Data considerations for Acceptance Individuals Chart

To ensure that your results are valid, consider the following guidelines when you collect data, perform the analysis, and interpret your results.

The data should be continuous: Continuous data are measurements that can take on any numeric value within a range of values along a continuous scale, including fractional or decimal values. Common examples include measurements such as length, weight, and temperature.
The data should be in time order: Because control charts detect changes over time, the order of the data is important. You should enter the data in the order it was collected, with the oldest data at the top of the worksheet.
The data should be collected at appropriate time intervals: Collect data at equally spaced time intervals, such as every hour, every shift, or every day. Select a time interval that is short enough that you can identify changes to the process soon after the changes occur.
The data should be individual observations that are not collected in subgroups: If you can collect data in subgroups, use Acceptance Xbar Chart.
The data should include at least 100 total observations: If you have fewer than the recommended number of observations, then you can still use the control chart, but the results are preliminary because the control limits might not be precise. If you use the chart regularly, re-estimate the standard deviation and the control limits after you collect the recommended number of observations.
The process variation should be stable: If the process variation is not stable, the estimate of process variation and the acceptance control limits might not be reliable.
The data should be moderately normal: If the data are very skewed, you could try a Box-Cox transformation to see if that corrects the nonnormal condition. If your process naturally produces nonnormal data and the transformation is effective, you can use the chart of the transformed data.
The observations should not be correlated with each other: If consecutive observations are correlated, the process standard deviation might be underestimated. As a result, the acceptance control limits might be too wide, and the chart might incorrectly show some out-of-control points as in control.