Terminology
The hypothesis test aids in business decisions through statistical analysis. Data crunchers can situate problems so that statistical theories can be tested concerning occurrences and correlations of data. When examining a population mean where the standard deviation is known, they'd use the Z test. When the standard deviation is not known, and the sample's SD is being used, the T test is then used. When population proportions, they use the z test for p. When a population variance is being examined from one sample, the x squared or standard deviation squared is preferred.
Research hypothesis- a statement of what the data analyst thinks will be the yield of a study or experiment. I.E. Younger workers are more likely to leave the company.
Statistical hypothesis- For the purpose of structuring a research hypothesis so that it can be scientifically tested, a statistical hypothesis is a formalized is arranged so that there are both a null hypothesis and an alternative hypothesis. The null hypothesis essentially says that nothing new is occurring and that the previous, former theory is right, and the process of examination is still valid. The alternative hypothesis illustrates that the new theory is correct and that the prior system is not functional. As an example, let's say a given package of Starbucks coffee is supposed to be 12 ounces. If the processes which ensure this consistent volume were in question, the null hypothesis would be that the packaging was accurate and that the containers had 12 ounces each. Ho would be 12 ounces per package- the null. Ha would be that the packages don't equal 12 ounces.
Substantive hypotheses- Significant to business people is different than data analysts or statisticians. To a business person, significant means usually a large quantity. A substantive result means the results of a study yields results that are significant to the decision maker.
A one-tailed test has an equal sign (=). It's directional, usually characterized by higher, lower, alternative theories have < or >. That which is not in alternative hypothesis is included in the null.
Two-tailed also has an equal sign (=). These are nondirectional, and terminology includes: same, different, equal, control and out of control. The alternative hypothesis can solely have a not equal to sign. The null/alternative hypotheses include all possible scenarios.
Black, K. (2017). Business Statistics: For Contemporary Decision Making, (9th Edition). Hoboken, NJ: Wiley.