How to conduct an experiment
Often growers wish to investigate new products or methods that might increase efficiency in some way. Experiments are performed to accomplish this. There are a few key things to remember when designing an experiment:
Look at enough plants – Formal statistical recommendations are to include between 20 and 30 plants in each experimental group (variable). This allows for the natural variation in genetic information.
Have a control –One treatment needs to be the ‘normal’ or control condition.
Grow all plants at the same time – If possible, each treatment should be applied and all the plants grown simultaneously. This will eliminate environmental variation (light, temperature, pest problems etc.) from confounding the results.
Label the treatments – each plant should be labeled as it is very easy to confuse plants from different treatments if they are temporarily moved. Sometimes the treatments are mixed in space (randomly) and it is essential that each plant have the treatment it received be easily identifiable.
One variable at a time – Investigate only one variable at a time. If two things are changed (for example, both growing temperature and cultivar), it is impossible to separate the effects of each variable and their contribution to the overall result. All conditions must remain the same except for the single variable being investigated.
Collect data – Determine ahead of time when and how you will evaluate the results of your experiment. On the pre-determined day, collect data in numerical form (height, weight, taste, appearance) and photo documentation. If data is to be collected multiple times, the same person should perform each data collection to eliminate the variability in technique from one experimenter to another.
Replicate the entire experiment – Observing a result after a single trial is not enough evidence to declare that result to be a real phenomenon. The experiment should be repeated at least 3 times before the result can be trusted (the first time could have been a fluke!).
Make sure all environmental conditions are the same for the replicate!
Take notes daily – It is tempting to think that you will remember small details at a later time – and you may for a few weeks – but a year or two after the completion of the experiment those small details will be forgotten. Dedicate a journal to daily notes and record observations and any problems you may encounter (ex. Temperature control failed, a plant dried out, insects were observed for the first time).
Take pictures weekly – This will allow for the determination of when and how much the treatments differ if the variable lends itself to visual inspection. Often a destructive harvest (i.e., weighing cut product) is not practical every week and a picture of the intact plant may help to provide more data.
Definitions
Variable – The condition you are changing. This may be a different cultivar, different growing technique, different environmental condition, or different growing system.
Treatment – Group of plants that share one of the variable conditions.
Control – The known or normal condition. This may be the ‘normal’ cultivar, growing technique, temperature, or growing system.
