Limitations and applications of statistical hypothesis testing1
Sampling Methodologies1
Stratified Random Sampling1
SRS2
Cluster Sampling2
Systematic Random Sampling2
Correlations3
Positive Correlation3
Negative correlation3
Minimal4
GROUP A4
If Positive Correlation4
If Negative Correlation4
If Minimal Correlation:4
GROUP B4
If Positive Correlation4
If Negative Correlation5
If Minimal Correlation:5
GROUP C5
If Positive Correlation5
If Negative Correlation5
If Minimal Correlation:5
GROUP D6
If Positive Correlation6
If Negative Correlation6
If Minimal Correlation:6
References7
Inferential Statistics
Limitations and applications of statistical hypothesis testing
There are two main limitations of hypothesis testing. The first is related to the sampling. In inferential statistics and hypothesis a researcher makes inferential recommendations about a population, and due to sampling the population is not completely studies. Because of this the statistician cannot be completely certain of his results. The second limitation is sometimes considered an extension to the first limitation. The second limitation is that a degree of uncertainty remains because at certain times the statistician has to make guesses on the methodology of the inferential test. These guesses are based on theory. Because of these guesses the complete surety over the certainty of the calculations and results is compromised.
Sampling Methodologies
Below are some of the sampling methodologies that can be applied to business situations.
Stratified Random Sampling
In this type of sampling the population is divided into homogenous subgroups and after that a simple random sample from each group is taken. The sub groups in which the population is divided are called strata.
The stratified random sampling not only represents the whole population but also represents the sub groups within the population through strata. In simple random sampling, a sample from the whole population is taken, but it does not show the exclusivity of subgroups within the population.
SRS
Simple random sampling is the simplest sampling method in statistics.
The simple random sample can be obtained by a variety of mechanical and computerized methods. The major factor for selecting a simple random sample is that the selector (whether a machine or a person) should be blindfolded against the population, there should not be any biasness for a particular value of n, because in a simple random sample every n value has equal probability of occurrence.
Cluster Sampling
Cluster sampling is a sampling technique in which the population is divided into separate groups or cluster. A random sample is selected from these clusters. Finally each observation from the randomly selected clusters is included in the sample.
Cluster sampling is done when the population is widely scattered or when the complete lists of members of population cannot be obtained ...