What are the basic units?

In a survey or experiment, measurements are made from discrete units — these could be people, farms, animals, plots of land, individual plants or many other types of unit.

Sometimes the type of unit that should be used in a project is clear (e.g. households in a survey or cows in an experiment about diets) but there are several possibilities in other projects. For example, in a crop experiment the individual plots that will be used for the different varieties could be of several different sizes and a decision must be made before the experiment is started.

There may also be some structure to the units. For example, a survey may collect information from a sample of households both at the household level and from the separate individuals within each household. The households may also be grouped within several different villages, forming another level. Similarly, the cows used in an experiment may be grouped into several herds.

The choice of units should depend on the objectives of the study.

What measurements should be made?

In a survey, information is collected from a sample of units. There is usually a trade-off between the quantity of information recorded, the cost of obtaining it and accuracy of the resulting data. If too many questions are asked in a questionnaire, the respondent will often either refuse to continue or give less accurate answers. Too often, researchers try to cover all possibilities with an unreasonable number of questions. Instead,

Considerable effort should be put into devising a minimal set of questions that will allow the objectives of the study to be attained.

In an experiment, there is sometimes a single obvious response measurement from each experimental unit (e.g. crop yield per square metre, concentration of impurities, exam mark), but often there are several variables that can be considered as response measurements.

For example, in a study of how a fertiliser affects growth of tomato plants, how do you measure growth?

There are many possibilities here and a decision must be made on which measurement or measurements will best address the objectives of the experiment.

A biologist should decide on which measurement is most important from a biological perspective.

Controlled variables in an experiment

There is an extra complication in an experiment — some characteristics of each unit should be varied by the researcher. Thought needs to be given to which variables need to be controlled and what settings should be used for these variables in different units.

In an agricultural experiment, do we only want to assess the difference in yields for three crop varieties, or do we simultaneously want to determine the effects of different levels of application of fertiliser? (And if so, what levels should be used?)

The controlled variables and their settings should be chosen to meet the objectives of the project.