24. Design of Experiments Flashcards
Experimental Design
A formal plan that details for conducting an experiment
Effect
The relationship between a factor and a response variable.
Response variable
The output variable that shows the observed results or value of an experimental treatment.
Observed value
A particular value of a response variable
Factor
An Independent variable or assignable cause
Noise Factor
An independent variable that is difficult or too expensive to control
Level
The setting or assignment of a factor at a specific value.
Design Space
The multidimensional region of possible treatment combinations formed by the selected factors and their levels.
Experimental Error
The variation that occurs in the response variable
Experimental Unit
The smallest entity receiving a particular treatment.
Treatment
Specifit setting or combination of factor levels.
Experimental Run
A single performance of the experiment
Factorial Design
Each complete trial or replication of the experiment all possible combinations of the levels of the factors are investigated.
Nested Design
Levels of one factor are nested under the levels of another factor.
Power and Sample Size
Power increases as the sample size increases (probability of a type II error, beta, decreases).
Repetition
The measurement of a response variable more than once under similar conditions.
Replication
The entire experiment is performed more than once for a given set of independent variables.
Replication reflects the sources of variability both within and between runs and adds DF.
Confounding (Aliasing)
Occurs when factors or interactions are not distinguishable from one another.
Order
The sequence in which the runs of an experiment will be conducted.
- Standard order
- Run order (random)
Blocking
A collection of experimental units more homogeneous than the full set of experimental units.
Block effect generally means that the method of blocking was appropriate and that an assignable cause has been found.
Main Effect
The impact or influence of a single factor on the mean of the response variable.
Factors with the greatest difference between + and - results have the greatest impact
Interaction Effect
When the influence of one factor on the response variable depends on one or more other factors
Balanced Design
All treatment combinations have the same number of observations
Resolution
The level of confounding in a fractional factorial design. Generally one more than the smallest-order interaction with which a main effect is confounded
Source Table
A computational table used to analyze an experimental design.
