In an education agency’s data governance system, a range of responsibilities may be assigned to staff to manage the collection, maintenance, reporting, and use of data. This chapter provides a more granular view of the practical workings of the process, detailing the personnel and activities involved; and describing a sample set of groups, roles, and responsibilities your organization may choose to assign in its data governance structure. The information is presented generically and should not be considered the only possible data governance structure. Rather, it is a suggested framework that may be adapted based on your agency’s specific needs, staffing capacities, and available resources.
Figure 3 presents a list of individual roles and groups that may constitute a data governance structure. As depicted in the figure, the data governance coordinator and data stewards work as individuals at the "core" of the data governance process. Several important peripheral roles may initiate, support, inform, or draw from the process.
Several groups should also work to drive the initiative forward, identifying data issues and collectively creating responses. As discussed in chapter 2, the core groups include the Data Governance Committee and the Data Steward Working Groups. Two peripheral groups, the Data Policy Committee and the Data Request Review Board, also serve important functions. The discussion below consolidates many activities into these four groups, though your agency may assign some of these tasks to additional groups.
The following draws on the Tennessee Department of Education (draft 2008), DQC 2008, EIMAC 2008, and Rozelle 2006 and 2008.
Several players may be involved in a data governance structure, but a few roles are essential and should always be
included. Specifically, every data governance structure should have a data governance coordinator, a group of data
stewards, and a Data Management Committee. However, aside from these key functions, many of the roles and
responsibilities may be assigned differently than in the manner outlined below.
What is important is not necessarily the roles you define, the titles you choose, or how the responsibilities are
assigned to each individual. What matters is that all major responsibilities are clearly assigned and staff know
their duties. In designating and managing these roles, at least three guidelines should be kept in mind:
A high ranking executive such as the education chief, or other high-level executive should provide support for the data governance initiative. Some of this staff member’s specific data governance-related responsibilities may include
The chief information officer (CIO) usually initiates the data governance process within the organization; and should take the important early steps to establish the initiative and the team. Thereafter, the CIO should remain heavily involved in many of the structure’s ongoing activities. Some specific data governance responsibilities may include
One individual should lead the data governance process: the data governance coordinator, perhaps the most important player in the data governance structure. In fact, no agency should undertake data governance without this single overseer. The coordinator should drive the data governance agenda, direct data steward activities around data quality efforts, and ensure that data issues are resolved. Many data governance efforts fail due to a loss of momentum, and the data governance coordinator must work as a catalyst to keep the initiative moving (NASCIO 2008). This individual must focus on the mission of the group, keeping an eye on the organization’s overarching goals, following timelines and managing deliverables, reinforcing the principles of good data stewardship, and working with staff to determine how the agency can better handle its data. The coordinator must stay positive throughout the process because data governance involves significant culture change and this can be hard on staff morale: the constant scrutiny and the focus on improving agency practices can be disheartening. The coordinator must assure everyone that this is necessary and will benefit the agency in the long run. Some specific data governance responsibilities may include
LDS Lore: Guns and governance |
|
In its Sunday issue, the county newspaper reported that a local school had recorded ten firearm incidents during the previous school year. The following morning, the school was a ghost town. Frantic parents flooded the office with calls, demanding to know why they had not been notified of these incidents. The staff assured callers that the article was wrong and children could safely return to school: there had been no firearm incidents in the school’s history. Zero. A call to the newspaper revealed that the reporter found the information on the state’s website. Sure enough, after a few mouse clicks, the school principal located the error. But how could this have happened? At the state level, the problem was traced back to three possible staff members, but all three said that reviewing those particular data was someone else’s job this year. A call to the district found similar confusion: no one could figure out who had dropped the ball there either. Further sleuthing found that a simple typo by an overworked school office staffer had made its way through the quality assurance efforts of both the district and the state, and into the state’s LDS. One thing was sure: a pervasive lack of clear roles, responsibilities, and procedures had clearly caused the education community a lot of embarrassment, not to mention unnecessary anxiety at the school. |
The "data owner" |
|
Some agencies distinguish “data owners” from “data stewards.” The “owners” may be program area directors assigned particular,
higher level authority for specific sets of data, while much of the work related to managing those data is assigned to
the "steward."
|
Data stewards should be selected to manage the organization’s data, with a single steward charged with managing the data and held accountable for ensuring the quality of specific data elements within a particular program area. It is critical that every data element be the responsibility of a single data steward. Note that some agencies divide the steward’s role into two levels, including the additional role of data owner in their data governance structure as discussed in the box above (KSDE 2008). Some specific data governance responsibilities may include
The public information officer (PIO) identifies and communicates data of interest to the public, and responds to data requests from the press. The PIO should be a member of the Data Governance Committee to stay apprised of data issues, and knowledgeable about ownership and data-sharing procedures. However, PIOs are not data stewards because they are not responsible for any data. When responding to press requests, agency procedures must be followed and data should be released only by the appropriate data steward to reduce the risk of inconsistent or inaccurate information. The PIO may also contribute to the Data Governance Committee by sharing news of any emerging public interest in certain types of data. (This role may also be fulfilled by the communications officer or other staff member with public relations responsibilities.)
As they are the source of all data the state agency collects, schools and districts should actively participate in the data governance process to ensure that local perspective informs all decisions. For instance, district representatives will know whether a proposed change to a collection can reasonably be met, and may suggest alterations to facilitate district compliance. Local education agency (LEA) staff can also offer information about district contracts, software, workload, and costs. They can help the state create solutions that will improve its communications and relationships with districts, and effectively decrease the local reporting burden. In addition, participating in the state’s data governance process will help increase the districts’ awareness of the importance of data quality and provide a model that could be implemented at the local level. At least one LEA representative should sit on the Data Governance Committee, but several would be preferable for a more rounded local perspective (representing districts of various sizes, from different geographic regions, etc.). Some states create an external group of districts and others, such as noneducation state agencies, specifically to review proposed changes to data collections. Some specific data governance responsibilities may include
LEA representative selection |
|
State education agencies (SEA) may find it useful to recruit local education agency (LEA) representation in the governance structure to increase perspective. To identify possible candidates, ask state program area leaders to identify district staff who
|
LDS Lore: The effects of disorder trickle down |
|
The districts were accustomed to supplying the state with a lot of data. Sure, they had to provide the same information multiple times, with some of the elements reported in both student-level and aggregate forms. Sure, changes to data collection requirements were often decreed without prior notice, and confusion about what was required and when it was due. And no, they didn’t really get much out of the deal—sending data to the state, but never getting anything useful back. If they realized the data were wrong due to a local or a state mistake, local staff often didn’t know whom to contact; and even state staff had trouble directing them to someone who could help. Since the state’s system didn’t meet local data needs, the district spent its time and money maintaining its own records, while grumbling about the state’s shortcomings. Unbeknownst to the districts, however, the state was implementing a data governance process. Gradually, word got around that several local representatives were working with the state to improve operations. Districts cheered the first sign of change: an email to the districts listing all of the state’s data stewards. Finally, district staff knew whom to call if they had questions about a particular program area or submission! Next came a data collection calendar accompanied by a notice that several of the state’s data collections were being eliminated or pared down due to redundancy. District staff gave a collective sigh of relief: not only was some burden being lifted, but they finally had an authoritative source of collection due dates. Of course, data governance hadn’t solved every problem, but it was quickly and noticeably improving operations as well as the districts’ relationships with the state. At the state level, goodwill and better data from the districts were confirming that their efforts were bearing fruit. |
The federal reporting coordinator oversees the federal reporting activities across the agency and, as such, should be involved in the data governance process. Some specific data governance responsibilities may include
In essence, the business analyst resides at the border between the business and technology sides of the agency and, within the organizational structure, may be in either area. Acting as a link between the two domains, the analyst harvests, assembles, and translates business needs into foundational technical specifications. One of the agency’s business analysts should actively participate in the data governance process. Some specific data governance responsibilities may include
Business first, then technology |
|
Program areas should lead the data governance process. Good data governance forces business staff to think more deeply about their needs. If the content area experts don’t think through the data’s purpose, definitions, and standards, IT will inevitably make assumptions in order to get the job done. If their assumptions are wrong—a common occurrence that is not the fault of IT—data quality will suffer and undue burden could be imposed on school districts. For instance, if a business rule is defined incorrectly (should a student who doesn’t earn a high school diploma but enrolls in college be counted as a dropout?) or if the option set for an element does not meet business needs (allowing “Null” in error, for instance), the resulting data may be problematic or even useless. In creating new specifications or amending existing ones, program area staff should work out the business solution to the very last detail, leaving no room for guesswork. Only then should IT be asked to determine how to implement the solution with technology. |
The following draws on the Tennessee Department of Education (draft 2008), DQC 2008, EIMAC 2008, and Rozelle 2006 and 2008, Chatis Consulting, and KSDE 2007.
In addition to the individual roles involved in the data governance structure, several groups of policymakers, data managers, and other stakeholders should be convened to address data issues collectively. When data issues affect multiple program areas, all affected stakeholders should be included to formulate the best response. These groups also help foster coordination and shared decisionmaking, ensuring that the agency approaches data issues consistently across program areas and over time.
The Data Policy Committee is a group of executive management stakeholders that may include the education chief (or other high-level staff member), the chief information officer, the data governance coordinator, and executive leaders from each program area that has data stewards. Rather than creating a new committee, it would be preferable to identify an existing group that includes these members and insert a standing time slot during regular meetings to focus on data governance issues. The relationship between this group and the Data Governance Committee (see below) is the critical link between leadership and those working directly with the data. The Data Policy Committee provides high-level sponsorship of the initiative as well as leverage for implementing major data-related decisions affecting multiple program areas. This committee also puts the executive “stamp of approval” on new or amended policies. Additionally, decisions that are contentious or beyond the authority of the Data Governance Committee should be referred to this committee for an authoritative resolution. For example, if staff resists a decision made by the Data Governance Committee, the Data Policy Committee can provide the authority to enforce it. Policies that significantly change the organization’s handling of data and data collections should also be discussed by this committee. This relatively small group need not meet as frequently as the Data Governance Committee, perhaps once every other month. Some specific data governance responsibilities may include
Data collection and reporting calendar |
|
Led by the data governance coordinator, the Data Governance Committee should create and maintain a comprehensive,
up-to-date calendar of data collection and reporting. The calendar should document and detail all current and planned
data collections, including due dates, descriptions of the data and their uses, collection format, and the staff member
responsible for each submission. This resource should be made available to all staff and to the public.
|
The Data Governance Committee is the center of control in the data governance structure, accountable for the quality of all agency data. It is an enterprise-wide group of data stewards chaired by the data governance coordinator. Ideally, the committee will also include the chief information officer, the public information officer, the lead business analyst, and at least one district representative (though involving multiple districts is preferable). Depending on the scope of the desired data system, this group may also include representatives from other external organizations, such as higher education; and other agencies, such as labor and child services. By including a variety of stakeholders, this group facilitates the collaboration necessary to address shared data issues. Meetings should occur frequently, perhaps monthly or even more if necessary. An effort should be made to make this group act as a resource for data stewards, rather than an obligation. Some specific data governance responsibilities may include
If the Data Governance Committee does not include much representation from districts and other affected stakeholders, the agency may adopt other strategies for collecting their feedback. For instance, it may hold in-person or online meetings with districts and other affected groups to periodically review proposed data collection changes. It is beneficial to include both program area and technology staff from each agency in these meetings. Alternatively, a separate group of state staff and affected external stakeholders may be formed to take on the responsibility of considering and approving proposed changes.
The Data Steward Working Groups are formed by the Data Governance Committee to resolve critical data issues that span more than one program area. Basically teams of data stewards, the groups work together to collectively identify the source of each issue and develop a solution that addresses each program area’s needs. While all members of a working group should contribute to formulating solutions, one steward should be responsible for ensuring that the team creates and implements an effective solution. (See appendix D for guidelines for Data Steward Working Group activities.) As Chatis Consulting explains, specific data governance responsibilities may include
LDS Lore: The committee becomes a resource |
|
When the agency began its data governance initiative, Gary was tapped as his program area’s data steward. He was less
than thrilled to find out he’d need to attend a monthly meeting with other stewards and staff from other program areas.
At the first meeting, he must have rolled his eyes a dozen times. “What a waste of time,” he thought. He had a ton of
work to do and needed to call one of the districts as its data were on the verge of being late for the third cycle in
a row.
|
Change management |
|
While getting a handle on the agency’s data early on in the LDS development process is very important, as time goes by,
it becomes clear that managing the changes made to the system is also crucial and challenging. As a result, some
have developed sophisticated approaches to manage these changes. System changes sometimes impose additional costs
and burdens on districts, and introduce data quality challenges. And poor management of change can also be a source
of tension between a state agency and its districts. For these reasons, a systematic approach to managing change
with the involvement of representatives from school districts and other affected stakeholder groups is essential
for the long-term success of the LDS.
A lesson from Texas
The Texas Education Agency has been collecting student-level data for nearly two decades. Early on, the agency realized how
difficult and costly districts found complying with changes to the state data collection (new data elements, altered codes
sets, etc.). The state quickly grasped the need to systematically manage the changes by involving districts and other
stakeholders in the process, and effectively communicating all modifications across the enterprise.
|
A Data Request Review Board may be formed to manage the data-sharing process and to handle data requests, which are likely to increase dramatically once the agency starts collecting student-level longitudinal data. This group of directors, legal counsel staff, and data stewards (or data owners) should meet monthly. Some specific data governance responsibilities may include