Page 99 -

4.5 Analysis of variance 85

When using SPSS to run the analysis, the data need to be carefully arranged to
avoid potential errors. The data points contributed by the same participant need to
be listed in the same row. It is recommended that you repeat the same pattern when
arranging the columns (see Table 4.14).

Table 4.14 Sample Data for Two-Way, Repeated Measures ANOVA Test

Transcription Composition
Standard Prediction Speech Standard Prediction Speech
Participant 1 245 246 178 256 265 189
Participant 2 236 213 289 269 232 321
Participant 3 321 265 222 333 254 202
Participant 4 212 189 189 246 199 198
Participant 5 267 201 245 259 194 278
Participant 6 334 197 311 357 221 341
Participant 7 287 289 267 301 302 279
Participant 8 259 224 197 278 243 229

Table 4.15 presents the simplified summary table for the two-way, repeated mea-
sures ANOVA test. The task type has a significant impact on the time spent to com-
plete the task (F(1, 7) = 14.217, p < 0.01). There is no significant difference among
the three text entry methods (F(2, 14) = 2.923, n.s.). The interaction effect between
the two independent variables is not significant either (F(2, 14) = 0.759, n.s.).

Table 4.15 Result of the Two-Way, Repeated Measures ANOVA Test

Sum of
Source Square df Mean Square F Significance
Task type 2745.187 1 2745.187 14.217 0.007
Error (task type) 1351.646 7 193.092
Entry method 17,564.625 2 8782.313 2.923 0.087
Error (entry 42,067.708 14 3004.836
method)
Task type*entry 114.875 2 57.438 0.759 0.486
method
Error (task 1058.792 14 75.628
type*entry method)

4.5.4 ANOVA FOR SPLIT-PLOT DESIGN
Sometimes you may choose a study design that involves both between-group fac-
tors and within-group factors. In the text entry study, you may recruit two groups of
participants. One group completes transcription tasks using all three data-entry
methods. The other group completes composition tasks using all three data-entry
methods (see Table 4.16). In this case, the type of task is a between-group factor and