Federal Source Code Study Series Part 6: Structural Dimensions

By Joe Castle Ph.D. and The Code.gov Team

As summarized in Part 1 of this series, the study explored four organizational factors believed to be hindering or aiding agency publication of OSS. The organizational factors were cultural beliefs, public engagement, structural dimensions, and organizational location and each will be explored in an upcoming series post. We continue here with structural dimensions.

Structural Dimensions

Social structure in organizations provides arrangements for gathering information to achieve a collective outcome, the very reason for organizational life. Social structures in organizations pertain to relationships among people, positions, and organizational units. Researchers who study organizations often focus on the configuration of technology-structure and agreed that centralization, formalization, differentiation, and coordination were the major theoretical dimensions of structure, and each dimension, when isolated, affected organizations’ responses to environmental disturbance.

Each structural dimension theoretical element is defined here.

• Centralization described where decision-making took place in an organization. In more centralized organizations, decisions were made by those at the top of the overall hierarchy; while, in more decentralized settings, many organizational participants from differing levels and parts of the organization were involved in decision-making.
• Formalization included the organization’s use of written rules, job descriptions, procedures, and communications. Formalization was an attempt to make behavior predictable through explicit procedures (e.g., rules, standard operating procedures), and the lack of formalization allowed for unit flexibility and spontaneity.
• Differentiation entailed the separation of work by expertise or the diversity of skills within and across units.
• Coordination was the information exchange accomplished among units through communication.

This study focused on structural dimensions as potentially affecting the frequency of units’ OSS publication as the executive order can be viewed as an environmental disturbance.

Interview Guide & Structural Dimensions

Recall from Part 1 of this series, there were 25 participants from software development units in 20 CFO Act agencies. Some agencies had multiple participants, but no participant was from the same SW development unit.

The interview guide consisted of the following questions and statements pertaining to structural dimensions.

• General: How do structural dimensions, specifically technology-structure, influence OSS publication?
• Formalization: Describe how priorities and tasks are created and assigned in your unit. (Are they established as a result of roles, rules, and procedures or through group participation?)
• Centralization: Describe decision-making in your unit pertaining to software development.
• Differentiation: Describe how your unit is organized. (Division of labor, variation of skills.)
• Coordination: Describe your unit’s communication and collaboration efforts when developing software. (Formal and informal communication, frequency.)

Findings

The evidence supported generation of four categories of decision making (centralization), understanding capabilities (formalization), comprised of diverse skills (differentiation), and coordinating tasks (coordination). Unit decision making occurred in either a directed or a participatory manner. Directed decisions often came from upper management, customers outside the unit, and contractors through specifications in the contract. Participatory decision making occurred with unit collaboration supported with electronic tools.

Understanding capabilities provided focus on documented and undocumented procedures and policies. Interestingly, units that more frequently published OSS relied on formalization based on professional documentation of professional organizations. All units relied on documentation to publish OSS. If there was no policy in place, then the unit most likely did not publish OSS and if there was a policy in place, a unit would most likely publish OSS. Having a policy is one of the most important unit artifacts to allow for implementation of technology and broader agency and government policy.

Units comprised of individuals with diverse skills allowed for more publication of OSS and this was often due to some units having individuals who have coded at one point in their career. FSCP implementation was made possible with collaboration and communication across the unit. Informal and through electronic tools was key.

Summary

Variations in four structural dimensions — centralization, formalization, differentiation, and coordination — were associated with whether and how frequently units published OSS. Units publishing more frequently relied on less centralization with participatory decision-making, more formalization with documentation, more differentiation by having diverse skills within the unit, and more coordination with collaboration and use of electronic tools.

Coming Up

The next part of this series focuses on the organizational factor of organizational location and its effect on OSS publication.