Can a researcher perform statistical analysis? Sure, why not? These days, anyone who can point and click can get a stats package to spit out some results. ;-)
A better question would be can the researcher perform statistical analysis that is appropriate, sensible and defensible. You might enjoy Robert Abelson's book "Statistics as Principled Argument". It discusses some of these issues.
Can a researcher perform statistical analysis? Sure, why not? These days, anyone who can point and click can get a stats package to spit out some results. ;-)
A better question would be can the researcher perform statistical analysis that is appropriate, sensible and defensible. You might enjoy Robert Abelson's book "Statistics as Principled Argument". It discusses some of these issues.
Of course, why not! If you read a little about them, you can perform the basic statistical operations for your study, including the descriptive statistics as well as some of the inferential ones. So it really depends on what types of data you have, the design of your study, and the outcomes\results you want to yield. But you need to be careful though, and if in doubt, you should consult with a more experienced investigator or even a statistician.
Yes, you can. It depends on the questions to be answer. I am sure you can do testing either through parametric or non-parametric tests, however sample size determination is typical one. Some formula exits, you may see to it. It is in book titled Sampling Techniques by Cochran, pl.
yes, I think the researcher should know the statistical methods to analyze his data although he can ask for the help from his students or colleagues. The correct statistical design and analysis will decide the credibility of the research results.
Any researcher can perform statistical procedures in their research, it is not limited to social sciences or medical sciences, it is for all disciplines.
@ Ahmed, You may use "IBM SPSS Sample Power" for accurate sample estimation.
Ahmed, with your second question, I assume your are asking if you can perform the statistical analysis for your study.
In our country, there is no regulatory body the limits statistical job only to "professional statisticians". You will not be usurping anyone's profession, if you are in our country.
In any university through out the world, if a researcher and his/her adviser think and manifested that s/he is capable to perform statistical analysis, even the research panel cannot stop him/her from doing so.
As many statisticians (not all) in RG expressed, there are a lot of "blunderous" statistical analyses in papers even already published in journals, because sometimes, a little learning is a "down voting thing".
Surely any researcher can do the statistics for their research. It even helps to insights that you would not get when having the stats done by someone else.
Before publication, several publishers require a thorough statistical review, just to make sure that the right statistical procedures are used.
A researcher should perform statistical analysis for his / her study so that s/he can have end to end view on the whole research i.e. from research problem to conclusion. In between performing the right statistical analysis can steer to the correct direction of knowledge contribution.
However, I came across 2 groups of researchers not hands on performing the statistical analysis:
1) Senior researchers well-versed in the statistical analysis but due to multiple studies / engagements, s/he delegated to the junior researchers to perform the actual analysis. The senior researchers will only review, endorse or ask the juniors to perform other statistical tests for better analysis / interpretation.
2) Some researchers started off with basic statistical skills but too focus on their subject / contents & paying less attention to statistical analysis. After a while they lost touch on the latest statistical tools e.g. they might know SPSS base-module and might not know newer generation of multivariate data analysis tools like AMOS & SmartPLS etc.
I think is important for a researcher to know why & how to perform statistical analysis so that:
1) not relying on others to do the analysis for you (less handicapped, more competent)
2) if you are hands on, the learning curve is shorter when you need to expose / learn newer generation of statistical tools
3) handy if you are in academic field whereby you will need to supervise your students' theses & they will have questions on statistical analysis & interpretation
4) you can delegate to others to perform statistical analysis but must know how to review & interpret the results