Typically, the experimenter hypothesizes that one behavioral
variable operates so as to produce another. Changes in one variable can be
expected to cause changes in another, and the experimental hypothesis
is stated in these terms. In a
hypothesis
test, the null hypothesis, which the
experiment is designed to reject, is a statement that negates the meaning
in the experimental hypothesis, generally by specifying that there is no
relationship between the variables. The probability that an
error
is made in rejecting the null hypothesis is called the alpha risk.
Experiment Variables
The variable that the experimenter manipulates is called the
independent variable. The aspect of behavior, which is
measured and thought to be affected or influenced by the independent
variable is called the dependent variable. All other variables
are called nuisance variables, or confounds, and should be
controlled by standardizing them, making sure they do not vary between the
experimental and control groups. Nuisance variables can also be
controlled by making sure that whatever influence they do have is thoroughly
scattered so as to have no consistent effect on either group.
Experiment Subjects
The groups of subjects in an experiment are designed around the
independent variable. There are two types of groups:
experimental
and control groups. Usually the control group experiences the naturally occurring condition in which the
independent variable is not changed, and the experimental groups experience
conditions in which the independent variable is manipulated to reach
particular levels.
Data Analysis
Data that are collected during the experiment are summarized with
descriptive statistics, which describe the group included in the
study. Some examples of descriptive statistics include the
mean (the average of the scores) and median (the middlemost
score). Variability measures are also used in data summary
and include the range and standard deviation.
Unlike descriptive statistics, inferential statistics are tools used for determining whether the manipulation of the independent variable during the experiment was effective in substantially changing the dependent variable. A commonly used tool is the t-test, which evaluates the mean scores of the groups and tells you whether the difference between the groups is significant or simply due to a chance variation in the subjects’ behavior.
For more info on statistics, check this site out!