Statistics was my least favorite course as an engineering undergrad. I was totally convinced that the entire class was a complete waste of time. Who really uses this stuff anyway?

Ironically, I recently found myself teaching a Statistical Analysis of Laboratory Data course for the American Chemical Society (ACS) and was shocked to discover that I thoroughly enjoyed the topic.

When the opportunity to teach statistics first presented itself about a year ago, I was very skeptical. The main reason for my skepticism was that I always considered “statistics” to be sort of nebulous and non-practical thing - like turbulence models in CFD simulations, or the weather channel (I am convinced that I can predict weather better then the weather people).

My apologies, if most of you are offended by this statement, but being an engineer, I can only believe in things that are practical. Things that work - make sense to me. Statistics, the weather channel, and turbulence have never provided me with any pragmatic value…well until recently.

Funny thing is that I have studied and enjoyed many things that might be considered much more “theoretically fluffy” than the weather prediction. Two of my favorite courses in grad school were General Relativity and High-energy Astrophysics. Talk about “out-there” concepts. But what I learned in these courses helped me understand the world around me; I could visualize the concepts and relate them to practical experience in everyday life. This was largely due a wonderful UCSD physics professor (thank you Dr. Fuller!) who made these courses pragmatic and relevant to me.

On the other hand, when statistics was taught to me as an undergrad engineering student with no on-the-job experience, it was nothing more then a jumble of confusing probability equations that seemed like a complete waste of time. When I started doing research in grad school,  I became more and more acutely aware of how much “bad” data is out there in the scientific community. So many times I was advised by experienced researchers not to use a set of data from a certain university or research group, because that university or research group was known to produce data that was unreliable and/or inaccurate. Their data was published in peer reviewed journals, yet there was no standard for quantitatively measuring its uncertainty?

Eventually, I realized that there are standards that ALL researcher should adhere too when reporting research data in any scientific field. Those standards, which describe the reliability of any data set, are identified by well-established statistical analysis guidelines. After teaching this course, I was shocked to realize how untrained most researchers (including myself) are in reporting their scientific data.

I think back to all of the technical papers I read through in grad school and during my work in R&D, all published in peer-reviewed journals…In so many cases the authors barely even mentioned their sources of error, yet alone described how they eliminated systematic errors, identified statistical outliers, or optimized their number of measurements!

If you are not appalled by the above paragraph, you need to re-visit statistics. Great place to start is: Data Analysis for Chemistry by Hibbert and Gooding. This short and to-the-point book approaches statistics in a very practical way and is great for researchers in any field. Now, I admit that in the past I was very guilty of ignoring statistics when publishing my own data. But after my recent revelation, you can be sure that any data I report in the future will be thoroughly combed in every direction with a statistical comb.

…And, who knows, eventually I might actually start to believe in turbulence modeling…

Cheers,

Masha

My downloadable courses on Statistics will be up by the end of August 2011 on the ACS website: http://www.proed.acs.org/courses/search_results.cfm

For a bit of completely shameless self-promotion - here is what a student wrote to ACS about my most recent Statistics course:

“The course was GREAT!!!!!!!! Probably one of the best I’ve attended in the last 4 years. Very valuable, USEFUL information and laid out in such a way that’s completely understandable to all. The online format was very good and allowed great flexibility in terms of attendance. Questions were always answered quickly. The handouts and example sheets will be invaluable going forward. This course was definitely worth the money spent and I highly recommend it!

Masha is really well versed in the course content and provided a great series. Please thank her!”

-Eden

Associate Director, QA

One of my favorite things about following an entrepreneurial path is the spontaneity and unlimited creativity that one can subject his or herself to while on this road.

For those of us who like unbeaten paths, who enjoy working 20-hour days one week and laying around on the couch “decompressing” the following week,  for those who love adventure, are not too keen on routines, for those who are annoyed by fear – this can be a great path to take.

So it is fun for me to announce that MVP is changing its entrepreneurial direction, in order to become more focused on the educational aspects of science and engineering. Since computer modeling will only become more and more integrated into future R&D processes, MVP will still provide training on the Best Practices for computer modeling, but will now focus on providing learning opportunities and platforms for, not only current engineers and scientists, but for the upcoming R&D workforce as well.

Future researchers need to know how to create virtual laboratories and experiments that can provide high-quality scientific data. Raised in the world of video games and social media – they deserve to know how to utilize their natural ability to “be one” with the virtual world in their work as scientists and engineers. Our high schools and universities leave much to be desired, when it comes to this area of STEM education.

And so, the new goal of MVP Modeling Solutions is - to provide learning opportunities that would allow all researchers, whether they are the Baby Boomers or the Millennials, to utilize the virtual space of computers and the web, to conduct real scientific experiments that will allow us to keep discovering the world around us and to keep on creating it.

Thoughts? Please comment below!

Cheers,

Masha

Take a look at the new MVP website and courses offered!