Microsoft Excel For Data Analysis in Schools.

The call today is for schools to do more with less. Money for public education at all levels is tight. Even when allocations remain the same, inflation reduces their buying power. It may be tax limitation legislation in some states and unpredictable tax revenues in others. Whatever the limiting factor, public schools must make even better decisions with regard to money and programs. Accountability for student academic achievement is the issue getting the greatest attention today in schools. The public is demanding improved results for the public dollars expended for education.

The Demand for Data-Driven Decision Making

Schools are always trying new programs. Most often these programs are selected by those not directly involved in instruction. As such, those responsible for their implementation have no vested interest. Too often teachers take the attitude that `this too will come to pass.' Eventually the program disappears without fanfare, and in time, another takes its place. The demand for accountability is bringing this practice to an end. There just isn't the money to spend on the latest fad. Money must be spent on change that makes an impact on instruction and thus student achievement.

The implication is that there must be measurement of the impact of any change in order to evaluate it. Teachers, in their undergraduate training, usually get very little exposure to data analysis procedures. The closest course required is probably one in tests and measures, and it is usually designed around teacher-made tests and understanding standardized testing. Most states require administrators to have Master's degrees. Two agencies oversee university programs in school administration: National Council for Accreditation of Teacher Education (NCATE) and Interstate School Leaders Licensure Consortium (ISLLC). Both of these require approved programs to develop students' skills and knowledge associated with research. In the area of data collection and analysis, ISLLC Standard I states, in part:

   The administrator facilitates processes and engages in activities ensuring
   that a) assessment data related to students learning are used to develop
   the school vision and goals; b) relevant demographic data pertaining to
   students and their families are used to develop the school mission and
   goals.

In the area of data collection and analysis, NCATE Block 1, Area I states:

   2.1 Conduct needs assessment by collecting information on the students; on
   staff and the school environment; on families and community values,
   expectations and priorities; and on national and global conditions
   affecting schools.

Furthermore, NCATE Block 2, Area I, states:

   2.3 Engages staff in an ongoing study of current best practices and
   relevant research and demographic data, and analyzes their implication for
   school improvement.

   2.4 Analyze and interpret educational data, issues, and trends for boards,
   committees and other groups, outlining possible actions and their
   implications.

These aspects of research are not at a level usually associated with Master's level graduate studies. This research emphasis is not the review of literature, statement of hypotheses, sample selection, collection of data, data analysis, interpretation and generalization of results that are associated with a dissertation. Rather it is more the use of data that already exist in a school system to produce information for making better decisions on local issues. This is known variously as Action Research or "data driven" decision making.

Problems with Current (or Traditional) Statistical Processes

Graduate schools most often teach students to analyze data by the use of packages such as SPSS, SAS or OSIRIS. These are comprehensive packages providing flexibility in analysis from perhaps a hundred different perspectives. To properly take advantage of these programs requires three things:

1. The programs must be readily available. At one time, programs had to reside only on a mainframe. The increased power of microcomputers and condensed versions of the software have greatly improved their availability.

2. The user needs an extensive knowledge of statistics. Data analysis is not an element of an entry level math statistics course. Such courses address the topics of permutations, combinations, series, sequences, etc. Usually only upper level undergraduate and graduate courses provide information on descriptive and inferential statistics. Without such, the staggering array of choices will only confuse the user.

3. The user needs to be knowledgeable about the subtleties of the particular software. Even if the statistic to be employed is known, the setting of the parameters for that statistic may not be clear. To be able to do this often entails making a careful study or even taking a course specific to the software.

All three of these requirements are barriers to Master's students in School Administration. To be sure, microcomputers with a statistical analysis program will be available to graduate students somewhere on the university campus, perhaps in their departments. Universities recognize the needs of doctoral students as well as professors, so they provide the technology. However, if Master's students are to make use of these skills on their jobs, the analysis program must be available there. A single copy of one of these data analysis programs can cost two to three hundred dollars. Since only a rare individual will have the knowledge to use it and its use at the present time would be occasional, it is a difficult `sell' to superiors with oftentimes severely limited resources at the building level. Some companies do offer these programs to students at a reduced rate. Some universities provide remote access to their main frame's analysis software. Considering the demands on the time of school personnel, it is probably not likely that they would want to avail themselves of such options.

With regard to the second issue, it is not likely that the Master's student has had the course work to understand the kind and level of statistics that a sophisticated analysis program can offer. A person could take an introductory data analysis course. It would likely provide instruction on topics from mean and standard deviation to linear regression and maybe analysis of variance. All of these would be in such a statistics program, but the latter topics go far beyond NCATE and ISLLC expectations. Such a course is more appropriate for one getting a specialist or doctoral degree. Since a Master's in School Administration has so much in the way of professional information to provide its students, a separate course with the depth necessary to understand such sophisticated software options is probably not required.

The third issue, taking a course in the specific software, is also not likely. In addition to taking courses beyond the introductory level to provide the concepts, it could take a specialized course tailored to unique screens of the analysis package to make full use of an analysis package. Again, only a graduate student at the specialist or doctorate level takes a course designed to teach the subtleties of the package. It would certainly not be available to undergraduates and most Master's students.

A Manageable, Viable Alternative

The Master's program at The University of Southern Mississippi has woven into the curriculum a unit on data analysis using Microsoft Excel. Why Excel? First, it is readily available in local schools as part of Microsoft Office. If a school in the state wanted to purchase it for the office or even if a school employee wanted it for use in his/her home, Microsoft Office is available on the state approved purchase list for under fifty dollars. Second, it is easy to use. It is not necessary to learn or incorporate a complex labeling system before entering data. Surprisingly, Excel will handle approximately 500 records. Many descriptive statistics such as mean and standard deviation are available at not much more than a click of the mouse under the "fx" button. Furthermore, a variety of graphing functions is available under the graph icon.

How then is the program used? In a preliminary activity, students are asked to enter a small amount of data (less than 25 cells) and perform a number of functions. They are to find column totals, mean and standard deviation, and correlation coefficient for various columns by the highlight, point and click method. Instruction is given on the fundamentals behind each concept involved with each operation. Anonymous data from a well known standardized test of achievement for a selected elementary grade level for the past three years are given to students on a disc. They are then asked to perform the same operations (analyses) on the real data. Conclusions such as trend over time and stability of the spread of the population's test scores are made. Graphs are then made representing parts of the whole (pie graphs), bar graphs, and scatter plots. With a few more clicks, the program will even draw the linear regression line!

The students' responses could not be more gratifying. Making simple summary statements about data and graphical depictions (especially for the mathematically challenged in the group) without all the usual computational effort, gives students feelings of accomplishment not possible before. Later in the program, students are expected to collect data on some issue of importance at their own schools. Being teachers in a school system and with support from their practitioner mentors, they are allowed access to district information not normally available to the public. This data can be in the areas of discipline, students' test scores, school culture, etc. The data is then put on an Excel spreadsheet, usually with secretarial assistance. Discussions are held as to the questions the collected data can answer and what additional questions could be answered if additional information were collected. The students are taught to assess the statistical significance of differences in data sets. One of the Excel functions is the "Student t." Excel comes back with the probability associated with the difference (after specifying appropriate parameters), all this without the using the cumbersome Student t formula or the often confusing probability tables with their degrees of freedom. There must be prior instruction on the generally accepted minimum level of statistical significance provided to the students.

The bottom line for the use of Excel is that Master's level students leave the program with the ability to ask a reasonable research question, collect appropriate data, produce descriptive statistics (even rudimentary inferential statistics) and create diagrams suitable for presentation to superiors and the public. All of this is designed to increase the number and quality of data driven decisions within the school setting. The program's graduates should thus make better decisions with regard to a program's effectiveness than those based on gut feelings of just a few years ago.

Bibliography

Council of Chief State School Officers. (1996). Interstate School Leaders Licensure Consortium. Washington, D. C.: The Council of Chief State School Officers.

NCATE. (1997). Standards, Procedures, and Policies for the Accreditation of Professional Education Units. :Washington, D.C.: National Council of Accreditation for Teacher Education.

John J. Marshak, University of Southern Mississippi

After fourteen years teaching and fifteen years of public school administration, John J. Marshak, Ph.D., is now in his third year of teaching pre-service teachers and graduate level courses in school administration. <[e-mail]>.