Table of experimenter vs treatment
Let's say you are running an experiment to see how light affects plant growth. You will expose plants to either light or dark environments, and measure their growth after a week. You have three volunteers, Ali, Beth and Cy who will each manage a pair of plants. So six plants in all. Back to overview...
Entering your results like this:
This doesn't have the disadvantages of the table of averages - it's good that all sample measurements are listed individually.
However, this table format does not clearly represent the concept of samples and variables. Each sample (one plant) can be considered to have two attributes (or variable values):
- The experimenter (who did the experiment)
- The treatment (light or dark)
But the concept of "Treatment" has been spread over two columns.
A better way to do it is described here, where you clearly have one sample (or measurement) per row, and each column represents a (named) variable.
This is also called "Tidy Data". You can read more about Tidy Data here.