Ninth Edition CoverGraziano & Raulin
Research Methods (9th edition)

Chapter 10 Fill-in-the-Blank Questions
Single-Variable, Independent-Groups Designs

Now that you have familiarity with the new key terms of this chapter, it is time to begin using the terms in the context of the chapter material. See if you can fill-in-the-blanks in the following discussions. To get the correct answer, click on the blank.

  1. The term "experimental design" refers to both the __________ involved in the detailed planning of the experiment and the __________ --the detailed plan itself. The experimental design is a blueprint that should list all aspects of the study from beginning to end.

  2. The concept of __________ (the variability among scores) is critical in understanding experimentation. There must be variation if we are to study relationships among variables. However, we want to avoid unwanted or __________ variance, since it can threaten the validity of the study. We make the distinction between systematic, __________- __________ variance, and nonsystematic, __________-__________ variance. The systematic, between-groups variance can result from the effects of the independent variable ( __________ variance) or may be the result of uncontrolled confounding variables ( __________ variance). In testing research hypotheses we seek to maximize the __________ variance and control the __________ variance. The term __________ variance is often used to describe the nonsystematic, within-groups variability. The largest single contributor to error variance is the __________-__________ among participants. These are stable differences between participants on variables such as coordination, personality, motivation, and so on. Error variance can also be increased by __________ factors that affect some participants but not others and may change over time.

  3. In order to analyze the results of a study, we look at an F-ratio. This is literally a ratio of two __________. The numerator is a measure based on the __________-__________ variation, while the denominator is a measure based on the __________-__________ variation. Because of the way the F-ratio is set up, the numerator reflects both the __________ effects (experimental variance plus extraneous variance) and the effects of sampling error ( __________ variance). The denominator of the F-ratio is a function of __________ variance only. If there are no __________ effects, we would expect that the F- ratio would be approximately equal to 1.00.

  4. A general principle in experimentation is that we want to __________ experimental variance, control __________ variance, and minimize __________ variance. We can maximize experimental variance by making sure that the __________ variable really does vary. We may wish to include a __________-__________ to verify that the levels of the independent variable really are different from one another. We control __________ variance by making sure that the __________ variable manipulation is the only difference between experimental and control groups. This might be accomplished by __________ assigning participants to groups or conditions or by using some other procedure designed to equate the groups at the beginning of the study. Error variance can be minimized by using measuring instruments that are __________ and carefully controlling the conditions of measurement. The individual differences component of error variance can be controlled by using a __________-__________ design or by some type of matching procedure.

  5. To understand the advantages of experimental designs it is important to look at some of the weaknesses of __________ designs. In Chapter 10, four nonexperimental designs were discussed. The __________-__________- __________ ("after the fact") design involves observing participants' behavior and drawing conclusions about what might have caused the participants' behavior. There is no __________ of variables. The ex post facto design has almost no __________ over confounding factors. The single-group, posttest-only design includes an actual __________ of the variable. An independent variable is manipulated, and a single posttest measure of the dependent variable is taken. Although the manipulation of the independent variable provides a slight increment in control over the ex post facto design, this design still provides very little control over __________. The single-group, __________-__________ design provides slightly more control by adding a pretest measure. In this design we can be sure that there really was a __________ between pretest and posttest measures--whereas, in the two previous designs, change had to be inferred. However, this design still controls for very few sources of confounding. We know that there was change, but we do not know what __________ the change. For example, the observed change may be the result of normal developmental processes ( __________ ) or may be the result of other factors that operated during the time from pretest to posttest ( __________ ). Because these sources of confounding are not controlled by the single-group, pretest-posttest design, we cannot conclude that the independent variable was responsible for the observed change in the __________ variable. The last __________ research design discussed in Chapter 10 is the pretest-posttest, natural control-group design. This design is very close to an __________. The difference between this design and an experimental design is that participants are not assigned __________ to the experimental and control groups and, consequently, we cannot be confident that the groups are comparable at the start of the study. However, the inclusion of a __________-__________--even if participants are not randomly assigned to groups--does provide more control over confounding than in the previous designs.

  6. Experimental designs maximize the control of __________. The factors that distinguish experiments from nonexperimental design are the inclusion of __________ groups or conditions and __________ assignment of participants to groups or conditions. In Chapter 10, three basic experimental designs are described. The randomized, posttest-only, control-group design employs two __________ with random assignment of participants to the groups. The groups differ on the level of the __________ variable. A single measure of the __________ variable is taken after the independent variable manipulation. In this design, although we cannot be sure that there was a change from pretest to posttest, we can assume that the groups were comparable at the __________ of the study. Therefore, any observed __________ at posttest should be the result of the __________ variable manipulation. The randomized, __________-__________, control-group design adds a pretest to the previous design. participants are again __________ assigned to the groups with a measure taken both before and after the __________ variable manipulation. This design allows us to determine if the independent variable had an effect on the dependent variable and what effect it might have had. By comparing the __________ and __________, we can see what change occurred as a result of the independent variable. The multilevel, completely randomized, between-subjects design is an extension of the two previous designs. Instead of there being just __________ groups, this design includes several groups. A pretest may or may not be included in the design. Since this design is a simple extension of the two previous designs, its ability to __________ sources of confounding is similar to those designs. 

  7. There are a variety of statistical procedures available for evaluating data. The appropriate statistical procedure depends largely on the level of measurement of the __________ variable: The chi-square is typically used for __________ data; a Mann-Whitney U-test is typically used for ordered data; a ____-__________ or an analyses of variance ( __________ ) are the procedures typically used for score data. The most widely used procedure is the ANOVA. Unlike the t-test, the __________-___- __________ is flexible enough to handle any number of groups. In the ANOVA procedure, the total sum of squares is __________ into the between-groups sum of squares and the __________- __________ sum of squares. Variance estimates are computed by dividing the sums of squares by the appropriate __________-__________- __________ (df). The test statistic-- __________ --is computed by dividing the between-groups variance estimate by the __________-__________ variance estimate. In analysis of variance, the variance estimates are referred to as __________-__________. The results of an analysis of variance are organized for presentation in an ANOVA __________- __________. If there are no __________ differences between groups, the F-ratio should be approximately equal to 1.00. To the extent that there are systematic differences, the F-ratio will be __________ than 1.00. However, a significant F in an ANOVA indicates only that at least __________ group is different from at least one other group. To find out which groups are __________ different from each other, we need to do __________-__________ comparisons. If the specific comparisons are planned as part of the research, we refer to them as __________ comparisons. If the comparisons were not planned in advance, they are referred to as __________- __________ comparisons.

  8. Other experimental designs are available. We make the distinction between __________ -groups designs, which have different participants in each group, and __________ -groups designs, which have the same or closely matched participants in the different groups. We also make the distinction between single-variable designs, in which only one __________ variable is manipulated, and multivariable or __________ studies, in which more than one independent variable is manipulated.