36 Simkin, Keuchler, Savage, and StiverTABLE 2Sources of Sample DataClass Class Title Number of StudentsClass 1 Introductory Financial Accounting 46Class 2 Introductory Financial Accounting 46Class 3 Introductory Financial Accounting 45Class 4 Introductory Financial Accounting 45Class 5 Introductory Financial Accounting 46Class 6 Accounting Informati<strong>on</strong> Systems 54Class 7 Accounting Informati<strong>on</strong> Systems 39Class 8 Accounting Informati<strong>on</strong> Systems 37Class 9 Intermediate Financial Accounting 45Class 10 Intermediate Financial Accounting 34of students enrolled in each of them. To ensure independent sample observati<strong>on</strong>s, we excludedsec<strong>on</strong>dary test scores from the few students who were simultaneously or c<strong>on</strong>secutively enrolled intwo of these classes.Dependent and Independent VariablesThe dependent variable for the study was the (percentage) score <strong>on</strong> the CR porti<strong>on</strong> of eachclass examinati<strong>on</strong>. From the standpoint of this investigati<strong>on</strong>, the most important independentvariable was the student’s score <strong>on</strong> the MC porti<strong>on</strong> of the examinati<strong>on</strong>. As illustrated in Figure 2,the CR <str<strong>on</strong>g>questi<strong>on</strong>s</str<strong>on</strong>g> required students to create journal entries or perform similar tasks. To ensurec<strong>on</strong>sistency in grading CR answers, the same instructor manually graded all the CR <str<strong>on</strong>g>questi<strong>on</strong>s</str<strong>on</strong>g> <strong>on</strong> allexaminati<strong>on</strong>s for a given class, indicating the correct answer(s) to each CR questi<strong>on</strong> as well as a listof numerical penalties to assess for comm<strong>on</strong> errors.Another important independent variable <str<strong>on</strong>g>use</str<strong>on</strong>g>d in this study was “gender.” As noted above,a number of prior studies have detected a relati<strong>on</strong>ship between “gender” and “computer-relatedoutcomes” (Hamilt<strong>on</strong>, 1999; Gutek and Biks<strong>on</strong>, 1985). Accordingly, “gender” was included in thelinear regressi<strong>on</strong> model as a dummy variable, using “1” for females and “0” for males. Finally, theexaminati<strong>on</strong>s in each class had different, and differently-worded, <str<strong>on</strong>g>questi<strong>on</strong>s</str<strong>on</strong>g>, and the tests were takenby different students. Accordingly, we added dummy (0-1) variables to the regressi<strong>on</strong> equati<strong>on</strong> foreach class to account for these differences.In summary, the regressi<strong>on</strong> equati<strong>on</strong> tested in this study <str<strong>on</strong>g>use</str<strong>on</strong>g>d student performance <strong>on</strong> the CRporti<strong>on</strong> of a final or midterm examinati<strong>on</strong> as the dependent variable, and student performance <strong>on</strong>the MC porti<strong>on</strong> of the exam, gender, and semester dummy variables as independent variables. Table3 lists these independent variables and provides brief descripti<strong>on</strong>s of them.ResultsTable 4 displays the results from the regressi<strong>on</strong> analysis described above. Coefficient values,t-statistics, and probabilities are <strong>on</strong>ly shown for nine of the ten classes beca<str<strong>on</strong>g>use</str<strong>on</strong>g> we <str<strong>on</strong>g>use</str<strong>on</strong>g>d 0-1 dummy
<str<strong>on</strong>g>Why</str<strong>on</strong>g> Use Multiple Choice Questi<strong>on</strong>s 37TABLE 3Independent Variables for the Regressi<strong>on</strong> ModelIndependent VariableMCGenderClass (1, 2, etc.)Descripti<strong>on</strong>The (percentage) score <strong>on</strong> the <str<strong>on</strong>g>multiple</str<strong>on</strong>g>-<str<strong>on</strong>g>choice</str<strong>on</strong>g> porti<strong>on</strong> of anexaminati<strong>on</strong>A dummy variable: 0 = male, 1 = femaleA dummy variable (0-1) to account for the differentexaminati<strong>on</strong> <str<strong>on</strong>g>questi<strong>on</strong>s</str<strong>on</strong>g> given to each classTABLE 4Linear Regressi<strong>on</strong> ResultsCoefficient (â) t-statistic ProbabilityIntercept 47.11 13.89