Statistics Public Forums

Feedback on Resources: Differences btwn textbooks?

memilanuk
Aug 09, 2014
ReplyFlag

What are the differences between the three texts? Other than a bit less material in the Adv. High School version (no multiple linear regression), the ToC appears very similar for all three. Why the different editions?

David
Aug 10, 2014
ReplyFlag

Thanks for the question! Just to ensure there is no ambiguity: these are different *textbooks*, not different editions. All three textbooks will be supported and maintained moving forward.

Here are a couple quick notes on the differences between the three textbooks:

Introduction to Statistics with Randomization and Simulation uses randomization and simulation to introduce the concepts of inference in Chapter 2. This is a very different approach to the intro to inference provided in OpenIntro Statistics' Chapter 4. To match this new type of introduction, the categorical and numerical inference chapters have also been reversed and adapted to better suit this ordering.

The Advanced High School Statistics textbook is customized to focus on AP Statistics. With this new textbook, an AP Statistics teacher can be confident that all required topics are covered, since some such topics (e.g. stem-and-leaf plots) are not covered in other OpenIntro textbooks.

Best,

David

Here are a couple quick notes on the differences between the three textbooks:

Introduction to Statistics with Randomization and Simulation uses randomization and simulation to introduce the concepts of inference in Chapter 2. This is a very different approach to the intro to inference provided in OpenIntro Statistics' Chapter 4. To match this new type of introduction, the categorical and numerical inference chapters have also been reversed and adapted to better suit this ordering.

The Advanced High School Statistics textbook is customized to focus on AP Statistics. With this new textbook, an AP Statistics teacher can be confident that all required topics are covered, since some such topics (e.g. stem-and-leaf plots) are not covered in other OpenIntro textbooks.

Best,

David

David
Aug 22, 2015
ReplyFlag

Just an update here that there is now a brief description when selecting a textbook for the first time. Hover over the textbook covers on the following screen to get a short description:

https://www.openintro.org/stat/textbook.php?stat_book=reset

https://www.openintro.org/stat/textbook.php?stat_book=reset

xallanmillerx
Sep 05, 2015
ReplyFlag

I'm also interested in the difference between OpeinIntro Statistics and Introduction to Statistics with Randomization and Simulation. I wonder what the advantages or disadvantages are of "using randomization and simulation to introduce the concepts of inference in Chapter 2" of the latter. What's the rationale for this alternative approach?

David
Sep 05, 2015
ReplyFlag

*** Upside ***

A significant motivation for randomization and simulation is to improve the understanding of core statistical ideas. The belief is that randomization and simulation make it easier for students to internalize the meaning of a p-value and other core topics. For instance, rather than some formula dictating the description of the null distribution, the null distribution is something students can think about more fundamentally; it can be generated from a simulation, and we go through that simulation process many times in Chapter 2.

The use of randomization and simulation also allow students to not get tripped up in the math while learning inference concepts. Once these core ideas are reasonably cemented around the end of Chapter 2, ISRS brings in the CLT and uses the case studies in the chapter to show the CLT in action. This sets the stage for the more traditional methods in Chapters 3-6. Having these traditional methods rounding out the textbook also mean it should hit all the prerequisites for future courses in other subjects.

*** Downside ***

I think the biggest downsides is that it's something new to teach. This means there are fewer resources available, and it will probably be a bit rougher the first time round. ISRS also sticks probability in the appendix, which may be annoying for courses that have probability as a core component. (Probability exercises would also need to be grabbled from OpenIntro Statistics, but since it's free and open source, students could just print out the dozen or so pages of exercises.)

*** Aside ***

Randomization and simulation with proportions are less controversial than the percentile bootstrap, which is less robust than is commonly believed. For example, the t-interval and t-test give more reliable results when the sample size is about 30 or less. For this reason, the bootstrap isn't highlighted as prominently in ISRS as in some alternative intro statistics textbooks that don't touch on these percentile bootstrap challenges.

A significant motivation for randomization and simulation is to improve the understanding of core statistical ideas. The belief is that randomization and simulation make it easier for students to internalize the meaning of a p-value and other core topics. For instance, rather than some formula dictating the description of the null distribution, the null distribution is something students can think about more fundamentally; it can be generated from a simulation, and we go through that simulation process many times in Chapter 2.

The use of randomization and simulation also allow students to not get tripped up in the math while learning inference concepts. Once these core ideas are reasonably cemented around the end of Chapter 2, ISRS brings in the CLT and uses the case studies in the chapter to show the CLT in action. This sets the stage for the more traditional methods in Chapters 3-6. Having these traditional methods rounding out the textbook also mean it should hit all the prerequisites for future courses in other subjects.

*** Downside ***

I think the biggest downsides is that it's something new to teach. This means there are fewer resources available, and it will probably be a bit rougher the first time round. ISRS also sticks probability in the appendix, which may be annoying for courses that have probability as a core component. (Probability exercises would also need to be grabbled from OpenIntro Statistics, but since it's free and open source, students could just print out the dozen or so pages of exercises.)

*** Aside ***

Randomization and simulation with proportions are less controversial than the percentile bootstrap, which is less robust than is commonly believed. For example, the t-interval and t-test give more reliable results when the sample size is about 30 or less. For this reason, the bootstrap isn't highlighted as prominently in ISRS as in some alternative intro statistics textbooks that don't touch on these percentile bootstrap challenges.

apbray
Sep 06, 2015
ReplyFlag

To follow up with some further reading, Nathan Tintle from Dordt College has written quite a bit about the rationale behind using randomization and simulation. Tim Hesterberg will has also written a shorter version of the paper listed above that will be coming out soon in The American Statistician.

Tintle NL, Rogers A*, Chance B, Cobb G, Rossman A, Roy S, Swanson T, VanderStoep J (2014). â€śQuantitative evidence for the use of simulation and randomization in the introductory statistics courseâ€ť Proceedings of the ninth International Conference on Teaching Statistics. Flagstaff, Arizona.

http://iase-web.org/icots/9/proceedings/pdfs/ICOTS9_8A3_TINTLE.pdf

Tintle NL, Chance B, Cobb G, Roy S, Swanson T, VanderStoep J â€śUsing simulation-based methods in introductory statistics courses to impact the entire undergraduate curriculum: getting out of the bogs to reach the pinnacle of statistical thinkingâ€ť The American Statistician. Upcoming.

Tintle NL, Rogers A*, Chance B, Cobb G, Rossman A, Roy S, Swanson T, VanderStoep J (2014). â€śQuantitative evidence for the use of simulation and randomization in the introductory statistics courseâ€ť Proceedings of the ninth International Conference on Teaching Statistics. Flagstaff, Arizona.

http://iase-web.org/icots/9/proceedings/pdfs/ICOTS9_8A3_TINTLE.pdf

Tintle NL, Chance B, Cobb G, Roy S, Swanson T, VanderStoep J â€śUsing simulation-based methods in introductory statistics courses to impact the entire undergraduate curriculum: getting out of the bogs to reach the pinnacle of statistical thinkingâ€ť The American Statistician. Upcoming.

Alexasma
Sep 24, 2017
ReplyFlag

hello, i need a book or a file/ website that explains the forcast qualitative method "historical analogies" i really need it

To add a comment or subscribe, please sign in or register.

Your User ID will appear with your posts.