ABSTRACT
In Chapter 1 a distinction was drawn between experimental and correlational research: in an experiment we manipulate one variable and measure the consequential changes in another variable; in a correlational study we measure both variables-that is, we try to relate naturally occurring variations in one variable, say income, to naturally occurring variations in another, say intelligence. Thus experiments lead to the discovery of direct, causal connections between variables whereas correlations may be mediated by a string of intervening variables and do not necessarily reflect causation at all. None the less, correlational research can yield interesting and theoretically useful findings in circumstances where experimental manipulations would be impossible, e.g. where social and personality variables are being studied (see p. 6).
