Monday, 25 May 2015

Research Hypothesis (RH) versus Null Hypothesis (NH)


    The Null (nothing) hypothesis indicate no relations or we can say it states equality (H0:μ12).As we are not aware of literature and theories, so null is a good point to start with. It states that there is no relation between two variables. For example, there is no difference between returns of spot and futures prices of Reliance Ltd (i.e. both market provides equal returns).



    The Research hypothesis (or alternate hypothesis), in contrast, states certain relationship between variables, based on some theory or existing literature. Thus, RH are statements of inequality. For example, returns of spot market are greater than that of futures market for Reliance Ltd (or vice versa, based on some literature).

    If we observe in literature, we see that generally NH are not reported for brevity (though they are reported sometimes), whereas RH are documented explicitly and lucidly. NH are for populations and RH are for samples. The main purpose for this write-up is to let researchers know that whatever we want to show, prove, disprove, or propose (based on literature review) all come under research hypotheses or alternate hypothesis.

Taken from the book: 100 Questions (and Answers) About Statistics  by  Neil J. Salkind.

Salkind, N. J. (2014). 100 Questions. Los Angeles: SAGE Publications Inc.