Testing a Hypothesis Is to Establish One.

¶ … testing a hypothesis is to establish one. There should be a null hypothesis that the data can be used to test. Data acquisition is therefore the next step in testing the hypothesis. The data needs to relate directly to the hypothesis with a clear relationship that can be subjected to quantitative analysis. Quantitative analysis will then measure the relationship between the variables to determine whether or not the data fits with the null hypothesis. The null hypothesis is then either accepted or rejected on the basis of the analysis (Investopedia, 2013). Further, the null hypothesis should identify the dependent and independent variables. The dependent variables are those that will be measured in relation to change in the independent variables. Thus, it is the independent variable(s) that will be changed to measure the effect that change has on the dependent variable(s). There may also be an alternative hypothesis, which may simply be to reject the null hypothesis.

The method of quantitative analysis must be determined. There are a number of methods to choose from, and usually the choice must be supported by research that shows such a method is an appropriate means of testing that type of hypothesis using that specific type of data. Regression is usually used to measure correlations between independent and dependent variables. ANOVA is a technique that can be used with smaller data sets in Excel. For larger and more complex data sets, SPSS is typically used. A number of test statistics are derived from the data and then these are converted, in theory, to meaningful data regarding the hypothesis (No author, 2013).

The output of quantitative analysis will consist of a number of different measures, each of which has its own meaning that must be interpreted. These meanings include fit, statistical significant and confidence intervals. The different output figures are challenging to interpret, because there is very little information in English -- too many sources talk too much about the math and it becomes difficult to understand the difference between, say "confidence interval" and "significance," which sound like pretty much the same thing.

For basic, non-ANOVA analysis, the metrics are easier to understand. Comparing the means of two or more groups is a very basic quantitative technique, but one that is also relatively poor when trying to find the links between different variables. Means may be used where the underlying variables are very similar, but preferably a more robust form of quantitative analysis will be used.

The correlation between two variables is usually tested with some sort of regression analysis, ANOVA or other robust technique. These techniques use advanced mathematics to explain the strength of the correlation between two variables, and the confidence that the researcher can have with those results, given the number of data points that were used. Such a technique will typically either confirm the null hypothesis, or leave the researcher in a position of rejecting the null hypothesis.