the relationship between observable self-talk and ... · conditions. when comparing positive...

16
JOURNAL OF SPORT & EXERCISE PSYCHOLOGY, 1994,16,400-415 O 1994 Human Kinetics Publishers, Inc. The Relationship Between Observable Self-Talk and Competitive Junior Tennis Players' Match Performances Judy L. Van Raalte, Britton W. Brewer, Patricia M. Rivera, and Albert J. Petitpas Springfield College In sport psychology, there is broad interest in cognitive factors that affect sport performance. The purpose of this research was to examine one such factor, self-talk, in competitive sport performance. Twenty-four junior tennis players were observed during tournament matches. Their observable self- talk, gestures, and match scores were recorded. Players also described their positive, negative, and other thoughts on a postmatch questionnaire. A de- scriptive analysis of the self-talk and gestures that occurred during competi- tion was generated. It was found that negative self-talk was associated with losing and that players who reported believing in the utility of self-talk won more points than players who did not. These results suggest that self-talk influences competitive sport outcomes. The importance of "believing" in self-talk and the potential motivational and detrimental effects of negative self-talk on performance are discussed. Key words: sport performance, youth sport, nonverbal behavior The cognitive revolution that has swept psychology has also greatly influ- enced sport psychology (Whelan, Mahoney, & Meyers, 1991). In applied sport psychology research, much attention has been focused on the use of cognitive- behavioral strategies such as self-talk. The belief that self-talk affects competitive sport performance (e.g., Finn, 1985; Gould, Hodge, Peterson, & Giannini, 1989; Mahoney & Avener, 1977; Weinberg, Grove, & Jackson, 1992) has led to the implementation of techniques such as thought stopping, cognitive restructuring, and countering to reduce the occurrence of negative self-talk and increase the occurrence of positive self-talk (Kendall, Hrycaiko, Martin, & Kendall, 1990; Weinberg, 1988; Williams & Bunker, 1986). No systematic studies, however, have examined the observable self-talk (and gestures) that occur during competi- tion and explored the relationship between this self-talk and performance. Assessments of the relationship between self-reported self-talk and sport -- J.L. Van Raalte, B.W. Brewer, P.M. Rivera, and A.J. Petitpas are with the Center for Performance Enhancement and Applied Research, Department of Psychology at Spring- field College, Springfield, MA 01 109.

Upload: others

Post on 23-Jun-2020

6 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

JOURNAL OF SPORT & EXERCISE PSYCHOLOGY, 1994,16,400-415 O 1994 Human Kinetics Publishers, Inc.

The Relationship Between Observable Self-Talk and Competitive Junior Tennis Players' Match Performances

Judy L. Van Raalte, Britton W. Brewer, Patricia M. Rivera, and Albert J. Petitpas

Springfield College

In sport psychology, there is broad interest in cognitive factors that affect sport performance. The purpose of this research was to examine one such factor, self-talk, in competitive sport performance. Twenty-four junior tennis players were observed during tournament matches. Their observable self- talk, gestures, and match scores were recorded. Players also described their positive, negative, and other thoughts on a postmatch questionnaire. A de- scriptive analysis of the self-talk and gestures that occurred during competi- tion was generated. It was found that negative self-talk was associated with losing and that players who reported believing in the utility of self-talk won more points than players who did not. These results suggest that self-talk influences competitive sport outcomes. The importance of "believing" in self-talk and the potential motivational and detrimental effects of negative self-talk on performance are discussed.

Key words: sport performance, youth sport, nonverbal behavior

The cognitive revolution that has swept psychology has also greatly influ- enced sport psychology (Whelan, Mahoney, & Meyers, 1991). In applied sport psychology research, much attention has been focused on the use of cognitive- behavioral strategies such as self-talk. The belief that self-talk affects competitive sport performance (e.g., Finn, 1985; Gould, Hodge, Peterson, & Giannini, 1989; Mahoney & Avener, 1977; Weinberg, Grove, & Jackson, 1992) has led to the implementation of techniques such as thought stopping, cognitive restructuring, and countering to reduce the occurrence of negative self-talk and increase the occurrence of positive self-talk (Kendall, Hrycaiko, Martin, & Kendall, 1990; Weinberg, 1988; Williams & Bunker, 1986). No systematic studies, however, have examined the observable self-talk (and gestures) that occur during competi- tion and explored the relationship between this self-talk and performance.

Assessments of the relationship between self-reported self-talk and sport

--

J.L. Van Raalte, B.W. Brewer, P.M. Rivera, and A.J. Petitpas are with the Center for Performance Enhancement and Applied Research, Department of Psychology at Spring- field College, Springfield, MA 01 109.

Page 2: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

Self-Talk and Tennis Performance 1 401

performance have produced equivocal findings. Rotella, Gansneder, Ojala, and Billing (1980) found that more successful elite skiers did not differ from less successful skiers in terms of self-talk used. Highlen and Bennett (1983) found that elite divers who qualified for the Pan American Games reported using more content-based self-instructions during competition and less positive self-talk than did nonqualifiers. Wrestlers who qualified for the Pan American Games reported using more critical self-talk during competition than nonqualifiers (Highlen & Bennett, 1983). Dagrou, Gauvin, and Halliwell (1991) found that elite Ivory Coast athletes reported using the same types of self-talk during their best and worst performances. Interestingly, these results suggest that negative (or less positive) self-talk is associated with better, or at least no worse, performance than positive self-talk. It is important to note that the studies described above relied on athletes' self-reports of their self-talk.

Experimental research examining the effects of self-talk on performance indicates that negative self-talk is associated with worse performances (Dagrou, Gauvin, & Halliwell, 1992; Schill, Monroe, Evans, & Ramanaiah, 1978; Van Raalte et al., in press) and that positive self-talk is associated with better perfor- mances (Dagrou et al., 1992; Johnston-O'Connor & Kirschenbaum, 1986; Kirschenbaum, Ordman, Tomarken, & Holtzbauer, 1982; Rushall, 1984; Rushall, Hall, Roux, Sasseville, & Rushall, 1988; Van Raalte et al., in press) or no different performances (Schill et al., 1978; Weinberg, 1985) compared to control conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to use positive self-talk performed signifi- cantly better on a mirror tracing task than did subjects who used negative self- talk. Similarly, Dagrou et al. (1992), Johnston-O'Connor and Kirschenbaum (1986), Kirschenbaum et al. (1982), and Van Raalte et al. (in press) found that subjects who used positive self-talk performed significantly better than subjects who used negative self-talk when throwing darts, golfing, bowling, and throwing darts, respectively.

In summary, experimental and nonexperimental investigations of the rela- tionship between self-talk and performance have produced discrepant findings. One explanation for these discrepancies is that negative self-talk may be more detrimental to performance of new skills (as in experimental studies) than to performance of well-learned tasks (as in field and interview studies). Another explanation may be that retrospective self-reports of cognitions experienced dur- ing sport performance are unreliable and subject to recall biases (Brewer, Van Raalte, Linder, &Van Raalte, 1991). This problem with nonexperimental research can be circumvented by assessing overt behavior during sport performance, thereby retaining the richness of naturalistic sport settings.

Purpose

The primary purposes of this project were (a) to generate a descriptive analysis of the observable self-talk and gestures used by competitive junior tennis players and (b) to examine the relationship between observable self-talk and sport performance. Tennis was selected for this study because tennis players frequently display self-talk and gestures during competition and because the antecedents and behavioral outcomes (points won and lost) of self-talk and

Page 3: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

402 / Van Raalte, Brewer, Rivera, and Petitpas

gestures can be recorded. On the basis of past experimental research, which is not subject to recall biases, it was hypothesized that negative self-talk and gestures would be associated with poorer tennis performance and that positive self-talk would be associated with better tennis performance.

Method

Recruitment of Subjects

Junior tennis players competing in United States Tennis Association (USTA) sanctioned tournaments were recruited for the study when they checked in at the tournament registration desk. All players were given a description of the study and were asked to sign a form indicating their willingness to participate in the study. A parent or guardian also signed the consent forms of players under 18 years of age.

Data were collected in first-round matches at two tournaments. The first tournament was a girls' 18-and-under National Qualifier tournament. Of the 12 players approached to participate in the study, 11 agreed to participate and signed consent forms prior to their matches. The second data collection was at a boys and girls' 16-and-under Challenger-level tournament. Of the 20 players approached to participate in the study, 17 signed consent forms indicating their willingness to participate. Data were collected only in those matches where both players agreed to participate in the study. Data were collected on a total of 24 players.

Subjects

Of the 24 tennis players (12 females and 12 males) included in this study, 14 were in the 16-and-under age group (58%) and 10 were in the 18-and-under age group (42%). Players were an average of 15.43 years of age. It was an experienced group of players, as 17 of them (71%) had held USTA sectional rankings in the previous year. Seven of the players (29%) were seeded players.

The matches varied in terms of how close players were in ability. This is reflected in the length of matches played. Matches averaged 75 min in length, with the shortest match lasting 35 min and the longest match (the only one that required three sets) lasting 118 min.

Measures

Self-Talk and Gestures Rating Scale (STAGRS). The STAGRS was de- signed to enable raters to unobtrusively record players' observable self-talk and gestures while simultaneously recording the score of their matches. The STAGRS consists of a scorekeeping component based on the Friend at Court (FAC) score- card marking procedures (Barr, Gerdes, Haitbrink, Schwan, & Snyder, 1993) and a self-talk and gestures rating component. The self-talk and gestures rating component was developed by observing junior tennis tournament matches and recording the various self-talk and gestures that occurred during the matches.' The initial set of STAGRS rating categories contained more negative than positive categories because more types of negative than types of positive self-talk and gestures were observed. Final adjustments to the STAGRS were made to facilitate

Page 4: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

Self-Talk and Tennis Performance / 403

the complex process of recording point scores and self-talk and gestures during competitive tennis matches.

The Self-Talk and Gestures Rating Scale includes the following 14 rating categories:

B = Ball abuse C = Compliment opponent F = Fist pump G = "Oh God," frustration H = Hit oneself I = Instructional self-talk L = Laugh (in frustration) M = Practice the stroke motion N = Negative self-talk 0 = Opponent abuse (e.g., questioning line calls) P = Positive self-talk R = Racquet abuse S = Self-talk (e.g., unclear, foreign language) Other = Does not fit the above categories

When using the STAGRS, raters watch each point and the players' behavior until the next serve is about to begin. Raters then record the previous point score and any self-talk or gestures that have occurred. This process allows raters to observe between-point self-talk and gestures.

The scorekeeping component of the STAGRS requires raters to record all points, faults, aces, games, sets, and service breaks on the STAGRS (see Figure 1). In accord with FAC recommendations, points are recorded with a slash, aces with an A, faults with a dot on the center line, and service breaks with an X in the left hand column (Barr et al., 1993). Raters keep a running tally of the match score in the right hand column of the STAGRS.

When raters have extra time (e.g., during changeovers), they also record specific self-talk statements and gestures on an attached sheet of paper. It would be ideal to record all the players' statements during the matches, but this is not feasible given the time constraints of continuous play and the need for the raters to be unobtrusive observers. The STAGRS allows raters to record as much of the players' behaviors as possible while maintaining accurate score keeping.

The interrater reliability of the STAGRS coding and score-keeping system was assessed by conducting a series of studies (Van Raalte, Brewer, Rivera, & Petitpas, 1994) similar to those conducted by Smith, Smoll, and Hunt (1977) when they developed the Coaching Behavior Assessment System (CBAS). In the first study, 7 raters viewed a videotaped sequence of 39 randomly ordered discrete player self-talk and gestures. Each of the 13 STAGRS content categories (all categories except "other") was represented three times in the videotape. Scoring accuracy was defined in terms of agreement with the scoring of the behaviors made by the first author. The number of scoring errors ranged from 0 to 3 with a mean of 1.7 errors per rater. This yielded an average agreement with criterion scoring of 96%. It should be noted that all rating errors occurred in the general self-talk category (S), which some raters coded as frustration (G). It is possible that these errors were due to problems with the training videotape

Page 5: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

'aleas Zu!~ea saJnlsa3 pue yleL-laS - a~n%!fi

sndl!iad puv ' v ~ a n ! ~ 'naMaq 'ailvvy u v ~ / pop

Page 6: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

Self-Talk and Tennis Performance / 405

rather than to a lack of understanding of the rating categories. This possibility was addressed in the second and third reliability studies.

The second and third reliability studies were designed to assess the reliabil- ity of both the raters and the specific rating categories in field settings. In these studies, interrater reliability was assessed by calculating the average correlation between the STAGRS scores of pairs of raters watching the same tennis match. Four pairs of raters watched practice tennis matches in each study. The correlation coefficients between the coding frequencies of rater pairs across the 14 STAGRS categories averaged .87 (ranging from .74 to .96) and .98 (ranging from .95 to .99) in the second and third studies, respectively. In the two studies, raters observed behaviors from 11 of the 14 STAGRS categories. Perhaps because the tennis players were conscious of being observed, and therefore censored their behavior, no ball abuse, opponent abuse, or positive self-talk was observed. In the third study, analysis of correlations between coding frequencies of rater pairs for each of the STAGRS categories indicated significant interrater reliability coefficients ( p < .05) for all of the categories observed, including the general self-talk category (S), across the two studies (rs ranging from .91 to .99).

Postmatch Questionnaire. Players were asked to report the match out- come, to think about what they said to themselves during the match, and to answer the following questions:

What positive thoughts did you have? What negative thoughts did you have? What else (if anything) were you thinking about during your match? (in- clude instructional thoughts such as "Move your feet") Did what you said to yourself during the match affect the outcome of the match? (yes and no response options) If yes, how did your self-talk affect the match?

Data Collection

Two raters sat at the net post and recorded players' first round match scores and self-talk (and gestures) on the STAGRS. After the players had completed the matches, they were given pens and asked to complete the postmatch questionnaire. Most players completed the questionnaires immediately. All questionnaires were returned to the raters at the players' convenience some time before they left the tennis club.

Research assistants were required to have at least 10 hours of training to qualify as a rater for this study (a copy of the STAGRS training manual can be obtained from the first author). Raters had at least 6 hours of "refresher" training prior to the second tournament. Training included identification of the STAGRS rating categories when watching a videotape of actors exhibiting tennis self-talk and gestures. Raters also practiced scoring both videotaped and live tennis matches using the STAGRS.

Data Coding

Data coders examined the STAGRS forms for each of the matches and recorded the self-talk and gestures that followed points won and points'lost, as

Page 7: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

406 / Van Raalte, Brewer, Rivera, and Petitpas

well as the outcome of the point following each self-talk or gesture recorded. Three independent raters reached 100% agreement in categorizing the self-talk and gestures into composite measures. The composite measure of positive self- talk and gestures was created by summing the occurrence of complimenting opponents (C), fist pumps (F), and positive self-talk (P). Similarly, the composite measure of negative self-talk and gestures was created by summing the occurrence of ball abuse (B), frustration (G), hitting oneself (H), laughing in frustration (L), negative self-talk (N), opponent abuse (O), and racquet abuse (R). A measure of instructional self-talk and gestures was created by summing the occurrence of instructional self-talk (I) and practice motions (M). Total self-talk was calcu- lated by adding the occurrences of negative self-talk (N), positive self-talk (P), and undefined self-talk (S). Length of match, total points, games, and sets won and lost were also calculated.

Results

STAGRS Descriptive Data

Although we had some concerns that the players would "censor" their behavior and that there would be little to observe due to the presence of raters on the court, this did not seem to be the case. All players exhibited at least some observable self-talk or gestures during their matches (means and standard deviations for the STAGRS categories are presented in Table 1). Use of self- talk and gestures ranged from 2 to 92 instances during the matches and occurred following approximately 30% of the points. Not surprisingly, more self-talk and gestures occurred in longer matches, r = .47, p < .02.

Table 1 Means and Standard Deviations of Self-Talk and Gestures Per Player

STAGRS categories M SD

B = ball abuse 1.08 1.67 C = compliment opponent 7.42 5.86 F = fist pump 1.46 2.67 G = "Oh God," frustration 17.46 10.76 H = hit oneself 2.83 4.63 I = instructional self-talk 8.42 11.57 L = laugh (in frustration) 0.50 1.14 M = practice the stroke motion 4.92 4.56 N = negative self-talk 10.38 7.78 0 = opponent abuse 0.38 0.97 P = positive self-talk 5.50 9.56 R = racquet abuse 2.08 2.70 S = self-talk 10.67 9.36 Other 1.54 2.04

Note. N = 24.

Page 8: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

Self-Talk and Tennis Performance / 407

The majority of players' positive self-talk and gestures followed points that they had won (66%). Twenty-two of the players (92%) complimented their opponent at least once during the match and 13 (54%) used positive self-talk at least once. Positive self-talk included such things as saying, "Yes!" after winning points and saying, "Come on!" and "Let's go!" for motivation. Ten of the players (42%) pumped their fists after points. Some of the players' behavior seemed instructional, as 21 of the players (88%) practiced the correct swing or motion when the ball was no longer in play, and 19 of the players (79%) verbalized self-instructions, such as "Move your feet." The majority of players' instructional self-talk (79%) followed points that they had lost.

Overall, however, the self-talk was negative, as 21 players (88%) used negative self-talk 13 or more times during their matches. In contrast, only 5 players (21%) used positive or instructional self-talk 13 or more times. The majority of players' negative self-talk and gestures followed points that they lost (96%). Twenty-three of the players (96%) used negative self-talk, 6 players (25%) laughed in anger or frustration, and all players (100%) exhibited frustration through body language during their matches. Negative self-talk included state- ments such as, "Oh my God!" "You are slow!" "You are lazy!" "Why can't I serve?" "What are you shooting at?" "God, I'm tight!" "I'm so sloppy today!" and "Oh God, that's horrible!" Seven players (17%) abused their oppo- nents verbally, 12 players (50%) hit themselves after they lost points, 11 players (46%) abused the balls, and 14 (58%) abused their racquets by doing such things as throwing them or hitting them against the court curtains.

Self-Talk, Gestures, and Tennis Performance

A MANOVA was conducted to compare match winners and losers in terms of positive, negative, and instructional self-talk and gestures. A significant multivariate effect was found, Wilks's lambda = .66, F(3, 20) = 3.42, p < .05, and univariate effects subsequently were examined. Results indicated that winners used significantly fewer negative self-talk and gestures (M = 23.08, SD = 12.63) than did losers (M = 40.67, SD = 17.49), F(1, 22) = 7.97, p < .02 (ES = 1.15). Winners and losers did not differ in terms of the positive or instructional self- talk and gestures used.

It is possible that match winners used less negative self-talk than did match losers because the winners lost fewer points and therefore had fewer reasons to express negative sentiments. Therefore, analyses of covariance (ANCOVAs) were conducted to compare winners and losers in terms of self-talk used and in terms of the effects of their self-talk and gestures on subsequent points played with total lost points or total won points serving as the covariate.

A MANCOVA comparing match winners and match losers in terms of positive, negative, and instructional self-talk while controlling for total lost points indicated that match losers did not significantly differ from winners in terms of self-talk, Wilks's lambda = .79, F(3, 19) = 1.68, p > .05. This suggests that the differences between match losers and match winners may not be based on differences in the way that they generate negative self-talk.

An ANCOVA comparing match winners and match losers on points lost following negative self-talk while controlling for total lost points indicated that match losers were significantly more likely to lose a point following negative

Page 9: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

408 / Van Raalte, Brewer, Rivera, and Petitpas

self-talk than were match winners, F(1, 19) = 4.57, p < .05. This suggests that match losers and match winners may differ in the way that they respond to negative self-talk. There were no significant differences between winners and losers in their point outcomes following positive self-talk while controlling for total won points.

A partial correlation of the number of points lost following negative self- talk with the number of sets lost while statistically controlling for the total number of points lost during the match indicated similar results, r = .46, p < .02. Subjects who lost points following negative self-talk also tended to lose sets, even when statistically controlling for the total number of points that they lost during the match. Subjects who won points following negative self-talk also tended to lose sets, r = .41, p < .03 (controlling for total points won). This finding suggests that negative self-talk may be harmful to performance in general. Further, the more lost points that were followed by negative self-talk, the more likely the players were to lose sets (even when statistically controlling for the total number of points that they lost), r = .35, p < .05.

Belief in Self-Talk and Match Outcome

The postrnatch questionnaires were completed and returned by 23 players (96%). Sixteen players (70%) said yes in response to the question, "Did what you said to yourself during the match affect the outcome of the match?" These self-talk "believers" did not differ from the "nonbelievers" in terms of the amount of self-talk that they used during their matches, F(1, 18) = 0.00, p > .05. An ANCOVA controlling for total points played, however, indicated that self- talk believers won significantly more points (M = 64.06) than did nonbelievers (M = 42.33), F(1, 18) = 5.72, p < .03.

Self-Reported Self-Talk

A qualitative analysis of players' self-reported self-talk was conducted. Initial evaluation of questionnaire responses indicated that subjects' responses were at similar levels of depth and ~pecificity.~ For example, in response to the question, "What positive thoughts did you have?" subjects responded with such statements as "You can win," "Come on, I can do it," and "Just concentrate on each point." Further analyses were conducted by following the procedures described by CGtC, Salmela, Baria, and Russell (1993). These procedures involved examination of the data by two investigators who identified themes that emerged from the data. The two investigators then independently categorized all responses, Interrater agreement was above 95% for all responses categorized. Consensus was reached by discussion on the remaining responses.

As shown in Table 2, a total of 8 themes were extracted from responses to the question regarding positive self-talk. These themes were then organized into three higher order themes. Players reported motivational, calming, and strategic functions of positive self-talk. Most of the positive self-talk, however, was motiva- tional, consisting of self-affirmations and energizing statements such as "I can hit this shot, give me another," "I won't miss," "Drop shot is on, you're being patient," and "You know you can do this." It seems likely that positive self- talk served a functional purpose for the players (see below for further discussion of this point).

Page 10: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

Self-Talk and Tennis Performance / 409

Table 2 Themes Developed From Responses to Open-Ended Question on Positive Self-Talk

Raw data themes No. No.

responses Higher order themes responses

Affirmations (I can win, etc.) 19 Motivation (come on, let's go) 12 Self-motivation 3 1

Concentrate (relax, play loose) 5 Don't . . . worry, give up, etc. 2 t Calming 10 Keep ball in play 2 Opponent is hitting good shots 1

General strategy Hit to opponent's weakness

1 1 Strategy

-

Note. Twenty subjects responded to this question.

In terms of negative self-talk, 13 themes were extracted and organized into 3 higher order themes (see Table 3). The negative self-talk seemed to primarily reflect the players' frustration with themselves as they said such things as "I am really bad," "Wow! My second serve stinks," "I suck, I stink," and "You stupid dimwit, you're so dumb." Fear of losing and self-instruction were also themes that emerged from players' negative self-talk responses.

Other self-talk identified by players fell into seven themes that were organized into three higher order themes: self-instruction, calming, and extra- neous thoughts (see Table 4). The bulk of other self-talk seemed to focus on self-instruction, with players reporting use of statements such as "Move your feet," "Hit through the ball," and "Change the pace." Like the positive self-talk, calming and self-instructional statements may serve an instrumental function that enhances tennis performance. In terms of extraneous thoughts, one athlete reported thinking about her boyfriend and going to a party, while another reported thinking about a song and noticing that the people playing on the next court were "really good." The extraneous thoughts, like the negative self-talk, may distract the players from their matches and diminish performance.

Players were also asked, "Did what you said to yourself during the match affect the outcome of the match? If yes, how did your self-talk affect the match?" The answers of the 16 players (70%) who said yes were categorized into nine themes and three higher order themes, which are presented in Table 5. Interest- ingly, the themes for how self-talk helped match performance mirrored the themes that emerged when evaluating players' positive self-talk. Players noted the calming and motivational aspects of self-talk. The self-talk that hurt match performance was generally negative and included such statements as, "I lost concentration, and he had full control of the match," and "My negative thoughts took my concentration away from the points."

Page 11: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

410 / Van Raalte, Brewer, Rivera, and Petitpas

Table 3 Themes Developed From Responses to Open-Ended Question on Negative Self-Talk

Raw data themes No. No.

responses Higher order themes responses

Bad shot, bad shot selection You are playing . . . bad, stupid Why . . . miss, doing this, etc. I am . . . a poor player, etc. Swearing More negative things Too tentative Harped on points

Talk about opponent Know you will lose Pressure, nerves

Footwork How to hit the ball

1 Negative

2 f Fear of losing 5 1

1 1 Self-instruction 2

Note. Twenty subjects responded to this question.

Table 4 Themes Developed From Responses to Open-Ended Question on Other Self-Talk

Raw data themes No. No.

responses Higher order themes responses

Body mechanics How to hit the ball Court, net, baseline

:: self-instruction 39 4

Concentrate (patient, etc.) 8 Don't let her get back into it 1 Calming

Extraneous tennis thoughts 4 Extraneous other thoughts 4 Other thoughts 8

Note. Twenty-one subjects responded to this question.

Page 12: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

Self-Talk and Tennis Performance / 41 1

Table 5 Themes Developed From Responses to Open-Ended Question on Self-Talk Affecting Match Play

Raw data themes No. No.

responses Higher order themes responses

Used to advantage, play better 4 Kept me moving, going 3 t Self-motivation 9 Proper mechanics 1 After being down, I won! 1

Negative thoughts distracted If I had believed in myself . . . I let her back in the match

Kept me concentrating Kept me calm

Negative 1

Calming

Note. Sixteen subjects responded to this question.

Discussion

Observable self-talk and gestures occurred frequently in the junior tennis matches examined in this study. Both positive and negative self-talk and gestures were readily observed by on-court raters. The preponderance of self-talk and gestures was negative. Although not a perfect record of all on-court behavior, these data provide the most complete information collected to date on tennis players' perceptions of their self-talk and their observable behaviors during com- petition.

The results of this research provide partial support for the findings of both self-report and experimental research. In accord with self-report research, positive self-talk was not associated with better performance. This lack of a relationship between positive self-talk and performance may indicate that positive self-talk is ineffective. It also is possible, however, that subjects used positive self-talk internally, and therefore, their positive self-talk was not observed by the raters. Consistent with experimental research, the results indicated that negative self- talk was associated with worse performances.

Although the relationship between negative self-talk and losing matches may not be surprising to sport psychology advocates and mental training experts, different perceptions may exist in the general tennis playing public. For example, in the March 1993 issue of Tennis magazine, the following letter appeared in the "Your Serve" column:

I would like to suggest that you rename Jim Loehr's articfes, "Comedy Comer." His "Toughen Up With Self-Talk" was certainly good for a laugh. Does he really think that this pseudo-psychology would have a lasting effect on a grown person? This attempt to change behavior by making suggestions to yourself during a match is absurd. (Buser, p. 9)

Page 13: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

412 / Van Raalte, Brewer, Rivera, and Petitpas

The doubts about the utility of self-talk interventions for tennis players held by this Tennis reader were mirrored by some of the players in this study. Interestingly, the players who reported that they did not believe in the utility of self-talk won significantly fewer points than players who reported believing in the efficacy of self-talk. These results could reflect a difference in the way "believers" and "nonbelievers" use self-talk. Perhaps believers try to use self- talk to their advantage and therefore reap performance benefits from self-talk. It also is possible that self-talk believers are more involved in both the mental and physical aspects of tennis and are better players than the nonbelievers. Future research assessing the relationship between believing in the efficacy of self-talk and winning points is needed to address these and other possibilities.

In observing other specific behaviors, it was interesting to note that players were much more likely to compliment their opponents than to abuse their oppo- nents. The raters noted, however, that some of the compliments may have been strategic self-talk. Saying "excellent shot" may be a way to try to get an opponent to try a risky, low percentage shot in the future, or a way of relieving self- criticism (i.e., "It wasn't my fault for missing it, it was just a great shot").

Some players who used negative self-talk and gestures won their matches, possibly because some of the so-called negative self-talk was actually motivational reflecting players' belief in their own competence. To some players, "That stinks" may be a way of saying to themselves, "Come on, you are a good player and can do better." In support of this argument, Goodhart (1986) found that subjects who used negative self-talk were motivated to avoid a negative outcome and tried harder on an anagram task than subjects who were asked to use positive self-talk. Horn and Lox (1993) suggested that performers who receive criticism, which presumably includes negative self-talk, actually perceive themselves as more competent than performers who receive neutral responses to failure.

Future Directions

This research was conducted in a field setting, and the constraints of the setting limited the sample size and precluded assessment of causal relationships between self-talk and performance. It may be that negative self-talk leads to losing, but it also is possible that losing leads to negative self-talk. We attempted to deal with this causality problem by statistically controlling for the number of points lost (and won when appropriate) when conducting analyses. Further experimental research with larger samples is needed to determine the direction of causality in the relationship between self-talk and match performance.

It should be noted that the information collected by the raters on the STAGRS in this study consisted only of players' public self-talk. The postmatch questionnaire was an attempt to tap into players internal self-talk. Analysis of the questionnaire responses provided some support for the relationship between public self-talk and performance observed by the raters. Negative self-talk was a relatively common behavior, and players associated negative self-talk with poor performances. Although significant relationships were not found between public positive self-talk and performance, the players reported that their positive self- talk was associated with better performances. Perhaps the effects of positive self- talk were not observed because players kept much of their positive self-talk

Page 14: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

Self-Talk and Tennis Performance / 413

private. On the other hand, it seems likely that the self-talk and gestures observed by the raters reflected a proportion of the players' internal experiences. Players probably kept some of their negative and some of their positive self-talk private. There is no reason to believe that players kept a disproportionate amount of their positive self-talk private. Players' reports of the utility of positive self-talk may reflect recall biases (Brewer et al., 1991).

Nonetheless, it may be useful to explore the relationship between both positive and negative internal self-talk and performance. Having players wear microphones during competitions to record their self-talk (e.g., Stadulis, Mac- Cracken, Levan, & Fender, 1993) might allow for a more complete assessment of players' verbalizations. Use of the retrospective approach developed by Van Noord (1984), in which players are asked to indicate the self-talk that occurred during their matches, could enable researchers to tap players' internal self-talk. Assessment of physiological (e.g., heart rate) and psychological (e.g., mood) variables associated with match play would also contribute to a more sophisticated analysis of the relationship between self-talk and match performance.

References

Ban, B., Gerdes, S., Haitbrink, D., Schwan, S., & Snyder, J. (Eds.) (1993). Friend at court. New York: Zimman.

Brewer, B.W., Van Raalte, J.L., Linder, D.E., & Van Raalte, N.S. (1991). Peak performance and the perils of retrospective introspection. Journal of Sport & Exercise Psychology, 13, 227-238.

Buser, J. (1993, March). [Letter to the editor]. Tennis, p. 9. CBtC, J., Salmela, J.H., Baria, A., Russell, S.J. (1993). Organizing and interpreting unstruc-

tured qualitative data. The Sport Psychologist, 7, 127-137. Dagrou, E., Gauvin, L., & Halliwell, W. (1991). La preparation mentale des athlbtes

ivoiriens: Pratiques courantes et perspectives de recherche [Mental preparation of Ivory Coast athletes: Current practice and research perspectives]. International Journal of Sport Psychology, 22, 15-34.

Dagrou, E., Gauvin, L., & Halliwell, W. (1992). Effets du langage positif, negatif, et neutre sur la performance motrice [Effects of positive, negative, and neutral self- talk on motor performance]. Canadian Journal of Sports Sciences, 17, 145-147.

Finn, J.A. (1985). Competitive excellence: It's a matter of mind and body. The Physician and Sportsmedicine, 13, 6 1-75.

Goodhart, D.E. (1986). The effects of positive and negative thinking on performance in an achievement situation. Journal of Personality and Social Psychology, 51, 117- 124.

Gould, D., Hodge, K., Peterson, K., & Giannini, J. (1989). An exploratory examination of strategies used by elite coaches to enhance self-efficacy in athletes. Journal of Sport & Exercise Psychology, 11, 128-140.

Highlen, P.S., & Bennett, B.B. (1983). Elite divers and wrestlers: A comparison between open- and closed-skill athletes. Journal of Sport Psychology, 5, 390-409.

Horn, T.S., & Lox, C. (1993). The self-fulfilling prophesy theory: When coaches' expecta- tions become reality. In J.M. Williams (Ed.), Applied sport psychology: Personal growth to peak performance (2nd ed., pp. 69-82). Mountain View, CA: Mayfield.

Johnston-O'Connor, E.J., & Kirschenbaum, D.S. (1986). Something succeeds like success:

Page 15: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

414 / Van Raalte, Brewer, Rivera, and Petitpas

Positive self-monitoring for unskilled golfers. Cognitive Therapy and Research, 10, 123-136.

Kendall, G., Hrycaiko, D., Martin, G.L., & Kendall, T. (1990). The effects of an imagery rehearsal, relaxation, and self-talk package on basketball game performance. Journal of Sport & Exercise Psychology, 12, 157-166.

Kirschenbaum, D.S., Ordman, A.M., Tomarken, A.J., & Holtzbauer, R. (1982). Effects of differential self-monitoring and level of mastery on sports performance: Brain power bowling. Cognitive Therapy and Research, 6, 335-342.

Mahoney, M.J., & Avener, M. (1977). Psychology of the elite athlete: An exploratory study. Cognitive Therapy and Research, 1, 135-141.

Rotella, R.J., Gansneder, B., Ojala, D., & Billings, J. (1980). Cognitions and coping strategies of elite skiers: An exploratory study of young developing athletes. Journal of Sport Psychology, 2, 350-354.

Rushall, B.S. (1984). Competition thinking. In W.F. Straub and J.M. Williams (Eds.), Cognitive sport psychology (pp. 55-62). Lansing, NY: Sport Science Associates.

Rushall, B.S., Hall, M., Roux, L., Sasseville, J., & Rushall, A.C. (1988). Effects of three types of thought content instructions on skiing performance. The Sport Psychologist, 2, 283-297.

Schill, T., Monroe, S., Evans, R., Ramanaiah, N. (1978). The effects of self-verbalizations on performance: A test of the rational-emotive position. Psychotherapy: Theory, Research and Practice, 15, 2-7.

Smith, R.E., Smoll, F.L., & Hunt, E. (1977). A system for the behavioral assessment of athletic coaches. Research Quarterly, 48, 401-407.

Stadulis, R., MacCracken, M., Levan, C., & Fender, L. (1993, June). Using both quantita- tive and qualitative methodology in sport psychology research: A competitive golf example. Paper presented at the Eighth World Congress of Sport Psychology, Lisbon, Portugal.

Van Noord, N.L. (1984). Development and evaluation of a self-talk assessment instrument for tennis players. Unpublished doctoral dissertation, Michigan State University, East Lansing, MI.

Van Raalte, J.I,., Brewer, B.W., Lewis, B.P., Linder, D.E., Wildman, G., & Kozimor, J. (in press). Cork! The effects of positive ahd negative self-talk on dart throwing performance. Journal of Sport Behavior.

Van Raalte, J.L., Brewer, B.W., Rivera, P.M., & Petitpas, A.J. (1994, August). A system for the behavioral assessment of self-talk and gestures during competitive tennis. Paper presented at the International Congress on Applied Research in Sports, Helsinki, Finland.

Weinberg, R.S. (1985). Relationship between self-efficacy and cognitive strategies in enhancing endurance performance. International Journal of Sport Psychology, 17, 280-292.

Weinberg, R.S. (1988). The mental advantage: Developing your psychological skills in tennis. Champaign, IL: Human Kinetics.

Weinberg, R.S., Grove, R., & Jackson, A. (1992). Strategies for building self-efficacy in tennis players: A comparative analysis of Australian and American coaches. The Sport Psychologist, 6, 3-13.

Whelm, J.P., Mahoney, M.J., & Meyers, A.W. (1991). Performance enhancement in sport: A cognitive behavioral domain. Behavior Therapy, 22, 307-327.

Page 16: The Relationship Between Observable Self-Talk and ... · conditions. When comparing positive self-talk to negative self-talk, Schill et al. (1978) found that subjects assigned to

Self-Talk and Tennis Performance / 415

Williams, J.M., & Bunker, L. (1986). Cognitive techniques for improving performance. In J.M. Williams (Ed.), Applied sport psychology: Personal growth to peak perfor- mance (pp. 235-255). Palo Alto, CA: Mayfield.

Notes

'In developing the STAGRS we found that much conversation during tennis matches was related to self-talk. Complimenting opponents (C) was associated with calming: "They hit a good shot, but it was just one, I can stay in the match." Opponent abuse ( 0 ) was associated with frustration: "I'm getting ripped off on line calls." We therefore felt it was important to include C and 0 as part of a broad definition of self-talk.

2Although interviews may have provided additional information about self-talk, we felt that interviews might also bias players' responses. Further, it was not feasible to interview all athletes due to the tournament format.

Acknowledgments

This research was supported by a grant from the United States Tennis Association. We gratefully acknowledge the assistance of the staff at The Grande Meadows tennis club (Barbara McMurray, Mike Meyers, and Edsel Ford) who allowed us to talk with each athlete during tournament check-in and provided support for this project. Further thanks are extended to the junior tennis players who volunteered to participate in this study and to the research assistants and raters, Brian Bartolini, Chris Buntrock, Joanne Daly, Nancy Diehl, Patrick Diehl, Ron Hokanson, Chris Izzo, Jeff Laubach, Bany Shrews- bury, Missy Tallant, and Bob Weinhold, who provided invaluable assistance in this project. Finally, we thank Karen Cogan, G.B. Giles, Darwyn Linder, Trent Petrie, and Ron Smith for their comments on an earlier version of this manuscript.

Manuscript submitted: Febmary 18, 1994 Revision received: June 20, 1994