Colleges and universities are at a crossroads with regard to students' data. They have more information at hand than ever before, but exploiting it to generate a significant change remains a challenge. A UCLA-MIT 2022 press study revealed that Higher education has trouble capturing and taking advantage of data for the impact. This digital disconnection is not only the result of obsolete systems; It is the complex network of cultural, organizational and infrastructural obstacles which leave many institutions rich in data but poor.
To discuss how institutions can transform raw data into real impact, Edsurge has met with Suzanne Carbonaro, vice-president of post-secondary education and programs 1edtech consortium (1edtech). With 27 years of experience in higher education and evaluation, she was a member of the faculty, has occupied leadership roles in evaluation and accreditation, and led the development of competence programs at the Philadelphia College of Pharmacy, the oldest pharmacy school in the country.
EDSURGE: What types of data do the higher education institutions find the most difficult to access and why?
Carbonaro: Despite the abundance of students' data, higher education establishments are faced with important access challenges and to use it effectively. The first problem is (the existence of) of data silos. Learning applications, student information systems (SIS), financial assistance platforms and test applications often work independently, without communication between them. While students move in a fluid way between these systems, their data does not. Each system exists on its own “island”, disconnected from the others and the files of holistic students.
Second, there is a bad signal / noise ratio. Even when learning applications share data, a large part is not structured or lack of context. For example, random data of random clicks often do not shed light on the learning path of a student. In addition, different systems can use inconsistent identifiers for the same student, which makes it difficult to increase or connect data to platforms.
Third, the cost of solving these problems is prohibitive for many institutions. Diskelling this jungle of data often requires external consultants or expensive tools that many colleges and universities simply cannot afford.
What key obstacles prevent institutions from obtaining and using this data effectively?
Institutions are fighting with more than technical challenges; They are also faced with cultural and organizational barriers. Professors often feel judged by analyzes or decisions taken on the basis of incomplete data. This distrust can hinder membership to adopt new tools or processes.
Privacy problems also play a role. Institutions must ensure that requests meet rigorous confidentiality, security and accessibility standards before adoption. For example, AI tools should use data responsible, improving learning results without storing sensitive information in owner data lakes controlled by the supplier.
Finally, institutions often do not know what questions to ask about their learners in advance. Without clear objectives or managers for the use of data, they risk collecting information that is not usable or waiting for too long to intervene when students need support.
How can the adoption of open standards help institutions access and take advantage of this data more usually?
Open standards serve as the basis for the resolution of these challenges. Think about it like plumbing in a house: a standardized infrastructure allows you to connect any tap or device transparently. In the same way, Interoperability standards as Backstage analytics And Interoperability of learning tools (LTI) Make sure EDTECH tools can work together without disruption when institutions change suppliers or adopt new technologies.
For example, open standards allow institutions to follow significant data on learner's events – such as clicks, time spent on tasks or questions posed in AI tools – and contextualized alongside other information on holistic students. This structured approach eliminates silos and makes the data usable in real time.
In 1edTech, we create open standards that connect disparate systems to coherent ecosystems. These standards allow institutions to modify suppliers if necessary without losing access to critical data or operations disturbance.
Can you share specific examples of improving data access had a positive impact on the success of the learner?
In the teaching of the pharmacy, the faculty has aligned the results of the programs with exam issues via a test platform, allowing real-time monitoring of student performance. By quickly analyzing this data, we have identified students who have struggled with specific fundamental knowledge areas. The connection of this information to the attributes of the learners helped us support students from different secondary schools or colleges who needed additional help before the exam. This also allowed collaboration with these institutions to strengthen critical concepts for future cohorts.
Another example comes from our work using Learner's complete files (CLR). By linking the pharmacy skills to key missions between courses in modules and allowing students to see their performance in time almost real via CLR dashboards, we allowed them to appropriate their learning trips. Students and their mentors could see the trends in the lesson months – not just the grades – and make informed decisions in the place where to concentrate their efforts.
Currently, we are working on a subsidy of $ 20 million from the National Science Foundation with the Georgia Institute of Technology and other institutions to study the impact of seven different AI assistants deployed in online courses aimed at supporting adult learners. Initially, this project was faced with challenges due to disparate AI applications issuing different data flows in separate visualization tools, without any way to combine data for longitudinal discovery. By implementing a tripod of open data standards – Edu -dep, LTI and analysis of calidators – we have unified these systems in a single cohesive pipeline which provides contextualized information on the learner's commitment.
AI applications go from tutors supporting the gaps in fundamental knowledge to social connections facilitators designed for online learners who could otherwise feel isolated. By consolidating these tools in a single reference architecture using open standards, we allowed institutions like Georgia Tech to evolve their efforts while maintaining flexibility between platforms.
What can other institutions do now to access this data in a way that provides significant information?
Institutions can take immediate measures to improve access to usable data:
- Ask open standards: When publication (requests for proposals) or the purchase of new tools, clearly indicate that suppliers must provide data in standardized formats such as calidators analysis rather than unstructured CSV files.
- Use predefined pipelines: For kindergarten districts to the 12th year and other post -secondary institutions without resources to build their own infrastructure, access executives of open standards, such as learning data reference architecture (LDRA).
- Focus on real-time data: The collection of data -based data, such as which has used which tools and for how long, combined with other key measures, such as results -based evaluation data, allows institutional stakeholders to be proactive to support their learners rather than waiting for weeks for ideas that can already be exceeded.
- Ask the right questions: Instead of collecting data in a reactive way or using the intestine -based decision -making, start by identifying what you want to know about your learners in advance so that you can personalize their learning and identify the support services they need for their success.
By taking these measures NOWInstitutions can create a base for more efficient decision -making and support for learners.
What work should still be done?
Although progress has been made in the construction of open architectures and pipelines like LDRA, there is still a lot of work in advance:
- Prepare confidence: The faculty must ensure that the analysis is supposed to support – not to judge – their teaching practices.
- Professional development: Professors and administrators must understand why interoperability is important and how it benefits learners.
- Privacy standards: Institutions must continue rigorously verification applications for confidentiality and security problems while guaranteeing accessibility for all users.
- Scaling solutions: Models like LDRA must be extended beyond pilot programs in large-scale implementations in various educational contexts.
1edTech is a united community committed to reaching an open, reliable and innovative education technology ecosystem that meets the needs throughout the life of each learner. We unite the community of education to build an integrated base of open standards which allow educational technologies to better work for everyone – reducing complexity, accelerating innovation and expanding the possibilities for learners around the world.
