The present study aimed to identify how self-concept, social support and coping style can act as protective factors against the potentially deleterious effects of negative performance in competitive sport. A cohort of swimmers (N = 272) competing at the Australian Age National Championships was examined to discriminate between three performance-related outcomes - initially successful performance, resilient performance (initial failure, followed by subsequent success) and non-resilient performance (initial failure followed by subsequent failure). A discriminant function analysis revealed two main discriminant functions. The first discriminated resilient performers from the other two groups. Resilient performers showed higher self-perceptions of physical endurance, but lower perceptions of perceived social support from significant others than the other two groups. The second discriminant function discriminated initially successful performers from resilient and non-resilient performers. The initially successful performers scored more highly than the other groups on the coping with adversity and peaking under pressure subscales of the Athletic Coping Skills Inventory. Importantly, this study demonstrates a relationship between psychological constructs and a measurable performance outcome. It is suggested that a high concept of physical endurance, good self-perceptions for peaking under pressure and coping with adversity, and a level of independence from social support are important factors in swimming performance.

Introduction

       Athletic careers are filled with fluctuations of fortune that allow athletes to experience both the thrill of victory and the agony of defeat within their own career. By example we can reflect on the career of American swimmer Pablo Morales. Morales was one of the swimming stars of the 1984 Olympic Games in Los Angeles. Although he won three medals in Los Angeles an individual title eluded him, as he was upset in the 100-meter butterfly final by German swimmer Michael Gross. Four years later, in 1988, Morales failed to make the U.S. team. However, he didn't quit there. After a brief retirement from the sport he returned in 1992 to make the U.S. Olympic team and went on to capture the gold in the 100-meter butterfly at the Barcelona Olympic Games. Kathy Love Ormsby, who in 1986 was favoured to win the 10,000 meters at the NCAA track meet, displayed a very different response to failure by an athlete at the elite level of competition. After falling to fourth place during the championship race, Ormsby dropped out of the competition and attempted suicide by jumping off a bridge (see Baumeister, 1991, pp. 89-90). Although the underlying reasons for Ormsby’s maladaptive response to poor performance are not evident, it is apparent that this is not a desired behavioural response for an athlete at any level of competition.

       These stories are but two extreme and heralded examples of individual response to poor athletic performance. Given the uncertainty of outcome that is the nature of sporting events, what leads individuals to produce dramatically different responses to failure? We all experience failure at one time or another. Indeed, this experience is endemic within sport. Yet, why do some people possess the ability to bounce back from adversity, while others simply give up or drop out? It is clear that the consequences of failure can be either motivating or disruptive depending on how one explains and subsequently deals with that failure.

       One approach to understanding the differential abilities of people to cope with the same negative experience has been the examination of moderator variables, characteristics of persons or their environments that make them more or less vulnerable to the negative effects of stressful life events (Anthony, 1987). Resilience is one such moderator that has helped us to understand why one person reacts with negative symptoms to an objectively minor event when someone else may not experience distress even in the face of apparent major disruption. According to Rutter (1987), resilience is a set of protective factors that modify, ameliorate, or alter a person's response to some environmental hazard that predisposes one to a maladaptive outcome. Braddock, Royster, Winfield, and Hawkins, (1991) employed the term in a similarly general way by conceptualizing resilience as an individual's positive response to situations of stress and adversity. This perspective presents resilience as a protective mechanism that is thought to emerge from specific personality features, such as self-esteem, or from aspects of social support and adaptive coping resources and strategies. In general, resilient persons are believed to possess the quality of rebounding and carrying on, an ability to bounce back and get on with life after adversity (Dyer & McGuinness, 1996; Richardson, Neiger, Jensen, & Kumpfer, 1990).

       Resilience has been proposed as an appropriate concept that reflects the dynamic and fluid nature of the processes that act to modify responses to psychological risk (Rutter, 1985; Rutter, 1987). Rutter (1987) emphasizes the need to develop an understanding of individual responses to adverse life circumstances based on the identification of protective processes - which are developing and changing - as opposed to the isolation of fixed factors that counter risk.

       Research regarding the protective factors contributing to a resilient response is equivocal. Cedarblad, Dahun, Hagnell, and Hanson, (1995) argued that protective factors could be considered to fall into two categories, namely individual dispositions and environmental factors. In contrast, a review of research into stress-resistant individuals by Garmezy (1987) identified three broad sets of variables whereby the environmental factors were further divided into family factors and other external support mechanisms. Hence, Garmezy identified personality features such as self-esteem, social factors such as family and peer support, and the availability of these systems to reinforce an individual's coping ability. These factors contribute to the competencies of individuals to develop problem-solving strategies needed to achieve the goodness of fit between situational demands and environmental contingencies. As pointed out by Block and Block, (1980), if individuals are able to invoke an available repertoire of problem-solving strategies, adaptive flexibility will be the result.

       The study of success after failure in sport is not new. Taylor (1988) offered an in-depth examination of performance slumps in sport and delineates slumps from occasional drops in performance. Performance slumps are behavioural manifestations of extended poor performance. Although the source of the extended period of poor performance can be caused by physical or technical changes to the individual, efforts have been made from the psychological perspective to identify the ways that individuals adaptively cope with extended periods of poor performance. Although interesting, Taylor's insight into performance slumps is offered with no empirical evidence in support. Grove and Heard (1997) showed that dispositional optimism and trait sport confidence were positively related to the adaptive use of problem-focused strategies and negatively related to the maladaptive use of emotion-focused strategies. Underlying the study of psychology of slumps is the desire to rectify extended poor performance by the correct use of psychological skills or by inoculating the individual participant against the negative consequences of the initial poor performance.

       There is evidence that psychological aspects associated with a single poor performance may beget future poor performance. Seligman, Nolen-Hoeksema, Thorton, and Moe-Thorton (1990) used measures of attributional style to predict swimming performance after perceived failure. As part of the study the researchers manipulated performance feedback given to a group of college swimmers to induce perceptions of relative success and failure. They found that participants exhibiting negative or pessimistic attributional patterns displayed greater performance deterioration on performance efforts following perceived failure than did the optimistic participants.

       To date the use of the term resilience in the sport literature has been restricted to the realm of sport injuries (Smith, Smoll & Ptacek, 1990; Patterson, Smith, Everett & Ptacek, 1998). Resilience in this case has mostly been operationalized as the time required to return to competition, or participation. Smith, Smoll, and Ptacek, (1990) studied the moderating effects of social support and psychological coping skills on the total number of days of non-participation due to sport-related injury in a sample of adolescent athletes. The results indicated that individuals who had low levels of social support combined with low coping skills took longer to recover from injury compared to other combinations of variables. Similar results were found by Patterson et al., (1990) in a study on injury in ballet dancers. Results of their study demonstrated that psychosocial factors, including life stress events and social support, can affect injury vulnerability in the population studied.

       The question remains as to why some people are able to bounce back from negative events or performances to regain high levels of performance, whereas others go on to prolonged poor performance or, as in the case of Kathy Orsmby, much worse. In the present paper we take the lead from existing psychological literature that suggests positive self-concept, social support and coping style can act as protective factors against the potentially deleterious effects of negative occurrences. In our research we specifically sought to discriminate between three performance-related outcomes - initially successful performance, resilient performance (initial failure, followed by subsequent success) and non-resilient performance (initial failure followed by subsequent failure) at an elite level of age-group competitive swimming using existing measures of self-concept, social support, and coping style.

Method

Participants

       A sample of 272 competitors at the Australian Age Championship swimming competition provided data for use in the study. The participants represented 108 different swimming clubs originating from eight Australian states and territories and from New Zealand. The final sample consisted of age group competitive swimmers, ranging in age from 12 to 18 years (M = 14.91, SD = 1.7). The final sample was made up of 123 males (45.1%) and 149 females (54.9%). All participants provided informed consent in writing, co-signed by a parent or guardian, prior to initiation of the study.

Instrumentation

       A questionnaire was given to participants that contained items that assessed coping style, social support and self-concept. The final questionnaire included the Athletic Coping Skills Inventory (ACSI-28; Smith, Schutz, Smoll, & Ptacek, 1995), the Multidimensional Scale of Perceived Social Support (MSPSS; Zimet, Dahlem, Zimet, & Farley, 1988; Zimet, Powell, Farley, Werkman, & Berkoff, 1990) and the Physical Self-Description Questionnaire (PSDQ; Marsh, 1996; Marsh, Richards, Johnson, Roche, & Tremayne, 1994).

       Demographic data were also obtained from each participant regarding age, gender, home club, and city, state, and country of origin. All questionnaires were pilot tested to ensure clarity and ease of comprehension.

       Coping Style. The Athletic Coping Style Inventory (ACSI-28; Smith, Schutz, Smoll, & Ptacek, 1995) was used to assess individual coping style. The ACSI contains 28 items describing seven sport-specific subscales: Coping with Adversity (4 items), Peaking Under Pressure (4), Goal Setting/Mental Preparation (4), Concentration (4), Freedom from Worry (4), Confidence and Achievement Motivation (4), and Coachability (4). The athletes were required to rate how often they experienced the situations presented in each of the related questions using a 6-point Likert-type scale. Although the ACSI-28 in its original form used a 4-point scale, a common 6-point scale was adopted for all measures utilised in the current study in order to present a common response format to the participant. The ACSI-28 has previously been used in the study of resiliency in the context of sport injuries (Smith, Smoll, & Ptacek, 1990) and has established reliability and validity (Smith et al., 1995).

       The definitions associated with the sub-scales of the ACSI-28, as outlined by Smith & Christensen, (1995) are as follows:

• Coping with Adversity: Remains positive and enthusiastic even when things are going badly; remains calm and controlled; can quickly bounce back from mistakes and setbacks.
• Peaking Under Pressure: Is challenged rather than threatened by pressure situations and performs well under pressure; a clutch performer.
• Goal Setting/Mental Preparation: Sets and works towards specific performance goals; plans and mentally prepares him/herself for competition and clearly has a 'game plan' for the competition.
• Concentration: Not easily distracted; able to focus on the task at hand in both practice and competitive situations, even when adverse or unexpected events occur.
• Freedom from Worry: Does not put pressure on him/herself by worrying about performing poorly or making mistakes; does not worry about what others will think if he/she performs poorly.
• Confidence and Achievement Motivation: Is confident and positively motivated; consistently gives 100% during practice and competitions and works hard to improve his/her skills.
• Coachability: Open to and learns from instruction; accepts constructive criticism without taking it personally or becoming upset.

       Social Support: Individual perceptions of social support were assessed using the Multidimensional Scale of Perceived Social Support (MSPSS; Zimet, Dahlem, Zimet, & Farley, 1988; Zimet, Powell, Farley, Werkman, & Berkoff, 1990). The MSPSS has been designed to assess adequacy of social support from three specific sources, family, friends and significant others. The MSPSS has been shown to by psychometrically sound, with good reliability, factor validity and adequate construct validity (Zimet et al., 1988). The authors have specifically constructed the scale to be economical, thus providing for the use of the instrument in conjunction with other measures. The MSPSS consists of 12 items describing three different sub-scales: Family Support (4 items), Friend Support (4), and Significant Other Support (4).

       Self-Concept: To assess self-concept specific to the physical domain the Physical Self-Description Questionnaire (PSDQ) by Marsh et al., (1994) was used. The PSDQ is designed for use by adolescents age 12 years and older and utilises 70 items to assess 11 scales: Appearance (6 items), Body Fat (6), Co-ordination (6), Endurance (6), Esteem (8), Flexibility (6), Global Physical (6), Health (8), Physical Activity (6), Strength (6), and Sports Competence (6). The PSDQ has demonstrated adequate validity and reliability (Marsh et al., 1994; Marsh, 1996) and was initially developed on a sample of Australian adolescents, thus increasing the cultural validity of the selected measure. The descriptions of the nine specific components of physical self-concept and two global components have been described by Marsh and colleagues (1994) as follows:

• Appearance: Being good looking, having a nice face
• Body Fat: Not being overweight, not being too fat
• Co-ordination: Being good at co-ordinated movements, being able to do physical movements smoothly
• Endurance: Being able to run a long way without stopping, not tiring easily when exercising hard
• Flexibility: Being able to bend and turn your body easily in different directions
• Health: Not getting sick often, getting well quickly when you are sick
• Physical Activity: Being physically active, doing lots of physical activities regularly
• Strength: Being strong, having a powerful body with lots of muscles
• Sports Competence : Being good at sports, being athletic, having good sport skills
• Global Physical: Feeling positive about one's physical self
• Esteem: Overall positive feelings about self

Procedure

       Data were collected at the Australian Age Championship swimming competition in Adelaide, South Australia. The use of a national championship competition as the forum for data collection and performance observation provided an ecologically valid study of relative performance. Being a championship-level competition we could be relatively confident that the athletes' state of preparation would be aimed at performing to the highest level. Use of an in-season, as opposed to peak-season, competition would mean that the participating athletes could be at vastly different relative stages of preparation depending on their individual program. By collecting data at a national championship we were confident that the majority of participants were in peak season form with the implicit goal of posting seasonal or lifetime best performances. All participants completed the study questionnaire during the three-day period prior to the start of the championship competition.

       Following the collection of the completed questionnaires, participant performance was tracked over the course of the competition. Classification of the participant's behavioural category was made based on the comparison of their performance time to their entry time listed in the official competition program. The reference performance for each individual participant was his/her first swim of the championship competition. Participants who recorded a personal best time relative to their official entry time on the initial swim of the competition were classified as having displayed a successful performance (initial success). Swimmers who failed to achieve a time equal to or better than their entry time in their initial event but who posted a best time in their second event of the championship were classified as having displayed a resilient performance (initial failure followed by subsequent success). Swimmers were classified as having a non-resilient performance if they failed to achieve a best time on their initial swim of the competition and failed to achieve a best time in their second event of the competition (initial failure followed by subsequent failure).

Results

       Data were coded and entered for each participant. First the psychological data for each participant were entered and matched with the individual's performance classification (successful, resilient, non-resilient). Descriptive statistics (Table 1) and internal consistencies (Table 2) were then calculated for each of the 21 different sub-scales that were derived from the three selected inventories. The alpha coefficients ranged from a low of .66 for ACSI Coachability to a high of .91 for PSDQ Body Fat and PSDQ Global Physical. The acceptable levels of internal consistency of the ACSI measures provide some support for the use of the 6-point scale adopted within the current study.