Statistical Computing and Data Science in Introductory Statistics

Students in the sciences, economics, social sciences, and medicine take an introductory statistics course. And yet statistics can be notoriously difficult for instructors to teach and for students to learn. To help overcome these challenges, Gelman and Nolan have put together this fascinating and thought-provoking book. Based on years of teaching experience the book provides a wealth of demonstrations, activities, examples and projects that involve active student participation. Part I of the book presents a large selection of activities for introductory statistics courses and has chapters such as ‘First week of class’- with exercises to break the ice and get students talking; then descriptive statistics, graphics, linear regression, data collection (sampling and experimentation), probability, inference, and statistical communication. Part II gives tips on what works and what doesn’t, how to set up effective demonstrations, how to encourage students to participate in class and to work effectively in group projects. Course plans for introductory statistics, statistics for social scientists, and communication and graphics are provided. Part III presents material for more advanced courses on topics such as decision theory, Bayesian statistics, sampling, and data science.

Download Full-text

Teaching stats for data science

10.7287/peerj.preprints.3205v1 ◽

2017 ◽

Author(s):

Daniel T Kaplan

Keyword(s):

20Th Century ◽

Data Science ◽

Statistics Education ◽

Introductory Statistics ◽

Applied Statistics ◽

Causal Relationships ◽

Chi Squared

The familiar mathematical topics of introductory statistics --- means, proportions, t-tests, normal and t distributions, chi-squared, etc. --- are a product of the first half of the 20th century. Naturally, they reflect the statistical conditions of that era: scarce, e.g. n < 10, data originating in benchtop or agricultural experiments; algorithms communicated via algebraic formulas. Today, applied statistics relates to a different environment: software is the means of algorithmic communication, observational and "unplanned" data are interpreted for causal relationships, and data are large both in n and the number of variables. This change in situation calls for a thorough rethinking of the topics in and approach to statistics education. This paper presents a set of ten organizing blocks for intro stats that are better suited to today's environment.

Download Full-text

Teaching stats for data science

10.7287/peerj.preprints.3205 ◽

2017 ◽

Author(s):

Daniel T Kaplan

Keyword(s):

20Th Century ◽

Data Science ◽

Statistics Education ◽

Introductory Statistics ◽

Applied Statistics ◽

Causal Relationships ◽

Chi Squared

The familiar mathematical topics of introductory statistics --- means, proportions, t-tests, normal and t distributions, chi-squared, etc. --- are a product of the first half of the 20th century. Naturally, they reflect the statistical conditions of that era: scarce, e.g. n < 10, data originating in benchtop or agricultural experiments; algorithms communicated via algebraic formulas. Today, applied statistics relates to a different environment: software is the means of algorithmic communication, observational and "unplanned" data are interpreted for causal relationships, and data are large both in n and the number of variables. This change in situation calls for a thorough rethinking of the topics in and approach to statistics education. This paper presents a set of ten organizing blocks for intro stats that are better suited to today's environment.

Download Full-text