Control charts change data into information

In a “webinar” sponsored by the American Society for Quality this month, John Conyers and Robert Ewy will share the story of the Palatine, IL school district that was recognized by the 2003 Malcolm Baldrige Award. The session “provides a real understanding of the Baldrige Award and its application to K-12 schools,” according to ASQ. To learn more about this session, http://www.asq.org/ed/webinar/descriptions/charting_your_course.html

School districts are increasingly utilizing the Baldrige Criteria to assess their processes and promote excellence, but most have not harnessed the power of the control chart as they pursue their improvement efforts. Palatine’s Consolidated School District, on the other hand, began to use control charts in its quality journey, providing an example for educational programs throughout the country.

What is the power of control charts, and why are educators so often afraid to use them?

It may be that because the statistics behind control chart technology are sophisticated, using this important tool can seem daunting. It may be that teachers, accustomed to the model of the grade book, often find themselves keeping data over time, so the run chart seems to be sufficient in their charting efforts (See Jan-Feb. Quality School Line). http://www.pqsystems.com/School_Line/2005/01/index.htm

Whatever the reason, it is useful to examine how control charts can support improvement efforts in classrooms as they do in hospital emergency rooms, bank lobbies, the U.S. postal service, and most manufacturing organizations. Control charts give information about processes, and every system—whether it is a classroom, a pharmacy, or a factory—is composed of processes.

Understanding natural and special variation in processes is key to improvement. That is, all processes will have variability: think of orchestrating a classroom of students (or adults) to clap their hands at the exact same moment. It simply will not happen, for a variety of benign reasons. Or imagine a student who always gets 100 percent on weekly spelling tests. If the student misses a word one week, is this reason to become concerned? When a student occasionally forgets lunch money or forgets homework, is this a situation that demands immediate or dramatic attention? At what point does it become such a situation?

To think about variation in a different way, reflect on the room temperature in your classroom or office. How many times does someone say, “It’s way too hot in here” and proceed to change the thermostat? Another person finds the room too cool, and tampers again with the switch. This process, with which we are all familiar, is based on “point mentality:” a single data point causes us to change the system. It represents “overcontrol” of the system. This may be the reason that classroom thermostats are often locked by maintenance personnel! If data were collected over time, on the other hand, it would become clear whether the temperature is predictable and stable or whether it in fact does become “too hot” or “too cool” because of the system itself.

Control charts help to identify normal, natural variation in processes. They also show when a process has changed. If a student who has been consistently earning 95% or more on spelling tests begins to score only 80% on a regular basis, the control chart can help to verify that a process has changed for that student. Control charts help avoid both overcontrol and undercontrol. Undercontrol is the belief that nothing should be done to change the system because it will take care of itself.

As an example of using control charts to predict stability in a system, consider school bus pickup times.

Bus #43 is supposed to arrive at 2041 Old Town Road at 8:14 to pick up students who will be transported about 3 miles to Normandy Elementary School. It is scheduled to arrive at the school by 8:29.

The following is a check sheet reflecting actual daily pick-up times for Bus #43 on Old Town Road:

Oct. 7	8:15
Oct. 8	8:25
Oct. 9	8:20
Oct. 10	8:22
Oct. 11	8:24
Oct. 14	8:13
Oct. 15	8:20
Oct. 16	8:26
Oct. 17	8:19
Oct. 18	8:26
Oct. 21	8:14
Oct. 22	8:19
Oct. 23	8:20
Oct. 24	8:15
Oct. 25	8:22
Oct. 28	8:12
Oct. 29	8:17
Oct. 30	8:20
Oct. 31	8:33

To a parent who is waiting with a kindergartener in the pre-dawn hours of winter, this may seem to be wild variation in the process. The child who goes to the bus stop at 8:14 may have to wait as long as 15 minutes, according to this chart. Why is there so much variability?

Check sheets give a certain amount of information, but a control chart can give even more, since it will indicate whether the variation in this process is “out of control”—that is, outside of control limits that are created from the data itself. To create a control chart, the scheduled arrival time of 8:14 is represented on the left (y) axis as 0, with variation from that indicated by data points. On two occasions, the bus was ahead of schedule (below the target time), but the process is predictable, within the variation indicated by the control charts.

The software that created the following chart has also figured the control limits—the points at which it can be said that the process is “out of control,” or unstable. These limits (upper control limits or UCL, and lower control limits, or LCL) show that the bus arrival process is in fact stable. That is, variation is due to natural causes—becoming delayed at stoplights, taking time to scrape frost from the windows, waiting for dilatory children to board the bus, even helping younger passengers with their Halloween costumes on October 31 may all cause slight delays in the arrival of the bus.

When a process swings wildly out of control—in this case, when arrival time varies by 35 minutes or more—it demands further investigation to assess the sources for special cause variation or understand what is known as common cause variation. The bus arrival process is a stable one, and the kindergartner’s parents can feel confident that the bus will arrive within a few minutes of its target.

In the case of bus schedules, knowing that the process is stable will help avoid making changes that are not effective and that do not improve the process. You can imagine what would happen if the transportation coordinator responded to every parent complaint about a bus that had arrived a few minutes late. Making changes without knowing their effect on the overall stability of the process is a counter-productive endeavor.

If, on the other hand, the process did demonstrate instability—that is, if the data showed that it was out of control, fluctuating widely above and below the average time—those responsible for its arrival might study the reasons for such variation and examine ways that the schedule could become more consistent. A new policy indicating that a bus will wait for a wayward child no more than 3 minutes, for example, would eliminate long waits that delay the bus for the rest of the route.

Once it has been determined that the process is stable, steps can indeed be taken to reduce even the small amount of variation that exists—in other words, to narrow the control limits by introducing steps to speed up the process. Changing the route to avoid certain traffic lights, not waiting for those who are not at the bus stop at the designated time, or having the driver arrive at the bus garage early enough to deal with exigencies such as frosted windows or small maintenance issues, for example, might help to reduce the variation in the process.

Control charts give information about the stability of a system so that appropriate improvements can be made, rather than over-reacting to every variation in the data. This useful tool can support classroom learning, as well as administrative processes. It can give information about variation in student performance over time, for example, and can help to identify “special causes” of variation. If a student has been staying up late on Thursdays, for example, and his or her test scores on Friday are always outside the control limits, appropriate recommendations about Thursday nights can be made. Students can chart their own progress with respect to grades, homework times, or reading speeds, understanding that their progress will not be a steadily upward path, but will have variation over time. They will take responsibility for improving their processes themselves, and will take pride in the progress they make toward stable processes.

For more information or to schedule Mathematics Standards training for your school or district, call Soren today at 800-777-3020 x133.

---

Got stories?

Many of you have used the Process and Tools to improve classroom processes and support learning. We’re opening Quality School Line to a series of these stories, and invite you to send details of your experiences with the improvement tools. If we use your story, we will send you a copy of Alfie Kohn’s book, Punished by Rewards, a stimulating discussion of the role of external and internal motivation in student learning. All you need to do is describe how you’ve used a particular tool, tell us a little about your school (its location, number of students, and a little about its quality journey), and indicate your position in the school. We can identify you and your school or not, as you prefer. Your colleagues who are looking for concrete ways to apply these tools will thank you! If you have applications of your own that you’d be willing to share with other teachers e-mail them to K12@pqsystems.com