Sometimes, we have, in research, some good findings, but we can not compile or organize it, for lack of adequate statistical knowledge. Thus, how essential, you think, to be competent in statistics to advance our research? If so, how to acquire it?
Statistics helps us turn data into information; that is, data that have been interpreted, understood and are useful to the recipient (For more information Please See: https://www.mheducation.co.uk/openup/chapters/9780335227242.pdf
If you are dealing with data I think it would absolutely essential to have a good knowledge of statistics. If you and your team are not confident in your statistical skills I highly recommend finding a statistician as a collaborator. I also recommend taking a few stats classes at the university and Coursera, but you will definitely want to make friends with a statistician. Compiling and making sense of information is not trivial.
I agree with Jill, it is very important to have someone with at least a masters, and preferably a Ph.D. in statitics on your team. Many statistics departments offer consultancy services to researchers. It is very important to always check with them and ask them to review and suggest the statistical analyses for answering your research question.
Thank you Garcia for sharing your experienced views. However, my aspect in the question was how better it would be to have adequate statistical knowledge among the researchers themselves. Just because, I experienced some of the researchers in my life, who had quite sound statistical knowledge, even better than many professional statistician, which I hardly find in my surrounding at present. So, my question was how a researcher can significantly contribute if he himself is well-learned in statistics? And, how the researcher can acquire that required statistical skills?
Hello, Imdadul. I would say that statistical knowledge is vital among the researchers themselves. In my experience, even having an acceptable knowledge of descriptive and inferential statistics, as well as experimental design, I always consult with experts in statistics. And not only after obtaining the data; sometimes you have to take statistics into account even before designing your study. For example, if you wish to compare different measures between groups, there are minimums of subjects that you have to include in your groups in order for the statistical tests to be valid. As to where to get more training in statistics: 1) Enroll in courses offered by statistics or science departments, check if your institution has special courses for researchers, and if not, request one; 2) There are many MOOCS, offered by such platforms as Coursera and EdX, where you and your team may find what you are looking for; 3) Contact another university and check if they offer courses for researchers; nowadays this gets easier with online learning.
Thank you for the handful suggestions. I now completely agree with your views.Thant is, we should have a expert statistician in our team, but we should have required skills in statistics as researcher. Everywhere in the research will be different if the researcher himself has erudite knowledge in statistics.
It depends on the field of study. If you have to deal with data it is important. but if you have to build/propose a model knowledge of statistics is not required.
It is easy to say that a researcher must have good statistical skills. However, much of recent social science research is crap (see replication crisis in Psychology) partly because statistics half wits were using century old statistical procedures, following recipes from bachelor grade courses (e.g., if your residuals are not normal use non parametric and the whole NHST nonsense).
I'd rather have one dedicated statistical expert in a team than five people who have worked through a bachelor level course book. It is important, though, that such an expert:
- is familiar with research practice in the particular domain
- is consulted upfront (Google: Ronald Fisher post mortem)
I am one hundred percent agree with your opinion. In fact, it is very easy to say to have adequate competency in statistics, but I experienced many statisticians with lack of proper practical orientation in statistics regardless the researchers. Reversely, there must have few researchers, who possess extraordinary skills in applying their the statistical knowledge practically, which is really tough and obviously appreciable.
I think the result can be valid only after the suitable statistical test is applied. Also, the acceptance of any article by various journals includes the application of a statistical test. Statistics help us in understanding the data and also help us in identifying the result and mistakes.
Statistics helps us turn data into information; that is, data that have been interpreted, understood and are useful to the recipient (For more information Please See: https://www.mheducation.co.uk/openup/chapters/9780335227242.pdf