Alrighty! If you know half the gibberish I have written above, then welcome to the world of statistics and do read on ...
The last four weeks have been an exercise in frustration as I had to redo my research tests over and over and over again. I put the fault squarely on the shoulders of my statistics lecturer who taught us only parametric tests and did not teach us non parametric tests for ordinal data! (If you are still wondering what I am going on about, just google it!) In fact, we used parametric tests on ordinal data in her class which is a big NO NO for pure statisticians! What kind of education standards and lecturers are we paying for these days? Grrrr ....
Or perhaps I should put the blame on the researcher whose published work used parametric tests for ordinal data! The publisher who published his work should be blamed too! How can they publish work that use the wrong kind of tests??? What happened to data and research integrity? Makes me wonder if the research companies we hire to do consumer research actually pays attention to data integrity?!
And the worst of it all is that after all the tests and retests, my findings are still the SAME!!! Blardy hell .... four weeks of pure torture. Running tests after tests after tests and looking at the endless data. I am almost cross-eyed by now! And the results? It's the same! What the heck is wrong with this scene? Geee, can't the higher ups in the learned society simplify statistics? If it ain't broken, don't fix it! And don't invent something new to confuse the issue!
Goodness ... I don't think I want to see another Friedman or Kruskal-Wallis or Mann-Whitney or Wilcoxon or Spearman or ANOVA or t-tests in this lifetime!
Begone with thee!
Subscribe to:
Post Comments (Atom)
2 comments:
Hi, I found your blog by accident. I feel the same way too! I never want to do statistical tests in this lifetime as well! But just a friendly update, I read somewhere that parametric tests can be done on ordinal data as long as the sample size is large enough.
Thanks for dropping by.
"I read somewhere that parametric tests can be done on ordinal data as long as the sample size is large enough."
I read about that too but a lecturer friend insisted that for a pure research, non parametric tests should be used.
Post a Comment