Pitfalls Essay Example | Topics and Well Written Essays - 250 words

Pitfalls - Essay Example This primarily happens due to the selection of wrong statistical methods or wrong size of survey population or both. It is highly expensive to select a large survey population and conduct interviews/form fill-ups extensively. So samples are smaller than the target population. However, sophisticated methods like purposive sampling may help, but still it is important to accommodate as many respondents as possible in a given survey population. 2. Making conclusions from non random samples too can prove to be a dangerous tendency. Samples must me random. Suppose, a woolen garment manufacturer conducts a survey only in the colder countries of northern Europe. In this way, they will calculate high demands existing in the global market, if they don’t randomize. In order to randomize, they will have to interview customers in temperate regions (e.g. southern Europe) and arid regions (e.g. Sub Saharan Africa). Then only a realistic view of market demand for woolen garments can be obtained. Hence, randomization is necessary to prevent biasing, especially in the case we need globally applicable inferences. 3. Attaching importance to rare observations is a clever strategy but risky practice. If rare observations are given as much importance as random observations, general behavior of the majority of the target population may be interpreted wrongly. It is a far better idea to mark rare observations as exceptions so that they can be studied separately by the means of statistical segregation in a more controlled business environment. 4. Using poor survey methods is the worse mistake. Survey methods are crucial. For example, when research involves exploration of natural resources, data has to be collected with the help of researchers who have trekking/exploration/fieldwork expertise. In the case of intellectual property research, we would need researchers who can research on databases and patent archives. That will involve extensive Internet/library research expertise. In