III.) Replicated, Statistically Designed Field Trials

Question 1

The most reliable results will be obtained from field trials if they are conducted as scientifically designed, replicated experiments with controls.

Answer 1

Replicating the treatment allows for the measurement of variability across experimental units. It is the measurement of this variability that is the key to the results of a statistical test.

If variability among replicates is high, it is very unlikely that a trial will yield statistically significant results unless the differences among treatments tested is also very high.

Question 2

Experimental Design

*A successful experiment defines the problem clearly and precisely.

Answer 2

The number of treatments, the number of controlled and measured variables, and the number of replications to be used in the experiment are major factors in the experimental design.

Treatments are typically compared to a control, either an untreated plot or a conventional standard treatment acting as a control.

The number of replications is determined by the degree of precision sought (influenced by the type of pest, abundance of pest, and the magnitude of expected differences).
-The higher the precision, the larger quantity of replications.
-The more variables/block, also the larger quantity of replications required.
Replication results provide the statistical basis for demonstrating whether an observed difference is real or just due to chance.

Question 3

Blocking

Answer 3

A block is an area that includes one plot or replication of each experimental treatment. Treatments within a block should be arranged so that each treatment is similarly exposed to any environmental or biological factors that you believe might influence results.

Question 4

Randomization

Answer 4

Each treatments should be randomly situated within the block so that each experimental plot has an equal chance of receiving the particular treatment.

Randomization of treatments guarantees unbiased estimates and reduces experimental error.

Question 5

Control

Answer 5

The control is the treatment to which other treatments will be compared.

When testing management techniques the ideal control is a replicated untreated area, alternatively, a well-know pesticide or standard grower practice can represent the control as well.

Question 6

Non-experimental Field Trials

*Several things can be done to enhance the value of non replicated field trials such as selecting sites that are as similar as possible in soil type, drainage, microclimate, habitat, pest pressure, and management practices.

Answer 6

It is not always possible for PCA’s or growers to carry out replicated field trials that meet the standards of scientific rigor.

However field testing of new materials or methods can be useful, and growers should be encouraged to try out new methods in small areas first.

Without a control and statistical backing, however, there is no safe assumption that the same results shall prove true at another site or under different conditions.

Question 7

Evaluating Data from Field Trials

*Although treatments may tend to show different degrees of control, statistical analysis is necessary to show whether these differences are truly significant or can be explained by chance.

Answer 7

Some common results with research results are

• Absence of a control
• Lack of replications
• Large standard error or standard deviation (variability)
• No indication of statistical significance.
• Unable to block out variables
• Comparing disimilar treatments (apples to oranges)
• Working off of pseudoreplications as if they were statistically sound.