Courses in Statistics have changed over the years with changes in technology, changes in society and changes in demands on the work force. Current guideline for contemporary statistics courses (ASA and Garfield, Hogg, Schau and Whittinghill 2000) now emphasis the need for students to rely on computers using statistical software programs. There is no way that contemporary analysis of data can be done without some form of a computer/calculator being used. Although calculators can be used in introductory courses, they are inadequate for problem solving and solution presentations of real world data.

There are many schools that teach an introductory course in statistics using Excel as a computing tool.

The issue here is that spreadsheets are universally used in business, government, education, research, manufacturing and just about all other sectors. Spreadsheets are an essential backbone in every aspect of what people do. Typically they are employed to:

1. Gather, create and manage data of all kinds; numbers, text, symbols, observations, surveys, money, financial, accounting, designs, etc.

2. To create models and make calculations.

3. To disseminate and share information in a basic form across a wide expanse of users and contributors in a completely connected world.

Spreadsheets have become indispensable tools for getting the informational work done. They are empowering tools that are expressive and apparently simple, yet underneath very complex. Text and numbers can be intermingled. They can be subservient in that they facilitate peer-to-peer sharing, non-technical people can do analysis and share the data and results. They facilitate back-channel, behind the scenes communications.

Raden 2006.

They also create enormous problems with errors in data entry, errors in equations, misuse of data sections, and incorrect use of functions. (This is discussed in my other articles.)

Business, Engineering, Psychology, Accounting and other schools in colleges and universities find they have to teach the intricacies of using spreadsheets such as Excel, because of its universal use in all sectors of the world. Statistics in a broad sense involves every one of the three areas that Raden (2006) points out. Therefore, why not teach statistics in a manner that involves spreadsheets? The other side of the coin is, if Excel has to be taught, why not include teaching about the use of the statistical functions and routines in Excel?

The University of Reading (SSC) has emphasized the use of Excel in statistics and the use of spreadsheets for entering and tabulating data. Arshan (20007a) has a web site that covers all the essential Excel statistical capabilities with respect to business applications, as part of an MBA degree program. Arshan (2007b and 2007c) has other sites that show extensive use of spreadsheets and cell equations to handle business decisions and other problems.

The intent of teaching the use of Excel is to produce graduates who can use the powerful spreadsheet capabilities of Excel, have some understanding of statistical methods and can do quantitative reasoning with statistics. Some other schools combine Excel with other software programs such as Minitab. Levine (1999) and Pelosi (2000) are some the more frequently used business statistics textbooks that incorporate Excel.

Levine and Fan (2000) say "The strongest aspect of the book ("Practical Statistics by Example Using Microsoft Excel", Sincich, Levine and Stephan, Prentice Hall, 1999) is the introduction and incorporation of Excel for doing and learning statistics. The authors rightfully argue in the preface that ‘today an increasing number of individuals use spreadsheet application as the means to retrieve and analyze directly the data they need. Employers now are beginning to desire, if not demand, that their college-educated, entry-level employees have more than just a cursory awareness of spreadsheet applications. Most students are familiar with Excel and/or have easy access to it on a personal computer…" They also point out that there are real dollar and time advantages to universities when Excel is used as part of the course. It allows students to do homework on personal computers, reduces the load on school computer labs and saves the cost of the expensive commercial licenses for use on each pc.

De Levie (2005) finds that Excel is useful in science and engineering laboratory and application areas. "Excel is a powerful spreadsheet. Even though it was developed primarily for business applications, it contains many mathematical functions, and its ease of use and wide distribution make it a very powerful tool for scientists and engineers." Some of the textbooks are Billo (2001), Block (2000), de Levie (2001), de Levie (2004), Gotfried (2000), Liengme (2002) and Orvis (1996). The Visual Basic for Applications (VBA) feature with the ability to program specific reductions and analysis is important in engineering and scientific areas. The main statistical usage here is equation fitting and regression. Graphic display of data and the display of data fits to mathematical relationships is also important.

The use of spreadsheets as a means of explaining subtle ideas by doing simulations is another valuable teaching method (Horgan 1999). See the great spreadsheet by Jacob Eisner (Eisner 2007) to teach the forward-backward algorithm to solve a probability problem. Some of the more difficult ideas in experimentation, sampling, variability and power can be demonstrated.

This paper refers to the following textbooks, which all incorporate the use of Excel for solving statistical problems. This is just a sample of many other textbooks that incorporate Excel.

1. Larson and Farber 2003

2. Levine, Berenson and Stephen 1999

3. Lind, Marchal and Mason 2002

4. Moore and McCabe 2003

5. Pelosi and Sandifer 2000

6. Triola 2001.

Texts 2, 3 and 5 are focused on business and economic applications. Text 4 also shows how to use JMP, Minitab, S-PLUS, SPSS and SAS. I am familiar with the use of 1 in an introductory statistics course.

However including Excel in an introductory statistics course has its own problems. The inclusion of Excel is controversial and is not universally accepted. Some teachers in university statistical departments disparage the use of Excel as a tool for doing statistical calculations. Others have tried it and found severe problems, and have discontinued the practice. Some of their arguments and criticisms are:

• It takes teaching time to teach Excel. Most introductory statistics courses are very time limited to teaching the material in the textbook. Teaching Excel takes away from teaching statistics. It is a lot easier to teach problem solving and to test students using hand calculators then with computer software.

• The Excel default graphics do not fit standard statistical data displays, and takes considerable teachers and students time to change them to standard displays. Excel 2007 makes it more difficult to make changes with the complex expanded menu structures. Excel is not a "one-button" statistics package.

• Excel functions and routines do not fully support the subject matter or the problems in the textbook (from 40% to 70%. of textbook problems are directly supported). Excel does not support expanded applications in line with contemporary statistics. The ANOVA and regression routines are too primitive and size limited. The data analysis routines are too primitive. Many teachers take the view that Excel is not capable of serious data analysis. Excel cannot be used in advanced classes (McCullough, 2004).

• The field of statistics is always evolving, with new ideas and methods to analyze data. At the introductory statistics level, some of this is being introduced. Excel is essentially a locked-in-time (ca 1990) approach to statistics, and has not introduced any new functions or routines (re: Excel 2007), relative to what has evolved since then.

• Uncertainty about the reported errors, faults and inaccuracies in Excel. This is a very weak argument, since it is not raised on other commercial software.

• Excel is not usable for classroom quizzes, tests and examinations based on solving problems to test students for comprehension and understanding. Testing is still a paper-and-pencil process.

• Text-books that incorporate the use of Excel normally include: a CD with data sets, files relating to teaching (illustrations, slides, etc.) and an Excel Add-in. Publishers charge more for this combination, markedly increasing the cost of a textbook with the Excel features for students. This is a significant unrecoverable expense for the student. I have heard lots of complaints from students on this.

• Based on my findings, I recommend that if Excel has to be used, the Excel 2003 version be used, because of its simplicity and general freedom from errors. However if the school/university is focused on Excel 2007, then it should be used. Since the look-and-feel of Excel 2007 is so different from Excel 2003, and the future Windows use will be Vista, teaching Excel 2007 would be the choice.

Note A, describes some of the other reported comments on the use of Excel in a teaching situation. Of particular interest is the paper by Peter C. Bell, (Bell 2000) about his course in business statistics using Excel.

Author : David Heiser [dheiser594@gmail.com]