Statistics learning recommendations: Difference between revisions

From Cognito
No edit summary
 
(17 intermediate revisions by 2 users not shown)
Line 1: Line 1:
== Generalities ==


Statistics is a fundamental tool in the social sciences, medical sciences, actuarial science and finance. Aside from being useful professionally, some of the basic concepts of statistics are also crucial to quantitative literacy.
Statistics is a fundamental tool in the social sciences, medical sciences, actuarial science and finance. Aside from being useful professionally, some of the basic concepts of statistics are also crucial to quantitative literacy and making sense of the world in day to day life.


Some of the commonly used conceptual frameworks in statistics are misguided, and expositions of statistics are affected by this. For example, the definition of "statistical significance" is in some ways arbitrary, and is often not useful in practice, but is nevertheless given a central role in many expositions.  
You should be warned that statistics expositions are sometimes conceptually misguided. The focus on ''frequentist statistics'' rather than ''Bayesian statistics'' is arguably an example of this, though this has been [http://lesswrong.com/lw/jne/a_fervent_defense_of_frequentist_statistics/ disputed]. It's important to learn the standard perspectives in order to understand and communicate with people who use them. However, you should be critical in your reading about the subject – if some of the material doesn't make sense, it's possible that this is because it doesn't have good justification rather than because you're misunderstanding something.


When one is learning statistics, one is faced with a two-fold challenge:
Statistical intuition isn't the same as technical knowledge of statistical methods like factor analysis. One can have the technical knowledge without the intuition or the intuition without the technical knowledge. If you're planning on going into a line of work where statistics is used, you need to acquire the technical knowledge, but for general quantitative literacy, reading our recommendations for statistical intuition may be sufficient.


* One wants to learn the language and concepts that are commonly used, independently of how useful they are, so as to be able to understand and communicate people who use them.
== Recommendations ==
* One wants to understand the subject at a deeper level so that one knows which of the commonly used concepts are useful in which contexts.
 
=== Statistical intuition  ===
 
These are non-technical books that illustrate the concepts of statistics with real world examples. Because they illustrate the ideas in context, helpful for building statistical intuition.
 
* [http://www.amazon.com/Naked-Statistics-Stripping-Dread-Data-ebook/dp/B007Q6XLF2/ Naked Statistics: Stripping the Dread from the Data] by Charles Wheelan.
 
* [http://www.amazon.com/Signal-Noise-Art-Science-Prediction-ebook/dp/B007V65R54/ The Signal and the Noise] by Nate Silver. This book is organized around the theme of predicting the future.
 
* [http://www.amazon.com/The-Lady-Tasting-Tea-Revolutionized/dp/0805071342 The Lady Tasting Tea: How Statistics Revolutionized Science in the Twentieth Century] by David Salsburg. This provides an early history of statistics.


== Recommendations ==
=== Frequentist approach textbook ===
 
The most commonly taught approach to statistics is the ''frequentist'' approach.
 
* [http://www.amazon.com/Statistics-Plain-English-Third-Timothy-ebook/dp/B004RM9VSY/ Statistics in Plain English] by Timothy C. Urdan is a lucid book that covers a lot of material in only 200 pages. It's somewhat terse and some readers may find it to be short on examples.
* [http://www.amazon.com/Statistics-For-Dummies-Deborah-Rumsey/dp/0470911085/ Statistics for Dummies], and [http://www.amazon.com/Statistics-For-Dummies-Deborah-Rumsey/dp/0470911085/ Statistics II for Dummies] offer a more leisurely exposition. [http://www.amazon.com/Statistics-Workbook-Dummies-Deborah-Rumsey/dp/0764584669/ Statistics Workbook for Dummies] offers exercises.


=== Bayesian statistics ===


[http://www.amazon.com/Statistics-Plain-English-Third-Timothy-ebook/dp/B004RM9VSY/ Statistics in Plain English] by Timothy C. Urdan is a lucid book that's only 200 pages long. It requires no more mathematical background than high school algebra.  
''Bayesian statistics'' is thought by many people to be a superior alternative to frequentist statistics.


The book does not touch on Bayesian statistics. Chapter 8 of [http://www.amazon.com/Signal-Noise-Art-Science-Prediction-ebook/dp/B007V65R54/ The Signal and the Noise] by Nate Silver gives some history of Bayesian statistics and an exposition of the basics of the subject. (Silver's book is well-worth skimming in entirety for developing statistical intuition.)
Chapter 8 of [http://www.amazon.com/Signal-Noise-Art-Science-Prediction-ebook/dp/B007V65R54/ The Signal and the Noise] by Nate Silver gives the history of Bayesian statistics and an exposition of the basics of the subject.


Some blog posts  
Some blog posts that overlap with the content of Silver's book but that the reader may find helpful for slightly different perspectives are:


* [http://lesswrong.com/lw/774/a_history_of_bayes_theorem/ A History of Bayes' Theorem] by Luke Muehlhauser * [http://yudkowsky.net/rational/bayes An Intuitive Explanation of Bayes' Theorem] by Eliezer Yudkowsky.
* [http://lesswrong.com/lw/774/a_history_of_bayes_theorem/ A History of Bayes' Theorem] by Luke Muehlhauser.
* [http://yudkowsky.net/rational/bayes An Intuitive Explanation of Bayes' Theorem] by Eliezer Yudkowsky.
* [http://lesswrong.com/lw/2b0/bayes_theorem_illustrated_my_way/ Bayes' Theorem Illustrated (My Way)] by Komponisto.
* [http://lesswrong.com/lw/2b0/bayes_theorem_illustrated_my_way/ Bayes' Theorem Illustrated (My Way)] by Komponisto.
For a full technical introduction to Bayesian statistics, [http://www.amazon.com/Introduction-Bayesian-Statistics-William-Bolstad-ebook/dp/B00D8XUJA2 Introduction to Bayesian Statistics] by William Bolstad may be helpful, but the book is very expensive.
==== Programming statistics ====
If you enjoy programming, are looking to learn programming, or will be implementing statistical algorithms in your work, these books may be good choices:
* [http://www.amazon.com/Discovering-Statistics-Using-Andy-Field-ebook/dp/B00HPZ4VVM/ Discovering Statistics Using R] by Andy Field, Jeremy Miles and Zoe Field.
* [http://www.greenteapress.com/thinkbayes/thinkbayes.pdf Think Bayes: Bayesian Statistics Made Simple] by Allen Downey teaches statistics through Python programming.
* [http://www.amazon.com/Doing-Bayesian-Data-Analysis-Tutorial/dp/0123814855/ Doing Bayesian Data Analysis: A Tutorial with R and BUGS] by John Kruschke.
=== Online classes ===
If you find it easier to learn from lectures than from a textbook, we encourage you to check out free online courses. You can try out different lecturers until you find one who you find especially easy to learn from.
* [https://www.coursera.org/ Coursera] offers ~45 statistics courses. Their offerings include [https://www.coursera.org/course/stats1 Statistics One], an introductory course.
* [https://www.edx.org/ edX] offers ~20 statistics courses, including [https://www.edx.org/course/uc-berkeleyx/uc-berkeleyx-stat2-1x-introduction-1138 Introduction to Statistics: Descriptive Statistics], [https://www.edx.org/course/uc-berkeleyx/uc-berkeleyx-stat2-3x-introduction-1533 Introduction to Statistics: Inference] and [https://www.edx.org/course/uc-berkeleyx/uc-berkeleyx-stat2-2x-introduction-1534 Introduction to Statistics: Probability], which are associated with UC Berkeley.
* [https://www.udacity.com/ Udacity] offers [https://www.udacity.com/course/st101 Intro to Statistics] and [https://www.udacity.com/course/st095 Statistics].
* [http://ocw.mit.edu/index.htm MIT OCW] offers an [http://ocw.mit.edu/courses/mathematics/ math courses] that include "Introduction to Probability and Statistics" and "Statistics for Applications."
* Carnegie Mellon University's [http://oli.cmu.edu/ Open Learning Initiative] offers [https://oli.cmu.edu/jcourse/webui/guest/join.do?section=probstat a course on probability and statistics].
Statistics.com offers over 110 [http://www.statistics.com/course-catalog/ statistics courses]. The courses are expensive: $500 for a 4-week long course. We don't have any inside knowledge of the quality of the courses.
===Cartoon books===
The blog post [http://www.r-bloggers.com/the-most-comprehensive-review-of-comic-books-teaching-statistics/ The most comprehensive review of comic books teaching statistics] discusses and reviews a number of cartoon books teaching statistics. These can be ideal for many people who want to learn statistics informally or even for people who want to learn statistics formally and are looking for a way to supplement their main learning source.

Latest revision as of 22:20, 27 June 2014

Statistics is a fundamental tool in the social sciences, medical sciences, actuarial science and finance. Aside from being useful professionally, some of the basic concepts of statistics are also crucial to quantitative literacy and making sense of the world in day to day life.

You should be warned that statistics expositions are sometimes conceptually misguided. The focus on frequentist statistics rather than Bayesian statistics is arguably an example of this, though this has been disputed. It's important to learn the standard perspectives in order to understand and communicate with people who use them. However, you should be critical in your reading about the subject – if some of the material doesn't make sense, it's possible that this is because it doesn't have good justification rather than because you're misunderstanding something.

Statistical intuition isn't the same as technical knowledge of statistical methods like factor analysis. One can have the technical knowledge without the intuition or the intuition without the technical knowledge. If you're planning on going into a line of work where statistics is used, you need to acquire the technical knowledge, but for general quantitative literacy, reading our recommendations for statistical intuition may be sufficient.

Recommendations

Statistical intuition

These are non-technical books that illustrate the concepts of statistics with real world examples. Because they illustrate the ideas in context, helpful for building statistical intuition.

Frequentist approach textbook

The most commonly taught approach to statistics is the frequentist approach.

Bayesian statistics

Bayesian statistics is thought by many people to be a superior alternative to frequentist statistics.

Chapter 8 of The Signal and the Noise by Nate Silver gives the history of Bayesian statistics and an exposition of the basics of the subject.

Some blog posts that overlap with the content of Silver's book but that the reader may find helpful for slightly different perspectives are:

For a full technical introduction to Bayesian statistics, Introduction to Bayesian Statistics by William Bolstad may be helpful, but the book is very expensive.

Programming statistics

If you enjoy programming, are looking to learn programming, or will be implementing statistical algorithms in your work, these books may be good choices:

Online classes

If you find it easier to learn from lectures than from a textbook, we encourage you to check out free online courses. You can try out different lecturers until you find one who you find especially easy to learn from.

Statistics.com offers over 110 statistics courses. The courses are expensive: $500 for a 4-week long course. We don't have any inside knowledge of the quality of the courses.

Cartoon books

The blog post The most comprehensive review of comic books teaching statistics discusses and reviews a number of cartoon books teaching statistics. These can be ideal for many people who want to learn statistics informally or even for people who want to learn statistics formally and are looking for a way to supplement their main learning source.