The widely used and the ideal measure of dispersion is standard deviation. Standard deviation is used for determining the value of the dispersion of a set of given data. However, it is beyond my knowledge on whether this measure is suitable for determining the dispersion if a set of time series data are given. Thus, the question here is "Is standard deviation suitable for measuring the dispersion of a given time series data ?"
Time series design is a research alternative when the same sample acts as both control and experimental groups. Notably, the first thing researchers do is to run a CONDESCRIPTIVE TASK for measuring the mean values of participants' performance on treatment and non treatment situations as well as the standard of deviation . As such, showing the dispersion of scores from the mean by computing the value of standard deviation is equally applicable to the cases of time series design.
Standard deviation is no doubt universally accepted measure of dispersion. Acceptance by people and the truth are different. Acceptance depends upon peoples' opinion. On the other hand truth is independent of peoples' opinion. What is accepted on the basis of opinions of the people may not be identical with what is the truth. Therefore, universal acceptance of the standard deviation as measure of dispersion does not imply that it is suitable measure of dispersion of time series data. Thus, it is to be justified whether standard deviation is suitable for measuring the dispersion of a given time series data.
You might want to try the standard error in percentage instead.
Dear Sindhuraj Mukherjee,
Standard error, is a measure of dispersion of an estimator (more specifically, dispersion of the possible estimated values yielded by the estimator). It is not a measure of dispersion of data. Thus standard error corresponds to estimator but not to data. Accordingly, standard error can not be a measure of dispersion of time series data.
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping