2
5. Determine the scale of the vertical axis. The scale should extend from a number 20% larger
than the largest value to a number 20% smaller than the smallest value. Label the axis in
equal intervals between these two numbers.
6. Plot the data values in the sequence in which they occurred.
7. Draw lines to connect the points on the graph.
8. Calculate the median of the plotted numbers and draw the line on the graph.
9. Title the chart, and note the goal line and the sample size.
10. Annotate the chart, indicating when key tests of change were initiated.
Note: A good technique in creating a run chart to learn about the impact of changes is to
identify an initial median using baseline data. If the baseline data come from a process
exhibiting no signals (shift, trend, runs, astronomical data point), extend or ‘freeze’ this initial
median into the future. By using baseline data and extending the median into the future, new
data are not allowed to influence the initial median. Any changes in the new data stand out
against the baseline median more clearly and this allows for more accurate detection of signals
of improvement (the probability-based rules are relative to this median value). This is
particularly important as you try to understand the impact of different changes to a system
over time.
Run Chart Analysis Rules to Detect Special Cause Variation
To avoid inaccurate analysis of run charts (and potential tampering/incorrect actions) three
probability-based rules are used to objectively analyze a run chart for evidence of non-random
(special cause) patterns in the data based on an α error of p<0.05. Rule 4 is based on identifying
a point that is highly unusual. With accurate analysis of your data you are able to determine
appropriate actions to achieve the improvement you are striving for.
Rule 1—Shift
Six or more consecutive points either all above or all below the median. Values that fall on the
median do not add to nor break a shift. Skip all values that fall on the median and continue
counting.
Rule 2—Trend
Five or more consecutive points all going up or all going down. If the value of 2 or more
consecutive points is the same, only count the 1st point and ignore the repeating values; like
values do not make or break a trend.
Rule 3—Runs
A non-random pattern is signaled by too few or too many runs, or crossings of the median line.
A run is a series of points in a row on one side of the median. If only chance is influencing the
process being measured with a run chart, then there should be a regularity at which data points
go above and below the median to satisfy this condition. Some points can fall exactly on the