1) The extent to which a measurement procedure yields the same value when brought into repeated contact with the same state of nature. Repeatability. A property of behavior that is described in terms of the dimensional quantity of countability.

2) Consistency of measurement. in applied behavior analysis, preferably, different observers estimate reliability by assessing the interobserver agreement, or coefficient of agreement between two or more independently scored records of data in the same episodes. The coefficient of agreement often is calculated as a percentage by dividing the number of agreements by the number of agreements plus disagreements, then multiplying the fraction by one hundred. Agreement measures should be reported for each phase of a within subject design when feasible. When estimating reliability of interval recording systems, if there are many unscored intervals, only scored intervals should be included in the calculations. Or two separate percentages of agreement can calculated-one for scored, one for unscored intervals.

3) The extent to which two independent observers agree on their observations, computed by the formula A/(A+D), in which A stands for the number of agreements and D stands for the number of disagreements. A quotient of 90% or more is acceptable unless the response definition is a new arid difficult one, in which case 80% or more is acceptable.