Anomalies

Data points can be analyzed using the box plot portlet to determine normal values and identify outlier cases.  In this example, we analyzed how many contacts it was taking each stand (field period).  Outlier data is determined statistically by identifying the median point and calculating the top and bottom 25% from the median.  These plots are easy ways to determine cases outside of the norm and can show how the average may be skewed away from the median.

The data download function will show the anomalous cases.

The anomalies portlet provides a list of issues defined by study that should be reviewed to help improve completion or quality rates. 

One example of a common anomaly would be cases that have more than five contacts and are not yet completed. Another example is a list of interviews completed late at night or early in the morning. A supervisor can address each anomaly as they occur. The system maintains a record of each anomaly so patterns can be identified over time.