The Effects of Goal Orientation and Perceived Competence on Cognitive Interference During Tennis and Snooker Performance.

Hatzigeorgiadis, Antonis ; Biddle, Stuart

The main purpose of this study was to explore the effects of goal orientation and perceived competence on cognitive interference during sport performance. The sample consisted of 182 snooker and tennis players. The Thought Occurrence Questionnaire (TOQ) was used to measure cognitive interference. The Task and Ego Orientation in Sport Questionnaire (TEOSQ) was used to assess athletes' goal orientation, while perceptions of competence were evaluated by a three-item scale. Multiple Group Path Analysis was conducted to test the structure of the relationship between the constructs. Results revealed that while task orientation was negatively associated with 'thoughts of escape', irrespective of levels of perceived competence, the relationship between ego orientation and 'thoughts of escape' was moderated by perceived competence. congruent with theoretical predictions, in the low perceived competence group there was a positive relationship between ego orientation and 'thoughts of escape', while in the high competenc e group the association was not significant.

In sport psychology, and in achievement motivation in particular, goal orientation theory (Nicholls 1984, 1989) has attracted a remarkable amount of attention in recent years. However, despite its popularity, only a limited amount of research has examined relationships between goals and aspects of sport performance. Within this literature sport anxiety has been the main focus of attention (e.g. Gould, Ecklund, Petlichkoff, Petersen and Bump, 1991; Hall and Kerr, 1997; Martin and Gill, 1991; Swain and Jones, 1992; Vealey and Campbell, 1988).

In sport it is common hearing athletes attributing their success to their level and quality of concentration, or their failure to lack of concentration. Furthermore, there are many times when athletes admit that thoughts not connected to the task interfere with what should be their only focus, that is personal performance. Though an important aspect of performance, cognitive interference has been largely ignored in sport psychology research. In contrast, a great deal of research focusing on cognitive interference has been conducted in educational settings. Here, cognitive interference has been described as task irrelevant, self-preoccupied thinking including components of worry over performance (Sarason, Sarason and Pierce, 1990). This paper explores whether goal orientation is associated with thought occurrence during sport performance.

Cognitive Interference

Educational research has taken place on the antecedents and the effects of cognitive interference. Originally, it was suggested that cognitive interference is a result of test anxiety (Deffenbacher and Deitz 1978; Wine 1971), and a number of studies were conducted to test this relationship. Sarason and Stoops (1978), applying a task presented as an intelligence test, found that high test anxious individuals displayed higher levels of cognitive interference than those low in test anxiety. In particular, they reported that high test anxious individuals were preoccupied during the task by how poorly they were doing, how other people were coping, and what the examiner would think of them. Zats and Chassin (1983) reported that during an anagram task high test anxious children experienced qualitatively more negative and task-irrelevant thoughts than moderate and low anxious children, while in similar tasks others have found that high levels of test anxiety are associated with more frequent negative and interfering cognitions during performance (Bruch, Juster and Kaflowitz, 1983; Deffenbacher and Hazaleus, 1985; Gallassi, Frierson and Sharer, 1981).

However, because not all differences in cognitive interference that participants were experiencing could be attributed to test anxiety, researchers tried to identify other factors that might stimulate, or interact with, test anxiety in generating interfering thoughts. It was revealed that cognitive interference was related to levels of study skills (Culler and Holahan, 1980; Paulman and Kennelly, 1984), perceived ability (Arkin, Detchon and Maruyama, 1982), previous performance, perceived preparation, grade expectation, examination importance (Hunsley, 1987), task difficulty (Arkin et al., 1982), and the degree to which evaluation was involved (Zats and Chassin, 1985).

Progressively, research on test anxiety moved towards the proposition that it impairs performance through interference from self-preoccupying task worries and task irrelevant thoughts. Paulman and Kennelly (1984), examining information-processing deficits in a dual-task paradigm related to cognitive interference, found that elevated cognitive interference scores were significantly associated with lower test performance. Miculiner (1989), in a study concerning learned helplessness, examined performance differences between those high and low in cognitive interference in a memory and visual search task following unsolvable problems. He reported that the more frequent the participants' 'mind-wandering', the less accurate their performance. This supported the view that excessive engagement in self-concerned thoughts and in off-task cognitions are important antecedents of performance deficits. Finally, Hoffman (1993) reported that in a computer-based task, cognitive interference, and in particular task-related wor ries, predicted poor performance.

In the sport psychology literature very few studies have been concerned with cognitive interference. In one of these, Schwenkmezger and Laux (1986) reported that high trait anxious handball players responded to a highly stressful situation with increased worry, whereas that was not the case for players with low trait anxiety. Furthermore, they found that cognitive interference was negatively related to performance. Man, Stuchlikova and Kindlmann (1995), with a sample of football players, did not support the prediction that highly stressful situations would evoke elevations in cognitive interference. However, they reported that cognitive interference was predicted by athletes' cognitive anxiety. Taking into account the above findings it becomes apparent that the investigation of cognitive interference in sports could reveal interesting as well as important aspects as to how cognitions are related to sport performance.

Goal Orientations

To explain differences in behavior a vast amount of research in sport psychology has been devoted to the examination of motivational processes. Within this body of research achievement motivation has been given considerable attention. One of the main theories in this domain is the 'goal perspectives theory' (Nicholls, 1984, 1989), which has proved to be valuable in explaining cognitions, affect and behavior individuals display in achievement settings (Dweck, 1992). According to the theory, two main kinds of achievement orientation have been identified in terms of the way people define success. One concentrates on learning, mastery and self-improvement, where the task is the major focus (task orientation), whereas the other focuses on normatively based accomplishment and social comparison, where the self and the presentation of the self are the points of reference (ego orientation). Furthermore, according to the theory, perceptions of competence are of particular importance for individuals displaying a high e go orientation (Nicholls, 1984, 1989). In general, task orientation, in comparison with ego orientation, especially in the case of low perceived competence, has been shown to be connected with more adaptive patterns of cognitions (Duda and Nicholls, 1992; Hom, Duda, and Miller, 1993; Walling and Duda, 1995), affect (Boyd, Callaghan and Yin, 1991; Duda, Chi, Newton, Walling and Catley, 1995; Duda and Nicholls, 1992), and behavior (Graham and Nolan, 1991; Solmon and Boone, 1993).

In the sport psychology literature one of the performance aspects that has been related to achievement motivation is anxiety. However, results have not been consistent. As Hall and Kerr (1997) indicate, the reason might lie in the different ways researchers have conceptualized and measured the two constructs of goals and anxiety. Vealey and Campbell (1988), using a unidimensional measure of anxiety (SCAT) and Ewing's (1981) technique to assess goals, found a negative association between task orientation and anxiety, but no relation between anxiety and ego orientation. Gould et al. (1991), also using Ewing's conceptualization of goals, but a multidimensional measure of anxiety (CSAI, children version) reported that achievement goals failed to predict pre- or post-task state anxiety and indicated the need for more valid measures of goal orientation. Swain and Jones (1992), using the Sport Orientation Questionnaire (SOQ), reported a negative relation between pre-competition cognitive anxiety and orientation tow ards 'goal' (focus on personal standards), as compared to orientation towards 'win' (focus on winning). However, Martin and Gill (1991) did not support any relations between the SOQ subscales and cognitive anxiety.

A possible justification for the lack of consistency for the above findings is provided by recent advancements in goal orientation theory (Nicholls, 1984, 1989) that provides a more sound conceptualization of goal orientation and highlights the importance of perceived competence as an important moderator between goals and cognitions, affect and behavior. In relation to these advancements, Hall and Kerr (1997) found that, for low perceived competence athletes, ego orientation (measured 30 minutes prior to competition) was positively related to cognitive state anxiety two days, one day and 30 minutes prior to competition, whereas the relation between task orientation and cognitive anxiety (measured with the same temporal pattern) was negative. However, for the high competence group the results did not show consistent patterns.

Overall, what is of more interest is that from all studies examining goal orientation in relation to anxiety only the latter (Hall and Kerr, 1997) has tested the hypothesis that perceived competence moderates the relationship between ego orientation and anxiety.

Integrating Cognitive Interference and Goal Orientation

Considering the self-centered character of ego orientation in comparison to the task-centered character of task orientation, and the self-preoccupying nature of cognitive interference as opposed to a task-related focus, a link between achievement goal orientation and cognitive interference seems plausible and worthy of testing.

Dweck (1989) and Kanfer and Ackerman (1989) have proposed that an ego orientation, through increasing the likelihood that individuals focus too much attention on developing attributions regarding ability, detracts from task performance. Cognitive activities of ego oriented individuals use up resources that otherwise could be applied to the task, therefore hindering task performance. Moreover, taking into consideration the way perceived competence influences ego-oriented individuals' cognitions, it might be suggested that performance of ego oriented individuals with low perceived ability, compared to those with high perceived ability, would be more negatively affected by the cognitive activities brought about by an ego goal orientation.

Diener and Dweck (1980), experimenting on paper and pencil tests, asked the participants to verbalize their thoughts during task performance. Ego oriented children, contrary to task oriented ones, engaged in task irrelevant verbalizations, usually of a self-aggrandizing nature. Hoffman (1993) examined the relationship between ego orientations, cognitive interference and performance. Ego orientation was found to be related to task-relevant worries, but no effect was indicated for task-irrelevant thoughts. However, measures of task orientation were not included, and thus comparisons were not possible. In addition, perceived competence was not tested as a moderator of the relationship.

In the sport context, Newton and Duda (1993), experimenting with students across three bowling games, examined the relationship between goal orientation and performance cognitive content. The only significant result that emerged was that in one of the three games task orientation was found to be negatively correlated with performance worry, and positively correlated with keeping one's concentration. The lack of consistency across the three games was attributed to the small sample size and the noncompetitive environment in which the games took place. Furthermore, it should be noted that the measures of cognitive content were based on single-item responses.

The main purpose of this study, therefore, was to test the theoretical predictions of goal orientation theory in relation to cognitive interference in sport. Therefore, it was hypothesized that cognitive interference will be negatively associated with task orientation and with ego orientation for players with high perceived competence. However, it was hypothesized that cognitive interference will be positively associated with ego orientation for players low in perceived competence.

Method

Sample

Athletes representing two sports, snooker and tennis, were sampled. The total sample consisted of 182 players (134 males, 45 females, 3 not identified by sex) comprising 90 snooker players and 92 tennis players. The mean age of the total sample was 30.4 years (standard deviation 12.9, range 15 - 57). Mean ages were 28.51 years for snooker and for 32.16 years for tennis. All players were non-elite, club members and regular participants in local and regional competitions and leagues. Participants represented 24 clubs (14 snooker clubs and 10 tennis clubs) located in the south west of England.

Instrumentation

Cognitive interference. Little attention has been directed towards cognitive interference in the sports domain. Hence, it is not surprising that instruments to measure cognitive interference specifically for sports are not available. For this study, therefore, an instrument developed in the academic setting - the Thought Occurrence Questionnaire (TOQ; Sarason, Sarason, Keefe, Hayes and Shearin, 1986) - was employed. The TOQ is a 28-item questionnaire that assesses the frequency of thoughts that come to peoples' mind during task performance (see Table 1). It consists of three subscales that are measured on a 7-point Likert scale from 1 (never) to 7 (very often). The first has been labeled 'task relevant worries' and includes statements such as (while performing) "I think about how poorly I am doing", "I think about what someone will think of me", and "I think about my level of ability". The second subscale has been named 'task irrelevant thoughts', and includes items such as "I think about other activities", "I think about personal worries", or "I think about something that might happen in the future". The third subscale is called 'thoughts of escape' and includes statements such as "I think about quitting", "I think about how unhappy I am", and "I think about how I cannot stand it any more". Initial validation of the instrument in educational research has supported the factorial structure of the scale and revealed satisfactory internal consistency coefficients for its three subscales (Cronbach's alpha range: .84 to .91; see Sarason et al., 1986). To our knowledge, part of the TOQ has been used in two cases in sport psychology, however its psychometric properties have not been tested. In particular, Schwenkmezger and Laux (1986) and Man et al. (1991) used a shortened, single-factor version of the scale after modifying it based on evidence of face validity.

Goal orientation. In contrast to cognitive interference, goal orientation is a well-documented and widely researched area in sport. For this study, the Task and Ego Orientation in Sport Questionnaire (TEOSQ), a well-established and validated instrument in sport psychology, was used (see Duda, 1993). The TEOSQ comprises 13 items measured on 5-point scales from 1 (strongly disagree) to 5 (strongly agree), following the stem 'I feel most successful in snooker/tennis when ...'. Six of the items refer to task orientation (e.g. "I learn a new skill by trying hard"), while seven of the items refer to ego orientation (e.g. "I can do better than others").

Perceived competence. Three items were employed to examine the way athletes would estimate their level of perceived competence. The items were taken from the perceived competence subscale included in the sport version of the Intrinsic Motivation Inventory (IMI; McAuley, Duncan, and Tammen, 1989) and were adapted accordingly for each sport. Participants were asked to express their self-perceptions on 7-point scales.

Procedure

At the initial stages of sampling a meeting with each of the club managers was arranged. During this meeting the purpose of the study was discussed and a sample questionnaire was presented. Furthermore, the nature of participants required (i.e. regular competitors) was mentioned and an estimation of the number of participants that could be included in the sample. was obtained. In subsequent visits the questionnaire was distributed to the athletes. With regard to the TOQ participants were informed that our interest was related to thoughts they might have had during past competitive performances. Finally, participants were informed that confidentiality was guaranteed.

Purpose of the Study and Data Analysis

Considering that the complete TOQ has not been validated in a sport context, the purpose of the present study was twofold. First, to examine the validity of the TOQ as an adequate measure of cognitive interference in sport and, second, to identify whether goal orientation is associated with the cognitive interference athletes experience while performing. In addition, the role of perceived competence as a moderator of such a relationship was examined.

The instrument used to measure cognitive interference was tested for factorial validity through confirmatory factor analysis (CFA). All instruments were tested for reliability through measures of internal consistency (Cronbach's Alpha coefficient). Finally, path analysis, through structural equation modeling using EC, was performed to identify the structure of the relationships between goal orientation, perceived competence and cognitive interference.

Results

Factorial Validity for the Thought Occurrence Questionnaire (TOQ)

Information about the structure of the TOQ was taken from Sarason et al. (1986), where the results of the initial validation of the instrument in an academic setting, using exploratory factor analysis, were presented (see Method). Accordingly, a three-factor solution was tested. Evaluations of the assumptions of multivariate normality and linearity indicated that there were no outliers among the sample. The fit indices that emerged from the analysis are presented in Table 2 and indicate a rather poor fit. A closer investigation of the results showed that most of the items of the "task relevant worries" subscale had low factor loadings and relatively high measurement errors (see Table 1). Furthermore, examination of the residual matrix revealed that the error variances of certain items were found to be highly correlated.

According to Joreskog (1993, p. 298), "if the model is rejected by the data, the problem is to determine what is wrong with the model and how the model should be modified to fit the data better." As Joreskog (1993) proposes, to further investigate the model it can be useful to examine each of the constructs separately, then to combine the constructs in pairs and finally to evaluate the overall model. Therefore, each of the subscales was tested in isolation. The fit indices that emerged are presented in Table 2. The indices for the "task relevant worries" subscale indicate a poor fit. As in the previous analysis, the factor loadings were generally not satisfactory, while the measurement errors were relatively high (Table I). Furthermore, most of the items (seven out of nine) revealed high error intercorrelation, indicating that these items "measure something else or something in addition to the construct they are supposed to measure" (Joreskog 1993, p.297). Therefore, this particular subscale was excluded fro m further analyses.

The fit indices for the "task irrelevant thoughts" subscale were slightly better, however still below normal acceptable criteria (Table 2). Examination of the standardized residuals showed that four of the items had high error correlation. These items were excluded, while the rest were kept for analysis along with the third of the subscales whose results were largely satisfactory. In particular, the fit indices for the 'thoughts of escape' subscale were marginally acceptable (Table 2). Examination of the results revealed that only one of the items had a low loading and high measurement error (Table 1), while the standardized residuals showed that it was the same item whose error was correlated with other items' error. Subsequently, the subscale was retained after dropping the item that seemed to be problematic. The two remaining subscales were included in the final confirmatory analysis. The fit indices that emerged from the analysis are presented in Table 2 (B + C) and suggest that the modified instrument w as appropriate for further analyses (see Table 1 for factor loadings and measurement errors).

Reliability of the Instruments

Internal consistency of the subscales was measured with Cronbach's alpha coefficient. Concerning the TOQ, in accordance with the initial psychometric evaluation by Sarason et al. (1986), indices for both scales revealed highly reliable scales ('task-irrelevant thoughts': [alpha] = .89; 'thoughts of escape': [alpha] = .89). Analysis also revealed satisfactory reliability coefficients for the TEOSQ ('ego': [alpha] = .85; 'task': [alpha] = .72). Finally, the reliability coefficient for the scale used to measure perceived competence was also high ([alpha] = .88).

Descriptive Statistics and Correlations

Means, standard deviations and the correlation matrix of the variables are displayed in Table 3. This shows, in accordance with contemporary sport psychology research (e.g. Duda and Nicholls 1992; Fox, Goudas, Biddle, Duda and Armstrong 1994; Goudas, Biddle and Fox 1994), that the two goal orientations are essentially orthogonal. Task orientation was negatively correlated with 'thoughts of escape', but not significantly correlated with 'task irrelevant thoughts'. Correlations between ego orientation and the two cognitive interference subscales were close to zero. Finally, perceived competence was negatively correlated with 'thoughts of escape'.

Path Analysis

The major objective of the study was to explore the relationship between goal orientations and cognitive interference and test whether perceived competence affected this relationship. One of the ways to test for moderator effects is through multiple-group path analysis (structural equation modeling). The sample is divided into groups according to the hypothesized moderator and the path coefficients (testing the relationship between the independent and the dependent variables) for each group are calculated and compared to each other. According to Baron and Kenny (1986), calculating path coefficients is preferred to correlations because correlation coefficients are influenced by differences in variance in the independent variables as well as differences in the measurement error of the dependent variables.

Separate analyses were calculated for the tennis and snooker samples. Each of the samples was divided into high and low perceived competence groups. The division between high and low competence was based on mean scores for each sport sample; 4.45 for the tennis sample and 4.36 for the snooker sample. After the two samples were split, 44 of the tennis players were included in the high competence group and 48 in the low competence group. For the snooker sample, the high competence group comprised 51 players and the low competence group 39 players. Splitting the samples according to the mean of the overall sample would have resulted in an identical classification. One way ANOVA showed that the mean differences in perceived competence between the high and low groups were statistically significant (tennis F(1,90) = l65.47,p [less than].05; snooker F(l,88) = 153.68, p [less than].05).

Subsequently, the hypothesized structural model was tested for each of the four groups to test whether the model was meaningful. In accordance with goal orientation theory (Nicholls, 1984, 1989), it was hypothesized that for all high and low perceived competence groups task orientation would be negatively associated with cognitive interference. Furthermore, it was hypothesized that for the high competence groups ego orientation would have a negative association with cognitive interference, while for the low competence groups there would be a positive relationship between ego orientation and cognitive interference. The model that was tested for the four groups is presented in Figure 1.

The fit indices that emerged from the analyses indicated poor fit for the models. Examination of the residual matrices revealed that this was due to high correlation between the error variance of the two cognitive interference subscales in all models. In addition, the Wald test revealed that the paths connecting task and ego orientation with 'task irrelevant thoughts' (for all models) should be dropped. Dropping these paths would not improve the fit of the model, however it would result in a more parsimonious model. Examination of the models after dropping these paths revealed that the error variance of the 'thoughts of escape' subscale was highly correlated with the variance of the 'task irrelevant thoughts' subscale (in these models 'task irrelevant thoughts' did not have error variance since it was not predicted by any variable). The Lagrange Multiplier (LM) test revealed that setting the error variance of the 'thoughts of escape' scale to correlate with the variance of the 'task irrelevant thoughts' subs cale would improve the fit of the models significantly (for tennis: high competence group [X.sup.2](1) change = 4.26, p [less than].05, low competence group: [X.sup.2](1) change = 13.43, p [less than].05; for snooker: high competence group [X.sup.2](1) change = 4.66, p [less than].05, low competence group [X.sup.2](1) change = 6.90, p[less than].05)

As Bollen (1989) suggests, with respect to the assumption of isolation, residual variances must not be correlated with variances of independent variables, since in this case effects of the independent variables on the dependent are not isolated from effects of variables that have not been included in the study (omitted variables). However, in the present study, the estimation of the correlation between residual variance of the 'thoughts of escape' subscale with the variance of the 'task irrelevant thoughts' subscale is not considered to violate assumptions of isolation since 'task irrelevant thoughts' is not independent from 'thoughts of escape'. Rather, this significant correlation suggests that in this model, intervening variables explaining the co-variation between the two variances have been omitted. Considering the character of cognitive interference, which implies thoughts distracting to the concentration of the performer, it seems reasonable to suggest that there should be some factors [such as anxiet y (Zats and Chassin, 1985), or proneness to distraction (Sarason et al., 1986)] that might explain this co-variation. Therefore this particular modification was considered meaningful.

The modified models that were tested for the two samples are in Figure 2 (tennis) and Figure 3 (snooker), along with the standardized path coefficients. The fit indices that emerged indicated very good fit and are presented in Table 4. In support of the hypothesis, the paths connecting task orientation and 'thoughts of escape' were negative and significant for all high and low perceived competence groups. Regarding the relationship between ego orientation and 'thoughts of escape', for the low competence groups the paths were positive and significant, while for the high competence groups the paths were negative, but rather low.

Subsequently, multiple-group analysis was performed separately for each of the sport samples to test for invariance of regression weights across competence groups. In this analysis the paths connecting task and ego orientations with 'thoughts of escape' were set equal Figure 3. The modified path model for the snooker sample. The regular characters indicate the standardised coefficients for the low competence group, while the bold characters indicate the coefficients for the high competence group.

for the two perceived competence groups. In addition, the LM test was calculated, which in the case of multiple-sample analysis indicates whether improvement of fit can be achieved if the equality constraints are dropped. The LM test revealed that significant improvement of fit could be obtained for both samples if the paths connecting ego orientation and 'thoughts of escape' were freed, which means that these paths should not be considered equal for the two perceived competence groups. (for tennis: [X.sup.2](1)change = 3.77, p = .05; for snooker: [X.sup.2](1) change = 4.62, p [less than].05). In contrast, the LM indices for the paths connecting task orientation and 'thoughts of escape' showed that the constraints were correctly imposed (for tennis: [X.sup.2](l) change = 0.401, p [greater than].05; for snooker: [X.sup.2](1) change = 0.158, p [greater than].04). In the final analysis, the paths connecting ego orientations and 'thoughts of escape' were set free to be estimated. The fit indices that emerged for each sample are presented in Table 5, indicating that the models fit the data very well.

Discussion

Cognitive interference that athletes experience during performance was examined in relation to their goal orientations and their perceptions of competence. Results showed that task orientation was negatively associated with thoughts of escape from the task, while perceptions of competence were seen to be important in determining the relationship between ego orientation and escaping thoughts. In addition, it was indicated that advancements in the measurement of cognitive interference are required to further explore its nature and role in sports.

The Measurement of cognitive Interference and the Use of TOQ in Sport Settings

The results concerning the evaluation of the TOQ were not straightforward. The fit indices obtained from the original confirmatory factor analysis failed to reach acceptable levels. Each subscale was then tested separately and the analysis for the 'task relevant worries' subscale indicated poor results. A possible reason for this might be that while some items are clearly negative in nature (e.g. "I think about how poorly I am doing", or "I think about how often I feel confused"), others might be perceived by performers either as positive or negative (e.g. "I think about my level of ability", or "I think about what someone will think of me"). Other items might even be perceived as motivational (e.g. "I think about how I should be more careful", or "I think about how difficult what I am doing is"). Therefore this particular subscale was excluded from further analysis and requires investigation prior to its use in sport in future.

The results for the 'task irrelevant thoughts' subscale showed that certain items created problems ("I think about something that happened in the recent past", "I think about something that happened in the distant past", "I think about something that might happen in the future", "I think about something that makes me feel angry"). The fact that the errors of these items were found to be correlated indicates that they might have some additional characteristics in relation to the rest of the items. What seems to be the hidden construct behind these items when compared to the rest (e.g. "I think about other activities", "I think about members of my family", "I think about friends", or "I think about personal worries") is that they refer to situations less approximate to the environment of the individual, and therefore more distant in terms of memory. However, these are simply assumptions and further investigation is needed to better explore this aspect.

Finally, for the 'thoughts of escape' subscale the fit indices were satisfactory. However, one item did not seem to fit appropriately with the rest of the scale ("I think about how hard it is"). The reason that this particular item was dropped is that such a thought does not seem to be as clearly negative as the rest (e.g. "I think about quitting", "I think about how I cannot stand it any more"). Furthermore, this item revealed high correlations with the items from the 'task relevant worries' subscale.

Overall, therefore, a great deal remains to be investigated about the conceptualization and the measurement of cognitive interference in sports. However, our approach allowed us to use psychometrically sound sections of the original TOQ. Further work is, however, strongly recommended.

Goals and Cognitive Interference

Investigating the structure of the relationships between goal orientations, perceived competence and the cognitive interference subscales that were retained, path analysis revealed an association between goal orientations and 'thoughts of escape'. Goal orientation theory (Nicholls 1984, 1989) predicts that perceptions of competence are important in determining behavior and affect for ego oriented people, while for those with a task orientation these perceptions are not influential. In accordance with the theory, task orientation proved to be negatively related to thoughts of escape regardless of levels of perceived competence. In contrast, it was hypothesized that for ego orientation low perceived competence would be associated with elevated levels of interfering thoughts, while high perceived competence would have a negative association with such thoughts. The patterns of relationship were confirmed for the low competence groups, while a nonsignificant, although negative, relation emerged for the high compe tence groups.

The results of the path analysis are informative as far as the role of perceived competence is concerned, indicating patterns of behavior that have been predicted and supported by Nicholls (1984, 1989) and other researchers investigating the relationship between goal orientation and learned helplessness (e.g. Diener and Dweck 1978, 1980; Elliott and Dweck 1988). In particular, if a content analysis of the 'thoughts of escape' subscale is considered, patterns of learned helpless behavior can be identified. Research has shown helpless individuals are more likely to be less persistent and more prone to withdrawing from a situation when facing difficulties. Moreover, research investigating learned helpless responses has identified various factors that lead to such performance disruptions. Ames (1984) and Diener and Dweck (1978) have reported that low ability attributions have been found to lead to a loss of belief about the utility of effort. Moreover, Leggett (1986) and Nicholls (1984) found support for the not ion that withdrawal may have a defensive character when continued effort is considered to further document low ability. Finally, Weiner (1982) has reported that negative affect, such as anxiety and shame, can motivate escape attempts.

In relation to these findings, and considering the patterns of cognitions concerning the role of effort, the importance of subjective outcome, and the perceptions of success that have been identified for ego and task oriented individuals (see Nicholls 1984), the present results are in line with past research supporting the notion that task orientation is associated with behavioral stability and persistence, while ego orientation has been described as more 'fragile' in relation to these concepts. The stability for task oriented athletes might be attributed to the fact that perceptions of competence do not seem to influence behavioral outcomes, while the fragility associated with ego orientation might be attributed to its dependency on perceptions of competence. Furthermore, even the perceptions of competence for those high in ego orientation are, by themselves, more fragile than those for the task orientation. Task oriented individuals base their assessment of competence on oneself, by evaluating effort input and levels of mastery concerning the task, and making comparisons with previous performances. In contrast, for ego oriented individuals' assessment of competence is based on performance outcome, often win or loss, and social and normative comparison. Therefore, when a task cannot be accomplished, or when the opposition is better, perceptions of competence are lowered resulting in maladaptive patterns of behavior such as thoughts of escape from the task and actual task avoidance.

The present results are in line with contemporary research and advances in goal orientation theory, suggesting that ego orientation is not unidimensional, but can be described as self-enhancing or self-defeating. In particular, Elliot and Harackiewicz (1996) and Skaalvik (1997), attempting to integrate goal orientation theory with earlier theories of achievement motivation focusing on approach and avoidance orientation (e.g. Atkinson, 1957), argued that ego orientation should be examined in relation to the individuals' approach to the achievement context, that is whether individuals seek to demonstrate competence or to avoid demonstrating incompetence. Accordingly, Elliot and Harackiewicz (1996) proposed a trichotomous variant of achievement goal framework composed of a mastery (task) orientation and two performance (ego) orientations, namely, performance approach and performance avoidance. According to their predictions, for mastery and performance approach oriented individuals achievement settings are cons idered challenging and exciting, thus promoting affective and cognitive investment which is translated into concentration and task absorption. In contrast, for performance avoidance oriented individuals achievement contexts are perceived as threatening and therefore anxiety evoking, eliciting, in the face of failure, disruption of concentration, and attempts of escape from the situation either physically, or mentally. In relation to these predictions Elliot and Harackiewicz (1996) and Skaalvik (1997) reported that mastery and performance approach orientations were positively related to intrinsic motivation, whereas performance avoidance orientation was found to undermine intrinsic motivation. In line with these propositions the present results emphasize that ego orientation is not necessarily connected to negative motivational and behavioral patterns. However, it should not be overlooked that task orientation was more directly associated to adaptive patterns of cognitions and behavior.

'Task irrelevant thoughts' were not associated with goal orientation nor to perceived competence. A careful look at the results of previous studies investigating cognitive interference is revealing. Since 1978, when cognitive interference was first studied by means of structured questionnaires, research has mostly been conducted using either the TOQ or the Cognitive Interference Questionnaire (CIQ; an instrument evaluating cognitive interference as a state characteristic, which also includes a 'task irrelevant thoughts' scale identical to the TOQ; see Sarason et al. 1986). Interestingly, previous studies in academic settings have not discriminated between the different subscales, dividing their samples into high and low TOQ or CIQ groups based on total scores. Though not erroneous, this fails to show which kinds of intrusive thoughts can be predicted by different variables, or which are more influential in terms of performance.

Hoffman (1993), however, found 'task-related thoughts' to be associated with both ego orientation and performance, while the 'task-unrelated thoughts' were not significantly associated with these variables. Similarly, Miculiner (1989), examining the relationship between cognitive interference and performance in the academic setting, found that 'thoughts of escape' had a significant effect on performance, while the effect of 'task irrelevant thoughts' was not significant.

Considering these findings, therefore, the results of the present study are not surprising. Due to the very small amount of research using the TOQ and CIQ instruments, and in particular in sport, it does not seem appropriate to draw firm conclusions at this stage. However, a possible reason for the results of the present study is that task irrelevant thoughts are rather neutral in nature, especially in long-lasting sports such as snooker and tennis where concentration cannot be at its maximum for the whole of the event and participants need some time to 'switch off' and relax their minds, having thoughts that are not connected to the game itself.

Goals, Cognitive Interference and the Anxiety - Performance Relationship

Recent thinking in sport anxiety research suggests that cognitive and somatic anxiety might be helpful for some athletes. In other words, we need to discriminate intensity from direction of anxiety (Jones, 1995). This particular approach seems promising and has opened new research avenues in sport psychology. However, research has to answer how and why individuals experience anxiety as debilitative or facilitative. Wine (1980) has support that for some individuals cognitive anxiety is detrimental to performance as a result of cognitive resources being consumed in responses irrelevant to the solution of the task, while Jones, Swain and Hardy (1993) have suggested that for some individuals elevations in cognitive anxiety, up to a certain level, might have desirable effects by increasing motivation and facilitating concentration. Furthermore, Jones (1995) suggests that affect, self-confidence, perceived ability, and perceptions of control might be some of the antecedents determining whether anxiety is experienc ed as debilitative or facilitative.

The results of the present study indicate that goal orientations and cognitive interference in the form of 'thoughts of escape', may be important issues to consider in the sport anxiety/performance relationship. In particular, if such interference is considered to have detrimental effects on performance (as has been the case in the academic setting), and if the facilitative and debilitative directions of anxiety prove to be related to good and poor performance respectively, goal orientations predicting differences in the nature and frequency of cognitive interference might be one of the antecedents determining why individuals experience (cognitive) anxiety in a facilitative or a debilitative fashion.

It would be expected that task oriented individuals would perceive anxiety as facilitative and would be characterized by a lack of negative intrusive thinking, while ego oriented individuals, and in particular those with low perceptions of competence, would perceive anxiety as debilitative, experiencing amounts of interfering thoughts detrimental to performance. This perspective seems quite appealing and should encourage further investigation into how goals and cognitive interference might be related to anxiety and performance in sport.

Address Correspondence To: Professor Stuart Biddle, Department of Physical Education, Sports Science, & Recreation Management, Loughborough University, Loughborough, Leics LE11 3TU, UK. Fax: + 44 1509 223 971. e-mail: s.j.h.biddle@lboro.ac.uk. Antonis Hatzigeorgiadis is now also at Loughborough University.

References

Ames, C. (1984). Achievement attributions and self-instructions under competitive and individualistic goal structures. Journal of Educational Psychology, 76, 478-487.

Arkin, R. M., Detchon, C. S., & Maruyama, G.M. (1982). Roles of attribution, affect, and cognitive interference in test anxiety. Journal of Personality and Social Psychology, 43, 1111-1124.

Atkinson, J. W. (1957). Motivational determinants of risk-taking behavior. Psychological Review, 6, 359-372.

Baron, R. M., & Kenny, D. A. (1986). The moderator, mediator distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology 51, 1173-1182.

Bollen, K. A. (1989). Structural equation with latent variables. New York: Wiley.

Boyd, M. P., Callaghan, J., & Yin, Z. (1991). Ego-involvement and low competence in sport as a source of competitive trait anxiety. Paper presented at the North American Society for the Psychology of Sport and Physical Activity, Asilomar, CA.

Bruch, M., Juster, H., & Kaflowitz, N. (1983). Relationships of cognitive components of test anxiety to test performance: Implications for assessment and treatment. Journal of Counselling Psychology, 30, 527-536.

Culler, R. E., & Holahan, C. J. (1980). Test anxiety and academic performance: The effects of study related behaviors. Journal of Educational Psychology, 72, 16-20.

Deffenbacher, J. L., & Deitz, S. R. (1978). Effects of test anxiety on performance, worry and emotionality in natural occurring exams. Psychology in the Schools, 15, 446-450.

Deffenbacher, J. L., & Hazaleus, S. L. (1985). Cognitive, emotional, and physiological components of test anxiety. Gognitive Therapy and Research, 9, 169-180.

Diener, C. I., & Dweck C. S. (1978). An analysis of learned helplessness: Continuous changes in performance, strategy, and achievement cognitions following failure. Journal of Personality and Social Psychology, 36, 451-462.

Diener, C. I., & Dweck C. S. (1980). An analysis of learned helplessness: The processing of success. Journal of Personality and Social Psychology, 39, 940-952.

Duda, J. L. (1993). Goals: A social-cognitive approach to the study of achievement motivation in sport. In R. N. Singer, M. Murphey & L. K. Tennant (Eds), Handbook of research on sport psychology (pp. 421-436). New York: Macmillan Publishing Company.

Duda, J. L., Chi, L., Newton, M., Walling, M., & Catley, D. (1995). Task and ego orientation and intrinsic motivation in sport. International Journal of Sport Psychology, 26, 40-63.

Duda, J. L., & Nicholls, J. G. (1992). Dimensions of achievement motivation in school-work and sport. Journal of Educational Psychology, 84, 290-299.

Dweck, C. S. (1989). Motivation. In Lesgold, A. & Glaser, R. (Eds), Foundations for a psychology of education (pp. 87-136). Hillsdale, N.J.: Lawrence Erlbaum Associates.

Dweck, C. S. (1992). The study of goals in psychology. Psychological Science, 3, 165-167.

Elliot, A. J., & Harackiewicz, J. M. (1996). Approach and avoidance achievement goals and intrinsic motivation: A mediational analysis. Journal of Personality and Social Psychology, 70,461-475.

Elliott, E. S., & Dweck, C. S. (1988). Goals: An approach to motivation and achievement. Journal of Personality and Social Psychology, 54, 5-12.

Ewing, M. E. (1981). Achievement motivation and sport behavior of males and females. Unpublished doctoral dissertation, University of Illinois, Urbana-Champaign.

Fox, K., Goudas, M., Biddle, S. J. H., Duda, J., & Armstrong, N. (1994). Children's task and ego goal profiles in sport. British Journal of Educational Psychology, 64, 253-261.

Gallassi, J. P., Frierson, Jr. H. T., & Sharer, R. (1981). Behaviour of high, moderate, and low test anxious students during an actual test situation. Journal of Consulting and Clinical Psychology, 49, 51-62.

Goudas, M., Biddle, S. J. H., & Fox, K. (1994). Achievement goal orientation and intrinsic motivation in physical fitness testing with children. Pediatric Exercise Science, 6, 159-167.

Gould, D., Ecklund, R. C., Petlichkoff, L., Peterson, K., & Bump, L. (1991). Psychological predictors of state anxiety and performance in age-group wrestlers. Pediatric Exercise Science, 3, 198-208.

Graham, J., & Nolan, S. (1991). Motivational influence on cognition: Task involvement, ego involvement, and depth of information processing. Journal of Educational Psychology, 83, 187-194.

Hall, H. K., & Kerr, A. W. (1997). Motivational antecedents of pre-competitive anxiety in youth sport. The Sport Psychologist, 11, 24-42.

Hoffman, D. A. (1993). The influence of goal orientation on task performance: A substantively meaningful suppressor variable. Journal of Applied Social Psychology, 23, 1827-1846.

Hom, H. L., Duda, J. L., & Miller, A. (1993). Correlates of goal orientation among young athletes. Pediatric Exercise Science, 5, 168-176.

Hunsley, J. (1987). Internal dialogue during academic examinations. Cognitive Therapy and Research, 11,653-664.

Jones, J. G. (1995). More than just a game: Research developments and issues in competitive anxiety in sport. British Journal of Psychology, 86, 449-478.

Jones, J. G., Swain, A. B. J., & Hardy, L. (1993). Intensity and direction dimensions of competitive state anxiety and relationships with performance. Journal of Sports Sciences, 11, 525-532.

Joreskog, K. (1993). Testing structural equation models. In K. Bollen, & S. Long, (Eds), Testing structural equation models (pp. 294-316). London: Sage Inc.

Kanfer, R., & Ackerman, P. L. (1989). Motivation and cognitive abilities: An integrative/aptitude-treatment interaction approach to skill acquisition. Journal of Applied Psychology, 74, 657-690.

Leggett, E. L. (1986). Individual differences in effort-ability inference rules: Implications for causal judgements. Paper presented at the annual meeting of the Eastern Psychological Association, New York.

Man, F., Stuchlikova, I., & Kindlmann, P. (1995). Trait-state anxiety, worry, emotionality, and self-confidence in top-level soccer players. The Sport Psychologist, 9, 212-224.

Martin, J. J., & Gill, D. L. (1991). The relationships among competitive orientation, sport-confidence, self-efficacy, anxiety, and performance. Journal of Sport and Exercise Psychology, 13, 149-159.

Miculiner, M. (1989). Cognitive interference and learned helplessness: The effect of off-task cognitions on performance following unsolvable problems. Journal of Personality and Social Psychology, 57, 129-135.

Newton, M., & Duda, J. L. (1993). The relationship of task and ego orientation to performance cognitive content, affect and attributions in bowling. Journal of Sport Behavior, 16, 209-220.

Nicholls, J. G. (1984). Achievement motivation: Conceptions of ability, subjective experience, task choice, and performance. Psychological Review, 91, 328-346.

Nicholls, J. G. (1989). The competitive ethos and democratic education. Cambridge, MA: Harvard University Press.

Paulman, R. G., & Kennelly, K. J. (1984). Test anxiety and ineffective test taking: Different names, same constructs? Journal of Educational Psychology, 76, 279-288.

Sarason, I. G., Sarason, B. R., Keefe, D. E., Hayes, B. E., & Shearin, E. N. (1986). Cognitive interference: Situational determinants and traitlike characteristics. Journal of Personality and Social Psychology, 51, 215-226.

Samson, I. G., Samson, B. R., & Pierce, G. R. (1990). Anxiety, cognitive interference and performance. In Booth-Butterfield, M. (Ed.), Communication, cognition, and anxiety. (Special issue). Journal of Social Behavior and Personality, 5, 1-18.

Sarason, I. G., & Stoops, R. (1978). Test anxiety and the passage of time. Journal of Consulting and Clinical Psychology, 46, 102-109.

Schwenkmezger, P., & Laux, L. (1986). Trait anxiety, worry, and emotionality in athletic competition. In C. D. Spielberger & R. Diaz-Guerrero (Eds), Cross-cultural anxiety, Vol. 3. Washington, D. C.: Hemisphere.

Skaalvik, E. M. (1997). Self-enhancing and self-defeating ego orientation: Relations with task and avoidance orientation, achievement, self-perceptions, and anxiety. Journal of Educational Psychology, 89, 71-81.

Solmon, M. A., & Boone, J. (1993). The impact of students goal orientation in physical education classes. Research Quarterly for Exercise and Sport, 64, 418-424

Swain, A., & Jones, G. (1992). Relationships between sport achievement orientation and competitive state anxiety. The Sport Psychologist, 6, 42-54.

Vealey, R. S., & Campbell, J. L. (1988). Achievement goals of adolescent skaters: Impact on self-confidence, anxiety, and performance. Journal of Adolescent Research, 3, 227-243.

Walling, M. D., & Duda, J. L. (1995). Goals and their associations with beliefs about success in and perceptions of the purposes of physical education. Journal of Teaching in Physical Education, 14, 140-156.

Weiner, B. (1982). An attribution theory of motivation and emotion. In H. Krohne & L. Laux (Eds), Achievement, stress, and anxiety (pp. 223-245). Washington, DC: Hemisphere.

Wine, J. D. (1971). Test anxiety and direction of attention. Psychological Bulletin, 76, 92-104.

Wine, J. D. (1980). Cognitive-attentional theory of test anxiety. In I. G. Sarason (Ed.), Test anxiety: Theory research, and applications (pp. 349-385). Hillsdale, N.J.: Erlbaum.

Zats, S., & Chassin, L. (1983). Cognitions of test-anxious children. Journal of Consulting Clinical Psychology, 51, 526-534.

Zats, S., & Chassin, L. (1985). Cognitions of test-anxious children under naturalistic test-taking conditions. Journal of Consulting and Clinical Psychology, 53, 393-401.

 Confirmatory Factor Analysis for the TOQ
While performing I think about: 1 [*] 2
how poorly I am doing .389 (.921) .439 (.899)
what someone will thing of me .647 (.762) .711 (.703)
how I should be more careful .464 (.886) .535 (.845)
how well others can do on what
 I am trying to do .524 (.852) .583 (.813)
how difficult what I am doing is .472 (.882) .483 (.876)
my level of ability .372 (.928) .353 (.936)
the purpose of what I am doing .303 (.953) .233 (.972)
how I would feel if I were told
 how I performed .523 (.852) .448 (.894)
how often I feel confused .497 (.868) .349 (.937)
other activities .640 (.769)
members of my family .657 (.754)
friends .720 (.694)
something that makes me feel guilty .769 (.639)
personal worries .825 (.565)
something that makes me feel tense .747 (.665)
something that makes me feel angry .650 (.760)
something that happened earlier in the day .829 (.559)
something that happened in the recent past .830 (.558)
something that happened in the distant past .739 (.673)
something that might happen in the future .686 (.727)
stopping .725 (.689)
how unhappy I am .615 (.788)
how hard it is .475 (.880)
how I cannot stand it any more .793 (.609)
quitting .836 (.549)
running away .727 (.687)
taking something (e.g., pills, a drink) to make
 it easier .741 (.672)
going to bed or to sleep .766 (.643)
While performing I think about: 3 4
how poorly I am doing
what someone will thing of me
how I should be more careful
how well others can do on what
 I am trying to do
how difficult what I am doing is
my level of ability
the purpose of what I am doing
how I would feel if I were told
 how I performed
how often I feel confused
other activities .641 (.768)
members of my family .664 (.748)
friends .726 (.687)
something that makes me feel guilty .771 (.636)
personal worries .830 (.558)
something that makes me feel tense .746 (.666)
something that makes me feel angry .651 (.759)
something that happened earlier in the day .830 (.557)
something that happened in the recent past .825 (.565)
something that happened in the distant past .730 (.684)
something that might happen in the future .679 (.734)
stopping .721 (.693)
how unhappy I am .596 (.803)
how hard it is .470 (.883)
how I cannot stand it any more .795 (.606)
quitting .843 (.538)
running away .726 (.688)
taking something (e.g., pills, a drink) to make
 it easier .752 (.660)
going to bed or to sleep .761 (.649)
While performing I think about: 5
how poorly I am doing
what someone will thing of me
how I should be more careful
how well others can do on what
 I am trying to do
how difficult what I am doing is
my level of ability
the purpose of what I am doing
how I would feel if I were told
 how I performed
how often I feel confused
other activities .667 (.745)
members of my family .675 (.752)
friends .717 (.697)
something that makes me feel guilty .816 (.578)
personal worries .863 (.506)
something that makes me feel tense .765 (.644)
something that makes me feel angry -
something that happened earlier in the day .757 (.654)
something that happened in the recent past -
something that happened in the distant past -
something that might happen in the future -
stopping .703 (.711)
how unhappy I am .601 (.798)
how hard it is -
how I cannot stand it any more .784 (.621)
quitting .869 (.495)
running away .736 (.677)
taking something (e.g., pills, a drink) to make
 it easier .756 (.655)
going to bed or to sleep .746 (.666)

Note. (*.)1. The initial analysis for the TOQ; 2. The analysis for the 'task relevant worries' subscale; 3. The analysis for the 'task irrelevant thoughts' subscale; 4. The analysis for the 'thoughts of escape' subscale; 5. The analysis for the two latter subscales after the modifications. The first numbers indicate the factor loadings while the ones in the parentheses indicate the measurement errors for all measurement models that were tested.

 The Fit Indices for all the Measurement Models of the TOQ
FIT INDEX ANALYSIS
 Task-relevant Task-irrelevant Thoughts
 TOQ worries thoughts(B) of escape(C)
Chi-square/degrees
 of freedom 904/347 99.5/27 280.5/44 90/20
Non-Normed Fit Index .768 .625 .782 .867
Comparative Fit Index .787 .719 .825 .905
Goodness of Fit Index .735 .890 .775 .893
FIT INDEX
 B + C
Chi-square/degrees
 of freedom 115/53
Non-Normed Fit Index .936
Comparative Fit Index .949
Goodness of Fit Index .907
 Descriptive Statistics and Correlations for all Variables
 Descriptive Statistics Correlations
 mean standard ego perceived
variable deviation competence
Task 3.85 .59 .10 .42 [*]
Ego 2.80 .97 .30 [*]
Perceived competence 4.40 1.30
Task-irrelevant thoughts 2.98 1.40
Thoughts of escape 2.45 1.44
 task-irrevelant thoughts of
variable thoughts escape
Task -.11 -.41 [*]
Ego -.04 .08
Perceived competence -.14 -.29 [*]
Task-irrelevant thoughts .40 [*]
Thoughts of escape
(*.)p [less than] .001
 The Fit Indices of the Modified Path Models for the Two Perceived
 Competence Groups of Each Sample
FIT INDEX TENNIS SNOOKER
 Low competence High competence Low competence
Chi-square/degrees
 of freedom .473/3 3.758/3 3.241/3
Non-Normed Fit Index 1.157 .835 .974
Comparative Fit Index 1.000 .918 .987
Goodness of Fit Index .995 .960 .964
FIT INDEX
 High competence
Chi-square/degrees
 of freedom 3.360/3
Non-Normed Fit Index .914
Comparative Fit Index .957
Goodness of Fit Index .967
 Fit Indices and Standardized Path Coefficients for
 Final Multiple-Sample Analysis
Fit Indices TENNIS SNOOKER
Chi-square/degrees of freedom 4.280/7 6.703/7
Non-Normed Fit Index 1.112 1.019
Comparative Fit Index 1.000 1.000
Goodness of Fit Index .978 .965
Note. The paths between 'task' and 'thoughts of escape'
were constrained to be equal, while the paths between
'ego' and 'thoughts of escape' were free to be estimated.