Chapter 9
Experimental Research

(Reminder: Don’t forget to utilize the concept maps and study questions as you study this and the other chapters.)

 

In this chapter we talk about what experiments are, we talk about how to control for extraneous variables, and we talk about two sets of experimental designs (weak designs and strong designs).

 

(Note: In the next chapter we will talk about middle of the road experimental designs; they are better than the weak designs discussed in this chapter, and they are not as good as the strong designs discussed in this chapter. The middle of the road, or medium quality designs are called quasi-experimental designs.)

 

It is important for you to remember that whenever an experimental research study is conducted the researcher's interest is always in determining cause and effect

 

 

The Experiment

Here is our definition of an experiment: The experiment is a situation in which a researcher objectively observes phenomena which are made to occur in a strictly controlled situation where one or more variables are varied and the others are kept constant.

 

 

 

Independent Variable Manipulation

The independent variable is the variable that is assumed to be the cause of the effect.  It is the variable that the researcher varies or manipulates in a specific way in order to learn its impact on the outcome variable.

 

Ways of Manipulating the Independent Variable
In Figure 9.1 (on page 266) you can see three different ways to manipulate the independent variable. Here is that figure reproduced for your convenience:

 

 

 

 

 

Control of Confounding Variables

Potential confounding variables can be controlled for by using of one or more of a variety of techniques that eliminate the differential influence an extraneous variable may have for the comparison groups in a research study.

 

Remember this important point: You want all of your comparison groups to be similar to each other (on all characteristics or variables) at the start of an experiment. Then, after manipulating the independent variable you will be better able to attribute the difference observed at the posttest to the independent variable because one group got a treatment and the other group did not.


Now we will discuss these six techniques that are used to control for confounding variables: random assignment, matching, holding the extraneous variable constant, building the extraneous variable into the research design, counterbalancing, and analysis of covariance.

 

Random Assignment
Random assignment is the most important technique that can be used to control confounding variables because it has the ability to control for both known and unknown confounding extraneous variables.  Because of this characteristic, you should randomly assign whenever and wherever possible.


 

You must be careful not to confuse random assignment with random selection! The two techniques differ in purpose. (Note: I strongly recommend that you re-read the section titled Random Selection and Ransom Assignment on pages 216-217; it is only three paragraphs long, but will help you with this very important distinction!)

·        The purpose of random selection is to generate a sample that represents a larger population. This topic was covered in our earlier chapter on Sampling (Chapter 7).

·        The purpose of random assignment is to take a sample (usually a convenience sample) and use the process of randomization to divide it into two or more groups that represent each other. That is, you use random assignment to create probabilistically “equivalent” groups.

·        Note that random selection (randomly selecting a sample from a population) helps ensure external validity, and  random assignment (randomly dividing a set of people into multiple groups) helps ensure internal validity.

·        Because the primary goal is experimental research is to establish firm evidence of cause and effect, random assignment is more important than random selection in experimental research. It that is counterintuitive to you, then please reread it as many times as is necessary.
 

Random assignment controls for the problem of differential influence (that was discussed earlier). It does they by insuring that each participant has an equal chance of being assigned to each comparison group.

 

Here is one way to carry out random assignment that we included in the first edition of our textbook:

 

 

 

 

Another way to conduct random assignment is to assign each person in your sample a number and then use a random assignment computer program. Here is one: http://www.graphpad.com/quickcalcs/randomize1.cfm

 


Matching
Matching controls for confounding extraneous variables by equating the comparison groups on one or more variables that are correlated with the dependent variable.

 

Holding the Extraneous Variable Constant

This technique controls for confounding extraneous variables by insuring that the participants in the different treatment groups have the same amount or type on a variable. 


Building the Extraneous Variable into the Research Design

This technique takes a confounding extraneous variable and makes it an additional independent variable in your research study.

·        For example, you might decide to include females and males in your research study.

·        This technique is especially useful when you want to study any effect that the  potentially confounding extraneous variable might have (i.e., you will be able to study the effect of your original independent variable as well as the additional variable(s) that you built into your design.


Counterbalancing

Counterbalancing is a technique used to control for sequencing effects (the two sequencing effects are order effects and carry-over effects).


Analysis of Covariance
Analysis of covariance (ANCOVA) is a statistical control technique that is used to statistically equate groups that differ on a pretest or some other variable.  

 

Experimental Research Designs

A research design is the outline, plan, or strategy that you are going to use to obtain an answer to your research question.  Research designs can be weak or strong (or quasi which are moderately strong; that is, in between the weak and the strong designs) depending on the extent to which they control for the influence of confounding variables.

 

Weak Experimental Research Designs
Some research designs are considered weak because they do not control for the influence of many confounding variables.

 


 

The one-group posttest-only design is a very weak research design where one group of research participants receives an experimental treatment and is then post tested on the dependent variable.

Because of the problems with this design it generally gives little evidence as to the effect of the treatment condition.

 

The next design is the one-group pretest-posttest design. Here is a depiction of it:

 

 

The next of the weak experimental research designs is the posttest-only design with nonequivalent groups.

 

 

 

For a summary of the threats to validity for the weak experimental designs, you should study Table 9.1 on page 277.

 


Strong Experimental Research Designs
A research design is considered to be a "strong research design" if it controls for the influence of confounding extraneous variables.  This is typically accomplished by including one or more control techniques into the research design. 

 

The first strong experimental design is the pretest-posttest control-group design. Here is a picture of it in its basic form:

 

 

 

The next strong experimental research design is the posttest-only control group design. Here is a picture of it:

 


 

The posttest-only control group design is a research design in which the research participants are randomly assigned to an experimental and control group and then post tested on the dependent variable after the experimental group has received the experimental treatment condition.

 

The next strong experimental research design is the factorial design. For a depiction of this design, please go to page 281 and look at it in Table 9.2. 

 

The layout for a factorial design with two independent variables (Type of instruction and level of anxiety) is shown in Figure 9.14 (p.287) and here for your convenience.

 

 

 

The next strong experimental research design is the repeated-measures design. Here is a picture of it in its basic form with counterbalancing:

 

 

The last strong experimental research design discussed in this chapter is the factorial design based on a mixed model. Here is a picture of this design when it has two independent variables:

 

 

 

As you study the designs in this chapter, two tables will be of maximum help.

 

Here are copies of these two tables for your convenience.