What are levels and treatments?
In experimental research, the values of the independent variables you will study are called levels.
You will be the one to choose which values (and how many) you will use for your independent variables.
Tips for choosing values for levels:
1. Follow the levels in a reference. If you have reason to believe that the levels used in a reference paper are effective in your own research, you may use them. Make sure to put a citation in your methodology.
2. Choose a high value and low value. If your research is exploratory, or the type where you can't find a good reference for levels, it's easier to set a low value: usually, this is the default value or something close to it.
If you need levels for temperature, choose the low value 30°C because this is near room temperature meaning it won't be difficult to adjust the temperature to this value. Afterwards, choose a high value such as 60°C. If you need a third level, get the middle value: 45°C. You can continue to take middle values if you want more levels.
3. Choose levels that are not very near each other. If you choose levels, try to space them out as far as possible. There is a possibility that the changes in the dependent variable between two close levels (say, 50°C and 60°C) are so small that you cannot detect their effect.
4. Don't study a lot of levels if not necessary. It's possible to only choose two or three levels in your research. Always remember that research is expensive in terms of money, time and effort. Do not include many levels especially if your methods are complicated.If you have more than one independent variable, the combination of levels that you will apply to one run of your experiment is called a treatment. Like levels, avoid having too many treatments.
There are certain experimental designs that could generate great insights even if there are fewer treatments needed to be done. Examples of this are fractional factorial design, balanced incomplete block design, and central composite design. The simplest is a completely randomized design, but the insights you will get from it might also be simple.
REMEMBER:
A level is a value for an independent variable that will be studied.
A treatment is a combination of levels of independent variables which you will apply to an experiment.
Avoid having too many levels/treatments because they are expensive in terms of money, time, and effort.
Tanim Talino 