Newsletter of the Section on Statistical Education Section of the American Statistical Association

Contents of Volume 2 Number 2:

The last issue of this newsletter, the Winter 1996 issue, was mailed in February 1996. It was incorrectly labeled as Volume 2, Number 2. It should have been labeled as Volume 2, Number 1. If you did not receive that issue please contact Carol Joyce Blumberg either by mail or electronically and it will be sent to you. Please include a complete mailing address. You can also access that issue via the World Wide Web (see article on Page 2 of this issue of the newsletter).

The Chicago Statistical Education meetings look great, with something for everyone. Make your plans to attend now, if you haven't already.

A couple of the new initiatives being highlighted this year will be of particular interest to those of us in the field of statistical education, because they'll be bringing ASA's ideas to much broader groups than our current members. One initiative is a practical plan to get the appropriate ASA journals (Chance and STATS) into the public schools. We'll hear more about the details of this in August. The second is the ASA Web page. We are beginning to have enough of the supporting pages up that browsers can obtain quite a lot of information that would have been difficult to obtain as quickly before.

I hope to see many of you in Chicago.

The annual business meeting of the Section on Statistical Section will be held on Wednesday, August 7 at 6:00 p.m.. Please see the final Joint Statistical Meetings program for the location.

Web Version Note: The meeting will be held in the San Francisco room of the Hyatt hotel.

Although those who have attended these annual meetings of Isolated Statisticians in the past are mostly academicians, anyone who feels isolated is most welcome. Please see the final Joint Statistical Meetings program for the meeting location. When the Isolated Statisticians meeting is finished, small groups will be formed for those wishing to go to dinner. For further information, please contact:

Should we try to teach Bayesian statistics to undergraduates? What software programs are being designed to help students learn statistics rather than just do statistics? How can we evaluate the success or failure of new approaches to teaching statistics? These are just a few of the questions that will be addressed in sessions sponsored by the Section on Statistical Education at the 1996 Joint Statistical Meetings to be held on August 4-8 in Chicago.

The Statistical Education sessions kick off with a bang on Sunday afternoon at 4:00 with an invited session on teaching Bayesian statistics to undergraduates. Don Berry and Jim Albert present the Bayesian viewpoint, while David Moore offers reasons to hesitate and Jeff Witmer discusses pros and cons. The middle of the week (Tuesday at 10:30 and Wednesday at 8:30) sees two invited sessions on technology, as George Cobb, Dennis Pearl, Paul Velleman, Bill Meeker, Dennis Wackerly, Steve Cohen, and David Lane all present projects which use software to help students learn about statistics. Sessions conclude on Thursday morning at 10:30 with Joan Garfield, Carol Blumberg, Joel Greenhouse, and Marsha Lovett participating in a panel discussion on questions of evaluating reform projects in statistics education.

Another highlight of the meeting will be Monday's roundtable lunch discussion groups, organized by Roxy Peck. Our Section is sponsoring eight tables on the following topics:

• Using Case Studies Effectively in Introductory Statistics Courses, with Norean Radke-Sharpe

• Benchmark Lessons and the World Wide Web: Tools for Teaching Statistics, with Andrew Schaffner

• Encouraging K-12 Statistical Education, with David Fluharty

• Integrating the Use of Mathematical Software in the Undergraduate Math-Stat Curriculum, with Richard Rossi

• Distance Education in Statistics, with Robert Stephenson

• Statistics for the Twenty-First Century: Technology, Economics, and Approach, with Arnold Goodman

• Computer Demonstrations in Introductory Statistics Lectures: Experiences and Examples, with Linda Collins

• Isolated Statisticians: Evaluation and Tenure, with Tim Hesterberg

Other highlights of the meeting will include an invited poster session on case studies from the Academe/Industry Collaboration Project (Mon. at 2:00) and special contributed paper sessions on trends in Introductory Applied Statistics Courses (Wed. at 2:00) and on Isolated Statisticians in Education (Wed. at 10:30). Still other special sessions of interest include the College Bowl (Tues. at 10:30 and Wed. at 10:30), Special Presentation by the winners of the Undergraduate Data Analysis Contest (Thur. at 8:30), and a Special Panel Discussion on effective teaching in industry and academia (Wed. at 8:30). A plethora of contributed paper sessions and poster sessions round out the many ways in which innovative ideas regarding the teaching of statistics will be exchanged at this meeting. You can consult the May issue of Amstat News for preliminary program information on dates and times of these sessions as well as the titles and speakers for individual presentations.

I thank the many people who volunteered to organize sessions or give presentations for the Statistics Education session. My only regret as Program Chair is that the large number of excellent abstracts that I received left me no choice but to schedule several sessions at the same times. I apologize in advance to those who will bemoan the difficult decisions that this dilemma poses.

I encourage you to attend the Chicago meeting and am confident that you will encounter many very fine sessions on Statistical Education.

For further information contact:

This year is the tenth anniversary of the American Statistics Team Project Competition. The competition is held for three grade categories: 4-6, 7-9, and 10-12. It is a wonderful opportunity for students to pursue a question they find interesting through all phases of the investigation: design, data collection, analysis, conclusions, and presentation. To be successful, students must integrate skills from several disciplines, including English and mathematics.

This year a number of exciting projects were submitted in each age category. For grades 4 to 6, there were two winners. One team wanted to determine whether stores (all associated with the same major grocery store chain) located in areas of their county with lower per capita income sold a higher or lower percentage of recycled paper towels than those stores located in higher per capita income areas of the county. The students carefully explained the difficulties encountered in collecting the data, presented numerous graphs, and used a sign test to help them conclude that there was no significant difference in the percentage of recycled paper towels sold in low and high per capita income areas. The other winning team questioned whether ears grow throughout a person's life and, if so, whether the growth rate differs with gender. After measuring 340 ears of people ranging in age from one to 86, they formed age groupings of 0 to 9, 10 to 19, etc. Using several plots, including a dot plot and a line plot, they concluded that ears do grow throughout life. Males' and females' ears grow about the same rate until the age of 20. After the age of 20, ears continue to grow slowly, but males' ears appear to grow a little faster than females' ears. Two other excellent projects were recognized with honorable mention awards. These also based conclusions upon inspection of graphs.

For the grade category 7 to 9, the team project winners chose to explore whether boys are better in science and mathematics than girls. Random samples of the semester averages of thirty male and thirty females students in each of the top mathematics and science classes in grades 9 to 12 were drawn. Graphs, including box-and-whisker plots, were used to visualize the data. Independent t-tests indicated no significant differences in the average grades of males and females in any of the classes. Honorable mention awards in this age category recognized two team projects.

If you have ever wondered how a cookie crumbles, the winning 10-12 team project provides some answers. Three different brands of chocolate chip cookies were ranked by 82 students participating in a taste test. A cookie crumbling apparatus was constructed and used to break 30 cookies of each brand. The number of cookie pieces the size of a chocolate chip or larger and the weight of the acceptable pieces were recorded. Graphs provided visual representations of the data. Using a chi-square test and t-tests, students showed that palatability increased as crumbliness increased. A very good project was awarded honorable mention.

Science projects often provide a good starting place for entries into this competition. However, the write-up would need to be changed so that the statistical aspects are highlighted and fully developed. Judges strive to follow the development of the project from conceptualization through conclusions. Some more statistically sophisticated projects, especially in grades 10 to 12, were not recognized, primarily because some basic steps in the process were not clearly explained. For example, is it clear how the sample was drawn? Were the students able to collect the data directly (as with the ears and cookies) or did they have to get some help (as with the paper towels and grades)? Data taken from a printed source are not regarded as highly as data collected by students. Projects focusing on illustrating statistical concepts are not appropriate.

If you have never worked with a school or a team for this competition, I would highly recommend you consider giving it a try. The students usually become engrossed in exploring their question. They fervently discuss whether there is a better way to view the data. In most cases, they bring it to a conclusion only because they must meet the April 15th deadline. They would really like to look at "just one more thing". The enthusiasm students have for pursuing the answer to their question makes this a rewarding experience for all involved.

For further information on the Team Project Competition as well as the ASA Poster Competition please contact:

The Royal Statistical Society Centre for Statistical Education, within the Mathematics Department of the University of Nottingham, was formally established on 1 August 1995. It isfunded (initially for five years) jointly by the University, the RSS and SPSS(UK) Ltd. In addition to financial sponsorship, however, it has the support of many other individuals and institutions concerned with statistical education.

The Centre's brief is to promote the improvement of statistical education and thinking at all levels, i.e. within education (primary to post-secondary), within employment (including the continuing professional development of statisticians, and all who use statistics as part of their professional lives or who work with statisticians), and within society at large. The Centre will be a focus for national and international collaboration, and will actively engage in statistical education research as well as stimulating and providing support for initiatives elsewhere.

The RSS Centre starts from a position of strength. Dr. Hawkins, its Director, is also President of the International Association for Statistical Education (IASE). She will see to it that the Centre establishes its own distinctive character. However, it can also build upon the excellent reputation enjoyed by the University of Sheffield Centre for Statistical Education until its closure in August 1995, after nearly 20 years of operation. With 5 years' guaranteed funding, the RSS Centre will have time to undertake work of considerable significance, and it starts with many national and international collaborations already in existence.

The remit of the RSS Centre is extremely broad and priorities will have to be identified. The new Centre will have several initial tasks that will involve identifying current needs at all levels. It is anticipated that it will commission material, when required, from experts across the whole spectrum of statistical education. It will cooperate closely with the RSS and its constituent committees, and with IASE and other international bodies, in the debate on statistical education. Particularly important will be an anticipated partnership with industry and commerce to develop in-service courses. As well as undertaking research in its own right, the Centre will encourage other Universities and research institutes to carry out research into statistical education at all levels and in all contexts. It will run conferences and courses, and contribute to appropriate national and international publications. Cooperation with the Department for Education and Employment, and the School Curriculum and Assessment Authority (SCAA), will be sought to improve statistical teaching and assessment.

The RSS Centre's first funded research will be a project entitled 'Matching Statistical Training and Assessment with Employment Needs'. This will be a two-year project, focusing on statistical education at the Higher Education level. Anne Hawkins will direct the project, in collaboration with a core project team comprising Peter Holmes and Gillian Constable (University of Sheffield), Neville Davies (Nottingham Trent University), and Margaret Rangecroft (Sheffield Hallam University). Several other groups and organizations have already committed themselves to being involved.

Initially, contacts will be forged with employers of graduates who are involved with statistics in their work (either as specialists or non-specialists), and with the graduate employees themselves. The next step will be to identify the nature of the mismatch between statistical training & assessment and employment needs. The project team will then (by means of publicity and workshops/seminars, etc.) draw interested parties into an ongoing network that will identify and disseminate good practice. The project team will also produce, pilot and evaluate materials for use by trainers. Some of the 'network' activity would be electronic, but not to the exclusion of potential participants without internet access. The ultimate objective would be to establish a network of trainers, employers and employees which would outlive the two year sponsorship period, and which would continue to engage in research, working towards improving the employability of graduates and making them more useful to employers.

For further information contact:

The most recent issue of the Journal of Educational and Behavioral Statistics (JEBS) is a special issue devoted to an issue many of us find to be extremely important: the teaching and learning of statistics. Thanks go to Jan de Leeuw, the editor of JEBS, for his interest in sponsoring this special issue.

As the editor of the special issue, I invited authors to submit papers on teaching or learning, with the one restriction that each manuscript describe empirical research on this issue. The papers were submitted for peer review and revision, and those that were eventually accepted cover a broad range of topics. All of the authors and I hope that many readers will find the contents of this special issue to be both thought-provoking and useful for those involved in (or simply interested in) the teaching and learning of statistics.

The issue begins with a paper on the evaluation of textbooks for statistics by Harwell, Herrick, Curtis, Mundfrom and Gold. The authors outline a scheme for rating statistics texts, based on the broader literature on text evaluation. They provide several instruments (printed in the issue) suitable for use by teachers, students, or independent "experts." The authors hope that readers will use these instruments with their own texts and classes, and communicate their results and suggestions for refinement of the evaluation scheme to the authors.

The next paper deals with instruction and learning. Cohen, Smith, Chechile, Burns and Tsai describe findings from an effort to teach and assess statistics learning with interactive computer software (ConStatS). They describe misconceptions held by students, based on an analysis of errors made by students using ConStatS. Their provocative claims should provide food for thought for all who teach statistics, whether using software or more traditional instructional methods.

In the third paper, Schram describes a meta-analysis of studies of gender differences in post-secondary statistics achievement. Many have wondered whether the male advantages observed in mathematics at lower grade levels persist in statistics classes in college or beyond. Schram finds small overall gender differences, but discovers that on average, females outperform males. However, the study results are not all consistent with a common gender effect, thus Schram explores several factors (including type of achievement outcome, department offering the course, and sex of the researcher) that explain variation in the sizes of gender differences.

The issue closes with an updated analysis of the literature on teaching statistics in education, with a focus on the literature as a resource for statistics instructors (Becker, 1996). Also I describe some of the electronic resources for teaching (as well as research on statistics instruction) now available on the Internet.

In March the Boston Chapter of the American Statistical Association (BCASA), with the support of the Section on Statistical Education and the Connecticut and Rhode Island Chapters, cosponsored a conference entitled "Trends in Introductory Applied Statistics Courses: Topics, Techniques, Technology." The goals of the conference were threefold. First, we hoped to bring together instructors of introductory applied statistics courses from a variety of disciplines and types of institutions to discuss recent trends in statistics education. Second, we thought that by bringing together such a varied group, along with some of the best-known statistics educators as speakers, we could get a good sense of where statistics education is today and where it appears to be headed in the future. Third, we thought that if our conference was a success, it could serve as a model for other regional conferences on teaching statistics.

There were approximately 150 participants. The BCASA originally envisioned a regional conference, and the majority of the participants did come from the New England area. But to our surprise, people came from as far away as Washington state, Wyoming, Puerto Rico, Canada, Virginia and Pennsylvania. About 20% of those in attendance were high school teachers, and 5% were from industry or research institutions. The rest were from four-year colleges and universities. Among the departments represented, in addition to mathematics and statistics, were business, economics, biostatistics and health sciences, psychology, sociology and education. In addition to the participants, there were several vendors including Minitab, Addison-Wesley Interactive, Springer-Verlag and W.H. Freeman.

The one-day conference started with welcoming remarks by Arthur Doyle, VP for Academic Affairs on behalf of Framingham (MA) State College where the conference was held; David Hoaglin, Vice-President of the American Statistical Association, on behalf of the ASA; and Robert Goldman, Boston Chapter President on behalf of the chapter. Following the welcomes, David S. Moore of Purdue University gave the Keynote Address. Prof. Moore's talk described some broad trends in statistics education including the synergy between content, pedagogy and technology in introductory statistics courses; the effects on statistics education of the movement to reform mathematics education at both the pre-college and college levels; changes in the way students, faculty and programs are assessed; and the coming pressures on faculty for more accountability and efficiency.

After the Keynote Address, participants chose among sessions on topics, techniques and technology. The speakers for the session on topics were George Cobb from Mt. Holyoke College and Jeff Witmer from Oberlin College. George talked about four kinds of understandings, four unifying themes, structured concept maps, and four under-taught topics. Jeff presented some "Themes and Activities for Introductory Statistics" that can be used to reinforce key topics such as EDA, data collection and sources of data, simulation, inference, checking assumptions and diagnostics.

Participants who attended the session on techniques heard Joan Garfield from University of Minnesota and Robin Lock from St. Lawrence University. Joan discussed "The CHANCE Approach to Teaching Statistics." She shared techniques she uses to engage and motivate students; class activities for particular topics; resources from the Chance data base, the Journal of Statistics Education and Activity Based Statistics; tips on using cooperative groups; and the role of assessment. Robin demonstrated "Some Things for Your Students to Do Other than Listen to You." These included using random rectangles to demonstrate sampling methods and penny flipping/spinning/tipping to demonstrate hypothesis testing.

The technology session featured Jackie Dietz from North Carolina State University and Paul Velleman from Cornell. Jackie took participants on a tour of the internet during her presentation on "Internet Resources for Teaching Statistics: Data sets, Software, Teaching Ideas, and Conversation," and gave them a list of addresses for many of the sites so that they can visit them again on their own. Paul talked about some strengths and weaknesses of instructional multimedia and gave the first public demonstration of the new multimedia-based presentation of the introductory statistics course that he is developing with Addison-Wesley Interactive.

After Round Table Luncheon discussions, participants from each of the three tracks - topics, techniques and technology - joined with others from the same track to come up with their consensus five best and five worst practices for teaching introductory statistics. These consensus lists were then reported back to the whole group. The consensus five best topics included exploratory data analysis, interpreting results, regression/correlation, inferences from data, and practical applications. The consensus five worst topics were formal probability, derivations, artificial examples, omission of real-life data, and ignoring checking assumptions.

Among five best techniques were using active and cooperative learning; incorporating appropriate technology; having students use oral and written communication; using real, relevant data including student-collected data; and presenting applications of statistics through articles, videos, guest speakers, site visits, and/or case studies. The list of five worst techniques included lack of connection between probability and the rest of the course, lack of variety in pedagogy and assessment, formula-driven instruction, use of data without a context, and trying to cover too much material.

In the technology track, the consensus five best practices included using technology to eliminate drudgery, finding and using realistic and relevant data sets, using multi-media, and doing in-class experiments and simulations. The five worst aspects of technology included lack of access to state-of-the-art technology, allowing technology to distract students from the statistical concepts, unreliable software or hardware, technology replacing human interaction, and using only computer assessment.

Based on the participants' evaluations of the conference and our own assessment of how well we met our three goals, the BCASA felt that the conference was very successful. The participants' ratings and comments were almost unanimously positive. The biggest complaint was that people wished they could have attended more than one of the sessions. As for our goals, we did get a diverse group of attendees, and through the lively round-table discussions at lunch and the following consensus-building sessions we feel that we did get some sense of where statistics education is today.

In order to share our findings about the current state and future direction of statistics education, and since this conference could serve as a model for other chapters interested in holding similar conferences on teaching statistics, the organizers - Bob Goldman from Simmons College, John McKenzie from Babson College, and Anne Sevin from Framingham State College - will be giving a special contributed paper session on the conference at the Joint Statistical Meetings in Chicago. In addition to presentations by the three organizers, Richard Goldstein will discuss statistical education and technology from the point of view of a consultant and expert on statistical software. Dick Scheaffer from University of Florida will serve as discussant. This session is scheduled for Wednesday, August 7 from 2:00 - 3:50 p.m.

For further information contact:

The Department of Mathematics, University of Wisconsin Oshkosh is proud to organize a conference for teachers of statistics. The conference will begin at 12:00 p.m. on Friday, March 21, and end at 5:00 p.m. on Saturday, March 22, 1997.

The purpose of the conference is to share the latest ideas and information on issues related to teaching statistics at all levels. The conference will feature invited presentations by leading statistics educators, and contributed paper sessions consistent with the mission of the conference. Authors wishing to contribute a 20 minute paper should submit an abstract (e-mail submission is preferred) by December 1, 1996. Conference programs with information concerning registration and housing will be available after January 1, 1997.

The scheduled keynote speakers include Professors Donald Bentley (Pomona College), Joan Garfield (University of Minnesota-Twin Cities), Dennis Gilliland (Michigan State University), Robert Wardrop (University of Wisconsin-Madison) and Jeffrey Witmer (Oberlin College).

Thinking of going to San Diego in January? If so, plan on making a presentation at the National MAA (Mathematical Association of America) Meeting to be held as part of the Joint Mathematics Meetings (co-sponsored by MAA and the American Mathematical Society (AMS)) in San Diego, January 8-11, 1997.

The MAA contributed paper sessions will include a session on statistics: Teaching the Practice of Statistics at All Levels. The session is organized by Chitra Gunawardena (University of Wisconsin Center Fox Valley), K.L.D. Gunawardena (University of Wisconsin Oshkosh) and Anne Sevin (Framingham State College).