This is the 2007 version. Click here for the 2017 chapter 01 table of contents.

Summary: Observational and Experimental Research

Observational research consists of data collection without experimentation. Data analysis may reveal patterns called correlations. Correlations can be used to make predictions, even if there is no accurate cause-effect analysis of the pattern. As a rule, computers using large amounts of correlational data can make predictions that are more accurate than expert humans.

The pitfalls of observational research involve biasing effects that introduce inaccuracies into data or make samples unrepresentative of a larger, target population. Measurement and observer effects occur when data is distorted by the very act of collecting it. Population validity problems occur when results obtained with one species or group of individuals are inaccuraete when applied to a different group. Leading questions, biasing context, and willingness of respondents to give opinions without any background knowledge are all factors that can distort the accuracy of a survey. Telephone polling has been distorted by push polls that attempt to influence opinion while pretending to gather data

Experimental research is used to test cause-effect claims. In experimental research, an independent variable is manipulated and dependent variables are measured. In a typical between-subjects experimental design, one group receives a genuine experimental manipulation while another group receives a placebo treatment. To prevent the placebo effect from producing misleading results, a single-blind design is used. To prevent bias from experimenter effects as well as placebo effects, a double-blind procedure is employed.

Observational research produces evidence of correlations, and correlations can be used to make predictions. This is true even if we do not know the reason for the correlation, as long as the correlation lasts into the future. The human brain is a correlation-detecting system, among other things. Human intuition is largely based on detecting correlations, providing us with expectations about what will happen in the future. Observational research, properly interpreted, extends or amplifies these powers of prediction.

Experimental research allows us to an analyze systems into their components and how they interact. This extends the human power of analysis and makes it possible for us to understand and repair natural systems.


Write to Dr. Dewey at psywww@gmail.com.

Don't see what you need? Psych Web has over 1,000 pages, so it may be elsewhere on the site. Do a site-specific Google search using the box below.

Custom Search

Copyright © 2007-2011 Russ Dewey