Pre-service physical education teachers’ implementation of “tgfu tennis”: An assessment of upper elementary students’ learning using a modified GPAI

Pre-service physical education teachers’ implementation of “tgfu tennis”: An assessment of upper elementary students’ learning using a modified GPAI

ABSTRACT

This article reports on a study of upper elementary students’ learning of three games components (base, decision-making, skill execution) as a result of their participation in an after-school tennis program called “TGfU Tennis.” Utilizing the Teaching Games for Understanding (TGfU) model, pre-service physical education teachers taught students (n = 54) from three separate schools. Video-recorded game play of 19 upper elementary students before and after tennis instruction was analyzed using a Game Performance Assessment Instrument (GPAI). Results indicate that mean indices for all three game components increased significantly between the baseline and post-intervention assessment dates. In light of these findings, suggestions are offered for future teaching and research related to TGfU.

INTRODUCTION

+ Teaching Games for Understanding (TGfU) is a games-centered teaching model that was initially intended to address traditional games teaching’s limitations, especially with respect to such things as techniques-based instruction and sustaining learners’ games interest (Bunker & Thorpe, 1982, 1986). Manifestations of these limitations, opined Bunker and Thorpe (1986), could be observed through various happenings, including learners’ limited psychomotor success, inadequate games understanding, poor decisionmaking capabilities, and overdependence on teachers.

Unlike traditional games teaching, in which learners are often taught a prescribed skill with limited-to-no mention or understanding of rationale or significance (Bunker & Thorpe, 1986), TGfU’s focus is on teaching the “why” before teaching the “how” of a game. Such a paradigmatic shift, Bunker and Thorpe (1986) suggest, enables learners’ increased games interest, enjoyment, and decision-making abilities. Traditional teaching, in which a “series of highly structured lessons rely on the teaching of skills and techniques” (Werner, Thorpe, & Bunker, 1996, p. 28), has previously been labeled “The Technical Model.” Within The Technical Model, once learners have mastered games skills, there is an expectation that these skills will be transferable to games and game-like scenarios (Werner et al.). Unfortunately, in practice, students’ application of their learning does not necessarily follow such a linear path.

To this, Holt, Ward, and Wallhead (2006) have previously suggested that a techniques-based approach results in students learning inflexible techniques resulting in an inability to transfer, and apply, their learning to game scenarios. While some might suggest that a minor games approach (as a variation of traditional practice) would circumvent this issue, like the full game approach, such teaching is still nonetheless “based on a facile analysis of games and what is required to play them effectively” (Launder, 2001, p. 28). TGfU as “The Tactical Model” presents an alternative that might also be recognizable as similar to some other recently popular terms for games-centered teaching approaches; in addition to TGfU these include Play Practice, Games Appreciation, and Games-Centered Learning (Butler, 2006; Griffin, Butler, Lombardo, & Nastasi, 2003; Hopper, 1998; Werner et al.).

In the almost 30 years which have passed since the publication of Bunker and Thorpe’s (1982) original model (see Figure 1), TGfU has become a “focus for researchers and teachers in several countries” (Griffin, Brooker, & Patton, 2005, p. 214) including Canada, the United States, and the United Kingdom. This acceptance within Western nations has also been similarly evidenced in other countries within South America, Africa, and Asia (Griffin et al.; Light, 2006). While all researchers and teachers within the field are yet to embrace TGfU, the increasing presence of TGfU within the literature suggests that it is here to stay.

Figure 1. Bunker and Thorpe’s (1982) Curriculum Model.

Figure 1. Pre-service physical education teachers’ implementation of “tgfu tennis”: An assessment of upper elementary students’ learning using a modified GPAI

Contenido disponible en el CD Colección Congresos nº 16

Although some researchers and teachers might willfully adopt TGfU as a model for games instruction, such an adoption nonetheless ought to be based on sound empirical research. That is, it is obviously inadequate for those who intuitively recognize the immediate and long-term benefits and potential of TGfU (i.e., for whom it simply “makes sense” or “feels right”) to adopt such a model without, for lack of a better phrase, “proof of the process.” However, despite theoretical and pedagogical aspects of the TGfU model having been discussed in research and professional contexts (Berkowitz, 1996; Holt, Strean, & Bengoechea, 2002; Thorpe, 1992), there remains a continued need for empirical research related to the merits of TGfU.

To this, and in order to improve upon the perceived legitimacy of TGfU, Griffin et al. (2003) recommend sound “data-based, not data-free development” (p. 215) is required. This, they suggest, will allow for researchers to contribute to, and improve upon, the “limited empirical support to back up an intuitive sense that [TGfU] works for students” (Griffin et al., p. 215). Initially, when the TGfU model first became a possible alternative for games instruction, research often compared one approach (e.g., tactical) to another (e.g., technical) as researchers set out to determine, in a polarizing fashion, “Which method is best?” (Harvey, Cushion, Wegis, & Massa-Gonzalez, 2010, p. 30). The results from these initial research studies (see Allison & Thorpe, 1997; Capel, 1991; Lawton, 1989; Turner & Martinek, 1992) have been equivocal (Holt et al., 2002).

Though undoubtedly true that “there is insufficient evidence to support the TGfU apologists [sic] claim that it is superior to other methods” (McMorris, 1998, p. 65), the research, nonetheless, suggests that TGfU deserves serious consideration. Common to many of these previously mentioned studies, for example, is the notion that a TGfU model has some positive potential, especially with respect to games sense and student affect. In response to previous comparative studies, Rink, French, and Graham (1996) have highlighted the inherent problems associated with discerning between tactical and technical instruction (e.g., the learning of tactics in a techniques-dominated class free of direct tactical instruction).

As the “pitting of technical against tactical approaches…was not the apparent purpose of the original TGfU model” (Holt et al., 2002), it is also somewhat curious that such comparative studies were given such earlier attention. Contrasting the frameworks of these previous investigations, Metzler (2005) contributes: I am suggesting that the question, Which model is best? is inappropriate because…that research design is simply invalid…. Domain priorities, engagement patterns, and learning outcomes vary across models, so trying to test how well two models promote different types of outcomes with processes they do not hold in common is the empirical equivalent of comparing apples to oranges (p. 190). Rink (2001) articulates an equally important point when she suggests, “when you spend all of your effort proving that a particular kind of teaching is better than another kind of teaching, you limit what you can learn about the very complex teaching/learning process” (p. 123).

AREA OF STUDY

Recognizing the need for further empirical evidence in support of TGfU (particularly without a continued focus on polarizing and comparative data), this study aimed to measure the impact of pre-service physical education (PE) teachers’ games instruction on the tactics and techniques of a group of upper elementary students. More specifically, pre-service PE teachers taught after-school tennis lessons (in a program titled “TGfU Tennis”) using a TGfU framework as it was previously introduced to them within a compulsory elementary physical education pedagogy course.

After three weeks of instruction related to TGfU philosophy, research, and application, small groups of pre-service PE teachers collaborated to design eight-lesson units in which they were required to apply Bunker and Thorpe’s (1982, 1986) curriculum model. Traditional methods have tended to concentrate on specific motor responses (techniques) while failing to consider the contextual nature of games (Bunker & Thorpe, 1986). Chandler (1996) explains that decontextualized skills are those learned in isolation from a game context, which, as a result, make little sense to games learners (as well as to spectators).

Pre-service PE teachers were therefore encouraged to avoid such decontextualized activities (especially decontextualized drills) so that they might “promote an understanding of the purpose of skills as well as the product of those skills” (Chandler, p. 49). The pre-service PE teachers’ lesson plan format required that participating students would begin lessons with contextualized instant activities, followed by game play, skill development activities, with a return to game play. Similar to as was earlier illustrated by Mandigo and Anderson (2003), pre-service PE teachers were required to include within their lesson plans potential reflective questions, appropriate generic and specific tactical problems, and authentic opportunities for decision-making.

Meeting the design goals of the TGfU model, pre-service PE teachers were also required to teach net/wall strategies and skills by allowing students to initially play modified games (rather than teaching tennis skills before allowing games play) (Doolittle & Girard, 1991). The pre-service PE teachers were further encouraged to meet the major assumptions about TGfU previously summarized by Griffin, Brooker, and Patton (2005); students were introduced to modified games that were representative of the advanced game form (i.e., modification through representation and exaggeration) and which included problems common to many net games through sampling (Griffin et al.). In order to meet the developmental needs of learners, pre-service PE teachers modified games through representation and in order to emphasize a particular skill and/or strategy, pre-service PE teachers modified games through exaggeration (Werner et al., 1996).

Tennis, perhaps as much as any other sport, “requires high skill levels for even moderate success” (Doolittle & Girard, 1991, p. 57). However, with the TGfU model, the pre-service PE teachers modified tennis to fit the students’ developmental levels and to emphasize fundamental concepts (Doolittle & Girard; Thorpe & Bunker, 1989). Recognizing that pre-service PE teachers often also harbour misconceptions about students “learning strategy on their own” (Rovegno, 1993; Rovegno & Bandhauer, 1994), and in the spirit of the TGfU model, lesson plans included explicit and specific attention to tactical problems and their solutions as outlined by Mandigo and Anderson (2003) (e.g., consistency – positioning, setting up for attack – shot placement, etc.).

METHODS

Setting and Participants

Sixteen pre-service PE teachers enrolled in an elementary physical education curriculum and instruction course taught an after-school tennis program at three local elementary schools as a service-learning component of their course. The pre-service PE teachers were all in the first year of a Bachelor of Education (BEd) degree program and all had previously completed a Bachelor of Human Kinetics (BHK) or Bachelor of Kinesiology (BKin) degree.

These after-school lessons were 90 minutes in length and were taught twice a week for four consecutive weeks (i.e., there were eight 90-minute lessons in all). While a total of 54 elementary students from grades 4 through 6 participated in the tennis program, only 19 students were also research participants (20 were actually involved but one student was absent on the final assessment day and so has been excluded from the data). The ten male students ranged in age from 9 to 12 years while the nine female students ranged in age from 10 to 12 years.

Instrument

TheGame Performance Assessment Instrument (GPAI), as a valid and reliable method for assessing game performance, provides researchers with a means of observing and coding performance behaviours (Oslin, Mitchell, & Griffin, 1998) such as appropriate movement and skill execution. Oslin et al.’s initial field testing of the GPAI for net/wall games was based on observations of volleyball and it included three of seven game components (adjust, decisions made, and skill execution) though some of the other components certainly applied to net/wall games; the remaining unobserved four components were base, support, cover, and guard/mark. Hopper’s (2003) piloting of the GPAI for other net/wall games (e.g., badminton and tennis) led him to adapt six of the GPAI components (base, decision-making, cover, adjust, skill execution, and support).

According to Hopper (2007), this adaptation was necessary due to the rapid nature of, and the related difficulty in assessing, games play in most net/wall games. In 2007, Hopper further explained the application of the GPAI for tennis instruction. Here, he detailed the application of the first five of these six components. A modified version of this GPAI (Oslin, 2005; Oslin et al., 1998) was used to assess students’ game performance behaviours before and after their participation in the after-school tennis program. Only three game components were included within this particular study; they were base, decision-making, and skill execution.

Data Collection

Prior to beginning the first tennis lesson, groups of four students engaged in a mini-tennis game in which their instructions were to cooperatively “rally” a ball for at least three hits before they were to attempt any “winning shots.” Modified equipment used for this initial assessment (and subsequent tennis lessons) included Wilson® Starter Foam Balls, a 45 cm Wilson® EZ tennis net, and children’s 60 cm tennis racquets.

The court dimensions were 16m by 7m. Students were given no further information regarding such possibilities as where to stand, where to move after striking the ball, or how to strike the ball. All five of these groups (two from School A, two from School B, and one from School C) included two female students and two male students. A research assistant (RA) video-recorded each mini-tennis game for a period of 12 minutes. After the completion of the final lesson, this same procedure was repeated. In the lone occasion when a student was absent for the final lesson, another student substituted for the game. This substitute was only required to maintain the doubles set-up; her/his performance was not assessed.

Data Analysis

The approximately 120 minutes of video-recordings were transferred to a MacBook Pro computer using the software program iMovie. Video-recordings of students were then observed by the principal investigator (PI) so that their performance in specific game components could be analyzed using the modified GPAI. In order to provide a comprehensive view of students’ overall performance (Oslin et al., 1998), base, decisionmaking, and skill execution were coded as appropriate/efficient or inappropriate/inefficient. More specific criteria for these three components are included in Table 1.

Table 1. GPAI component criteria.

Table 1. Pre-service physical education teachers’ implementation of “tgfu tennis”: An assessment of upper elementary students’ learning using a modified GPAI

Contenido disponible en el CD Colección Congresos nº 16

Borrowing from Oslin et al. (1998) and previously applied by Haneishi, Griffin, Seigel, and Shelton (2009), each of these components were indexed and given equal relative value in order to compute an overall game performance value (see Table 2). Because of the differing and relatively large numbers of movements made in both sessions, it was ideal to express these values as indices. This decision was made in keeping with Griffin, Mitchell, and Oslin’s (1997) suggestion while also being aware of the possible limitations as suggested by Harvey et al. (2010) and Memmert and Harvey (2008).

Table 2. GPAI component indices and formulae.

Table 2. Pre-service physical education teachers’ implementation of “tgfu tennis”: An assessment of upper elementary students’ learning using a modified GPAI

Contenido disponible en el CD Colección Congresos nº 16

To ensure objectivity in applying the GPAI for coding students’ game performance before and after their tennis instruction, a volunteer also viewed two of the video-recorded sessions and coded four of the students’ game performance (i.e., 21% of the total research participants). This volunteer was especially familiar with TGfU and the GPAI. For example, she/he had completed a review of over 25 articles related to TGfU and had experience teaching tennis using a TGfU approach.

RESULTS

Inter-observer Reliability Inter-observer agreement (IOA) for the two separate coders (i.e., the PI and the volunteer) was computed using a Pearson product moment correlation as well as through a common calculation of interobserver reliability. Pearson’s r-values were r = .85 (BI), r = .89 (DMI), r = .95 (SEI), and r = .93 (GPI). The especially high values for Pearson’s r are likely due to its limitation with respect to considering the relative position of values in pair-wise correlations. It is for this reason that IOA was also calculated using a common calculation method (see van der Mars, 1989). This IOA calculation was also completed for base, decisionmaking, skill execution, and overall game performance.

Calculations were made based on the observers’ agreements and possible agreements, as previously identified and utilized by Harvey et al. (2010). In these calculations (agreements ÷ [agreements + disagreements] × 100) (Caro, Roper, Young, & Dank, 1979; van der Mars, 1989), base had an IOA agreement of 88%, decision-making an IOA of 87%, skill execution an IOA of 97%, and collectively the IOA was 90%. Base and decision-making had acceptable IOA scores of greater than 80%. That skill execution had an IOA of 97% was expected; criteria for assessing SE was effectively productbased rather than process-based. That is, both observers reviewed video-recorded footage to see if the ball was successfully hit over the net and in bounds (rather than focusing on the form of the motor task); it is perhaps curious then that this agreement was not 100%.

Component Indices

The fewest number of game component observations for a student was 30 (decision-making and skill execution) and the most was 67 (base). The means and standard deviations for the indices for the three components (and the cumulative GPI) were computed (see Table 3). In all three indices for the GPAI components (BI, DMI, SEI) students demonstrated increases in their mean scores.

Table 3. Component indices’ baseline and post-intervention means.

Table 3. Pre-service physical education teachers’ implementation of “tgfu tennis”: An assessment of upper elementary students’ learning using a modified GPAI

Contenido disponible en el CD Colección Congresos nº 16

A repeated measures dependent samples two-tailed t test revealed significance for this change for all three of the component indices. The BI (t(18) = -5.5928, p = .001, ? = .05), DMI (t(18) = -3.120, p = .006, ? = .05), and SEI (t(18) = -6.568, p = .001, ? = .05) were 1.31 (SEI) to 2.87 (BI) higher after students completed their eight 90- minute lessons.

DISCUSSION

The increase from the students’ baseline means in all three GPAI components, though positive, require further discussion, especially with respect to a consideration of the meaning of index scores and limitations related to the design of this particular study. Initial indices ranged from M = 0.568 (BI) to M = 0.783 (DMI). An index of M = 0.568 suggests that for every 100 attempts, a student would be successful (i.e., appropriate or efficient) 36 times while an index of M = 0.783 suggests that for every 100 attempts, a student would be successful 44 times. When these indices are considered this way, one might more easily be able to consider the students’ performance in the three game components after their involvement in the eight-lesson tennis program. Post-intervention game performance assessments have indices that obviously suggest a significant improvement for base, decision-making, and skill execution.

Expressed as percentages (as previously described), students who participated in the tennis program would make appropriate base movements 63% of the time, appropriate decisions 56% of the time, and efficient skill executions 68% of the time. While the 68% success rate for the skill execution is lower than the 80% benchmark suggested by Rink (2010) and Doolittle (1995), it is promising that this 68% reflects a 30% improvement over the baseline mean. Furthermore, individual students’ SEI were as high as 3.884 (i.e., a success rate of 80%). While these elementary students improved in the three game components (taught and) observed, the study’s limitations command explicit consideration.

These limitations are related to the nature of the activity, the nature of the participants, the amount of instructional time, the focus of instruction, and the pre-service PE teachers’ implementation of TGfU. There was no tennis equipment (i.e., racquets, nets, foam tennis balls) at any of the three schools before the implementation of this research study. Informal interviews with school administration, physical education teachers, and participating students revealed that none of the students had participated in an in-school tennis program at any time (though at least one student had participated in a summer tennis program). Consequently, these students were basically being taught tennis for the first time. With such limited background in tennis, low baseline scores in all three game components might be expected (though, of course, positive transfer from other net and wall games could also occur).

Furthermore, their significant improvements in all three areas might also be expected because of this factor. Unlike the captive audience characterizing in-school physical education programs, the after-school tennis program only included students who chose to enroll in the program. Because of this factor, participating students could be understood to be especially interested, and motivated to participate, in the tennis program. Again, because of this factor, greater success might be anticipated. The eight 90-minute lessons certainly exceeds the amount of time traditionally dedicated to in-school physical education time within Canadian schools (Janzen, 2004). Students at all three of the participating schools have physical education for 30-minute periods, twice every six-day rotation. Put into perspective, in the four weeks (and 720 minutes) in which the students participated in the after-school tennis program, they could have as few as 180 minutes of physical education instructional time.

Quite obviously, the amount of time afforded to the tennis instruction grossly exceeds the amount of time students would be afforded to within four weeks of physical education classes. Consequently, suggesting that such results might similarly occur within “regular” physical education instructional would be unfair. Pre-service PE teachers knew that the elementary students’ game play was being observed for base, decision-making, and skill execution. With this awareness (and despite the parameters of the assignment), the pre-service PE teachers likely placed greater focus on these game components. Furthermore, as no observations were made of cover and adjust, it is impossible to suggest that students’ similarly improved in these areas. Another important piece of information learned from engaging in this study was related to the task students performed during their baseline and post-intervention assessments.

By requiring students to play in groups of four (i.e., doubles), there were unnecessarily additional people included in the video-recorded sessions. In fact, having students play a doubles mini-tennis game was most often a prohibitive structure as different ability levels prevented students from demonstrating their abilities and decision-making. Such observations have similarly been noted by Buck and Harrison (1990) in their observations of students’ game play in volleyball. There they found that low ability students would avoid play while expecting high ability players to “take” the ball; these same observations were witnessed when reviewing the video-recordings. Future researchers’ use of a GPAI to assess students’ game play in tennis should consequently consider engaging students in singles scenarios. Finally, while pre-service PE teachers were required to plan and teach using the TGfU model, it was not altogether possible to always label their instruction as a true implementation of The Tactical Model.

This occurred despite the professors’ three purposeful practices to help ensure TGfU implementation. First, the professors spent twelve hours of instructional time teaching and modeling the TGfU approach. Second, students’ lesson plans were assessed (prior to implementation) to ensure an understanding of teaching according to a TGfU model. Finally, the two professors’ presence at all of the teaching sessions was intended to allow them to support the pre-service PE teachers’ implementation of the TGfU model; this required that they engage in such activities as questioning, prompting and modeling. However, despite these three efforts, pre-service teachers would at times seem to “fall back” upon their experiences as students. Such actions were characterized by a return to some of the previously mentioned characteristics of The Technical Model.

Despite these limitations, this study nonetheless offers some empirical support for the potential of teaching tennis (as a net/wall game) using a TGfU approach. Tennis is perhaps one of the most difficult net/wall games to learn. That upper elementary students were able to demonstrate significant improvement in their game play in all three components as a result of their pre-service PE teachers’ instruction suggests that the model has obvious potential in school physical education programs. While this service learning project was intended to allow the pre-service PE teachers to have opportunities to prepare themselves to be successful teachers (who are knowledgeable and skilled with respect to TGfU), the previously mentioned limitations undoubtedly prohibit sweeping generalizations about TGfU’s application to the classroom context. However, it is strongly believed that continued study in authentic environments might continue to support the sorts of claims established herein.

CONCLUSION

While this research study has aimed to contribute to the call put forth by Griffin et al. (2003) for sound empirical research, continued research must also continue with focused attention to the results and limitations previously suggested. Because TGfU has been presented as a (rather than the) teaching model that might be incorporated into games instruction within physical education, similar research might aim to repeat aspects of this study while also aiming to more honestly reflect the modern physical education context. For example, by teaching captive audiences for shorter periods of time, or by having a teacher to student ratio no smaller than 1:20, conditions might be more similar to the authentic classroom environment.

While such attention to the naturalistic setting certainly presents challenges (e.g., teachers feeling de-skilled, scheduling constraints), such research is essential if we are to make inroads within pre-service and in-service education programs (Brooker, Kirk, Braiuka, & Bransgrove, 2000). Contributing to this challenge is the reality that one cannot simply “add” TGfU to existing practice as it requires that some teachers would need to undergo a radical philosophical adjustment (Butler, 1996). While Kirk and Claxton (as cited in Light, 2002) have suggested that TGfU continues to encounter resistance from in-service physical education teachers, Light has more recently suggested that pre-service PE teachers might become “amenable to the TGfU approach” (p. 299) if given opportunities to engage with it within their teacher education programs.

While the GPAI utilized for this survey only considered three of the game components, future research might also consider the two others suggested by Hopper (2007) for net and wall games. As ongoing and future research continues to contribute to the growing body of literature providing sound empirical support for TGfU as a games instruction model, increasing numbers of PE pedagogues, their university students, and in turn, public school students will continue to benefit.