• Publications

Statistics for Self-Marketing2

Pre-Test selection for Self-MarketingIn the first section of ‘Statistics for Self-Marketing’, we saw how to find significance quite everywhere out of big and irrelevant data combined with ignorance of statistical preconditions. Now it’s getting more sophisticated. We have a clear plan what to proof and searching for a scientific proof.

First you should see the difference to science. The main task is not to clarify the own doubts with the awareness of a fragile and unknown truth, but to do a business.

Unfortunately big parts of science follow this strategy either by individual self-marketing through ‘portfolio education’ or by institutional lecture marketing through the expansion of individualized scientific goals and fees.

An indispensable skill today is the capability to produce significant outcomes following a specified strategy. The statistical procedure is quite vice versa of finding universal significance. The dominant variable out of independent data is changed from relation to the relevant main unit. With this formulation, you easily see the statistical misconception – data is not independent any more by defining a guiding relevance. First you should made up your mind about the minimal size for significant outcomes. Then you multiply the minimal numbers of experimental trial and start the sessions. While running all sessions simultaneously, pre-tests in all experimental trails check the quality. With these pre-tests, all trials developing in a wrong direction can be identified. Now a bit creativity is needed to find good scientific reasons for elimination of these so-called special tests. In the end, you have an appropriate number of well evolved experimental testing. No doubt, the scientific outcome can hardly be reproduced, but this is a minor risk in times of forced publication, especially if testing is complex and expensive. No journal will insist on proofing the vast laboratory data set with an international research project, located around the world.

image:

PreTestSelection by dundotcan

S2n – ScienceSocial.net – ISSNĀ 2366-0104
Sharing Sciences – Science Blog – Science Features – Science Papers – with a fair share

share
(Visited 263 times)

Leave a Reply

  • S2n - ISSN 2366-0104