Statistical Data Collection Methods - Making Sense of the Mine Field

When use correctly, statistical data can be usedcause-and-effect relationships, how one thing
to improve an array of areas from efficiency, toaffects another.
lead time, and profit. But in order to make- Observational study - Observational studies also
improvements you need to know how the datatry to discover the cause and effect relations, but
has been collected initially. This article is aboutunlike experiments, they are not able to control
Statistical data collection methods.how subjects are assigned to groups.
There are four main Statistical data collectionAs it was already pointed out, every method has
methods:its own pros and cons, so one must be able to
- Censusknow and make a decision regarding which
- Sample surveymethod should be applied in a given situation.
- ExperimentThere are three factors that should affect this
- Observational studydecision and they are - resources, generalizability,
Each of these methods has it's own set ofcausal inference.
advantages and drawbacks, that's why one mustIf resources are the main factor, then obviously
be aware of all their characteristics to be able towith such a large population, a sample survey has
choose the right method according to thean advantage over census. If the sample survey
individual situation. Here is a brief definition of eachis well designed, then it can definitely provide
method:results that are really close to the actual figures
- Census - A census is a case study that acquires(high level of accuracy), and it will be done in a
data from every population member. For thequicker and cheaper manner, requiring less man
majority of cases, a census is not practical, duepower than a census.
to the large amount of time and cost required toGeneralizability stands for applying findings from a
conduct it.sample study to a larger population. Generalizability
- Sample Survey - A sample survey is a caserequires random selection. In case the participants
study that obtains data only from a subset of thein a study are randomly selected from a larger
entire population, not every member, as opposepopulation, it is appropriate to generalize study
to Census, so it's much more practical andresults to the larger population, otherwise it might
efficient to carry out, but the results might not beprovide accurate results.
that accurate. For best results using this method itStatistical data collection methods are essential for
may be appropriate sub-categorize your targetsustainable economics, social and environmental
group and take a sample set from eachdevelopment. We are living in the 'Information
sub-category. A basic example would be differentAge' where certain data sets are growing in size
ethnic groups.and complexity, reaching massive proportions,
- Experiment - The experiment is a controlledthat's why such data collection methods are so
study in which researchers try to understand theimportant.