Re: Languages don't matter. A mathematical refutation

["Jean François Martinez" <darkquark99@gmail.com>]

On Thursday, March 26, 2015 at 4:21:45 PM UTC+1, Dmitry A. Kazakov wrote:
> On Thu, 26 Mar 2015 06:43:01 -0700 (PDT), Maciej Sobczak wrote:
> 
> > Unless you prove that this had no influence on the results in question, I
> > refute the whole of your mathematical proof.
> 
> Yep, when *mathematical* statistics is used, then the burden of showing it
> applicable lies on the author's shoulders. In particular, the software
> metric used for some SW design must be shown to be a random variable. The
> elementary outcomes presented. Their independence explained etc.
> 
> > But more seriously, it would be interesting to flip languages every year
> > and see whether the advantages of Ada hold in terms of better results. Or
> > run two classes in parallel (with random assignments of students to each
> > class) and compare results over the years.
> 
> I doubt the process of random selection of a developing team from a pool of
> "indistinguishable" teams were a good basis for a statistical model of
> software development.
> 



> The bottom line: statistic were applied incorrectly and the results
> obtained carry no prediction power.
> 

Really?  Text books on statistics are choke a full of similar examples and exercises like comparing the proportion of defective products at manufacting plant A and B then checking the null hypothesis before deciding that A's products are better.   In our case we have a random variable and that is group 
quality.  This depends on quality of students but also on group dynamics.

And it is common in this field to assume taht your population is a sample
of an infinitely-sized conceptual population.  Anyway I submitted the mathematical aspects to a professional statistician and he saw no flaw to my
mathodology.  In fact he told me that similar methodologoies were used for 
ensuring fairness at the competitive exam for French engineer schools (ie avoiding a candidate being rejected because he had the misfortune of 
getting a tough examinator).

Then we leave the matermatical aspect of statistics and enter in the world of 
applied statistics like in the studies published by your friendly
bureau of statistics ,or whatever it is named in Germany or Russia, where after having established that the difference is significative ie rejecting the null hypothesis the statistician has to apply knowledge of the subject investigated (economics, scholar system, biology) and consider if there are other factors that could have influenced.  That is why for instance I was careful to note that students were of the same university:  if C students had been of University of Northern Iowa and Ada students from MIT, Standford or similar universities whose students have far higher SAT scores then that factor could have been the cause. At this point a statistician would do further studies in order to isolate its influence.  I admit there are points I did not consider (Ada students being of newer genetrations that had got pre-university experience) and others I did not discuss because investigating them would need to bother professor McCormick. 


> (Nonetheless, we know that Ada is a better language, anyway...)
> 


It is not that we know.  It is.  :-) But there are people who would be unmoved even if we could present them the kind of study I dream of  with several universities, alternating C and Ada in order to elminate the generation effect, first and last year students in order to eiiminate the slow learnng curve and so on 

Regards

Jean-François Martinez