Observe the data points with respect to the center, horizontal line. This center line, the mean, should lie roughly in the middle of all the data points. If it does not, this implies that the data is roughly biased in one direction, with a few outliers in the other direction. If this is the case, then it is possible that the observations for certain times are special cases, and should be individually analyzed.
Check if any data points lie outside the standard deviation borders. The horizontal lines at the top and bottom of the “I” chart represent the numerical values three standard deviations from the mean in the positive and negative directions, respectively. If any data points lie above the top horizontal line or below the lower horizontal line, then they are definitely special cases and should be individually analyzed.
Note the overall pattern of the “I” chart. It should appear random. If the chart does not appear random -- if there is a pattern such as a linear or parabolic pattern -- then the individual data points are time dependent, which means that the process being observed is not entirely random.
Observe the data points with respect to the center, horizontal line. This center line, the mean, should lie roughly in the middle of all the data points. If it does not, this implies that there are occasionally large spikes in the values. These spikes may indicate sudden improvements or deteriorations in the process. If you find such points, you should observe the differences between those data points that correspond to the data points in the “MR” chart.
Check if any data points lie outside the standard deviation borders. The horizontal lines at the top and bottom of the “MR” chart represent the numerical values three standard deviations from the mean in the positive and negative directions, respectively. Finding one or two data points outside these lines may not indicate anything. However, if you find many points outside these lines, there is a strong possibility that the process being observed is unstable.
Note the overall pattern of the “MR” chart. It should appear random. If the chart does not appear random -- if there is a pattern such as a linear or parabolic pattern -- then the differences between measurements are time dependent, which means that the process being observed is not entirely random. Usually a pattern in the “I” chart appears simultaneously with a pattern in the “MR” chart.