Surprisingly, naturalistic observational research is difficult to do well. Methods have been
constructed to guide the researcher, especially the beginning researcher, in the collection of data
under naturalistic conditions. Most importantly for this method, the very presence of the
investigator, as well as the data recording techniques that are used, must remain unobtrusive; that
is, no characteristics of the observation itself should have any effect on the behavior of the
organism under study.
Advantages of naturalistic observation include the collection of data that clearly reflect the
constraints of the organism's normal environment (human subjects, in particular, may respond
differently under laboratory conditions when they know that they are in an experiment), and
the collection of data in cases where experimental techniques would be impractical and/or
unethical (for example, in an exploration of the social interaction patterns of mentally retarded
children in the classroom).
One problem with naturalistic observation techniques is that different observers may draw
different conclusions from watching the same instance of behavior.
This potential problem is
often addressed in two ways. One safeguard against this problem is known as "working up
reliability". In the early stages of characterizing a behavioral phenomenon, the scientists working
on a project will get together to create definitions of the behaviors in question, definitions upon
which they can all agree and which they can implement when collecting data.
References for this material include: (www.uconn.edu).
Topic of Study: A four year old boy and his three year old male cousin as they get in to
trouble when they think no one is watching.
Method of Observation: Video camera set up without the two boys noticing. This
unobtrusive surveillance caused no outside control on the behavior of the boys.
Notes Collected: Both boys were playing together in the living room in front of adults...