The effects of peer instruction on students’ conceptual learning and motivation

Experimental Design and Procedure

Students self-select into courses on the basis of personal choice, subjects could not be randomly assigned to treatment and control groups, nor could equal numbers of students be enrolled in each section. This limitation was addressed by using a “quasi-experimental design”, as outlined by Campbell & Stanley (1963), Cook & Campbell (1979). Quasi-experimental designs assume that subjects cannot be randomly assigned to treatment or control groups, and thus, groups may be unequal as for as students’ gender, majors, ability, background, etc. (the classes themselves were, however, randomly selected as treatment or control groups). The experimental design and analysis chosen, the Solomon four-group design (Campbell & Stanley, 1963), attempts to account statistically for any dissimilarities between treatment and control groups, but this is indeed a limitation of quasi-experimental designs.

The Solomon four-group design involves assignment of subjects to four groups. Two of the groups are pretested, and two groups are not. One of the pretested groups and one of the unpretested groups are subjected to the experimental treatment. The other two groups serve as controls. All four groups are then posttested (Campbell & Stanley, 1963) (Table 1).

Table 1 Solomon Four-Group Design

Pretest	Treatment	Posttest
O	X	O
O		O
	X	O
		O
X, treatment; O, dependent variables.

The Solomon four-group design offers rigorous control of most sources of internal and external validity and allows for increased generalizability vs. other experimental designs, because the four design elements are paralleled (Campbell & Stanley, 1963; Cook & Campbell, 1979). The paralleled elements control for the possible effects of a pretest on students’ subsequent performance and determine both the main effects and interactions of testing. If the pretest cues the students, both pretest groups will have higher posttest scores than the groups that do not receive the pretest. If there is an interaction between the pretest and the experimental treatment, so that the pretest provides an advantage to those students who receive only the treatment, the pretest-treatment-posttest group will have higher posttest scores than the treatment-posttest group (Abraham & Cracolice, 1994; Campbell & Stanley, 1963; Cook & Campbell, 1979). So, this design allowed for the investigation of variables as well as interaction effects. All students were required to buy a student textbook, which serve as a template for basic course content, lecture notes, and review problems. The study was performed by the same instructor. All sections followed the same course outline and were taught similar paces.

The study was conducted in an algebra-based physics course (concerning Newtonian Concepts). The primary objective of the course was to have students be able to describe and explain the kinematics, first, second, and third law, superposition principle, and kinds of force.

This study was performed in the two-year college classroom in Turkey. A total of 132 students in algebra-based introductory physics were initially included when this study was initiated. However, due to an attrition rate that was not unexpected in this subject area, only 123 completed all requirements of the study. Because instruments were administered on different days at the beginning and end of the study due to scheduling and because some students were absent on the days the instruments were given, the numbers of students who took the pretest and posttests varied between sections. These numbers were statistically accounted for by the analytical procedures used.

The number of students who completed the study in the treatment groups included 62 students. In this group, 53.89% were male, 46.11% were female. A total of 61 students participated in the control groups. Like the treatment group, most of the students in this group were male (52.62%), with fewer female (47.38%). The course content was the same for all classes participating in the study. Other course structure variables, such as the syllabus, text, content, grading procedures, and exam structure/formats were held constant and did not deviate from previous course structure. The control groups were taught using the traditional, didactic lecture method of instruction (i.e., students listened as the instructor lectured on the content).

The treatment groups were taught using peer instruction mentioned in the “Introduction” section. The lecture emphasizing the concepts and the ideas behind the proof were implemented for 5–10 min while avoiding equations and derivations. After the short lecture period, the ConcepTest question was presented. The question was read to the students, making sure there was no misunderstanding about it. Then, they had 1 min to select an answer (more time allows them to fall back onto equations rather than think). Since it was important each student to answer individually, it was not allowed them to talk to one another. After about a minute, the students asked to record their answer and the voting was started. In this study, students used flashcards to answer the ConcepTest in the lectures instead of showing of hands. Flashcard was that each student had a set of six or more cards labeled A–F to signal the answer to a question. According to the number of the correct answer, the instructor decided whether start a discussion between students or not. If it was needed, students formed small groups by picking their friends where they actively discussed the answers for several minutes. Students tried to convince a neighbor of the rightness of that answer. The convince-your-neighbor discussions were always conducted with a few groups of students. Doing so allowed us to assess mistakes being made, to hear how students who have the right answer explain their reasoning, and also to minimize the effect of lack of instructional ability of students. Also the small group of students could join to the larger group if they were finished discussing or if they got stuck. After the discussion period, they were asked to record their revised answer. The instructor then reviewed the answer with the class as a whole. The entire process from beginning to end took approximately 10–15 min.

PI sections used three to four ConcepTests per 75-min class while the traditional section relied solely on lecturing. For a ConcepTest to be most effective, the question requiring higher-level thinking about a concept was selected so that students were not simply recalling something they read or using “trial and error” with equations. ConcepTests were also at an appropriate difficulty level so students were challenged but could reason to the answer with their existing knowledge.

Instruments

The data used in this study were collected with two ways. The first of them was Force Concept Inventory “FCI” (Hestenes, Wells, & Swackhammer, 1992) and the last was Motivated Strategies for Learning Questionnaire “MSLQ” (Pintrich, Smith, Garcia, & McKeachie, 1993). It was expected to find answers to all two research questions with the outcomes obtained from those statistical tools, respectively. The details of them were given as follows:

1. FCI (Hestenes et al., 1992) is a widely used tool to assess student’s knowledge about topics in Newtonian Mechanics. FCI test consisting of 29 multiple-choice questions related to force and motion concepts, was used as a pre-and posttest. Internal reliabilities (Kuder-Richardson 21) for these tests were calculated as 0.69 and 0.71, respectively. FCI questions were divided to the 6 categories which were originally used by Hestenes et al. (1992) (Table 2; note that questions can appear in more than one category).

2. Motivated Strategies for Learning Questionnaire “MSLQ” (Pintrich et al., 1993) is a self-report instrument designed to assess college students’ motivational beliefs and use of learning strategies. The instrument consists of motivation and learning strategies scales. The motivation scale proposes three general motivation constructs: value, expectancy, and affect. The motivation scale consists of thirty-one items. The reliability and validity of the scale was performed in Turkish (Karadeniz, Buyukozturk, Akgun, Cakmak, & Demirel, 2008).

To determine the effectiveness of peer instruction on students’ conceptual learning, motivation, and self-efficacy, the posttest scores (means) of the conceptual test and the means of MSLQ’s categories were statistically analyzed for differences between the groups using 2x2 factorial design (Albanese & Mitchell, 1993).

Table 2 Categorization of the Question on FCI (Hestenes et al., 1992)

	Category	Questions
1. Kinematics	Velocity discriminated from position	20
	Acceleration discriminated from velocity	21
	Constant acceleration entails parabolic orbit changing speed	23, 24 25
	Vector addition of velocities	7
2. First Law	with no force velocity direction constant speed constant	4, 6, 10 26 8, 27
2. First Law	with cancelling forces	18, 28
3. Second Law	Impulsive force	6, 7
3. Second Law	Constant force implies constant acceleration	24, 25
4. Third Law	for impulsive forces	2, 11
4. Third Law	for continuous forces	13, 14
5. Superposition Principle	Vector sum	19
5. Superposition Principle	Cancelling forces	9, 18, 28
6. Kinds of Force	Solid contact passive Impulsive Friction opposes motion	9, 12 15 29
	Fluid contact Air resistance buoyant (air pressure)	22 12
	Gravitation acceleration independent of weight parabolic trajectory	5, 9, 12, 17, 18, 22 1, 3 16, 23