The Effect of Data Gap on the Organizational Learning based on Data: Focusing on the Exploration and Exploitation Process

In the age of datafication, organizations have access to more information than ever before and increasingly rely on data to make decisions. However, few approaches analyze the impact of data gaps on organizational performance, specifically focusing on the data gaps that may arise due to differences of connectivity among organizations and individuals. In this study, I modeled data gaps -instances where a dataset is missing a specific element or social group- in two ways: (1) when an organization's ability to provide connectivity is limited, resulting in missing data needed to measure the intended outcome in terms of construct validity, and (2) when connectivity to specific members is lost, leading to missing data from a portion of the target population in terms of representativeness. March's (1991) model of organizational learning, which emphasizes exploration and exploitation, was applied to analyze the impact of these two types of data gaps on organizational knowledge and the average knowledge of members, both of which are outcomes from the organizational perspective, as well as on similarity, an outcome from the individual perspective. The results showed that both types of data gaps restricted exploration and reduced learning performance from the organizational perspective, with the construct validity data gap having a particularly strong impact. On the other hand, representativeness data gaps created a similarity gap between those included in the organization's dataset and those excluded from it during learning.