JOURNAL OF SPORT & EXERCISE PSYCHOLOGY, 1988, 10, 471478

Sport-Confidence and Competitive Orientation: An Addendum on Scoring Procedures

and Gender Differences

Robin S. Vealey Miami University

In an earlier study, an approach to the study of self-confidence was in- troduced in which the constructs sport-confidence and competitive orientation were conceptualized (Vealey, 1986). Sport-confidence was defined as the belief or degree of certainty individuals possess about their ability to be successful in sport. Competitive orientation was conceptualized to account for differences in how in- dividuals define success and may be defined as a tendency for them to base their satisfaction and feelings of competence either on winning (outcome orientation) or performing well (performance orientation).

Based on the distinction between personality traits and states, sport-confi- dence was separated into two constructs: a dispositional construct termed trait sport-confidence (SC-trait) and a state construct termed state sport-confidence (SC-state). To operationalize the constructs in this model, three instruments were developed: the Trait Sport-Confidence Inventory (TSCI), the State Sport- Confidence Inventory (SSCI), and the Competitive Orientation Inventory (COI). Through five phases of data collection, evidence of internal consistency, item discrimination, test-retest reliability (trait scales only), concurrent validity, and construct validity was accumulated for these instruments as valid operationaliza- tions of the constructs within the conceptual model (Vealey, 1986).

As an addendum to the original article (Vealey, 1986), this paper has two purposes. First, the intricate scoring procedures for the COI are clarified, a new procedure for computing a composite competitive orientation score is outlined, and norms for the new composite score are presented. Second, the analyses from the original study are extended to examine gender differences in the personality dispositions of SC-trait and competitive orientation based on the level of sport structure.

Computing the COI Score

Since the COI was introduced (Vealey, 1986), questions have been raised about the procedures involved in scoring the inventory. The original scoring

Requests for reprints should be submitted to Robin S. Vealey, Department of Physical Education, Health, and Sport Studies, Phiiips Hall, Miami University, Oxford, OH 45056.

472 / Vealey

method yielded two scores-a performance orientation score (COI-performance) and an outcome orientation score (COI-outcome). Initially, both scores were in- cluded in instrument validation and hypotheses testing based on the model of sport- confidence, but that has proven to be unwieldy and even unnecessary based on their strong interrelationship (r= - .82 to - .88 in various samples collected by the author). Also, the distinguishing feature of the COI is that it requires subjects to weigh both performance and outcome simultaneously so that the scoring for- mat is based on a continuum with performance orientation on one end and out- come orientation on the other. This format is different from Gill's (1986) Competitiveness Inventory in which goal and win orientations are separate con- structs. For these reasons, the scoring procedures are clarified and computation procedures for the COI composite score are outlined.

The COI uses a matrix format that contains 16 cells representing different situations in sport (see Vealey, 1986). Each cell represents a situation that com- bines a certain level of performance with a certain outcome. This matrix format forces subjects to weigh the value of both goals simultaneously. Subjects com- plete the inventory by assigning a number from 0 to 10 for each cell that represents how satisfied they would feel in that situation. Scoring the COI involves comput- ing the proportion of variance that is based on different outcomes (outcome score) and the proportion of variance that is based on differences in performance (perfor- mance score). Thus, the outcome score represents how much athletes' feelings of satisfaction vary based on whether they win or lose, and the performance score represents how much athletes' feelings of satisfaction vary based on whether they perform well or poorly.

To aid researchers, the scoring procedures for the COI are outlined in the Appendix. These procedures may be easily formatted into data transforma- tion statements in computer statistical packages (SPSS-X, SAS, etc.). In Step 1 the format for setting up the data is explained. In Step 2 the substeps involved in computing the performance score are explained. This requires (a) summing the values for each row, (b) plugging each athlete's COI responses into the basic sum of squares formula found in any statistics text, and (c) and (d) dividing the performance sum of squares by the total sum of squares so the performance score represents a proportion of the total variance. In Step 3 the substeps involved in computing the outcome score are explained. These substeps include the same procedures as in Step 2 except that the values are summed for each column.

In Step 4, substeps for computing the composite competitive orientation score are explained. As discussed previously, this step has been included to sim- plify data analyses and interpretation based on a performance-outcome orienta- tion continuum. That is, two scores are unnecessary when one score adequately operationalizes the competitive orientation construct. Because of the nature of the COI format in which subjects weigh performance and outcomes simulta- neously, the performance score can also be thought of as the inverse of the out- come score (Substep 4a). Thus, the performance score and the inverse of the outcome score are averaged (Substep 4b) to provide a composite competitive orien- tation score that represents performance orientation. Because the composite score represents variance, it will range from 0 to 1.0.

This composite score reflects a more accurate measure of competitive orien- tation as it accounts for performance score in relation to outcome score for each

Sport-Confidence: An Addendum 1 473

subject. For example, a subject with a performance score of .80 and an outcome score of .02 would have a composite competitive orientation score of .89. How- ever, a subject with a performance score of .80 and an outcome score of .16 would have a composite competitive orientation score of .82. Thus, even though both subjects have identical performance scores, differences in their outcome scores create differences in their final composite scores.

Thus, the COI now yields one competitive orientation score that represents performance orientation. The normative data distributions computed from sever- al data sets using the COI (N=692) are listed in Table 1. COI scores for elite athletes should be interpreted with caution, as only one sample representing one sport (gymnastics) is included. As indicated by Gill and Dzewaltowski (1988), the nature of different types of sports may influence athletes' orientations toward performance and outcome. Clearly, additional data from elite athletes in various sports is needed.

Table 1

Normative Data for the COI

Raw score 12-14 yrs High school College Elite

Percentile (n = 156) (n = 262) (n = 226) (n = 48)

The data from the original validation of the COI (Vealey, 1986) were re- analyzed using the new competitive orientation score and yielded similar find- ings, with a slight increase in some correlations. These findings emphasize that computing the composite competitive orientation score does not change the na- ture of the competitive orientation construct from its original conceptualization, but rather reduces the complexity of interpreting two separate scores and includ- ing both in statistical analyses and interpretation. Although scoring the COI may still appear complex and unwieldy, the steps listed in the Appendix can simply be written into the statistical computer program being used by the researcher. Hopefully, the clarification of these procedures and the creation of one compos- ite score will simplify the scoring and interpretation of the COI.

474 1 Vealey

Gender Differences in Sport-Confidence and Competitive Orientation

The second objective of this paper is to extend the analyses using the original validation study sample to examine gender differences in sport-confidence and competitive orientation. The rationale behind using the same data set for an addi- tional analysis is based on the uniqueness of the sample, which included 48 inter- national caliber athletes competing in a national competition that served as the Olympic trials for their sport. Gender differences on various personality con- structs are usually examined using high school, college, or even recreational ath- letes. Although research in sport psychology indicates that females are lower in self-confidence than males with regard to motor performance (Corbin, 1981 ; Cor- bin, Landers, Feltz, & Senior, 1983; Corbin & Nix, 1979; Duquin, 1986; Mar- tens, Burton, Vealey, Bump, & Smith, 1982; Ryan & Pryor, 1976), it seems that this may largely be the result of gender-role socialization. That is, because sport is largely perceived as a male activity and most sport behaviors are regard- ed as masculine behaviors, females may feel less competent, perceive they have less social support, and receive fewer awards for sport participation and success. However, it may be that elite female athletes differ from the norm with regard to how they perceive their sport experiences.

Findings in gender role research indicate that females in competitive sport possess more "masculine" (instrumental) personality characteristics than do fe- male nonparticipants (Colker & Widom, 1980; Del Rey & Sheppard, 1981; Har- ris & Jennings, 1977; Helrnreich & Spence, 1977). We could extend this to predict that females who have entered the elite phase of competition are less influenced by traditional gender roles and possess as much confidence as males. Thus, the inclusion of an elite sample of athletes along with college and high school samples permits the examination of how sport-confidence (SC-trait) and competitive orien- tation are manifested differently based on gender and the structural level of sport competition.

Method

Subjects for this study included 103 (53 female, 50 male) high school, 96 (71 female, 25 male) college, and 48 (20 female, 28 male) elite athletes (n=247). The high school and college athletes participated in tennis, track and field, base- balllsoftball, and basketball, and the elite athletes were gymnasts participating in the United States Gymnastics Federation National Meet. These are the same samples that were used in the original validation studies (Vealey , 1986). In a non- competitive situation (practice or team meeting), subjects were administered the TSCI and the COI. Inventories were administered by the investigator, and coaches were not present.

Results

Although the relationship between SC-trait and competitive orientation (r= .19) was statistically significant, this low correlation indicated a lack of shared variance between the variables. Also, these variables were found to be orthogonal in three other samples collected by the author. Finally, from a theoretical per- spective, SC-trait and competitive orientation are conceptualized as unrelated dis-

Sport-Confidence: An Addendum / 475

positional constructs that interact to predict SC-state. Thus, univariate analyses for each variable based on gender and participation level (high school, college, elite) were deemed appropriate.

As discussed previously, recent research indicates that the nature of specific sport types may influence the cognitive orientations of athletes (Gill & Dzewal- towski, 1988). Thus, preliminary one-way analyses of variance were conducted to determine if there were differences in SC-trait and competitive orientation based on sport type (individual vs. team). The results indicated that team sport athletes (M=82.84) were higher in SC-trait than were individual sport athletes (M=77.24), F(1, 182) = 8.67, but no significant differences emerged for competitive orienta- tion based on sport type.

Because the previous analysis indicated that SC-trait is influenced by sport type, the gender-by-participation-level analysis was conducted using athletes from only individual sports including gymnastics, track and field, and tennis. The TSCI scores of 46 high school and collegiate subjects from the original test-retest sample (Vealey, 1986) were added to the sample to fill all cells in the factorial design with individual sport athletes. A 2 X 3 (Gender X Participation Level) factorial analysis of variance for SC-trait indicated a significant interaction, F(l, 167) = 5.78, p<. 0 1 (see Figure 1). Student-Newman-Keuls post hoc analyses re- vealed that elite male athletes were higher in SC-trait than all other groups with the exception of elite female athletes. That is, no differences in SC-trait emerged between elite male and female athletes. Elite female athletes were higher in SC- trait than both high school groups and college female athletes. Finally, college male athletes were higher in SC-trait than college female athletes and both high school groups.

SC-TRAIT

Males

Females

Figure 1 - SC-trait as a function of gender and participation level.

60 -- SD = 16.31 SD = 15.59 N = 32 N = 29

I I I I I

High School College Elite

PARTICIPATION LEVEL

'ST I-TOI '6 ' J O ! A W ~ ~ 50 ltnunof .1xa$uo3 pods e u! ssamns jo L 3 w d x a pue uospduro3 p p o s '(9867) '3.m 'ulnbna ' S L I - S ~ T 'ZI 'IC8010y2ICsd uodsjo Itnunor ~znro~~tnua~uz waa~sa-jlas 01 salawe a p a j uy L u B o . ~ ~ u e py%o[oy3dsdjo dyqsuoyie~aa '(7867) ' S 'preddaqs 3g '-d 'Lax .(aa

'ZS-EP ' I 'ICgoloy2ICsd pods SO Ifnunor .uogyladuro:, xas-sson lave pue aiojaq u a i p ~ q 3 jo suoyppa~d ssasms pue sayxynyjx? p3ysdqd jo guyd&-xas ' ( 6 ~ 6 1 ) '3 ' x y ~ 3g ' '8' 3 ' U I ~ J O ~

'OTPLOP '95 'uods p a as?3~aq do5 X?.iaz -1rm0 yavasay .ssau[ngseoq a p .sn a3uapguoD jo y3e[ a p a d : s a l e q s a a 3 m -lopad U! samaJajj!p xas '(£861) 'x 'loNaS p '-?.a 'z lad " N ' a ' s iapug " 8 . 3 ' y q i o 3

'OLZ-S9Z 'E 'IC8010y3ICsd uodsjo Itnunor . u o g e r ~ ~ s a~p!~aduro:, e uy amapguoa-jlas %u!yaj -je sioieq se & ! ~ q e ~uauoddo pue '~uauoddo jo xas '13arqns jo xas '(1861) ' 8 . 3 'u'qlo3 ' ~ S - L P '9 'saloy m~ 'uawom p ~ m o ) sapwye pue uraa~sa-jps 'Qyuyupaj ' & y q -n3sepq :uo!~edys!tn?d 3gawe a@- jo saleIa.uo3 '(0861) 3 . 3 ' u r o p y ~ 3g '.a ' l a n o 3

-uoy1edy3yp~d jo IaaaI pue ~apua8 uodn pasvq uoyquapo aay1paduro3 p w 11-ea-3s jo si~nr1suo:, hypuosrad aqi uy sa3uaragyp ~ w o d u r r auros aleasuowap Lay L11yadoy 'a~!ld!r3sap L~arnd aJe Lpnls s n ~ ] jo s8uypug a q q8noqw -a1qeno~luo3 Lpuoslad a n i e q splep -uqs uo saqasurav a~enpna p w uo sn3oj 01 paurea1 aAeq salama alga leyl aq LEUI q .as01 lo u y ~ Laq JarDayM jo pealsuy urroj~ad bq I ~ M ~ o q uo uoy13ej -syies pue a3ualaduro:, jo sSuy1aaj q a q aseq salawe avla 1eq lsa8Sns w p asayl ' J ~ A ~ M O H .sqaMe 01 8uyL~syies aq p w paqsaqdura aq ppoqs 8uluuy~ wq ssa1 8ugou I E ~ Iaaj LaqL .sn3oj amoz~$no w u e q axom a~wu~ .~o j r ad a~uequa ua3 sn3oj a3wulropad e ~ e q sap! a q lda33e 01 asyal uogyaduro3 jo sIaAa1 laMo1 ayl le saq3eo3 pue salaIqe Lww ley) u! 3!uor! s! salame 1ooq3s q8yq pue a8a1 -103 ueq pa1uapo-a3ueuuojrad slow araM saiawe ailla 1eq 8u!pug aqL

-1anaI alyp w le 8uyaduro3 w o ~ j amap -guo:, jo sIaAa1 qB!y do~aaap Aaq raqaqM lo ( h y ~ q e pw) a3uapyuo:, ryayl jo asne3aq pods jo sIaaaI alga olu! pal~alas ale Lay1 raqaqm sy passarppe aq oi uoyl -sanb Buylsala~u! IIV .pods q a q 1e p~ssa33ns aq 01 h q ~ q e a q aAeq Laq aAaqaq L18u0.11~ 1uauraqoAuy pods jo paal a ~ y p w je sapuraj JEW 1sa88ns qep asaqL .salame apuraj p w apm alyp uaawaq a3uapguo3 u! punoj araM sa3ua~ajjyp ou 'salame a8a1103 put! 1ooq3s q81y u! luaredde aJaM pea-3s uy sasualajjrp rapua8 $noylw .a:,uaprjuo3-j~as u~ sa:,ua~ag!p rapua% 1-e 8rrp1ool uayM mno3:,e oluy u a m aq ppoqs salame jo pAa1 uoped!3vd ayl 1eq ~sa88ns s8u!pug asaqL

.~apua% uo paseq uoylquauo a~ylgaduro3 uy punq alaM sa3ua.1a~~yp ON 'salame (6s' =m) poq3s q8q ro (29- =m) a8ano3 wq paluauo a 3 m o j J a d alom aiaM ( 9 ~ - =m) s q a w alga l eq palmptq s a s A p mq )sod s ~ n a ~ - w u ~ ~ a ~ - ~ u a p m s ' 100->d ' 2 7 ' 6 = ( ~ ~ ~ '2)d ' p a 1 uog~dpy.~ed lo3 13ajja u ~ u r 1w3gp8ys e pale3ptq a ~ w p e a jo sysL@ut! ppo13ej (~ana? uoyledr -3ywd x r a p u a ~ ) g x e 'uoylquapo a~!y1aduro:, ~ o d .sysLpw uopquayxo a q -yaduro3 a q uy papnpuy alaM s a l a w pods m a 1 p w pnpyAypq qoq 'ad4 pods uo paseq pale3ypuy a$aM uoylwuayro a~gpaduro3 uy sa3ualagyp on a s n m a ~

Sport-Confidence: An Addendum / 477

Gill, D.L. (1986). Competitiveness among females and males in physical activity classes. Sex Roles, 15, 233-247.

Gill D.L., & Dzewaltowski, D. (1988). Competitive orientations among intercollegiate athletes: Is winning the only thing? m e Sport Psychologist, 2 , 212-221.

Harris, D.V., & Jennings, S.E. (1977). Self-perceptions of female distance runners. An- nals of the New York Academy of Sciences, 301, 808-815.

Helmreich, R., & Spence, J.T. (1977). Sex roles and achievement. In R.W. Christina & D.M. Landers (Eds.), Psychology of motor behavior and sport-1976 (Vol. 2, pp. 33-46). Champaign, IL: Human Kinetics.

Martens, R., Burton, D., Vealey, R.S., Bump, L.A., & Smith, D. (1982, May). Com- petitive State Anxiety Inventory-2. Symposium presented at the NASPSPA Confer- ence, College Park, MD.

Ryan, T.T., & Plyor, F.D. (1976). Sex cues in estimating and performing a simple and motor task. Perceptual and Motor Skills, 43, 547-552.

Vealey, R.S. (1986). Conceptualization of sport-confidence and competitive orientation: Preliminary investigation and instrument development. Jouml of Sport Psychol- ogy, 8, 221-246.

Appendix

1. Each COI cell is a separate variable and in your data list can be written as COIl to COI16 (see matrix below for cell numbers).

2. Compute the performance score. a. Sum the values for the cells in each row.

R1 =COIl +COI2+COI3 +COI4 R2 = COI5 + COI6 + COI7 + C018 R3=COI9+COI1O+COI11 +COI12 R4=COI13+COI14+COI15+COI16

b. Plug the computed row variables into the sum of squares equation. PERSS = ((Rl *R1)/4+(R2*R2)/4+(R3*R3)/4+(R4*R4)/4 -(((Rl +R2+R3

+R4)* (Rl+R2+R3+R4))/16) c. Compute the total sum of squares.

TOTSS = ((COIl *COIl) + (COI2*COI2) +(COI3*COI3) + (COI4*COI4) +(COI5 *COI5) + (COI6*COI6) + (COI7*COI7) + (COI8*COI8) +(C019*C019) + (COI1O*COI10) +(COIl 1*COI11) +(COI12*COI12) + (COI13*COI13) +(COI14*COI14)+(COI15*COI15) +(COI16*COI16)) -((Oil +R2+R3 +R4)*(R1 +R2+R3+R4))/16)

d. Divide the performance sum of squares by the total sum of squares to get the performance score. COI-PERFORMANCE=PERSS/TOTSS

(cont.)

478 / Vealey

Appendix (cont.)

3. Compute the outcome score. a. Sum the values for the cells in each column.

C1 =C011+C015+C019+C0113 C2=COI2+COI6+COI1O+COI14 C3=COI3+COI7+COIl1 +COI15 C4=COI4+COI8+COI12+COI16

b. Plug the computed column variables into the sum of squares equation. OUTSS=((Cl *Cl)/4+(C2*C2)/4+(C3*C3)/4+(C4*C4)/4-(((Cl +C2+C3 +

C4)* (Cl+ C2+C3+C4))116) c. You have already computed the total sum of squares (see above). d. Divide the outcome sum of squares by the total sum of squares to get the outcome

score. COI-OUTCOME =OUTSS/TOTSS

4. Compute the composite competitive orientation score. a. Compute the inverse of the outcome score.

INVERSE = 1 - COI-OUTCOME b. Compute the competitive orientation score as the average of the performance score

and the inverse of the outcome score. COI = (COI-PERFORMANCE +INVERSE)/2

Manuscript submitted: March 3, 1988 Revision received: May 19, 1988