Continuation of modifying article 1

Research Designs

edit

'One-Way designs

The simplest experimental design is a one-way design. In this type of design, there is one and only one independent variable. Furthermore, the simplest kind of one-way design is called two-group design. In a two-group design, there is only one independent variable and this variable has two levels. A two-group design mainly consists of an experimental group (a group that receives treatment) and a control group (a group that doesn’t receive treatment)[1]. In addition to two group designs, experimenters often make use of another kind of one-way design called the one-way, multiple groups design. This is another design in which there is only a single independent variable, but the independent variable takes on three or more levels[2]. This type of design is useful in studies such as those that measure perception. Although these types of designs may be simple, they do have limitations.

Factorial Designs

One major limitation of one-way designs is the fact that they allow researchers to look at only one independent variable at a time. The problem is that a great deal of human behavior is comprised of multiple variables acting together. Because of this, R.A Fisher popularized the use of factorial designs. Factorial designs are designs that contain two or more independent variables that are completely crossed. This means that every level of the independent variable appears in combination with every level of every other independent variable. There are a broad variety of factorial designs, so researchers have specific descriptions for the different designs. The label given to a factorial design specifies how many independent variables exist in the design and how many levels of each independent variable exist in the design. Therefore a 2x3 factorial design has two independent variables (because there are two numbers in the description), the first of which has two levels and the second having three levels.

Main Effects and Interactions

The simple straightforward effects of independent variables in factorial studies are referred to as main effects. Main effects are the factorial equivalent of the only kind of effect that you can detect in a one-way design. This refers to the overall effect of an independent variable, averaging across all levels of the other independent variables[3]. Main effects are simple. They only have to do with one variable. In addition to providing information about main effects, studies can also produce a second, very important kind of information called interactions. Interactions exist when the effect of one independent variable on a dependent variable depends on the level of a second independent variable.

Within-Subjects Designs

The two basic approaches to research design include between-subjects design and within-subjects design. Between-subjects designs are designs in which each participant serves in one and only one condition of an experiment. In contrast, within-subjects or repeated measures designs are those in which each participant serves in more than one or perhaps all of the conditions of a study[4]. Within-subjects have some huge advantages over between-subjects designs especially when it comes to complex factorial designs that have many conditions. Within-subjects designs eliminate person confounds. When researchers use this type of design, they eliminate person confounds in a much more direct approach. They ask the same people to serve in the different experimental conditions in which they happen to be interested. In a sense, these designs take advantage of the only perfect form of matching and in doing so, they totally eliminate person confounds. While there are advantages to this type of design, there are disadvantages as well. There are three closely related biases that are applicable to within-subjects designs. The first bias has to do with the fact that people’s psychological states change as they spend time working on one or more tasks. More specifically, sequence effects can pose serious problems. Sequence effects occur when the simple passage of time begins to take its toll on people’s responses. A second closely related problem has to do with carry-over effects. Carry-over effects occur when people’s responses to one stimulus in a study directly influence their responses to a second stimulus[5]. Another kind of carry-over effect can occur when participants knowingly or unknowingly learn something by performing an experimental task. When a participants’ experience with one task makes it easier for them to perform a different task that comes along later, they have benefited from practice effects. This is a problem because researchers cannot tell if people’s superior performance on the second task happened because of an experimental manipulation or because of simple practice.

References

edit
  1. ^ Kline, R. B. (2004). Effect Size Estimation in Multifactor Designs. In , Beyond significance testing: Reforming data analysis methods in behavioral research (pp. 203-243). Washington, DC US: American Psychological Association. doi:10.1037/10693-007
  2. ^ Kline, R. B. (2004). Effect Size Estimation in One-Way Designs. In , Beyond significance testing: Reforming data analysis methods in behavioral research (pp. 163-202). Washington, DC US: American Psychological Association. doi:10.1037/10693-006
  3. ^ Xu, L., Yang, F., Abula, A., & Qin, S. (2013). A parametric bootstrap approach for two-way ANOVA in presence of possible interactions with unequal variances. Journal Of Multivariate Analysis, 115172-180. doi:10.1016/j.jmva.2012.10.008
  4. ^ Charness, G., Gneezy, U., & Kuhn, M. A. (2012). Experimental methods: Between-subject and within-subject design. Journal Of Economic Behavior & Organization, 81(1), 1-8. doi:10.1016/j.jebo.2011.08.009
  5. ^ Brooks, J. L. (2012). Counterbalancing for serial order carryover effects in experimental condition orders. Psychological Methods, 17(4), 600-614. doi:10.1037/a0029310