Reading from text, Chapter 15, second part
Chapter 15: 15.6 - 15.7
- The utility of and rationale for fractional incomplete designs from 2k
- Construction of ... : One way to describe how to make a good fractional design
- Aliases ... : We say hypotheses are confounded / effects are aliased. Generating
- How aliases ... : Using the generating function to determine aliases
- Formal ... : Choose your generating function such that the main effects are aliased
with higher-order interactions and not other main effects. Start with the full design and
cut it down to fewer conditions, e.g. 4 factors using an eight-cell design.
Example: ABCD = I.
- Construction ... : This time, we add factors onto a smaller design, e.g. D & E
added to ABC. Still, we want the main effects to be aliased with higher-order
- Finding the SSs and constructing the ANOVA table. Need to pool some effects to
make the error term because of "S" instead of "G".