While debate on educationally beneficial outcomes for students in Physical Education has spanned across a variety of learning domains (affective, fitness, and social outcomes) (Casey and Goodyear, 2015), student performance remains positioned at the center of the research interests (Harvey and Jarret, 2014; Rink et al., 1996). Nevertheless, both cognitive (knowledge) and motor performance (skills) conditions contribute to students' performance improvements (Farias et al., 2015). Indeed, in any physical activity where cognitive and motor domains are recruited inextricably, the alignment between a student's knowledge repertoire and her/his motor response may not be as straightforward as expected (French and Thomas, 1987). In line with these premises, research has advocated that an examination of the cognitive domain outside the practice field is critical for a more comprehensive measure of student performance (Blomqvist et al., 2001).
Despite the fact that the ways through which students' performance can be expressed and examined are highly sport specific (Rink et al., 1996), less attention has been paid to individual sports involving closed skills (Metzler, 2011), such as gymnastics or athletics. Traditional methods of knowledge testing have targeted principally low-order cognitive aspects (knowledge, comprehension, and application) through paper-and-pencil assessment of factual information on team sports (declarative and procedural knowledge) (Rink et al., 1996). The emergence of instructional perspectives, which place high emphasis on the cognitive domain of performance (Bunker and Thorpe, 1982; Griffin et al., 1997), has innovated the methods toward video-based assessment of game understanding (what, when and why to do) in team (Blomqvist et al., 2005; Oslin et al., 1998) and single games (Blomqvist, Luhtanen, Laasko, and Keskinen, 2000). In athletics, there was also a shift from a focus on content knowledge, such as rules and technical terminology of technique, to assessment of cognitive aptitudes such as analysis, application, and evaluation, and student ability to provide feedback on motor performance through video-based observations (Hastie et al., 2013).
The examination of students' improvements has long been grounded on assumptions on how students learn best (Rink, 2001) and cannot be dissociated from the specific instructional approaches used to produce such outcomes (O'Sullivan, 2013). Such instructional approaches for teaching Physical Education have been located in a continuum of teacher's directedness separating more teacher-centered (related to skills-based approach to content) from more student-centered approaches (emphasis on cognitive processes and social interactions) (Metzler, 2011; Rink, 2001). A classic example of a teacher-centered approach is the Direct Instruction (DI) model in which the teacher acts as the sole instructional leader who takes all the decisions on content development, class management and student engagement patterns (Metzler, 2011). The main priority in DI is the psychomotor domain because its underlying assumption concerning student performance is that some level of proficiency in elementary skills is necessary before proficient engagement in more complex game situations or tasks can be achieved (Rink, 1993). Thus, DI is designed for creating immediate success and "development of movement patterns and skills performed by students" (Metzler, 2011, p. 43). The high structured learning tasks permits close supervision by the teacher who "critically observes and analyses the movement patterns and skills performed by students" and provides "high rates of feedback on performance" (Metzler, 2011, p. 43). The skills are learned through teacher-directed instruction, shaping and modeling. Social interactions and affective outcomes are not explicitly addressed by DI curriculum and it is known to appeal to low-order cognitive processes. Thus, students' cognitive processes are recruited when they receive information from the teacher and internalize that information (McMorris, 1998). Also, when they are taught directly on rules and terminology or receive instruction and feedback on skills criteria as a means "to help them learn motor-skill patterns more quickly and proficiently" (Metzler, 2011, p. 179).
Research within DI is equivocal regarding students' cognitive outcomes. Students improved procedural and declarative knowledge assessed through written tests in rugby (Browne et al., 2004), while no improvements were found in badminton (Lawton, 1989), soccer (Mitchell, Oslin, and Griffin, 1995), and volleyball (Pritchard et al., 2008).
On the opposite side of the directness continuum lie the student-centered approaches for teaching Physical Education. Here, it is rejected the notion of learning as transmission and internalization in favor of looking at learning as a cognitively and socially active construction by students within a complex and culturally situated process (Kirk and Macdonald, 1998). Learning is a guided discovery process with tasks organized to enact cooperative work, problem solving, critical reflection and face-to-face interaction (Dyson et al., 2004). The Sport Education (SE) model (Dugas, 1994) is a high expression of a student-centered approach. SE has a multidimensional perspective on student learning as it simultaneously considers psychomotor (competence), affective (enthusiasm), and cognitive (literacy) outcomes. The primary features of SE include seasons instead of short units ("more time to learn"), affiliation ("work toward common goals"), formal competition ("games become meaningful"), keeping records ("built-in feedback"), culminating events ("recognition of excellence"), and festivity ("celebration") (Van der Mars and Tannehill, 2010, p. 307). As the actual season progresses, the students take greater responsibility for the organization and conduct of the unit by adopting leadership and management positions (role playing such as referees, coaches, statistician, or sports director are assigned to students). The instruction evolves through student face-to-face interaction and most of task presentations for knowledge and skill development take place through peer teaching and student-led cooperative activities within team practice sessions (Dyson et al., 2004). Students are also responsible for monitoring their teammates' learning and are held accountable on learning achievements through on-going performance records. The season concludes with formal competition events. The existing research on knowledge and skill development, although scarce, showed improvements in students' knowledge development (Farias et al., 2015; Hastie et al., 2009).
Some research has offered either a comparative or a combined examination of DI and SE in a variety of team sports. The outcomes were however somewhat inconclusive. Pritchard et al. (2008) conducted a comparative analysis between DI and SE in volleyball. Results revealed no significant difference between models for skills and knowledge while SE was more efficient in promoting quality game play. In rugby (Browne et al., 2004), results indicated that students in both groups (DI and SE) made significant improvements in knowledge of rules and in game skills.
Recently, other studies have begun to respond to the requests for more research on non-game-like activities such as aquatics, gymnastics, or in the case of this study, in track and field athletics. Nevertheless, to date, this is a poorly investigated area (Hastie et al., 2011b; Wallhead and O'Sullivan, 2005). The students' knowledge in track and field has been studied by Hastie et al. (2013), namely on the shot put, hurdles, and triple jump. The results showed that, while both groups (DI and SE) improved significantly in technical performance (examination of specific movement components) and performance measures (time and horizontal distance records), only the SE group made significant improvements in content knowledge. Nevertheless, this study did not consider students' sex and skill-level, which probably could interfere on the learning students' outcomes. Indeed, some students' characteristics, such as sex and skill level have been shown to be factors that hold considerable influence in student opportunities to participate in the activities and, consequently, have an effect on learning outcomes (Araujo et al., 2014). The fact that Hastie et al. (2013) observed positive results to the class as a whole, does not mean necessarily that both groups (boys and girls; higher and lower skill-level students) were given the same opportunities to learn.
The present study is part of a larger research project, which investigated the impact of two instructional approaches (DI and SE) on students' learning outcomes in athletics. Based on the same original dataset, while Pereira et al. (2015) published their results on the pupils' technical performance, the current study focused on students' content knowledge. Specifically, Pereira et al. (2015) found that, "while SE and DI approaches could both lead to improvements in the development of technical performance in track and field" (p. 125), the performance measures were markedly distinct when student sex and skill level were taken into account in the analysis. Girls and less skilled students improved technical performance in SE but not in DI.
Therefore, given the limited information provided by prior research on Physical Education with respect to cognitive learning and its scarce focus on athletics, the purpose of the current investigation was to examine the effects of SE and DI on students' content knowledge in three track and field events (hurdles, triple jump, and...