The Basics of Experimentation I: Variables and Control
Lecture, Chapter 6
What are Variables?
lA variable is an event or behavior that can assume two or more values.
lIn order to understand the researcherís use of a variable, as well as to replicate the study, variables must be operationally defined.
lOperational definition = defining the IV, DV, and extraneous variables in terms of the operations needed to produce them.
Independent Variables (IVs)
lAn independent variable is the variable that the researcher manipulates to determine its influences on behavior.
lPhysiological IVs involve physical or biological changes to the participantsí natural state.
lExperience IVs involve manipulation of the amount or type of training or learning.
lStimulus IVís involve manipulation of an environmental aspect.
lParticipant IVís are aspects of the participants themselves, such as personality characteristics, that cannot be directly manipulated because they exist before the experiment is conducted.
lExtraneous variables are those factors that cause unwanted influence on the dependent variable.
lEither extraneous, or confounding, variables or the IV can be responsible for the results or changes in the DV.
lConfounding effects, which are often difficult to identify, can cause differences between the control and treatment groups to be unrealistically maximized or minimized.
Dependent Variables (DV)
lThe dependent variable is typically some type of behavior or response that changes as a result of the level of the IV the participant experiences; in essence, it depends on the IV.
lA good DV can be measured with validity (directly relating to and truly measuring the effect of the IV) and reliability (yielding same results if experiment were duplicated).
lIn certain situations, measuring two DVs may lend meaning to the results of the experiment (ex. reading test measuring both # errors and time completed).
lNuisance variables are unwanted causes of increased variability within groups.
lWhereas extraneous variables influence the responses of the entire sample, nuisance variables cause a wider spread of scores within the sample.
lEx. Measuring weight loss of obese women.
If sample was derived from a specific weight loss organization, the degree of motivation might be higher than average, serving as an extraneous variable.
Varying metabolism rates of participants within the sample serve as a nuisance variable.
lIn order to eliminate the influence of extraneous variables and minimize the effect of nuisance variables, certain control techniques should be used.
lRandom assignment to treatment groups ensures that each participant has an equal chance of being assigned to either group.
lElimination involves completely removing extraneous variables from an experiment.
lConstancy is used to reduce extraneous variables to a single level common to all participants (ex. keeping treatment conditions and experimenters the same).
lWhen all conditions cannot be held completely constant, experimenters can distribute, or balance, extraneous variables equally to all groups (ex. both groups taught by a male half the time and a female the other half of the time).
lSometimes, participants participate in more than one condition in an experiment. In such cases, counterbalancing is needed to control for order effects.
lOrder effects are the effects on participantsí responses that merely the position of a treatment in a series, as opposed to the treatment itself, determines.
lIn complete counterbalancing, all possible treatment sequences are presented. To determine # sequences needed, break down the number of treatments (n) into component parts and multiply them together.
l3 treatments = 3 x 2 x 1 = 6 possible sequences
lIn incomplete counterbalancing, only a portion all possible treatment sequences are used; in some cases, these are chosen randomly.