In preparing a survey to be administer to my university peers, it is necessary t
ID: 401997 • Letter: I
Question
In preparing a survey to be administer to my university peers, it is necessary to tabulate these data once collected. How should the data be tabulated and what statistical methos should I use to analyze these data for creating a solid foundation on wich to base my research report recommendations to the professor.What ismy null hypothesis? What ismy alternative hypothesis? Which questions should be examined together for correlation and which should examined singularly. (The final plan must be complete and put into steps for bulding the statistical case so the basic statistics folder under lessons for ideas.)Explanation / Answer
hypothesis is just a word.null hypothesisandalternative hypothesisare words from statistics that have a rather specific meaning.
one way to put this into context is that the hypothesis is "the suspicion you have that a particular correlation exists" -- it's "the reason for doing the experiment."
the null hypothesis is a statement directly related to the statistical data that you are collecting. In particular, the statistical data you collect is generally used to assess whether or not there is a "significant correlation" between two variables in your experiment. so you "null hypothesis" is that there is not a significant correlation between the two variables in questiion (unless somestatistical measureis exceeded by some amount related to the confidence level of your data.)
the alternative hypothesis is often a simple negation of the null hypothesis -- it "turns on" when the statistical measure for significance meets the significance criteria; it amount to a statement that "there is a signficant correlation between two variable in your experiment." This is where the notion of "rejecting the null hypothesis" comes from.
So in your example, your hypothesis is that there is a difference between plant species on each side of the hill (but you'd also have to identify what characteristics of the plants you're using to make that observation/assertion.) Your general null hypothesis would be "there is not a significant difference between the plants on each side of the hill." Then your general alternative hypothesis would be that there is a significant difference between the plants on each side of the hill, with respect to the characteristic traits that you assessed and ran your statistics against. But these hypotheses, as posed, are much too broad/general. You'd do better by fielding a null hypothesis like "there is not a significant difference in the average heights/size of the plants on each side of the hill." Then it becomes clear that your statistical measure of significant difference is with respect to plant height/size. Because you could also have a null hypothesis that "there is not a significant difference in the average number of flowers/fruit between the plants on each side of the hill." Then again, it is much clearer what you are measuring and much better bounds the meaning of "statistical significance."