15.9
Factorial Design
If a researcher is curious about a topic, such as people’s food preferences, they may consider a factorial analysis design—an experimental approach used when examining the effects of more than one factor, at least two independent variables, on one dependent variable.
For example, they can decide to manipulate two independent variables: food category and food temperature. Each factor consists of two levels: the ice cream and soup are served hot or cold. This case is referred to as a 2x2 factorial design, with favorability ratings operating as the single dependent variable.
As a result, the researcher can test two types of hypotheses. One type predicts the main effects, which assess the influence of conditions across each factor separately. For instance, to examine the main effect for food category, the favorability ratings of ice cream would be compared against soup.
Similarly, observing participants’ preferences between these foods when served warm versus cold relates to the main effect for temperature.
The other type of hypothesis—which is a major advantage of the approach—involves the assessment of interaction effects. These are observed when the effect of a factor is dependent on the level of the other factors in the experimental model.
For example, the researcher would predict that people prefer warm soup and cold ice cream.
Consequently, this procedure offers a more efficient and cost-effective means to test several combinations of two or more conditions at the same time. However, the application of the design may prove more challenging as the number of levels and factors increases.
Accordingly, researchers must take certain precautions both in terms of methodology and statistical analyses when interpreting complex experimental designs.
Factorial Analysis is an experimental design that applies Analysis of Variance (ANOVA) statistical procedures to examine a change in a dependent variable due to more than one independent variable, also known as factors. Changes in worker productivity can be reasoned, for example, to be influenced by salary and other conditions, such as skill level. One way to test this hypothesis is by categorizing salary into three levels (low, moderate, and high) and skills sets into two levels (entry level vs. experienced).
Naming
This scientific approach is designated a label that either underscores the number of factors or the number of conditions tested for each independent variable. The example experiment above would be described as a two-way factorial ANOVA, because it involves two independent variables. In respect to the number of levels considered for salary (low, moderate, and high) and skill sets (entry level and experienced), this same experiment is also designated as a 3 by 2 design, formally written as a 3 x 2 Factorial ANOVA. Of note, computing the product of 3 and 2 signifies that there is a total of 6 combinations of experimental conditions observed.
Investigative Advantages
Observing the effects of at least two independent variables is a more practical and economical approach. This averts the need to expend time and resources for separate experiments. Furthermore, collecting data for different combinations of conditions enables researchers to make a variety of assessments, including main and interaction effects.
Certain research questions may require understanding know how each factor may independently impact a dependent variable. For example, observed changes in worker productivity scores due to salary are separated from those due to skill level, to help determine the main effects for each. Results could potentially reveal that high productivity found in entry level employees may or may not apply to those who are more experienced. Likewise, low productivity that may be found in low salaried employees may or may not be evident with increased wages. The ability to recognize if results can be generalized to different circumstances or group features thus serves as another advantage for this type of design.
An interaction effect occurs when the influence of an independent variable on a given dependent variable depends on the level of other factors being examined. It might be found, for instance, that the impact of salary on worker productivity may be more pronounced for entry level as opposed to experienced employees. This type of analysis allows researchers to gain a deeper view of patterns that may emerge in the data set.
Implications of Factorial Analysis
Due to its flexibility and practicality, factorial analysis continues to be one of the most common experimental designs used across all disciplines. A recent study, for instance, investigated whether consumer behavior may depend on whether the product is utilitarian or hedonic. In addition to product type, researchers also included product image (close-up vs. wide shot) as a potential factor that influenced purchase decisions. These researchers further examined if type of persuasive technique, such as rational or emotional appeal would have an impact (Kim, Lee, & Choi, 2019). Hedonic products tended to gain more favorable attitudes when a wide shot image was accompanied by emotion inducing advertisements. Inverse findings were observed for utilitarian products in this 2 x 2 x 2 Factorial ANOVA, otherwise known as a three-way ANOVA.
If a researcher is curious about a topic, such as people’s food preferences, they may consider a factorial analysis design—an experimental approach used when examining the effects of more than one factor, at least two independent variables, on one dependent variable.
For example, they can decide to manipulate two independent variables: food category and food temperature. Each factor consists of two levels: the ice cream and soup are served hot or cold. This case is referred to as a 2x2 factorial design, with favorability ratings operating as the single dependent variable.
As a result, the researcher can test two types of hypotheses. One type predicts the main effects, which assess the influence of conditions across each factor separately. For instance, to examine the main effect for food category, the favorability ratings of ice cream would be compared against soup.
Similarly, observing participants’ preferences between these foods when served warm versus cold relates to the main effect for temperature.
The other type of hypothesis—which is a major advantage of the approach—involves the assessment of interaction effects. These are observed when the effect of a factor is dependent on the level of the other factors in the experimental model.
For example, the researcher would predict that people prefer warm soup and cold ice cream.
Consequently, this procedure offers a more efficient and cost-effective means to test several combinations of two or more conditions at the same time. However, the application of the design may prove more challenging as the number of levels and factors increases.
Accordingly, researchers must take certain precautions both in terms of methodology and statistical analyses when interpreting complex experimental designs.
From Chapter 15:
Now Playing
15.9 : Factorial Design
Research Methods
13.8K Views
15.1 : The Scientific Method
Research Methods
57.8K Views
15.2 : Case Studies
Research Methods
11.4K Views
15.3 : Naturalistic Observations
Research Methods
15.4K Views
15.4 : Surveys
Research Methods
14.5K Views
15.5 : Archival Research
Research Methods
15.6K Views
15.6 : Longitudinal Research
Research Methods
11.8K Views
15.7 : Cross-Sectional Research
Research Methods
11.6K Views
15.8 : Group Design
Research Methods
9.4K Views
15.10 : The Placebo Effect
Research Methods
6.0K Views
15.11 : Blind Procedures
Research Methods
10.9K Views
15.12 : Ethics in Research
Research Methods
23.9K Views
15.13 : Correlations
Research Methods
34.9K Views
15.14 : Cause and Effect
Research Methods
10.6K Views
15.15 : Reliability and Validity
Research Methods
13.0K Views
See More