The Physical Self-Concept. What does it mean to be Old: What does it mean to be Young? Garry E. Richards and Kate Johnson 1 University of Western Sydney, Australia1 Jeannine Stiller University of Leipzig, Germany2

Paper presented at NZARE AARE, Auckland, New Zealand November 2003 RIC03771

1

The Physical Self-concept. What does it mean to be old: What does it mean to be young? Garry E. Richards and Kate Johnson 1 University of Western Sydney, Australia1 Jeannine Stiller2 University of Leipzig, Germany2 Does the concept of physical self remain stable across age? Or does the concept of physical self change as the body wearies and changes in shape? Do the components emphasised in our evaluation of our physical selves change with age or remain stable? Without evaluating such questions it is inappropriate to talk about Physical Self-Concept as though it is the same thing, with just different scores, for the elderly and for the young. This research evaluates the Physical Self-Concept of two distinctly different groups, the elderly (over 60 years of age) and a group of young people (aged 11-15 years). It also evaluates the relative importance of the various dimensions of Physical Self-Concept. Data were collected in Sydney, Australia, using the Physical Self-Concept Scale (Richards, 1987), from volunteer older people (average age 69 years) and a representative group of high school students (average age 13.5years). Internal reliabilities and confirmatory factor analyses (CFA) were used to determine if the instrument was suitable in terms of reliability and factorial structure for each of the groups. Sound psychometric properties were found for the Richards Physical Self Concept Instrument across age and gender. Analyses showed factor scores are differentiated by sex and by age for some factors. Remarkable similarities and only some significant differences were found for older and younger and also male and female groups. Introduction Does the physical self-concept remain stable across ages… or does the inevitability of physical change weigh heavily on our self-perceptions as we age. Developmental psychology and more specialised areas such as the study of self-concept have given most importance to the study of children, adolescents and young people and there has been little research taken into the area of aged people. This paper looks at a cohort of people aged over 60 (mean age 69 years) and looks at the measurement of physical self-concept for this age group and then also looks at a younger age group (mean age 13.5 years) and compares the two. A main focus within gerontological research is on the effects of the aging process on the body’s functioning (Birren & Schaie, 2001). Posner (1995) describes this process as “bodily declines” that includes diminution in such areas as athletics and related motor capabilities, reflexes, physical strength, and stamina (Posner 1995, p. 18). Within this topic it is surprising that there is a paucity of specific physical self-concept measuring instruments constructed within a validity framework according to the psychological implications of these physical changes (Pinquart, 1998; Whitbourne & Primus, 1996). From a theoretical perspective the physical self-concept is proposed (see full note below) to be one part of the self-concept. Shavelson, Hubner and Stanton (1976) were the first who described a self-concept structure that is similar to a hierarchically based personality structure and can be tested empirically. In general their work, expanded and consolidated by Marsh& Shavelson (1985), (Shavelson & Marsh 1986) has led to models representing a global perception of the self as a person (general self-concept) at the apex with more specific factors

2

at the base. It is assumed that this structure applies to all ages and it also tends to assume equal importance of all factors. In the light of this general theory, what do we know about the structure of self-concept in the elderly, so far? Pinquart (1998) supposes that self-concept in the elderly must split into three basic dimensions such as social, mental and physical self-concept (see figure 1). (There are of course other models… eg Marsh (1990) includes Academic as one of the higher order factors.. as do many others. If Pinquart’s convincing research were to be taken as a guide it would appear that the physical self-concept may play a pivotal role in the overall self-concept of the elderly. Within this age group however the perception of one’s own body, the physical self-concept, is relatively unexplored. Pinquart suggests it is split into physical appearance, physical health and physical ability; with overall physical self-esteem situated at the apex. Unfortunately there is a dearth of empirical evidence to support this theoretical structure. It is suggested that Pinquart’s model, in terms of physical self concept may be inadequate, particularly in that it presents only three sub scales, image, abilities and health and does not appear to capture the multiple sub factors. It is for this reason that the Richards Self-Concept Scale was used in this investigation. Figure 1: Hypothetical self-concept in the elderly (modified by Pinquart, 1998)

global self-concept e.g.: I don’t lay any claims to my live. I’m focusing the things remaining me.

social self -concept Sometimes I feel lonely.

mental self-concept

physical self-concept

For my age I have good mental abilities.

My body is flexible.

social competence

social integration

action flexibility

I feel ensurely.

My social relations getting worser.

I slowly adapt in new situations.

Degree of interests

memory

For my age Sometimes I I’m interested have in many problems with things. my memory.

body image

physical abilities

perceived health

I’m good looking.

I have a lot of power in my body.

I get sick o lot.

Measuring gerontological physical self-concept is one of the particular challenges in physical self-concept research. There isn’t a well-established and validated tool that can be used to assess the physical self-concept of the older population. So one focus of this paper is on a construct validation approach to the measurement of a multidimensional physical selfconcept especially in the aging field. One purpose of the present study is to evaluate the Richards Physical Self-Concept Scale (RPSC), (Richards, 1987) for these purposes. A construct validation approach was undertaken to determine the suitability and factorial structure of the instrument for older people. Methods Procedure The RPSCS has seven factors: Body Build, Appearance, Health, Action, Coordination, Strength which all have face validity to their factor descriptor. The seventh factor is named Satisfaction. It comprises 5 items which are indicative and representative of the other factors

3

but which are presented in a form to elicit the degree to which the person is satisfied with his or her physical self in a broad way. Each factor has five items presented as statements and responded to on an eight point Likert scale with “Like Me” being represented by 8 and “Not Like Me” represented by 1. The RPSC has demonstrated throughout various data collections that its reliability and factor structure was robust over gender and a wide age range (Marsh et al., 1994; Richards, 1987). Confirmatory factor analysis showed sound factorial structure with a TLI of .915. A summary of internal reliabilities (Cronbach Alpha) for each of its sub scales and the total RPSCS based on other data collected to date are shown in table 2 below. Table 2: Physical Self Concept Scales Internal Reliabilities Factor Males aged Females aged 11-84 11-90 (N=1533) (N=1688) Bodybuild .88 .90 Appearance .93 .93 Health .79 .85 Action .88 .89 Coordination .89 .87 Strength .88 .85 Satisfaction .85 .80 Total .92 .92 In a recent study, Stiller, J., Marsh, H. W., Richards, G. E., & Alferman, D.(2003) also undertook an evaluation of the RPSCS using a construct validation approach using data collected from elderly participants. Sound psychometric properties were again found, with high scale internal reliabilities (coefficient alpha ranging from .788 to .941. Utilising confirmatory factor analysis, model factor loadings ranged from .587 (action) to .954 (appearance). The overall fit was good ( 2 (df = 168) = 307.68***; RNI = .956; TLI = .945; RMSEA = .060).

Hence with a large-scale general population study and with a specific aged population study both showing sound psychometric properties for the RPSCS it was with some confidence that the RPSCS was used for this particular investigation into the elderly and for use in the comparison with younger people. One of the most critical and least addressed issues in the study of self-concept is the relative importance or salience of sub factors in determining the real self-perceptions of people. It is simply not reasonable to make all sub scales summative in order to get a singular self-concept rating, and yet this is the most common practice. The point is made by a simple illustration. I know a man who has a very high and positive self-concept; he is crippled and has a misshapen face… On any questionnaire requiring him to rate himself on coordination, agility, or speed, he would sensibly rate himself low, if other rating were asked of him for facial appearance or attractiveness he would also rate himself low. If these scales were then added he would clearly be rated as having a low physical self- concept. However, this man has a strong focus on health and strength, goes to the gym religiously and has a magnificent physique and is in good health. Give him questions on these and add them up and he will

4

evidence a very high physical self-concept. Combine the scores of all of these scales and he will show a summative score of a medium or mediocre physical self-concept. If you interview him you will quickly discover that he has a very high physical self-concept. He has entirely sublimated his facial disfigurement and crippled legs as areas where nothing can be done about them…his whole focus is on what he can do… and there he has made great achievements... and has anchored his physical self-concept there and discounted all other issues. Hence the internal frame of reference can in this case be seen to be paramount. However in many cases external frame of reference may be either as important or even more so. Participants This research was conducted as a cross-sectional study of two age groupings. Data were collected in Sydney, Australia. The old age sample consisted of 169 female and 71 male volunteers aged 60 to 90 (M = 69.02; SD = 7.53) who completed the RPSCS. The younger age sample consisted of 621 female and 392 male high school students aged 11-16 (M = 13.5; SD = .962) who also completed the RPSCS. Data Analysis Internal reliabilities and confirmatory factor analyses (CFA) were calculated to determine the goodness of fit to the putative models. CFAs were conducted with LISREL 7 (Jöreskog & Sörbom, 1989) using maximum likelihood (ML) estimation. In order to assess goodness of fit for the instrument fro the young and the older age group samples and for the total sample we considered chi-square ( 2), and the Tucker-Lewis Index (TLI) . Results The TLI for the older sample was .93 (n=154) The TLI for the younger sample was .94 (n = 1013) The TLI for the total sample was .94 High scale internal reliabilities (coefficient alpha = .781 - .940 : See table 3 below) were similar across both age groups and are consistent with other administrations of the RPSCS. Table 3: Physical Self Concept Scales Internal Reliabilities Factor Young aged Older aged 55-90 11-16 (N=228) (N=1013) Bodybuild .87 .91 Appearance .92 .94 Health .81 .78 Action .89 .87 Coordination .85 .90 Strength .84 .86 Satisfaction .81 .81 Total .92 .92

5

Inter-scale correlations were calculated and are presented in table 4 below. Table 4: Physical Self Concept Inter-factor correlations overall and for younger and older people. BODYBLD APPEAR HEALTH ACTION BODYBUILD yng old APPEAR yng old HEALTH

yng old

COORD

STRENGTH SATIS

.00 .00 .636 .391

1.00 1.00

. 33 . 54

. 77 .030

.00 .00

.263 .306

.201 .157

.403 .376

1.00 1.00

.689 .287

.540 .393

.283 .4 3

.48 .586

.00 .00

.492 .309

.434 .322

.205 .326

.338 .428

.564 .710

1.00 1.00

.384 .355

.27 .257

.060 .23

.085 .385

.40 .37

.309 .386

ACTION

yng old COORD yng

old STRENGTH yng

old SATIS yng

old

.00 .00

The differences in inter-factor correlations reveal some interesting insights into young and older people’s conceptualisations of their physical selves. It can be observed from table 5 below that the factor “Satisfaction” (i.e. the overall satisfaction one has with one’s physical self) is consistently and significantly lower than all other factors and this holds for both young and old and both male and female. (With no significant differences between old/young/male/female). However the relative power of each factor to contribute to this self-view may differ over age. For example, as can be seen in table 4 above, older people have a higher correlation between Satisfaction and Action (r =.385) compared to younger people (r =.085) suggesting that being more active is likely to bring them more overall satisfaction with their physical self concepts. Older people also have higher correlations between satisfaction and Health than young people (r = .231 and r = .06 respectively). This is interesting in that although when compared to younger people the correlation is higher, the actual correlation between health and satisfaction is only very modest. This result is reinforced by the finding illustrated in table 5 where, seemingly against expectation, older people in the sample (average age 69) actually rate their health not only higher than any other factor, but in real terms very high i.e. nearly 7 on an 8 point scale), and, actually rate their health self concept significantly higher than the younger sample (p