Activity Sampling is a statistical technique that can be used as a means for collecting data. It is defined by BS 3138:41008 as:
A technique in which a large number of observations are
made over a period of time of one group of machines, processes
or workers. Each observation records what is happening at that
instant and the percentage of observations recorded for a particular
activity or delay is a measure of the percentage of time during
which that activity or delay occurs.
It is normally used for collecting information on the percentages of time spent on activities, without the need to devote the time that would otherwise be required for any continuous observation.
One of the great advantages of this technique is that it enables lengthy activities or groups of activities to be studied economically and in a way that produces statistically accurate data.
Activity Sampling can be carried out at random intervals or fixed intervals. Random activity sampling is where the intervals between observations are selected at random e.g. from a table of random numbers. Fixed interval activity sampling is where the same interval exists between observations. A decision will need to be made on which of these two approaches is to be chosen. A fixed interval is usually chosen where activities are performed by a person or group of people who have a degree of control over what they do and when they do it. Random intervals will normally be used where there are a series of automated tasks or activities as part of a process, that are have to be performed in a pre established regular pattern. If fixed interval sampling were to be used in this situation there is a danger that the sampling point would continue to occur at the same point in the activity cycle.
Remember, that activity sampling is used for assessing the percentage of time spent on activities.
Because activity sampling conforms to the binomial distribution it is possible to use a calculation to determine how many observations will be needed to operate within specified limits of accuracy.
The formula for the number of observations is as follows:
= 4 x p x (100 - p)
Where p is the estimated % time spent on the activity
Where L is the limit of error, expressed as a %
Once the above calculation has been completed the observations can begin and activities are recorded at the agreed time intervals. When they have been completed a further calculation can be used to determine the error rate, as follows:
Error Rate = ± 2 x √( p x (100 - p) )
This is very much an overview to the topic of activity sampling, with a definition of what it is, its advantage over continuous observation and the formulae that can be used to establish the confidence levels that can be obtained.
More information can be obtained from links to this site.