In the science field, scientists collect data all the time to make analysis and conclusions in the various fields. In the medical field for example, researchers use research data to find out about the type of medication they should administer. Therefore, it is very important for scientist to use effective methods of collecting data.
Since all subjects cannot be studied, scientists select a sample of subject from a group of interest to conduct their studies. The results from the sample selected, are assumed to represent the characteristics of the whole group. Samples are usually collected in a random manner to at least ensure every aspect of the group is captured.
Collecting just a sample of subject makes sense in some circumstances. For example, if a researcher is interested in finding out the effect of an advertisement on T.V in a certain town, it will be tedious and extremely expensive to study every individual in the town.
The best thing to do is to randomly pick a few people from the population and study them. In many cases, the result will reflect the effect on the general population of the town. However, choosing subjects by random method might not actually be random. Unknowingly, the researcher might pick subjects who do not represent the group of interest.
For example, a medical team in USA might be interested in finding out the side effects of a certain treatment drug they are developing. It is apparent that the team will pick about 1000 random individuals of different ages throughout the country. Also, they probably will ensure that both genders are well represented. Let us say the result will indicate that 90% of them vomited.
In many ways, these results are inconclusive. First of all, only a thousand subjects were examined and probably the drug will be used by over a million patients throughout the world. Factors that will affect the side effects of this drug will most likely depend on medical background, heredity and even geographical location. Out of the thousand picked, there is a big chance that all these factors were not taken into account or reflected on the results.
It would not be big surprise if only 7% of the patients administered with the drug vomited. Probably, the largest percentage of the patients would exhibit far much worse side effects like death. This will obviously be a catastrophic mistake made by the medical team. Other than the method having a huge magnitude of error, the randomized experiments could have expensive methods of repair or render subjects to lethal circumstances (HECKARD, 2006).
Despite of this, there are other methods which can be used to collect data with much accuracy. Perhaps one of the most accurate methods of sampling, especially one involving treatment methods with a placebo effect, is the snowball technique. In this method, information is collected from a minute pool that has the desired characteristics. The people in this group are then asked to identify other people who posses similar desired characteristics.
The researcher then moves on to the referred group and does the same thing until he collects enough data to make a conclusive report or, until there are no changes in the data. In this method, the researcher will be able to include crucial subjects that would have otherwise been left out. Since this method deals with specific type of group, the results are much more conclusive than most of the sampling methods present. Other than its accuracy, this method is cheap and has minimal damage risk to the subjects (Sharon, 2009).
Works Cited
HECKARD. (2006). Statistical Ideas and Methods. New York: Thomson .
Sharon, L. L. (2009). Sampling: Design and Analysis. Seattle: Duxbury Press.