For factorial designs I usually see the independent variables being tested in levels (e.g., high/low). In my experiment, though, the effects of two different independent variables can only be tested this way: either present or absent. I am about to program the following conditions: IV1 absent and IV2 present; IV1 absent and IV2 absent; IV1 present and IV2 present; IV1 present and IV2 absent. Would this be appropriate?
Hi Pedro! Yes, this method is suitable for investigating your system. However, using a binary approach ("present" and "absent") will limit your analysis to estimating main effects, not graphical response surfaces. I recommend planning your variables quantitatively, such as "0 to X ºC", "0 to Y mol/L", etc. This approach allows for the inclusion of central points in your design, facilitating error estimation in statistical analysis and enabling the use of response surfaces. Note that this quantitative approach also incorporates the levels "absent" and "present" by passing through an intermediary level. I hope my response helps you. Best regards.
- Badges
- Science topic
- Similar topics
- Papers