What do you think about the effect size v. significance test debate: should we stick with significance levels, or replace them by effect size indices and confidence intervals?
What will be an ideal response?
There is currently quite a debate about this in the statistics community. What pretty much everyone agrees on is that it is not sufficient to rely only on significance levels, as they
are determined in part by sample size and therefore are a poor guide of the strength of a relationship. Therefore we should in all cases calculate an effect size measure as well as
a significance test. Some researchers say that, in part because significance tests are so often misinterpreted as effect size measures, and because of the arbitrary nature of the cut off points, it would be better to not use significance tests at all. Most researchers, however, feel that they still have a useful role to play alongside effect size measures.
You might also like to view...
The culture in which an individual is raised shapes gender roles
Indicate whether the statement is true or false
Select the type of research methodology which is most appropriate for the question "What are the characteristics of schools that have high morale?"
A. Ethnographic research B. Correlational research C. Historical research D. Survey research
Which of the following philosophers saw play as exuberant or extra energy?
A. Kant B. Locke C. Schiller D. Dewey
Jennifer was born 7 weeks prematurely. At that time she weighted 4.3 lbs and developed respiratory distress syndrome. Six months later she still had signs of her premature complications and was frequently hospitalized for bronchopulmonary dysplasia (BPD). The BPD is related to problems that began during this period:
a. Perinatal b. Natal c. Postnatal d. All of the above