Herrera and Kapur begin there piece on “Improving Data Quality: Actors, Incentives and Capabilities” with a brief definition on data construction: an operation performed by data actors according to their own incentives and capabilities. Through their use of the adjective ‘own’ I came to interpret that the actors behind data quality and data construction were the primary actors behind the problems and therefore of manifesting solutions to such problems. However I came to realize throughout the article that the actors are subordinate to incentives and capabilities that aren’t falling under the category of being their ‘own’. Data actors don’t have control or ownership over factors such as norms or which countries are rich and poor and therefore provide accessible or feasible data collection. Lower down on the data supply chain we see factors such as fear, punishment and saying yes or no to material gain; which data actors can have more ownership and control over in the data construction punishment. It means lowering possible benefits to the data actor though; but I do believe that the data actor has the ability to make such decisions because these incentives and capabilities are fostered by him/herself. They are of his/her own manifestation. What the data actor can stand to personally gain or what opportunity cost values the actor places on different options. Yet further along the data supply chain there is a decrease in the data actors ability to manifest their own incentives and capabilities and therefore the problems associated with data set construction begin to fall out of the hands of the actor.

What I am left wondering is if we can better improve data quality at a lower level on the data supply chain? Thoughts or comments?