Effects of Problem Solving Activities and Calculators on Problem Solving and Computation in Grade Four.

The purpose was to investigate the effects of a planned sequence of problem solving activities and the use of calculators on problem solving and computational performance of fourth grade students. Procedure. Twenty-four intact classrooms of 538 fourth grade students from schools in Region 4, Detroit Public Schools, were r and omly selected for this study. Classrooms were assigned to two treatment group and a control group to equalize mean scores on a st and ardized problem solving test. Both treatments were taught by the regular classroom teachers. The Calculator treatment (Cal) consisted of a series of sixteen problem solving lessons with calculators used. The Non Calculator treatment (Non Cal) used the same series of sixteen lessons with paper and pencil for computing. The pre- and posttest, prepared by the author, contained 10 story problems and 8 computational exercises. Students used only paper and pencil on these tests. A Calculator test with calculators allowed, consisted of 19 difficult story problems and 4 computational exercises, involving large numbers. Pre and post Attitude Surveys were given. Analyses of variance and covariance and Scheffe paired comparison were used to analyze the data. Findings. On the ten-item posttest on problem solving, Cal was favored over Control (p = .0000) and Non Cal surpassed Control (p = .0014). Using the problem solving pretest as a covariate, adjusted means were Cal (6.03), Non Cal (5.80), and Control (5.12). Cal and Non Cal gained 0.87 ad 0.84 respectively while Control changed -0.18. On the eight-item posttest on computation, Cal was favored over Non Cal (p = .0000) and over Control (p = .0122). Using the computation pretest as a covariate, adjusted means were Cal (4.67), Non Cal (4.57), and Control (4.08). Cal and Non Cal gained 0.14 and 0.51 respectively while Control changed -0.16. There was a significant difference on the posttest total scores, favoring Cal over Non Cal (p = .0043) and Cal over Control (p = .0014). Using the pretest total scores as a covariate, the adjusted means were: Cal (10.63), Non Cal (10.48), and Control (9.18). Cal and Non Cal gained 1.01 and 1.42 respectively while Control changed -0.34. On the Calculator Total Test Scores, Cal surpassed Non Cal (p = .0000). Using the pretest total scores as a covariate, adjusted means were: Cal (6.18), Control (6.03), and Non Cal (4.94). Cal attempted 76.2% of the possible problems while Control tried 64.7% and Non Cal worked 57.5%. There were no significant differences on attitudes towards solving story problems. There were no within treatment sex differences. An item analysis revealed that success on problem solving was comparable to success on computation with similar computations for addition, multiplication, and division. The "How many more?" subtraction story problems were the most difficult and subtraction with zero in the tens place was the hardest computation item. There appeared to be help in learning new operations or in reinforcing skills recently learned by the use of the calculator. Conclusions. Time spent working on story problems increased problem solving ability and computational facility. Calculator use did not decrease computation skills. Students needed special help when working with large numbers on calculators. A calculator did not in itself assure success with large numbers. The calculator was useful if the problems were within the scope of the child's paper and pencil ability. The percent of attempts declined when computational difficulty was beyond grade four. The Cal students were less afraid to attempt difficult problems than were the students who had not used a calculator previously.PhDCurriculum developmentUniversity of Michiganhttp://deepblue.lib.umich.edu/bitstream/2027.42/157890/1/8025655.pd