Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
hypothesis versus problem statement
Message posted by kandyce on November 23, 1999 at 12:00 AM (ET)
question: are hypotheses formulated only in experimental research? my design is more exploratory--should i therefore ask a question instead of formulate a hypothesis?
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: hypothesis versus problem statement
Message posted by JG on November 24, 1999 at 12:00 AM (ET)
Hypothesis simply means assuption and in science that is all we ever have. Certainty is a theological and not a scientific concept. In science everything can change at any time. For instance, Einstein suggested that Newton's laws are a special case of more general laws and certain esoteric observations - none of which could have been done or even imagined in Newton's time suggest that Einstein may be right. Some recent data suggests that even Einstein's ideas may need to be modified when looking at the universe as a whole.
Re: hypothesis versus problem statement
Message posted by Matt Frank on November 24, 1999 at 12:00 AM (ET)
In the initial phases of determining what the relationships are between certain variables, before good theory has evolved, it is good science to do exploratory data analyses on any data collected. To that end, by all means, if there is no good theory for your area of interest, you should frame the issues in the broadest possible terms, and let the data tell you what the relationships are. Later, you can conduct hypotheses tests on other data, from the same, or similar populations, to determine the exent and generalizability of the relationships you may find in the exploratory phase. There are several books on exploratory data analyses.
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.