1. What's next in experimental design?
We've reached the end of this course, but there's still much more to consider with respect to designing experiments.
2. What's next?
The most immediate example is that there are a huge number of other factorial experiment types out there. We only looked at 2^k factorial experiments, when in fact you can design factorial experiments that include factor variables with more than 2 levels. Of course, all of the combining involved in a lot of factors and factor levels means you might also start to consider a fractional factorial design, where you only consider some fraction of the number of possible combinations.
You could continue to expand your knowledge by designing experiments with random factors or could utilize nested or split plot designs, or even a lattice design. There are many other types of designs and plenty of classical texts and other resources if you're interested in deepening your knowledge.
What's most important is that design is a valued and integrated part of the experimental process. From A/B tests to studies on people's health, education, or loan fitness, as we considered here, and even beyond, the steps of an experiment should be thought out, the outcome and how to measure it considered, and the possible factors affecting the outcome thought through.
In reality, there's always going to be something you'll miss, some unmeasured confounder that introduces noise into your model. Hopefully, what you've picked up throughout this course are some ideas to help reduce that noise and begin to create clear experiments built on the principles of reproducibility, randomization, and replication.
3. Go forth and design experiments!
Now it's your turn. Go forth in the world and design experiments!