The factorial experimental design is a test whose design encompasses of at least two factors, each with discrete likely values or levels and whose experimental units take on all conceivable combinations of these levels over every such factor. A full factorial design may likewise be known as a fully crossed design. The experiment enables the effects of interactions between factors on the response variable and the investigator to learn the consequence of each factor on the reaction variable.
For the majority of factorial experimental designs, each factor has just two levels. An illustration; with two factors each taking two levels, a factorial experiment would have four treatment combinations totally, and is typically called a 2×2 factorial design.
On the off chance that the number of combination in a full factorial design is too high to be logistically feasible, a fractional factorial plan might be done, in which a portion of the possible combinations (normally at least half) are omitted.
Factorial experiments permit subtle manipulations of a larger number of interdependent variables. While the technique has limitations, it is a useful method for streamlining exploration and giving important, measurable statistical methods highlight any correlations.
Two Basic Levels of Factorial Experimental Design
- Full factorial: includes no less than one trial for every possible combination of factors and levels.
- Partial or fractional factorial: includes no less than one trial for a few, but not every, possible combination of factors and levels.
Which one you pick relies upon your particular experiments, the data being gathered, and the budgetary/time limitations.
In a full factorial experimental design, no less than one trial is included for every single possible combination of factors and levels. This broad approach makes it impossible for any communications to be missed as all factor interactions are accounted for. The careful quality of this approach, in any case, makes it very costly and tedious for experiments with various factors– and this increases exponentially with the number of factors and levels.
Since full factorial design experiments are frequently time-and cost-prohibitive when various treatment factors are included, many individuals utilize partial or fractional factorial experimental designs. These designs assess just a subset of the possible permutation of factors and levels. For the most part, a fractional factorial experimental design resembles a full factorial plan for fewer factors, with additional factor columns included (however no additional rows).
Utilizing partial factorial experimental design makes experiments less expensive and quicker to run, yet can likewise muddle interactions between factors.
Steps of Factorial Experimental Design
- Define the problem
- Determine the objective
- Design the experiment
- Conduct the experiments and collect the data
- Analyze the data
- Interpret the outcomes
- Verify the predicted outcomes
These steps don’t guarantee a successful experiment. However, they do constrain the experimenter to continue logically. All experiments conducted in this way give helpful information, although some may require a second experiment to accomplish the desired improvement. Designing an effective trial regularly requires a team of people comfortable with the procedure or the design.
The Pros and Cons of Factorial Experimental Design
- Factorial designs are to a great degree helpful to clinicians and field scientist as a preliminary study, enabling them to judge whether there is a link between variables while decreasing the possibility of experimental error and complex variables.
- The factorial plan, as well as simplifying the process and making research less expensive, permits many levels of analysis. And featuring the relationships between variables.
- It allows the effect of manipulating a single variable to be isolated and analyzed independently.
- The main demerit is the trouble of experimenting with more than two variables or many levels.
- The factorial experimental design has to be arranged carefully, as an error in one of the levels, or in the general operationalization, will imperil a great amount of work.
Other than these minor derogations, a factorial experimental design is a backbone of abundant scientific disciplines, delivering awesome outcomes in the field.
The aim of the experiment design method is to give required information cost-effectively. Likewise, factors are identified which don’t affect the quality of the process or product yet do give extra cost savings. Reproducible outcomes are the key strength of this method. These procedures can enhance quality without causing capital and material cost increments.