Each of the four data sources we used groups organizations into subcategories, but each does so differently, according to its own system. These native systems do not match up.

To overcome this issue, we created a new classification system of 14 subtypes that cuts across all four sources. Individual organizations are then allocated to these subtypes in the dataset.

It is important to note that these subtypes are not mutually exclusive, meaning organizations can be classified as more than one at once. Nor are they collectively exhaustive: while nearly all providers have at least one subtype, a few providers don't correspond to any of our new subtypes. This is because of limitations in the data sources' descriptions of providers. Organizations in these subcategories are in the ‘To be determined’ bubble in the chart to the right.

Nonetheless, our new subtypes system represents a valuable step towards creating a truly integrated view of workforce training providers in the US.