 |
Chapter 3 - Standardization of the Scales Stratification of the Sample
The mean age of the normative sample was 23.8 years (S.D. = 10). In a positively skewed distribution, the modal age of respondents was 18 years, with a median of 20 years. Table 3.1 presents the age distribution for the normative sample. Gender (CSP, LSP) - The normative sample included 512 female athletes and 545 male athletes. A significant difference in age was noted for the two groups. The male respondents had a mean age of 24.6 years, and the female respondents had a mean age of 22.96 years (t=2.68, p<.01). Race - The normative sample included a variety of ethnic groups. A total of 844 respondents (79.8%) identified themselves as White. There were 103 Black athletes (9.7%), and 110 other non-White respondents (10.4%). Level - The majority of athletes in the normative sample were involved in sports competition at the collegiate or university level. Other groups of athletes included those competing at the high school level, as well as those involved at the masters and professional levels. The distribution for level of competition is presented in Table 3.2. Sport - The normative sample included athletes from a wide variety of sports, with both team and individual sports being represented. The distribution of respondents across sports is presented in Table 3.3. Scale Construction The following section documents the development of the 24 scales of the CSP and 20 scales of the LSP, including the evaluation of scale reliability and the establishment of scoring norms. Derivation of Raw Scores - As has been described earlier, the scales on the CSP and LSP were created by the technique of internal consistency analysis. In this approach, items are worded to reflect a theoretical trait or characteristic. Each item is assigned a weight by the author, in the belief that strong endorsement of the item, by the respondent, in the direction assumed by the author, reflects the presence of the given trait. For example, a strong positive endorsement to Item 97 of the CSP ("My primary source of motivation is my own inner joy from doing well.") is believed to reflect strong Intrinsic Motivation. Items were designed to reflect each of the 24 traits to be measured by the CSP and each of the 20 traits to be measured by the LSP
After administering this pool of items to the normative sample, it is possible to evaluate how well the items correlate with items of similar content. Inter-item correlations can be determined, as can alpha co-efficients. Alpha co-efficients are a measure of the internal consistency of a scale, in essence, a measure of how well the items hang together. As such, alpha co-efficients reflect on the reliability of the scale. Alpha co-efficients for the CSP and LSP are presented in Tables 3.4a and 3.4b. As the above tables illustrate, most scales were found to have acceptable levels of internal consistency, based on the original four items used to evaluate each athletic trait, while some scales required modification (specifically, the dropping of one item), in order to show acceptable reliability. Even with correction, the alpha co-efficients for the LSP are generally lower than those of the CSP, likely reflecting the difficulty inherent in the measurement of learning, but remaining a focus for evaluation in future versions of the instrument. Scale Means - After correcting for items that are reverse-scored (38 on the CSP, and 24 on the LSP), means were computed for each of the scales. As t-tests, to be discussed shortly, showed, several statistically significant differences on scale scores between female respondents and male respondents, it was necessary to develop separate norms for these two groups. In Tables 3.5a and 3.5b, the scale means (and standard deviations) for the CSP and LSP are presented for the male athletes and female athletes. In Tables 3.6a and 3.6b, the scale means (and standard deviations) for the CSP and LSP are presented for the three racial groups (White, Black, and other non-White). In Tables 3.7a and 3.7b, the scale means (and standard deviations) for the CSP and LSP are presented for different levels of competition. Factors - Equamax factor extraction revealed five factors underlie the scales of the CSP and four factors that underlie the scales of the LSP. The rotated factor matrices are presented in Tables 3.8a and 3.8b. Factor loadings below .40 are not shown. The factor matrices showed only minor variations between male athletes and female athletes, and across racial groups. In most cases, the core factor structure remained intact, but some cross loading of scales (for example, Performance Under Preassure) occured. Factor Labels - Labels were generated for each of the nine factors derived from the factor analyses. These were as follows:
. Scale Intercorrelations - Scale intercorrelations for the CSP and the LSP are prenented in Tables 3.9a and 3.9b Alpha Co-Efficients - Alpha co-efficients for the five fators of the CSP and four factors of the LSP are presented in Table 3.10. CONTENTS
| INTRODUCTION |
CHAPTERS:
|
||||||||||||||||||||||||
 |
|||||||||||||||||||||||||
he
CSP and LSP norms presented in this Manual are based on a sample of 1057
athletes, ranging in age from 13 years to 60 years, who participate at
various levels of sports competition. While a stratification based on
U.S. Census reports might be useful for developing norms for tests, the
reality of sports participation is that people involved in elite level
competitive sports do not necessarily mimic gender, race, and age distributions
found in the general population. As a result, the norms for the CSP and
LSP were derived from a sample that differs significantly from the general
population.
