Statistical significance Essay

If the outcome of a particular experiment cannot be explained by chance but must be ascribed to a particular cause, then statistical significance exists.

If a population has an average IQ of 100, for example, the logic behind statistical significance can be described as follows. The expected mean for a random sample of 100 people should be close to 100, such as 98.9 or 101.2, but not precisely 100 due to errors like sampling error, measurement error, etc. However, if this group has received some type of educational intervention that was intended to boost their IQ, then a significance test done on this population is expected to produce a sample mean that is different from the population. Thus in this case the statistical significance arising can be attributed to the education intervention and not to chance (Ziliak, 2008).

Nominal and Ordinal Variable

Nominal variable are variables that have neither quantitative value nor any numerical/statistical significance. Examples of nominal variables are; a person's gender which is either male or female or the colour of someone's eyes which could be brown, black, blue or green. Such variables are mutually exclusive meaning that they never overlap. Ordinal variables are variables in which the significance of the values lies in their order and the difference between each value is not known. For instance, when asked how satisfied one feels, the levels of satisfaction could be very satisfied, satisfied, slightly satisfied and not satisfied. In this case it is easy to see how much one is satisfied but we cannot measure the difference in satisfaction between very satisfied and satisfied (Martella, 2013).

One Tailed and Two Tailed Test of significance

One tailed test of significance is used to determine if the test parameter is less or greater than the critical value but not both. Prior to the test being done a direction has to be chosen since one tailed tests will only convey the effect of change in one direction only and not the other. For example, if u were to carry a one tailed test on a generic drug against a brand name, one could measure the effectiveness of the drugs and determine that the generic, though less effective, is cheaper and therefore buy it. The two tailed test of significance is used when one has to determine if there is a difference between two means. This difference could be positive or negative and this test takes that into account. For instance, in the above example one could choose to do a two tailed test that will tell one how much less effective the generic drug is to the brand name and it could be found to be far much less effective that one chooses to pay more for the brand name (Rosenthal, 2011).

References

Martella, R. C. (2013). Understanding and Interpreting Educational Research. New York, NY: Guilford Press.

Rosenthal, J. A. (2011). Statistics and Data Interpretation for Social Work. London, UK: Springer Publishing Company.

Ziliak, S. T. (2008). The Cult of Statistical Significance: How the Standard Error Costs Us Jobs, Justice, and Lives. New York: University of Michigan Press.