I have not forgotten about you all. Unfortunately, I have been studying two subjects which are not worth blogging about: Linear Algebra and C. I’m still in the middle of them, but I can see light at the end of the tunnel. They are both leading to interesting projects: quantum computing and I will be cracking DES soon. Do not despair!
Here is a very interesting look at the split-brain experiments by CGP Grey: https://www.youtube.com/watch?v=wfYbgdo8e-8
Always a good feeling when wrapping up a book or a course. This is the first of my course reviews. Apparently, I take them as a hobby.
HarvardX: PH207x Health in Numbers: Quantitative Methods in Clinical & Public Health Research
Available on edX.
What it teaches:
Harvard’s introductory course to biostatistics, which is statistics that focuses on public health models and analysis, and epidemiology, which is an investigation of the causes of diseases and conditions and their correlations. Epidemiology uses various case studies as frameworks to set up valid analysis for biostatistics.
Harvard originally taught these subjects in two different courses, and this particular course is a fusion. The course also teaches how to use the
statistical software package Stata for assistance in your analyses. It also introduces guest lecturers who lead the class in critiquing published case papers.
PH207x was planned as a 12 week course. If you already have a knowledge of statistics, you can probably fly through it. Most of the online homework questions refer to basic calculations and concepts.
Input[1]:
Basic fluency of algebra. Computer literacy.
Output:
Ability to understand basic statistics [2].
Ability to critique papers on epidemiological studies.
Ability to assist in biostatistics and epidemiology data analysis.
Ability to use basic Stata analysis.
Mixers:
Principles of Biostatistics (co-written by one of the lecturers, reviewed earlier in this blog)
Stata software package. Version 12 was used in the class, 14 was used for this review. The differences are mostly trivial[3].
Why You Should Take This Course:
First, if you have no prior knowledge of statistics, learning the subject is very informative. Statistics allows you to account for variability and probability, and to make predictions. The field of statistics gives you a set of tools which can be just as interesting and valuable as those acquired from calculus. Second, you will be able to call bullshit on many claims given by people not just on statistics, but on health and nutrition as well. You’ll know what makes a valid analysis and when to ask for evidence. Third, it was fun to see what a proper statistical program can do.
Do I recommend it?
Yes. The two instructors are well-versed and communicate the material of their lectures very well. The textbook is lucid and not overly formal, which is fine for applied mathematics. I had no problem understanding the material. For a very small investment: algebra and computer literacy, you may acquire many useful rewards. Most especially, statistical analysis is a tool that isn’t just limited to public health, but a wide array of sciences, and I highly recommend learning it.
[1] What skills you need to properly process the course.
[2] The course will bring you up to linear and logistic regression, then recommends that you take another course in regression to progress.
[3] A calculator was relocated, and pseudo-random seeds will not be the same.
Scattered in my readings are interesting historical tidbits. I enjoy history, because it allows you to learn from other people’s mistakes. I don’t have a favorite place or time in history, but I like the history of inventions and ideas. How people approached and devised a solution to a problem, either scientific, or military, or even just in general. I also enjoy being corrected (or correcting others.. heh) in misapprehensions about history.
So here’s a question: how old is life insurance?
I shan’t lead you on: 2600 years, give or take. The Romans and Greeks had guilds which a person would pay into, and they would take care of funeral expenses and stipends for the family of the deceased. The concept of insurance in general goes back to the Code of Hammurabi, which gave a form of maritime insurance to those who take a loan and lose a ship at sea. If you think about it, the idea of property insurance is just a redistribution of risk from taking out a loan. Loans are a result of an individual or organization having a surfeit of money and wishing to invest. Money itself is a shorthand for transfer of goods and labor, which comes from a settled, agricultural society[1]. So property insurance itself isn’t that complicated, it just naturally arises from money. Life insurance is a little more complex, but it too naturally arises from property insurance.
[1] I neglected to mention trade as a factor, which is a whole topic in itself: measuring investment against the risk of maritime loss. I do not exaggerate when I say that trade is perhaps one of the hugest factors in world history, and is often hidden as a motivation in high school textbooks.
There’s an interesting game called Katamari Damacy where you, the Prince, are charged to make katamari, which is Japanese for “a jumble”, “mass”, or “cluster”. You roll small things which snowball into larger ones and thus into huge katamari and you present them, under a deadline, to The King of All Cosmos, who then makes it into a new star.
That’s if you make the deadline. The King is quite cross with you if you don’t make it. One of those little speeches goes (from memory).
“There you are, always in the middle of things, never quite finishing anything!!!
Blogs are like being in the middle of things. So rarely do they start from a proper beginning, with a proper introduction. At least, perhaps the good ones do. People like to talk about themselves, which is often Not Very Interesting. I’d rather talk about other things and inform the reader prior to when they need information about how I view things or operate. Here and there I’ll leave clues perhaps to who I am, but its really not that important.
So this first post is a review of the book Introduction into Biostatistics by Marcello Pagano and Kimberlee Gauvreau. This is the second edition. I suppose I should inform you[1] that I have a habit of studying textbooks and courses. There is an excellent resource for courses called EdX, with top level university courses. Introduction to Biostatistics (henceforth abbreviated to ItB) is the textbook for the
course, which is the first course of the archives.
I prefer to go through the entire textbook and do all of the problems before taking a course. Hermione would be proud. I tried reading the textbook and doing all of the problems with the course, but some instructors jump around the textbook a bit, which is fine for them but distracting for me.
I do not have a medical or statistics background. I do have a Baccalaureates in Electrical Engineering. So my mathematics background is multivariable calculus, ordinary differential equations, probability, discrete math (which I enjoyed the most), and some linear algebra. Probability is taught to EE’s as a component to separating noise in radio communications, but we never quite got to that part in the course, which makes me cross. We also never got around to using ODEs in electronics as well. *grumble emoji would go here*
Prior to reading this textbook, I knew what means, medians, and modes were, but I couldn’t rightfully tell you what a p value was. I am absolutely delighted by this textbook, which serves as an introduction to a large set of new and fascinating tools. The book is composed of 22 chapters, each ranging from 20 to 40 pages long, and ends with 10 to 20 questions. I never had a problem understanding the material, and only towards the end, on the topic of linear regression, is when the text started to gloss a bit over information, which is fine for an introductory textbook. Linear regression is also where I started leaning on software packages for analysis, since performing linear regression on a data set of 100 entries seemed unnecessarily tedious. For most of the book I was fine using online web pages for making graphs and lengthy calculations, though I tried to do some things by hand for understanding’s sake.
Introduction into Biostatistics uses real information and has many references and sources, which is a bonus. The real data made me feel like I was doing actual work and not playing with made up numbers. I did not have access to an answer key (or even half a key as most textbooks provide), but at no point did I feel lost. I might have an unresolved error here or there, but the nature of the problems was that I felt I could always answer them with the information given in the chapter. The book expects you to use either Minitab or Stata with it and does not teach you how to use these packages, so if you’re not computer savvy, you may have some confusion.
Overall I am very happy with what I learned and the material wasn’t difficult at all. I consider statistics to be a very welcome addition to my toolbox of problem solving. I look forward to taking the EdX course, which combines biostatistics with epidemiology and Pagano is one of the instructors of the course.
[1] I’m thinking of a name for the reader, Gentle Reader is taken by Miss Manners. Cecil Adams of The Straight Dope used the Teeming Millions to refer to his readership. I’ll think about an original name, but let’s keep the ball rolling. Suggestions are welcome.