Perform a thorough analysis. In doing your qualitative research, it is important to gather as much information as possible. This will help to best describe observed behaviors and differentiate between them. If you expect to later perform a reliability test, determine what specific factors of your study are vulnerable. Examples might include integrity of students, background details that disregard participants, environmental factors at the time of observation or anything else potentially influential.
Create a numerical key for factors affecting reliability. For example, a score of 1 indicates no reliability threat, 2 is a modest threat and 3 is likely or certainly unreliable. You might need to compare personal research with empirical data to determine reliability of observations. Otherwise, use available information and an educated opinion to assign values.
Perform quantitative analysis on your qualitative research. You or a supervisor might have a goal for reliability. Standard deviations, based on the bell curve, are commonly used in statistics to measure coherence of groups, consistency and reliability. If 68% of reliability scores are acceptable, data is reliable within one standard deviation. To reach two standard deviations, 95% of scores must be reliable. Over 99% must be reliable for data to be reliable within three standard deviations.