Development of a CSAI-2 short form for assessing competitive state anxiety during and immediately prior to competition.

Cox, Richard H. ; Russell, William D. ; Robb, Marshall 等

The literature surrounding the study of the relationship between sports performance and anxiety has relied primarily upon precompetitive measures of anxiety as opposed to actual measures of anxiety taken during competition. Recent exceptions to this observation include studies by Krane, Joyce, and Rafeld (1994) in which measurements of anxiety were assessed immediately prior to softball batting performance, and Schedlowski and Tewes (1992) in which perceived anxiety prior to and during parachute jumping were assessed retroactively and through monitoring of heart rate.

Historically, Spielberger's (1983) State Anxiety Inventory (SAI) was utilized by many researchers to obtain an estimate of precompetitive state anxiety in athletes. In an effort to shorten the time necessary to administer the SAI, Martens (1977) developed the Competitive State Anxiety Inventory (CSAI), which is a shortened version of the SAI. Use of the shortened CSAI was short lived, however, due to the development of multidimensional tests of state anxiety such as the Competitive State Anxiety Inventory-2 (CSAI-2) (Martens, Vealey, & Burton, 1990).

The CSAI-2 provided a measurement of both cognitive and somatic competitive state anxiety. This distinction has had significance in helping sport psychologists better understand the relationship between anxiety and performance. Specifically, this distinction is based on a multidimensional theory of competitive state anxiety from which the CSAI-2 was created. Multidimensional theory posits that a distinction needs to be made between cognitive state anxiety and somatic state anxiety, and that their relative independence warrants separate measurement (Martens, et al., 1990). The CSAI-2, however, is 27 items in length, requires approximately 5-min to administer, and is far too long to use as a nonintrusive measure of anxiety either immediately prior to performance or during breaks in performance (e.g., between tennis serves, etc.). The need to find viable methods to ascertain competitive state anxiety either immediately prior to or during competition is consistent with McAuley's (1985) recommendation that future research attempt to assess state anxiety during competition.

Therefore, the purpose of the present investigation was to develop a nonintrusive instrument for measuring state anxiety immediately prior to and/or during competition. To accomplish this, individual items from the CSAI-2 anxiety subscales were stepped into a multiple regression analysis to determine the best 3-item prediction model for both somatic and cognitive anxiety. As a secondary purpose, the resultant rating scales were subsequently field tested to determine their predictive relationship to subscales on the CSAI-2 parent test. In this respect, the present investigation was similar to research reported by Shacham (1983) and Grove and Prapavessis (1992) relative to the development of shortened versions of the Profile of Mood States (POMS) (McNair, Lorr, & Droppleman, 1971). Even more germane is the recent work by Dean, Whelan, and Meyers (1990) and McClung (1992) relative to the development of the "Incredibly Short POMS."

In developing a short version of the CSAI-2, the authors had in mind a perceived exertion-like scale (Borg, 1973, 1982) in which subjects would rate themselves relative to a single statement of somatic or cognitive state anxiety, as opposed to a set of items. This would require the selection of a single item, or the blending of selected items from a CSAI-2 9-item subscale to form a single aggregated statement. In retrospect, Borg's Rating of Perceived Exertion (RPE) scale has proven to be incredibly effective in measuring physiological stress and perceived exertion during exercise (Dunbar, Robertson, Baun, Blandin, Metz, Burdett, & Goss, 1992).

Phase 1

Method

Participants. Data for phase one of this research were obtained from 492 male and female individuals participating in intramural basketball or volleyball at a large Midwest university. The average age of 138 male volleyball players was 20.0 years (SD = 1.6), while the average age of 88 female volleyball players was 19.4 years (SD = 1.7). The average age of 155 male basketball players was 20.9 years (SD = 2.3), while the average age of 111 female basketball players was 19.4 years (SD = 1.1).

Instrument. The instrument used for collecting anxiety data was the Competitive State Anxiety Inventory - 2 (CSAI-2). The CSAI-2 is composed of 27 items. Nine of the 27 items measure cognitive state anxiety, nine measure somatic state anxiety, and nine measure competitive self-confidence. Rigorous empirical study has demonstrated the CSAI-2 to be a reliable and valid measure of multidimensional state anxiety (Martens, et al., 1990).

Procedures. During round-robin and play-off basketball and volleyball competition, starting members of selected teams completed the CSAI-2 approximately 15-min before the beginning of the game/match. During round-robin competition, records were kept so that selected team members were tested only once. Due to the team elimination nature of play-off games, it was not feasible to avoid an occasional repeated measurement situation in which an athlete was measured more than once. Measurements were considered to be independent of each other, however, because of the varying conditions that prevail prior to each new single elimination play-off round. This is a position that has been taken repeatedly in the sport psychology literature when measuring anxiety (Raglin & Turner, 1993; Turner & Raglin, 1996).

Prior to the administration of the CSAI-2, participants were briefed on the inventory, and told that it was to record their feelings prior to competition. Participants were asked to read the instructions on the inventory, answer the questions carefully and honestly. Included in the written instructions was an anti-social desirability statement recommended by Martens, et al. (1990). Finally, participants were assured that their responses were confidential and anonymous. The CSAI-2 was administered to (a) 100 male and 100 female basketball players prior to round-robin games, (b) 100 male and 72 female volleyball players prior to round-robin matches, (c) 55 male and 11 female basketball players prior to play-off games, and (d) 38 male and 16 female volleyball players prior to play-off matches.

Data Analysis. Prior to data analysis, inventories were scored using procedures outlined by Martens, et al. (1990). Scoring yielded three independent scores representing an athlete's precompetitive somatic state anxiety, cognitive state anxiety, and self-confidence. Only the cognitive and somatic state anxiety subscale scores were utilized in subsequent analyses.

For all analyses, subjects' CSAI-2 subscale score on somatic or cognitive state anxiety served as the dependent variable while the nine item scores associated with somatic and cognitive state anxiety served as predictor variables. Maximum [R.sup.2] stepwise multiple regression procedures were utilized to determine which predictor variables were most highly correlated with the dependent variable and which contributed most to the predictive power of a three variable model. The first three variables to be stepped into the model (and remain) were selected to form the basis of a single item competitive state anxiety rating scale for somatic and cognitive state anxiety.

Results

The first three items to be stepped into the three variable model for somatic state anxiety were:

1. I feel nervous (item 2).

2. My body feels tight (item 26).

3. I feel tense in my stomach (item 11).

Together these three items accounted for 84% of the variance in the three variable model. By itself, the first item accounted for 61% of the variance, which increased to 76% when the second variable was stepped in.

The first three items to be stepped into the three variable model for cognitive state anxiety were:

1. I'm concerned about performing poorly (item 16).

2. I am concerned about this competition (item 1).

3. I'm concerned that others will be disappointed with my performance (item 22).

Together these three items accounted for 80% of the variance in the three variable model. By itself, the first item accounted for 51% of the variance, which increased to 69% when the second variable was stepped in.

The simple correlations among the CSAI-2 items for somatic and cognitive state anxiety were of low to moderate magnitude, with the majority of the correlations accounting for less than 25% of the variance of any other item. Only four of the somatic item intercorrelations and three of the cognitive item intercorrelations were greater than .50. Conversely, all of the simple correlations between somatic and cognitive items with total score for somatic or cognitive state anxiety exhibited correlations greater than .50. Simple correlations among anxiety items and with total state anxiety scores are displayed in Tables 1 and 2.

[TABULAR DATA FOR TABLE 2 OMITTED]

[TABULAR DATA FOR TABLE 2 OMITTED]

Discussion

Based upon the results of phase 1 of this investigation, two independent anxiety rating scales were developed. The first, the Anxiety Rating Scale-Somatic (ARS-S), was designed to estimate somatic state anxiety immediately prior to or during brief breaks in sports competition, while the second, the Anxiety Rating Scale-Cognitive (ARS-C), was designed to do the same thing for cognitive state anxiety. The resulting statement for somatic state anxiety would ask the athlete to rate how he/she feels right now relative to the statement: "I feel nervous, my body feels tight and/or my stomach tense." The statement for cognitive state anxiety would ask the athlete to rate how he/she feels right now relative to the statement: "I feel concerned about performing poorly and that others will be disappointed with my performance." As illustrated in Tables 3 and 4, the two single item anxiety statements are associated with 7-point Likert scales. The selection of a 7-point scale as opposed to a shorter or longer scale was based upon a pilot study contrasting the utility of using a short or long Likert scale (Cox, Russell, & Robb, 1994). The results of the pilot study revealed that the 7-point Likert scale yield the largest correlation coefficients with CSAI-2 subscales. The decision to form a single blended or aggregated statement from the 3-item prediction model was also based on the same pilot study. Results of pilot testing revealed that a composite statement was more predictive of the appropriate CSAI-2 subscale than was a single unambiguous statement (i.e., the item that correlated the highest with the appropriate subscale).

Phase 2

In order to test the concurrent validity of the Anxiety Rating Scale-Somatic (ARS-S) and the Anxiety Rating Scale-Cognitive (ARS-C), it was desirable that they be tested in a field setting. The selected field setting was intramural basketball competition.

Method

Participants. Data for this research were obtained from 279 male and female individuals participating in intramural basketball at a major Midwest university. The average age of the 238 male participants was 21.0 years (SD = 2.5), while the average age of the 41 female participants was 19.5 years (SD = 1.4). The sparsity of female intramural basketball participants contributed to the disparity in numbers of male and female respondents. Basketball participants in phase 2 were independent of basketball participants in phase 1 in that data for phase 2 were collected a year later.

Instruments. The instruments used for collecting anxiety data was the Competitive State Anxiety Inventory - 2 (CSAI-2), and the recently developed anxiety rating scales (ARS-S, ARS-C) for competitive cognitive and somatic state anxiety (see Tables 3 & 4).

Procedures. During round-robin and play-off basketball competition, starting members of selected basketball teams completed the CSAI-2 and the somatic and cognitive anxiety rating scales approximately 15-min before tip-off. During round-robin competition, records were kept so that selected teams were tested only once. Again, due to the team elimination nature of play-off games, it was not feasible to avoid an occasional repeated measurement situation in which an athlete was measured more than once. Measurements were considered to be independent of each other, however, because of the varying conditions that prevail prior to each new single elimination play-off round.

As with phase one, subjects were briefed regarding the nature of inventories, asked to read questions carefully and respond in like manner. As in phase 1, an anti-social desirability statement recommended by Martens, et al. (1990) was included in the written instructions. After responding to general information questions, subjects completed the two competitive Anxiety Rating Scales (ARS-S, ARS-C) followed by the 27-item CSAI-2. Observations were obtained from 117 males prior to round-robin games, 18 females prior to round-robin games, 121 males prior to play-off games, and 23 females prior to play-off games.

Data Analysis. Somatic and cognitive subscales from the CSAI-2 were regressed against and correlated with anxiety rating scale scores for somatic and cognitive state anxiety. Data points were also plotted and inspected. In all analyses, the CSAI-2 subscales served as the dependent variable while the anxiety rating scale scores served as the independent variable.

Results

When regressed against competitive somatic state anxiety (CSAI-2), the anxiety rating scale for somatic state anxiety (ARS-S) accounted for 45% (r = .67) of the variance of the CSAI-2 subscale for somatic anxiety. The regression analysis resulted in a Y intercept of 9.35 and a slope (beta) of 2.75, t(277) = 14.99, p [less than] .0001. With this equation it could be predicted that given a hypothetical somatic state anxiety rating score of 4, the CSAI-2 somatic subscale would be 20.3 [+ or -] 7.9 95% of the time (Jackson, 1989).

When regressed against competitive cognitive state anxiety (CSAI-2), the anxiety rating scale for cognitive state anxiety (ARS-C) accounted for 40% (r = .63) of the variance of the CSAI-2 subscale for cognitive anxiety. The regression analysis resulted in a Y intercept of 11.76 and a slope (beta) of 2.27, t(277) = 13.67, p [less than] .0001). With this equation it could be predicted that given a hypothetical cognitive state anxiety rating score of 4, the CSAI-2 cognitive subscale would be 20.8 [+ or -] 7.8 95% of the time (Jackson, 1989).

As can be observed in Table 5, the size of the correlations for both somatic and cognitive anxiety are similar for all subjects combined, males, round-robin, and play-off situations. The correlation for females, however, is considerably larger than for males for both somatic and cognitive state anxiety.

Discussion

While the correlations between the ARS-S and ARS-C with CSAI-2 subscales are of a respectable size, they are not as large as hoped for (with the possible exception of female subjects). The most likely explanation for correlations that explain less than 50% of the variance is the intrusive nature of the CSAI-2. Intramural athletes, excited about participating in a sporting event, tend not to be as attentive to instructions and to written inventories as they might be. They are more interested in getting out on the court to play. This is, of course, a limitation of field based research (Thomas & Nelson, 1990). It does, however, provide intuitive support for the need to develop nonintrusive instruments for measuring competitive state anxiety immediately prior to competition.

From the data displayed in Table 5, it appears that the ARS-S and ARS-C are better predictors of competitive state anxiety in females than for males. This observation, however, may be premature due to the small sample size for females. Future research in this area must focus upon obtaining approximately equal as well as large numbers of male and female participants.

[TABULAR DATA FOR TABLE 5 OMITTED]

Overall Discussion

While the results of these two investigations support the conclusion that the ARS-S and ARS-C are valid predictors of state anxiety in a competitive situation, there is still much work to be done. The Anxiety Rating Scales must be further tested with individual sport athletes as well as a greater number of female basketball players. It would also be desirable to compare the relative predictive validity of the ARS-S and ARS-C with other brief measures of competitive state anxiety that have been developed. Consequently, in future research we intend to compare the predictive and concurrent validity of the Mental Readiness Form (MRF), as revised by Krane (1994), with the Anxiety Rating Scale (ARS) in the same competitive environment. Finally, it is important to note that because participants for this study were college age intramural athletes, that caution should be taken in generalizing the results to elite or highly skilled athletes.

Table 3

Phase One Derived Anxiety Rating Scale for Measuring Competitive Somatic State Anxiety (ARS-S).

Relative to the upcoming competition, and relative to the following statement, rate how you feel right now (circle the appropriate number):

I feel nervous, my body feels tight and/or my stomach tense

1 Not at all 2 A little bit 3 Somewhat 4 Moderately so 5 Quite a bit 6 Very much so 7 Intensely so

Table 4

Phase One Derived Anxiety Rating Scale for Measuring Competitive Cognitive State Anxiety (ARS-C).

Relative to the upcoming competition, and relative to the following statement, rate how you feel right now (circle the appropriate number):

I feel concerned about performing poorly and that others will be disappointed with my performance

Not at all

2 A little bit 3 Somewhat 4 Moderately so 5 Quite a bit 6 Very much so 7 Intensely so

References

Borg, G. A. V. (1973). Perceived exertion: A note on "history" and methods. Medicine and Science in Sports, 5, 90-93.

Borg, G. A. V. (1982). Psychophysical bases of perceived exertion. Medicine and Science in Sports and Exercise, 14, 377-381.

Cox, R. H., Russell, W. D., & Robb, M. (1994). Five phases in the development of an instrument for assessing competitive state anxiety during and prior to competition. Unpublished manuscript, University of Missouri - Columbia.

Dean, J. E., Whelan, J. P., & Meyers, A. W. (1990, September). An incredibly quick way to assess mood states: The incredibly short POMS. Paper presented at the conference of the Association for the advancement of Applied Sport Psychology, San Antonio, TX.

Dunbar, C. C., Robertson, R. J., Baun, R., Blandin, M. F., Metz, K., Burdett, R., & Goss, F. L. (1992). The validity of regulating exercise intensity by ratings of perceived exertion. Medicine and Science in Sports and Exercise, 24, 94-99.

Grove, J. R., & Prapavessis, H. (1992). Preliminary evidence for the reliability and validity of an abbreviated Profile of Mood States. International Journal of Sport Psychology, 23, 93-104.

Jackson, A. S. (1989). Application of regression analysis to exercise science. In M. J. Safrit & T. M. Wood (Eds.), Measurement Concepts in physical education and exercise science (p. 190). Champaign, IL: Human Kinetics Books.

Krane, V. (1994). The mental readiness form as a measure of competitive state anxiety. The Sport Psychologist, 8, 189-202.

Krane, V., Joyce, D., & Rafeld, J. (1994). Competitive anxiety, situation criticality, and softball performance. The Sport Psychologist, 8, 58-73.

McNair, D. M., Lorr, M., & Droppleman, L. F. (1971). Profile of Mood States manual. San Diego, CA: Educational and Industrial Testing Service.

Martens, R. (1977). Sport Competition Anxiety Test. Champaign, IL: Human Kinetics Publishers.

Martens, R., Vealey, R. S., & Burton, D. (1990). Competitive anxiety in sport. Champaign, IL: Human Kinetics Publishers.

McAuley, E. (1985). State - anxiety: Antecedent or result of sport performance. Journal of Sport Behavior, 8, 71 - 77.

McClung, L. R. (1992). Validation study of the Incredibly Short Profile of Mood States. Unpublished master's thesis, California State University, Sacramento.

Raglin, J. S., & Turner, P. E. (1993). Anxiety and performance in track and field athletes: A comparison of the inverted-U hypothesis with zone of optimal function theory. Personality and Individual Differences, 14, 163-171.

Schedlowski, M., & Tewes, U. (1992). Physiological arousal and perception of bodily state during parachute jumping. Psychophysiology, 29, 95-103.

Shacham, S. (1983). A shortened version of the profile of mood states. Journal of Personality Assessment, 47, 305 - 306.

Spielberger C. D. (1983). Manual for the state-trait anxiety inventory (Form-Y). Palo Alto, CA: Consulting Psychologists Press.

Thomas, J. R., & Nelson, J. K. (1990). Research methods in physical activity (2nd ed.). Champaign, IL: Human Kinetics Books.

Turner, P. E., & Raglin, J. S. (1996). Variability in precompetition anxiety and performance in college track and field athletes. Medicine and Science in Sports and Exercise, 28 (3), 378-385.

Richard H Cox, William D. Russell, & Marshall Robb are in the Department of Health and Exercise Sciences, University of Missouri - Columbia.