0:22

A run chart is very simple.

Â On the x-axis, we have data in some sort of chronological order,

Â for example Monday, Tuesday, Wednesday,

Â it could be January, February, March, and on the y-axis we put our measure,

Â I'm just gonna call it M, measure of interest.

Â And that could be a percent, it could be a count, could be money.

Â 0:49

We get the data in chronological order, and we plot them.

Â We connect the dots with a line.

Â And then we need to start figuring out how to interpret this chart.

Â Well, we do that with several simple steps.

Â The first step is to put a center line through the middle of the plotted dots.

Â And this center line, sometimes called CL, for

Â a run chart, is what's called the median.

Â 1:26

We have the data plotted in time sequence, we have our center line,

Â a form of the average, if you will, but it's the median for the run chart.

Â Now what we're going to do is define a run.

Â What is a run?

Â Now, a run is one or more data points on the same side of the median.

Â 1:48

So here we have one data point, here we have one.

Â One, one, one.

Â But here we have one, two, three, four data points.

Â A run can be one or more data points on the same side of the center line.

Â That's the key.

Â It's as the data flip and flop back and forth across the center line,

Â we count how many dots end up in a little cluster.

Â Here we have one, one, one.

Â Here we have three.

Â Maybe two here, one, one.

Â So we get the number of runs, and

Â then we're going to be able to interpret the chart.

Â And we do that by using a series of simple run chart rules.

Â Now, there are many run chart rules that people have used over time.

Â And you will see different people using different rules.

Â Here at the IHI, we have what we call the four simple run chart rules.

Â They are basically, a shift in

Â the data, a trend in the data,

Â whether you have too many or

Â too few runs, and finally,

Â an astronomical data point.

Â And let me explain each of these quickly.

Â A shift in the data is when you have too much of the data hanging above or

Â below this median center line.

Â And the way we make this determination is if you have six or more data

Â points hanging in a group above or below the center line, that's an indication

Â of a shift, that the data had moved to a level and stayed there too long.

Â So you have random and all of a sudden you have one, two, three, four, five,

Â six, and then it goes random again.

Â This run of six data points in a row above the center line signals a shift.

Â And the data have hung there for

Â too long when they should be just randomly flipping and flopping.

Â And you can see that there could be a shift downward as well.

Â The second one is a trend.

Â And while some people think that this is a downward trend, or

Â this is an upward trend, two data points does not make a trend.

Â What we're looking for to get a statistical trend in the data is to have

Â five data points constantly going up or constantly going down.

Â 4:19

Now, if you had data points that went up, up, up, repeat,

Â repeat, repeat, but kept going up, you don't count the repeats, but as

Â long as it continued its upward journey, or downward journey, it's still a trend.

Â If it went up, up, up, equal,

Â equal, then dropped, then the trend would be canceled, all right?

Â But a trend, and this is one a lot of people struggle with, five or

Â more data points constantly going up, or constantly going down.

Â 4:54

What you do is you find out how many runs you have on your chart, and

Â then you look up in this table for the total number of data points, and

Â what was the low number of runs and the high number of runs?

Â And for a given number of data points, say, 20 data points,

Â it'll tell that you should have no fewer than x number of runs, and no more than y.

Â 5:20

And the idea here is that if data are randomly arrayed,

Â you should see just some sort of random flipping and flopping back and forth.

Â If you get data again that are hanging, on one side or

Â the other, and you only have two runs in your data,

Â you're gonna have not enough data that forms essentially a normal distribution.

Â So this table which has been figured out mathematically for years,

Â is designed to tell you how much variation there should be in a given set of data.

Â So if you had 15 data points, 20, 30,

Â it will tell you the lower and upper boundaries of the number of runs.

Â The final test, or rule, if you will, is whether or

Â not we have an astronomical data point.

Â Now, this is a judgement call,

Â something I refer to as the interocular test of significance.

Â We have data that are going along, and then all of a sudden,

Â wonk, we've got this huge spike, and we wonder why.

Â Well, often times two things.

Â One, we could have collected the wrong data, that for some reason,

Â data got into our data set that shouldn't have been there,

Â cuz here's where the bulk of the data typically fall.

Â Or, in fact something, special was going on on that day.

Â If this is food trays being deliver to the medical units, and

Â this is the day that the elevator people came and

Â shut down three banks of elevators, when all the food trays backed up.

Â And if you're looking a percent of food trays delivered on time,

Â you'd see this big spike.

Â Well, an astronomical data point is not a statistical determination on a run chart,

Â it's an eyeball test.

Â And it's guidance that you should either look at your data or

Â put the data on a control chart, which will be in a subsequent session to find

Â out if in fact that is truly different than the rest of the data.

Â So there you have it.

Â The run chart in a nutshell.

Â You've got the elements, x- and y-axis, the median, plot the data over time,

Â figure the number of runs, and then apply the four simple run chart rules.

Â