, whereas RT2 and RT3 improved also coping planning and intrinsic motivation related to both physical activity and training. Furthermore, RT2 and RT3 improved their intrinsic motivation related to training also compared to RT1. In support of these findings, effect sizes suggest that training twice-or 3 times-a-week was related to greater changes in these characteristics over the entire intervention period. It is likely that the structure of the intervention influenced these findings and should be remembered when interpreting the results: all training groups participated in RT twice-a-week during the first 3 months and then split into different frequencies. This meant that the once-a-week group was forced to reduce their training frequency. Therefore, it is not possible to separate the effects of training frequency per se from the effects of reducing training frequency. In future studies, it should further be investigated whether it is the recommended RT frequency of at least twice-a-week 4 to derive also the most motivational benefits. However, our results suggest that greater resistance training frequency contributes little to improved motivational and volitional characteristics since we did not observed differences from month-3 to month-9 measurements.
Almost half of the participants (46%) continued RT during the 1-year follow-up at least once-a-week. The continuation rate is similar to the studies by Geirsdottir et al., 39 in which ~18% of participants continued RT once-or twice-a-week and ~24% ≥three times-a-week, and Inaba et al., 40 in which 43% continued RT at least once-a-week. In both studies the continuation was measured by self-reports. A limitation of the present study is that participants were not able to continue RT in the same gym used during the intervention. Finding a new place to continue RT could have been a significant barrier for some participants and reduced the continuance rates. In any case, it is a positive finding that almost half of the participants continued training at least once-a-week, since even participating in RT only once-a-week leads to important physical improvements for older individuals compared to not participating at all. 26,41 However, another limitation of this study is that the continuance frequencies are based on selfreports and information about participants' actual gym visits was not available. For example, Van Roie et al. 19 obtained information regarding gym visits from the fitness center and F I G U R E 2 Key findings: relative changes (±standard error estimation) in motivational characteristics between baseline and month-3 (A), between month-3 and month-9 (B), and between baseline and month-9 (C). TR, Training; PA, Physical Activity found that only 20% of participants continued RT during a 4-month follow-up and none of them met the recommended twice-a-week frequency. However, even though self-reports may both over-or underestimate the actual amount of physical activity, 42 even a single question could be a useful method to estimate physical activity. 43 The present study provided some interesting findings regarding motivational and volitional characteristics and RT continuation. Firstly, contrary to previous research, 15,18 the level of the characteristics was not related to RT continuation. A possible reason for that could be that the participants of this study had relatively high values in these characteristics already at baseline (eg compared to a sample of younger physically inactive Finnish adults 28 ). In addition, it is possible that participants may have had high motivation for exercise in general but they preferred physical activities other than RT. This may not have been observable in the present study because the measurements used were related to non-specific physical activity or exercise. Indeed, when exercise preferences were asked from participants at follow-up 2, over half of them mentioned some type of aerobic exercise and only 16% mentioned RT as their favorite type of exercise. Unfortunately, the information about frequency to participate in other physical activities during the follow-up period was not gathered in the present study. Secondly, increases in exercise self-efficacy and intrinsic motivation related to training during the intervention were related to continuance of RT during follow-up. Individuals with high exercise self-efficacy have better strategies and are ready to put more effort on exercising and overcoming barriers. It is possible that during regular participation in RT for 9 months, some participants realized that they could overcome the potential barriers and that contributed to continuing RT also after the intervention. This finding supports the results of previous studies in relation to self-efficacy change and exercise continuance. 18,44 Regarding intrinsic motivation, it is possible that simply those who enjoy RT continued training after the intervention. Thirdly, there were differences only between twice-a-week continuers and non-continuers, not between once-a-week continuers and non-continuers. Moreover, twice-a-week continuers had greater increases in intrinsic motivation also compared to once-a-week continuers. This is an important finding for future studies to investigate because RT twice-a-week is superior to once-a-week for strength gains. 45 According to the results of the presents study, it can be suggested that by improving motivational and   Level at post-intervention, mean (standard deviation). b Change during the intervention (from month-0 to month-9), mean (standard deviation). c Significant difference between non-continuers and twice-a-week continuers analyzed by ANOVA (P=.021) and Bonferroni post-hoc tests. d significant difference between non-continuers vs. twice-a-week continuers and once-a-week continuers vs twice-a-week continuers (P=.002).
volitional characteristics related to exercise it could be possible to also increase RT participation. It has been shown that RT programs with behavioral counseling are more effective for the adoption of RT and exercise self-efficacy than basic programs including only exercise. 46,47 Therefore, physical activity interventions should also aim to improve these characteristics in order to promote the continuance of behavior after the intervention. When generalizing the results of this study, it should be taken into account that first, only 23% of the original random sample replied to the invitation letters. As mentioned above, it is possible that the study sample was already highly motivated to participate in exercise/RT. Secondly, the sample consisted of healthy older adults aged 65-75 who did not meet the physical activity guidelines for either leisure time aerobic exercise or RT at baseline. Therefore, the results may not necessarily be generalized to different patient groups, all ages, or more physically active individuals. It should be noted that the amount of leisure time aerobic exercise and RT background at baseline were measured with single questions, thus the information about specific exercise patterns before the intervention is not available. Thirdly, as a secondary analysis of a randomized controlled trial, this study may have been unpowered to detect differences in motivational and volitional characteristics with this sample size.

| PERSPECTIVES
As resistance training is beneficial and recommended for older adults, 1-4 but only a minority regularly participate (~10%) in it, 6 it is essential to identify ways to promote participation in resistance training activities. The present study showed that a resistance training intervention can improve exercise-related motivational and volitional characteristics and these improvements were related to continuing selfdirected resistance training after the intervention. Therefore, if interventions can target and improve exercise motivation and self-efficacy, then it may be possible to increase participation rates in regular resistance training in healthy older populations.