When a state Extension specialist or Extension educator makes a presentation, the individual will occasionally make reference to “statistical significance” or some variant that alludes to statistical analysis and its use in determining treatment differences. So what is meant by statistical significance?

Why should a producer, consultant, or retailer care about statistics? Can the average of treatment effects be used alone to evaluate differences? These are often asked questions that need clarification. Over the next several CORN Newsletters, we will attempt to explain why researchers use statistics as a tool, why statistics are useful and necessary, and why split field unreplicated information is difficult to draw conclusions from.

Useful Terms

Observation– a measurement that is made for some output(s) of interest (yield, plant stand, nutrient status, disease incidence, insect infestation, etc.).

Treatment– controlled input factor applied to an experimental plot (seeding rate, fertilization rate, insecticide application, fungicide application, etc.) that will hypothetically have an impact on an output(s) of interest. Obviously multiple treatments are needed to conduct a well-designed experiment.

Error – deviation of a measurement from the true value (usually caused by variation in the measuring device being used or uncontrolled factors that affect the measurement directly).

Experimental error –differences in observations from treatments due to environmental conditions that can not be controlled by the experimenter (differences in soil texture, topography, soil compaction, rainfall, nutrient status, disease infestation, etc.).

Statistics 101

Any observation made within an experiment has a certain amount of error associated with it. In fact, there is error associated with any measurement. Statistics allows us to quantify and assess this error. If only a single observation is made can you get an estimate of error? Some would argue this point with a complicated mathematical model, but for the sake of argument we will agree that a single observation will not provide an estimate of error. In order to estimate the error of an observation, multiple observations need to be made. We call these multiple observations replications. Replication allows us to estimate the experimental error associated with the observations that are being made (and evaluation of the tools we are using to measure the observation). In a field experiment, the observations can be confounded with a multitude of soil and environmental factors therefore we must replicate the treatments across the landscape. To ensure the estimates of experimental error for each treatment are unbiased (not systematically influenced by underlying environmental conditions (i.e. soil type, topography, etc.)), the replications should be randomly placed within the field. We have just discovered the two most important things (in our humble estimation) in statistics – 1) to estimate the experimental error of treatments requires replication and 2) to ensure an unbiased estimate of experimental error requires randomization of the treatments.

Statistical Significance

This is often mentioned but seldom explained in an extension venue. When an experiment is conducted (properly replicated and randomized), the experimental error (average variance within treatment observations) is computed and used to assess whether or not treatments differ “significantly” from one another. Statistics is based on probability, and researchers select what level of probability constitutes significance. The probability level selected is solely at the discretion of the researcher. The scientific community general prefers a probability level of 95%. So a researcher can state with 95% probability that one treatment is different from another. If the 95% probability criterion is met, then the treatments are “significantly” different. This is where some gray area enters into research, what is the appropriate probability level? Each researcher has their own set of criteria. The next time you hear a speaker discussing some research data, think about what level of probability is being used to evaluate treatment differences.

Statistics is a tool that allows researchers to assess the error associated with conducting an experiment and to separate real treatment differences from differences caused by uncontrollable environmental factors. Researchers can separate the grain from the chaff as it were. Like any tool, it must be used properly to be effective (replication and randomization). Statistics is not a hard and fast science. It does require some direction from the researcher conducting the experiment.

In the next CORN Newsletter article, we will continue our discussion on statistics and look at some actual data.

By Robert Mullen, Maurice Watson, OSU

Leave a Comment

Name (required)

Email (will not be published) (required)

Website