First, a brief discussion of Reliability and Validity of Observation research
Reliability - how reliable is your measure. Will you get the same results if you conduct your research at another time? Would someone else get the same results if they conducted the research?
Things to watch out for, and ways to improve reliability.
1. Independent Observer - would they come to the
same conclusion.
Greater precision in defining the categories leads to greater reliability
4. The less the researcher interferes with the participant, the less the actions differ because of the researcher's presence - the more reliable.
Participation sometimes makes the researcher less noticeable
Take notes after the observation session
Show the participants your notes so they are not intimidated by them
Choose a situation in which people are so busy or intense that they don't notice
5. Boredom in the observation session can lead to bad observations. Or satiation with the topic.
Observe in shorter periods - but the sequences may be missed
Take breaks
Validity - how valid is your conception of the situation, and how generalizable are your results.
This is the strength of qualitative research - it
is more likely to be valid than quantitative research. But there are things
to be careful of.
1. Adequate sampling
2. Precision of observation
3. Subtle changes in environment and persons observed
Internal validity - Did the methods that you
used color the results? Were there other factors that interfered?
Also: Can we draw cause and effect conclusions from
the results?
External validity - How generalizable are the results.
Face Validity
Statistical Measures of validity:
Type I Error = Reject the hypothesis when it is
true
Type II Error = Accept the hypothesis when it is
false
GROUNDED THEORY
Grounded theory is the process of developing a theory from systematically obtained data. It was developed by Glaser and Strauss in the late 1960s - to counter the charge that qualitative data is not