• Featured
  • 11.17.20

The Trouble with Inconsistent or Incorrect Data: A Higher Education Perspective

  • by Sonia Schaible Brandon, University of Northern Colorado

Data governance structures in higher education can lead to a lack of understanding of the value of data within an institution. According to the most recent American College President Study, college and university presidents’ tenures continue to decline and institutions may face different administrations with contrary views concerning the value of data. Some may be very focused on data while others may only see data as a means to process paychecks and register students. For the latter, there is an increased risk that institutional data will become compromised and inconsistent. It is important for institutional researchers to recognize this possibility as one works towards maximizing the value of the data. 

Costs of Compromised Data 

Even if upper administration does not fully value data for decision making, institutional researchers should. As the country moves towards what is being called the “enrollment cliff,” all institutions should be wary of the changing demographics and how the distribution and decline of high school graduates will affect future admission pools (Grawe, 2018). Clean and reliable data produce more accurate statistical models and forecasts, and the information produced for enrollment management administrators is more trustworthy. Processed data, without a quality data governance framework, can very quickly produce bad information that doesn’t bear any fruit and is not useful in a competitive environment. Bad data can literally endanger an institution’s survival. 

Indicators of Compromised Data 

How does one go into an institutional research office and assess the health of an institution’s data? Often, staff who work closely with the data are better poised to identify issues. Perhaps the biggest issues are easiest to identify. Many times, data problems stem from a desire to “democratize” the data, and institutions will allow a broad range of staff to access and modify data without oversight or accountability. This can be a tough to address as data inconsistencies can run deep. There are also key clues to watch out for in an initial scan of the data:

1. Does the institution have data snapshots with multiple transformed tables not supported by the warehousing of native tables? Failing to snapshot native tables risks the data dictionary not being accurate or data being used inappropriately/beyond the intended meaning.

2. Does the institution have multiple versions of snapshot tables that contain “identical” information? This is where inconsistencies in reporting can often be found, particularly if native tables are not also being snapshot.

3. Is the institution using CASE statements to update transformed fields? If this is happening, it becomes extremely difficult to keep the data clean.   

4. Are transformed fields maintaining their meanings? For example, which of the following means Pell Eligible? A student who is offered Pell, one where Pell is accepted, or is it the student who is paid Pell? Frankly, the only student who has been deemed eligible and passed all verification is the one who was paid Pell. 

5. Do different analysts across the institution get different results from the data? This scenario makes it nearly impossible for administrators to trust the data and indicates a need to evaluate and improve existing data governance structures.

6. Are data inaccuracies being fixed in native tables or other snapshots? Data integrity checks and transparency are critical for maintaining reliable data. 

    Start a Data Governance Process, NOW! 

    Institutional researchers are the best poised to coordinate an initiative to evaluate and improve data governance at the institution. Create the data governance committee with key stakeholders, then gain the support of decision makers, and finally implement data governance procedures as institutional policy. Assuming administration is supportive of a data governance initiative these steps can be followed to get control of institutional data. If after assessing the landscape of institutional data, any of these clues have been found, it is time to act. If a data governance process is in place at the institution, periodic reviews are always helpful. In either case, proper care and curation of the data through data governance processes poises institutional researchers to produce accurate and actionable information. 


    American Council of Education. (2017). American College President Study. Retrieved from https://www.aceacps.org/. 

    Carruthers, C. & Jackson, P. (2018). The Chief Data Officer’s Playbook. Facet Publishing. 

    Grawe, N. D. (2018). Demographics and the Demand for Higher Education. Johns Hopkins University Press.