The Presence of an Exercise Instructor or Group Members Affects Physical Self-Concept and Physical Self-Esteem: An Experimental Study in the Elderly ()
1. Introduction
There seems to be no doubt, that one’s self-concept and self-esteem are important for health [1] . Moreover, it is apparent that the presence of others in health-related environments (e.g., the exercise environment) may impact self-concept and self-esteem [2] . The mere presence of others in the exercise environment may, however, be less relevant in its impact on an individual’s self-concept and esteem than the issue of “who” such individuals are. Presently, however, it is unclear whether all individuals operating within the exercise environment have an equal impact on the self-concept and esteem levels of exercise participants. Thus, in the context of an exercise program, the purpose of the present experiment is to examine the relative impact of the presence of certain individuals-group members, an instructor, or both—on the physical self-concept and self-esteem of elderly individuals participating in an 8-week exercise program. Specifically, we assume that the well-investigated effect of exercise on self-concept and self-esteem [3] depend not only on whether or not others are present, but on whom such individuals are.
In terms of the rationale for the present study two main considerations are important to address. First, selfconcept and self-esteem have emerged as two of the strongest predictors of mental health and subjective wellbeing [1] [4] . Individuals evaluating themselves in a positive way report higher levels of mental health (e.g., emotional stability, hardiness, controllability) and lower levels of mental disorders (e.g., depression, anxiety; [2] ). Second, physical self-esteem in the elderly is of special interest, since research has shown that both general self-esteem and physical self-esteem declines markedly from the 60s to the 80s [5] [6] . This decline of self-esteem in older persons may not only impair health itself, but have a strong link to unhealthy behaviors (e.g., a sedentary life style [7] ), which in turn represent additional risks for health.
Physical self-concept and physical self-esteem. Physical self-esteem (often used synonymously with physical self-worth) is strongly related—albeit distinct from—physical self-concept. While self-concept is a self-description, self-esteem is self-evaluative in nature. More specifically, physical self-concept includes a description and a self-perception (e.g., self-descriptions of individual physical properties, attributes and behaviors). Physical self-esteem on the other hand, reflects an individual’s evaluation of his or her own physical attributes [3] . A further distinction between physical self-concept and physical self-esteem is that the former can be differentiated into different aspects of the physical self (e.g., endurance, strength), whereas the latter is generally conceived of as an overall evaluation of the various components of one’s physical self-concept [8] .
These conceptual considerations led Sonstroem and Morgan [9] to develop a hierarchical model in which the authors argued that descriptions or beliefs about physical competencies (i.e., physical self-concept) are a prerequisite for the enhancement of physical self-esteem. Despite these theoretical distinctions between physical selfconcept and physical self-esteem, empirically researchers have found strong overlaps between the two constructs [8] [10] .
Exercise and enhancement of physical self-concept and self-esteem. The assumption that exercise leads to an enhancement in physical self-concept and/or self-esteem has been empirically demonstrated (see [11] for a review in older adults). Most of the research in this area, however, has been conducted with middle age adults or youth [3] . With the exception of unpublished doctoral studies, Fox [3] found no published randomized controlled studies investigating the effects of exercise on physical self-concept or esteem in the elderly. One exception is a study by McAuley and colleagues who reported exercise induced physical self-esteem benefits in the elderly [12] . In this study participants took part in either a walking program or a toning exercise program, three times per week over a 6-month period. With both exercise groups the authors found increases in self-reported domain-specific esteem (domains: body, strength, condition), general physical esteem, and global esteem. In conclusion, research on the effects of exercise on elderly individuals’ physical self-esteem and physical selfconcept remains noticeably deficient [5] [13] . As indicated, given the decline in physical self-esteem among elder adults and the likely implications of such a decline for a reduction in physical activity, creating exercise environments that optimize elder adults’ self-concept and esteem is of clear importance.
Physical self-esteem, self-concept, and the exercise environment. With regard to the present study, it is important to ask, why would the presence of others in the exercise environment (e.g., an exercise program for the elderly) have an influence on physical self-concept and self-esteem development? The answer to this question may be better understood by referring to Cohen’s [14] conceptual considerations on the link between social relationships and health outcomes (e.g., enhancements of self-concept). In particular, the constructs of social integration and social support may be two important mechanisms that help explain the impact of the exercise environment on the development of elder adults’ physical self-concept and esteem, the subject of the present investigation. Social integration refers to the feeling of belongingness and being part of the group. In this regard, groups are functional since they help to “fulfill a need for positive self-esteem by allowing one to construct a social identity which is evaluatively positive” [15] . Hogg’s [15] conceptualization of social identity is especially germane for athletes and exercisers, many of whom develop positive self-evaluations through their affiliation with a particular sport, team or exercise group [16] [17] .
Social support on the other hand, refers to emotional and instrumental support in coping with external demands. Specifically, related others help the individual directly (i.e., support in solving tasks) or indirectly (i.e., reducing demands) to act effectively and successfully. In the domain of exercise, such perceived effects, success, or competencies lead to an enhancement of physical self-efficacy, control beliefs [4] , and self-evaluation [9] . Furthermore, for the elderly, McAuley showed that social support during an exercise program had positive effects on well-being [18] .
In summary, the links between the type of exercise environment and the development of physical self-concept and esteem may be explained by principally different mechanisms, namely social integration or social support.. Regardless of the mechanisms by which (i.e., “how”) the presence of others influences physical self-concept and esteem among elderly individuals involved in a physical activity program, researchers have yet to investigate the issue of “who” influences physical self-concept and self-esteem to the greatest extent—the instructor or other exercise group members. Given that exercisers may choose to exercise alone, in groups, or in structured programs led by a fitness leader, examining the impact of different social environments on elder adults’ self-concept and self-esteem has important implications for their well-being and sustained exercise involvement. Therefore, we designed an experiment to test the effect of two social conditions on physical self-concept and esteem changes, namely an “instructor” condition (i.e., with and without an instructor) and a “group” condition (i.e., exercising alone or as part of a group). Furthermore, regarding the hierarchical construction of the physical self-esteem, as dependent variables, we also tested distinct and specific parts of the physical self-concept (e.g., endurance, strength) as well as global physical self-esteem (i.e., self-evaluation of general athleticism). In general, we expected the presence of others—be it a group instructor and group members-to have a positive impact on the development of elder adults’ physical self-concept and esteem. However, since previous research has yet to address the issue of which individuals (instructor versus group) or which constellation of individuals (instructor and group, instructor/no-group, no-instructor/group, no-instructor/no-group) has the strongest effects on self-concept and esteem, we did not advance specific hypotheses on this issue.
2. Method
2.1. Participants
In total, 92 elderly individuals (57 females and 35 males) agreed to participate. Dropouts during the walking program (17 persons) and missing data (nine persons) resulted in a final sample size of 66 participants (46 females and 20 males) ranging in age from 59 to 81 years (M = 65.42, SD = 4.35). Except for one participant, all were physically active in some type of leisure time activity (e.g., hiking, gardening) at least once per week. Only nine persons, however, were previously active in regular exercise activities (e.g. in a sports club) at least once per week at the time of the investigation. Forty-five participants took medication regularly and 42 reported having a chronic disease (e.g., cardiovascular disease, rheumatic disease).
2.2. Treatment
All participants took part in a walking program after being instructed in correct walking techniques and important training principles. In particular, all participants received training information regarding warm-up procedures, how to find the optimal training intensity, and signals of overreaching (e.g., complaints). In terms of the optimal training intensity, participants were educated on how to assess their pulse control (130 - 150 bpm) and a moderate rate of subjectively perceived exertion (RPE = 12 - 14; [19] ). The program lasted eight weeks with an average of three training sessions per week. The walking program took place in a natural setting (forested parkland adjacent to the university sports centre). The training intensity in the group condition was determined based on an average of the physical fitness levels of group members. Each training session lasted approximately 30 to 45 minutes with an advised training intensity aimed at improving the basic endurance levels of participants. For ethical and medical reasons, members of the study team were available during the first two weeks of the exercise program to answer any questions the participants had, regardless of the experimental condition.
Treatment design. The presence of an instructor and a group during the walking sessions was manipulated via two between-subject-factors—factor “instructor” and factor “group”. Each factor was subdivided into two levels (2 × 2 experimental conditions) with one level reflecting an “absence” and one level reflecting a “presence”.
Manipulation of instructor presence. The factor “instructor” implied either the presence or absence of a professional fitness instructor. Among those training with an instructor and a group, participants trained three times per week, led by an instructor who implemented a predefined exercise regimen (all instructors of the study had been briefed for the exercise program before the beginning of the study). Due to organizational issues, participants who walked individually but with an instructor (instructor/no-group condition) were required to walk without the presence of the instructor in 15% - 20% of the exercise sessions. For participants in the no-instructor condition, such individuals determined their training dates and training area on their own. In terms of training duration and training intensity, participants in the no-instructor condition were instructed to consider the training principles highlighted in the treatment section described above.
Manipulation of group presence. The factor “group” implied either the presence or the absence of other study participants during the walking sessions. In the group condition, participants walked together with other participants in groups of four to eight members. The group sizes varied as a result of participant drop-out over the course of the 8-week program as well as organizational issues (e.g., conflicting obligations among study participants). In the no-group condition, participants walked independently without peers.
2.3. Measures
Physical self-concept. Physical self-concept was assessed using a German adaptation of the Physical Self-Description Questionnaire (PSDQ; [20] ). For the self-description of physical competencies, the German version (PSK; [21] ) contains five subscales: coordination (e.g. “I find it easy to control my movements”; α = 0.86), strength (e.g. “I am good at lifting heavy objects”; α = 0.89), flexibility (e.g. “My body is flexible”; α = 0.88), endurance (e.g. “I could be physically active for a long period of time without getting tired”; α = 0.85) and speediness (e.g. “Over a short distance I can run fast”; α = 0.71), with response options ranging from 1 (is not true) to 4 (is true) on a 4-point Likert type scale. The items of each factor were aggregated by computing a mean value for each factor. Reliability and validity of the German version have been demonstrated previously [21] .
Physical self-esteem. To assess the general self-evaluation of physical competencies (i.e., physical self-esteem) the subscale “general athleticism” of the PSK was used. This subscale contains six items, that evaluate one’s general physical abilities (e.g. “Other people think that I am good in sports”; α = 0.85). It should be mentioned, that Stiller, Würth, and Alfermann [21] did not develop this subscale for the assessment of self-esteem. However, the subscale contains many items with a strong evaluative connotation (e.g., “I am good in most sport activities”, “I am better in sports than most of my friends”), which per definition (see introduction) relates to self-esteem rather than self-concept. Response options ranged from 1 (is not true) to 4 (is true) on a 4-point Likert type scale. The items were aggregated by computing a mean value for the factor “general athleticism”.
2.4. Procedure
Participants were recruited via newspaper advertisements offering a specific walking-program free of charge aimed at enhancing the endurance capability of the elderly. For ethical and medical reasons, willing participants were obligated to prove their physical fitness through a health certificate completed by a physician. After randomly assigning eligible participants to one of the four experimental conditions, an introductory meeting was held prior to the commencement of the 8-week-program in order to prepare the sample for the study. Baseline physical self-concept and physical self-esteem measures were completed at this initial meeting. A few days after the introductory meeting, each experimental condition had its initial walking-session, followed by an eight week walking program as described above in the treatment section. After the final walking-session, physical selfconcept and physical self-esteem were measured once again. Thus, physical self-concept and physical self-esteem were measured twice (pre and post) resulting in a 2 × 2 × 2-study-design.
2.5. Data Analysis
Data were analyzed using SPSS 22.0. The analysis comprised descriptive and inferential examination of the study variables. In terms of inferential statistics, separate 2 by 2 ANOVAs with repeated measures (2 × 2 × 2- design) were computed for each of the six dependent variables including: coordination, strength, flexibility, endurance, speediness, general athleticism. The significance level was at p = 0.05.
3. Results
3.1. Descriptive Statistics
At the onset of the study, the total mean value for general athleticism across all groups was slightly below the midpoint of the 4-point Likert type scale (M = 2.26, SD = 0.58). This finding indicates that, overall, the sample had a moderate to low level of self-reported general athleticism. However, the individual values observed ranged from 1 to 4, indicating a fairly broad range in perceived athleticism. In terms of specific aspects of physical selfconcept, the total mean values of endurance (M = 1.95, SD = 0.58) and speediness (M = 2.18, SD = 0.61) were lower than general athleticism. Coordination (M = 2.53, SD = 0.57), strength (M = 2.32, SD = 0.68) and flexibility (M = 2.63, SD = 0.63), however, displayed higher mean values than general athleticism.
After the intervention, general athleticism (M = 2.47, SD = 0.60) was higher than before; it appeared at the midpoint of the scale with the individual values ranging from 1 to 3.67. The specific aspects of physical self-concept were higher as well. Again, endurance (M = 2.24, SD = 0.64) and speediness were below general athleticism, whereas coordination (M = 2.73, SD = 0.50), strength (M = 2.52, SD = 0.67) and flexibility (M = 2.79, SD = 0.63) were higher than general athleticism. For detailed information concerning the descriptive statistics of each experimental condition, see Table1
In terms of the bivariate correlations between the dependent variables, each correlation was significant, except for the correlation between strength and flexibility at the post-measurement (see Table 2). In particular, the correlations between general athleticism and the other variables were invariably large, falling above 0.50 in all instances.
3.2. Analyses of Variance
Physical self-concept. Regarding the factor coordination, the ANOVA resulted in a significant within-subject factor: F(1, 62) = 19.43, p < 0.001 , η2 = 0.24. As shown in Table 1, self-rated coordination increased with time. There were no significant main effects for the instructor: F(1, 62) = 0.16, p = 0.691, η2 < 0.01, or group condition: F(1, 62) = 0.76, p = 0.386, η2 = 0.01, and no significant two-way-interactions for time x instructor: F(1, 62) = 0.07, p = 0.800, η2 < 0.01; time x group: F(1, 62) = 2.84, p = 0.097, η2 = 0.04; or instructor x group: F(1, 62) < 0.01, p = 0.997, η2 < 0.01.
However, we found a significant three-way-interaction for time x instructor x group: F(1, 62) = 8.26, p < 0.006, η2 = 0.12. To further examine this finding, four subsequent ANOVAs were computed. Two ANOVAs were conducted to analyze the two-way-interaction of time x instructor, resulting in a non-significant two-wayinteraction for the individual condition, F(1, 32) = 3.44, p = 0.073 , η2 = 0.10, and a significant two-way-interaction for the group condition, F(1, 30) = 4.90, p = 0.035 , η2 = 0.14. As displayed in Figure 1, the latter findings suggests, that the increase of coordination was higher when the group, but not the instructor was present, compared to both the group and the instructor being present (for means see Table 1). Two further ANOVAs showed a non-significant time x group interaction for the no-instructor condition, F(1, 30) = 0.54, p = 0.467, η2 = 0.02, and a significant time x group interaction for the instructor condition, F(1, 30) = 14.33, p =0.001, η2 = 0.31, indicating that coordination improved more for study participants walking individually with an instructor compared to those walking in a group with an instructor (see Figure 2).
Regarding strength, the ANOVA resulted in a significant within-subject factor with self-reported strength being higher after the intervention: F(1, 61) = 13.05, p = 0.001, η2 = 0.18. No main effects for the instructor: F(1, 61) = 1.13, p = 0.291, η2 = 0.02; or group conditions: F(1, 61) = 1.80, p = 0.139, η2 = 0.04 were found. Moreo