Animal Behavior

Summary of information on variables and experiments:


A characteristic of an organism that can be observed or measured, and that differs from one individual to the next in an observable fashion, is a variable. There are many kinds of variables in the science of biology.

• Qualitative variables or attributes are mutually exclusive, non-overlapping states of physiological, morphological, behavioral, or ecological categories. For instance, the distinction between male and female is a qualitative one, as is the difference between larva and adult.

• Quantitative variables are based on counts or measurements. Discrete, meristic, or discontinuous quantitative variables take on only whole number values. Number of individuals in a group, number of ectoparasites on an individual animal, and number of eggs laid by a female are all examples of discrete quantitative variables. Continuous quantitative variables can theoretically take on any value between two fixed points. The accuracy of the observation depends on the sensitivity of the measuring device. Here the true value of the variable cannot be known with certainty, but can only be estimated. The measured flight distances by Trimerotropis saxatilis is an example of a continuous quantitative variable.

 Many of our observations in biology are made through a special process called an experiment. The term "experiment" has a particular meaning to a scientist, which is usually poorly understood by non-scientists. The "experimental unit" is the entity whose responses to environmental stimuli are observed during the experiment. This may be an individual in the biological sense, but could also be a collection of individuals (a social group, a nest full of eggs) or could be a part of an individual (a muscle group, a sense organ, a nerve cell). The responses of experimental units to environmental stimuli are our variables.

What is the essence of a true scientific experiment?

• Manipulation: One or more independent variables is altered by the experimenter so that the effect on dependent variables (in our case a variety of kinds of "behavioral assays") can be observed. Frequently an experiment consists of simultaneous manipulations of more than one independent variable. A group of experimental units (see replication) is subjected to each kind of manipulation or treatment.

• Control: One group of experimental units is not subjected to the manipulation; the independent variable for this group is held in a more or less "natural" condition. Sometimes more than one control group is necessary. For instance, if a researcher is studying the effect of a gonadal hormone on aggressiveness, different treatment groups might be injected with different dose levels of the hormone, one control group would be injected with the solvent (also called the "vehicle") in which the hormone is dissolved, and a third group (the "sham control") is handled or disturbed as the other groups are, but is not given any injection.

• Randomization: The assignment of experimental units to treatment groups is at random with respect to characteristics such as gender, body size, reproductive condition, age, etc. This is done to minimize bias in the results due to variables that are not being controlled by the experimenter.

• Replication: Each treatment or control group consists of more than one experimental unit. This minimizes bias in the results due to variation among individual experimental units in their response to the independent variable. Replication is necessary to generate means and variances which are necessary for parametric statistical tests (see handout on "Statistics for Novices").

• Repeatability: The experiment is done and is reported in such a way that the experiment can be repeated by other researchers to verify the results.