-
Research, Statistics and Mathematics Educators in Nigeria: Effect Size Perspective by K.A. Adeniji (2016)
5 Comments-
Research paper which argues against the over-reliance on null hypothesis significance testing (NHST) in statistical research. For some background on how the null hypothesis test works, click here: https://www.thoughtco.com/definition-of-null-hypothesis-and-examples-605436.
-
This is a huge problem. In part it derives from an incomplete comprehension of Statistics and I use the entire subject intentionally. Many ‘master’ (read memorize) formulas and do not grasp (due to poor teaching) the underlying philosophy of this aspect of the reasoning process, i.e., Statistics. It’s not grasped because its not taught. The very teaching of Statistics is NOT taught. So someone who has a somewhat of a functional grasp of how to apply different tests of significance ends up ‘teaching’ a subject. Let me stop. Like this article.-
I would only add – in regards to the philosophy point – that it also affects how researchers and academics argue their findings. Without an understanding of other metrics like effect size, there’s a lot of “distinction without a difference.” A result being “statistically significant” might not mean anything at all.
-
Unpopular perspective: Truth be told given the segmented nature of the Western Canon and all of its offshoots regardless of labels, and the reductionist character of the scientific method in all of its empiricism, even if they were using say multiple regression analysis and the multiple coefficient of determination (R square) in terms of hypothesis testing, they aren’t asking worthwhile questions. Are we seeking answers to the right question(s)?
-
You’re in a better position to answer that than I am Okunini. But seeing as how so much of economics relies on regression analysis and even more complex mathematics, and most of their results are useless, my guess is that researchers in the West are not asking the right questions.
-
-
-
-
-
