(1a) ratio. (1b) ordinal. (1c) nominal. (1d) ratio. (1e) interval. (1f) nominal. (1g) ordinal (2a) ordinal. (2b) ratio. (2c) nominal. (2d) interval. (3) Independent variable = type of problem display, type independent variable, bivalent independent variable, manipulated between subjects. (4) With a between-subjects design, the primary concern is with group differences. Thus, one might be concerned about group differences for interest in math and ability (which can also affect interest). If group differences exist in math interest before the study, they cloud the test of the independent variable effect on the dependent variable. (5) The external validity can be improved with more realistic math problems and classroom situations. A field experiment can be done by designing different texts that present math problems in the different formats and by asking students to use them in the classroom. Tests can also be designed for classroom use with the different presentation formats. (6) regression toward the mean. (7) double-blind. (8) the Hawthorne effect. (9) external. (10) Attrition occurs when subjects drop out of a study before it is completed. The problem with attrition is that the subjects who drop out may be characteristically different from subjects who remain in the study. This can result in data that only apply to certain members of the group being studied, which can limit the conclusions a researcher can draw from the results of the study. See Chapter 6 for more discussion of bias in subject samples.