Unlike the experiment, observational methods do not allow for the control and manipulation of variables.  They produce information about the covariance of behavioral variables - that is, they allow the researcher to say that two variables are connected in such a way that when one changes, the other changes.  These methods do not provide evidence about causal relationships between variables, however.

In direct observation, the researcher observes and records what people are doing.  The researcher must be careful that he/she is not causing an observer effect, in which the subjects behave differently because they are being watched.  In surveys and questionnaires, people are asked to say what they are doing, feeling, or thinking and the researcher records what they report.  Surveys can take place as mail surveys, personal interviews, or by telephone.  The survey process includes:

1. developing and pretesting the research plan
2. collecting data
3. analyzing data
4. preparing a report.

The analysis of observational data may include calculating a correlation coefficient, which measures the degree to which one variable is related to another.  The strength of a correlation ranges from 0.00 (no relationship) to 1.00 (perfect relationship).  Very infrequently, however, do we find correlations over 0.90.  The strength of a correlation depends on the number of participants in a survey.  Smaller correlations are significant if they come from large samples.  A positive correlation, indicated as a positive coefficient, shows that when one variable increases in value, the other increases as well.  A negative correlation, shown as a negative coefficient, indicates that when one variable increases in value, the other decreases.