Issues in Implementing LOGO
The original philosophy of LOGO stressed the importance of personal discovery within a responsive exploratory environment. Many teachers appear to have interpreted the initial guidelines for working with LOGO as advocating little planning or intervention on their part (Keller, 1990). As a result, some teachers introduced students only briefly to commands or techniques and then allowed them to explore on their own. Observation of students’ behaviors under these circumstances indicated that they found the LOGO environment to be motivating. However, whatever the expectations, most younger students working in LOGO do not spontaneously engage in the general problem-solving skills thought to be prompted by programming environments. Without specific guidance, many students would find some set of commands that produced an interesting pattern and fiddle with this pattern by entering slightly different values for the key variables (Littlefield et al., 1988). They might continue variations on the same theme for several class periods using a trial-and-error approach. Researchers noted that students seemed highly motivated and intrigued with the interesting results their programs would produce, but that they were frequently unable to predict what a specific version of the program would do before they ran it or explain why particular results were produced. These observations suggest that many students engage in only weak problem-solving practices when they are allowed complete freedom to explore. In terms of the principles we discussed in Chapter 2, it seems that the external activity of programming within a discovery environment often does not engage internal processes thought to be important components of problem solving.
A Matter of Style
Even when students are task oriented, some educators believe they tend to work in a way that may be successful in the short run but does not lead to advanced problem-solving skills in the long run. Observation suggests that most young LOGO programmers use a style described as "product oriented," "brute force," and "linear" (Kurland, Clement, Mawby, & Pea, 1987). In this approach, the student has some desired screen effect in mind and generates a sequence of individual commands (linear approach) to achieve the desired effect (product oriented). The highly interactive nature of LOGO programming almost rewards this kind of behavior; the student can note bugs and insert or change individual commands very easily (by trial and error or brute force). A more disciplined approach would emphasize careful analysis of the problem to identify subproblems, the generation of procedures to solve the subproblems, and the construction of superprocedures to integrate the procedures. Some teachers require students to sketch out a plan on paper before working at the computer. This approach is an attempt to encourage students to think the task through before they begin to enter code, like asking students to construct an outline before writing.
More Effective Programming Instruction
These observations of students’ programming have implications for teachers and schools that want to use LOGO or other languages to develop problem-solving skills.
First, the analysis suggests that educators take a more realistic look at what they expect LOGO to accomplish. At best, students in elementary and middle school settings are likely to spend thirty to fifty hours in a year programming (Pea & Kurland, 1987b). Schools have many instructional responsibilities, and it may not be practical to assume that they can increase this time commitment substantially. We are not describing the formal programming courses most likely to be available at the secondary level. Perhaps it is necessary to accept the reality that students will not become very proficient as programmers in thirty to fifty hours, especially when they use much of the time in self-directed discovery. The solution may be to switch away from pure discovery experiences. This is the approach taken by educators who advocate more structured and mediated (but not lock-step) approaches to LOGO instruction.
Second, it appears that the development of problem-solving skills requires guidance. Increasing the amount of structure in the learning environment seems to improve student knowledge of LOGO, but it does not appear to be sufficient to develop general problem-solving skills (Littlefield et al., 1988). It appears that more general skills can be developed when teachers employ mediated instruction (Keller, 1990; Littlefield et al., 1988). In a mediated instruction approach, the teacher works directly to develop the thinking skills or strategies associated with the academic task the student is performing. The teacher must take care to establish that he or she is as concerned with the development of important thinking processes as with more visible products. When the content taught by mediated instruction is programming, key thinking processes might include planning, breaking complex problems into smaller problems, and using a systematic approach to identify and fix bugs. In contrast, the products are the program code and the result of the program (a graphic design when using turtlegraphics).
Students need to be told that planning and other cognitive activities are important, taught how to perform these skills, and monitored to make certain they use the skills. Because of the broad educational expectations associated with programming, it might also be useful to ascertain that students understand that the targeted cognitive behaviors may be processes involved in general problem solving. Planning, breaking down complex problems, and debugging are useful processes in many arenas, but it appears that the transfer of these skills to general applications is not automatic. Thus, mediated instruction might also concern the conscious transfer of skills learned in one domain to other settings and other problems. Teachers might ask students to think about how planning or debugging skills emphasized in programming might apply to writing or preparing a speech. Students must understand that they are learning skills they will be expected to apply in a variety of areas.
Mediated instruction involves a number of techniques. First, the teacher makes critical strategies explicit. Then he or she can mention specific skills and demonstrate them through think-aloud techniques. The teacher can demonstrate how to debug a faulty program and discuss how to bridge thinking skills learned in programming to other areas. For example, Littlefield et al. (1988) discuss applying processes from programming to plan a class party: What are the components of the problem, and what has to be done to prepare for each component? Second, mediated instruction attempts to involve students in thinking about and analyzing their own thinking and behavior. This can be a very different process from telling students what to do or evaluating the product of what they have done. Teachers seem to communicate differently when teaching LOGO than when teaching other subject matter. They provide fewer instructions and ask more questions (Emihovich & Miller, 1988). The questions often are not requests for information but are intended to get the students to think about their own behavior. A sequence of such questions might include: What did you tell the turtle to do? What did you want it to do? How are you going to fix it? (Au, Horton, & Ryba, 1987; Clements & Gullo, 1984). The use of questions is also opportunistic; teachers must generate the appropriate questions in the appropriate situations.