Example of Hypothesis

Sometimes scientists cannot gain enough control over the situation to allow them to conduct a true experiment. Here's an example: Suppose that you wanted to test the hypothesis that children who ingest lead-based paint (common in older homes, especially in low-income urban housing) run an increased risk of learning disabilities. You couldn't do an experiment to verify this hypothesis. Why? In an experiment you would have to manipulate the independent variable--which would mean giving toxic material to a group of children. Obviously, this would be harmful and unethical.

Fortunately, you can find a way around the problem--but at the expense of some control over the research conditions. The solution takes the form of a correlational study. In correlational research you, in effect, look for a "natural experiment" that has already occurred by chance in the world outside the laboratory. So, in a correlational study on the effects of ingesting lead-based paint, you might look for a group of children who had already been exposed to leaded paint. Then, you would compare them to another group who had not been exposed. As a further control, you should try to match the groups so that they are comparable in every conceivable respect (such as age, family income, and gender)--except for their exposure to leaded paint.

Correlational study

A form of research in which the relationship between variables is studied, but without the experimental manipulation of an independent variable. Correlational studies cannot determine cause-and-effect relationships.

The big drawback of a correlational study is that you can never be confident that the groups are really comparable, because you did not randomly assign people to test groups or manipulate the independent variable. In fact, the groups may differ on some important variables (such as access to health care or nutrition) that you may have overlooked. Thus, you cannot say with certainty that the condition of interest was the cause of the effects you observed. So, even if you observe more learning disabilities among children who were exposed to lead-based paint, you cannot conclude that exposure to the paint caused the disabilities. The most you can say is that lead-based paint is correlated or associated with learning disabilities. Scientists often put the general principle this way: Correlation does not necessarily mean causation. In fact, confusing correlation with causation is one of the most common critical thinking errors, and is an example of a fallacy in reasoning

Researchers usually express the degree of correlation as a number known as the correlation coefficient, often symbolized in formulas by the letter r. The size of the correlation coefficient summarizes the relationship between the two variables: It can range from a negative number (as low as -1.0) to a positive number (as high as +1.0). We won't go into the details