D:\687319001.doc Page 1 of 3 FIRST INCLASS ANOVA EXAMPLE WITH WRITEUP Let's say we have three different ways of teaching columnar addition. One way emphasizes rote memorization of a sequence of steps, one way uses manipulatives such as Popsicle sticks, and the last way uses a calculator. We teach 3 groups of 2nd graders, 5 to a group for a whole semester, then give a posttest of 20 addition problems. Here are the scores of each group: ROTE 7 5 3 4 1 Sum of X = 20 Mean = 4 MANIP 5 9 12 12 7 Sum of X = 45 Mean = 9 CALCULAT 21 15 17 18 14 Sum of X = 85 Mean = 17 1. Write the research hypothesis in English. In the populations from which the samples were drawn, there are differences in at least one pair of mean math scores across the three classes (rote method, mainipulatives method, and calculator method). 2. Write the null hypothesis in English: In the populations from which the samples were drawn, there are no differences in mean math scores across the three classes (rote method, mainipulatives method, and calculator method). 3. Write the name of the statistical test you think is appropriate. One way ANOVA followed by a multiple comparison test. 4. Write a narrative paragraph or two summarizing the results, assuming no table in the article. Be sure to indicate whether one class scored higher than any other as well as how each class did relative to the others. Three second-grade classes were taught columnar addition for a semester by three different methods. After instruction, a one-way analysis of variance (ANOVA) was conducted on the mean addition scores of the three groups of participants. One group was taught using the rote memory method (M = 4.00, SD = 2.24), one group was taught with manipulatives (M = 9.00, SD = 3.08), and one group was taught with the calculator method (M = 17.00, SD = 2.74). This analysis produced a significant analysis of variance (F(2,12) = 29.32, p <.001), indicating that there were differences among these means. Eta squared was .83, indicating a strong effect size. Multiple comparisons with the Tukey HSD test revealed that differences exist among all pairwise comparison of means with the mean for the calculator group being highest, the mean for the manipulatives group in the middle, and the mean for the rote memory group lowest (p < .05). Thus, the null hypothesis of no differences among the means was rejected. This study produced support for the idea that the best method for teaching columnar addition to second graders is the calculator method, while the poorest method is rote memory. 8. Write a narrative paragraph or two summarizing the results, but including one or more tables. Three second-grade classes were taught columnar addition for a semester by three different methods. After instruction, a one-way analysis of variance (ANOVA) was conducted on the mean addition D:\687319001.doc Page 2 of 3 scores of the three groups of participants. Table 1 presents the means and standard deviations of these three groups. Class: The old, 4th edition of the APA manual called for the following to be placed in text at this point, with the table itself on a separate sheet at the end of the manuscript: _________________________ Insert Table 1 About Here _________________________ Some journals still require this. Consult their author guidelines. The new, 5th edition has done away with this. You should still place each table on a separate sheet at the end of the manuscript, but you need not use "Insert Table About Here." A one-way analysis of variance (ANOVA) was calculated on these mean scores. Table 2 presents the source table for this analysis. Again, the old manual would require: _________________________ Insert Table 2 About Here _________________________ Inspection of this table reveals a significant F ratio, indicating that the three means were not equal. Eta squared was .83, indicating a strong effect size. Tukey's test for Honestly Significant Differences (HSD) was then calculated to determine which means differed. This test revealed significant differences among all pairs of means, with the mean for the calculator group being highest, the mean for the manipulatives group in the middle, and the mean for the rote memory group lowest. Thus, the null hypothesis was rejected, and the research hypothesis for the study was accepted. This study produced support for the idea that the best method for teaching columnar addition to second graders is the calculator method, while the poorest method is rote memory. NOTE: THE TABLES GO LAST IN THE MANUSCRIPT, AFTER THE REFERENCE PAGE/S, ONE TABLE TO A PAGE, DOUBLE-SPACED, NO VERTICAL LINES, PAGES CONTINUOUSLY NUMBERED WITH THE REST OF THE MANUSCRIPT. Table 1 Mean Number of Correct Answers for Three Groups of Second Grade Students on a Test of Columnar Addition __________________________________________________________________________________ Group n M SD __________________________________________________________________________________ Rote Memory 5 4.00 2.24 Manipulatives 5 9.00 3.08 Calculator 5 17.00 2.74 __________________________________________________________________________________ D:\687319001.doc Page 3 of 3 Table 2 Analysis of Variance of Math Scores After Three Different Methods of Instruction ________________________________________________________________________ Source SS df MS F p _______________________________________________________________ Between Within Total 430.30 2 215.00 88.00 12 7.33 518.00 14 29.32 .001 ________________________________________________________________________