Develop a holistic data collection strategy that includes non-academic measures of impact.
While academic improvement is the primary purpose of a tutoring program, it is not the only goal. Programs need to collect data across multiple dimensions to ensure that they are serving all students equitably, for example, and find ways to qualitatively evaluate student experiences with tutors, not just student academic growth.
Programs should collect feedback from all stakeholders (students, families, teachers, and administrators) to understand and improve program impact at all levels. While achievement data and feedback from school partners is critical, programs should always include student voices when evaluating program impact: tutoring programs exist to serve students, after all, not parents or teachers or administrators.
Set specific benchmarks with expected dates to help stay on track.
Programs should set benchmarks with expected dates for all measures — not just for student growth, but also for aspects like student/tutor/teacher/parent satisfaction. Routinely reviewing data and comparing it to benchmarks helps programs understand where they are on-track or off-track; this is critical for establishing a data-to-action cycle of insights and iterative improvements.
Align routine assessments with session targets (and, ideally, with classroom curriculum).
Well-aligned, routine assessments can help programs quickly identify student knowledge gaps and target upcoming sessions to meet specific student needs as they emerge.
For formative assessments to result in more student learning, tutors need time and support to review the assessment and formulate a plan to address each student’s needs.
Develop systems for visualizing data for stakeholders.
Programs should develop in-house capability for distilling data so that information can be presented in a digestible and actionable format. Some programs may have databases and utilize software such as Tableau to visualize data, while other programs that operate at a smaller scale may find it sufficient to store data in well-designed Google spreadsheets.
Ultimately, the method chosen for visualizing data should allow for users to sort the data and easily extract insights.
Programs should regularly gather feedback on their data collection and visualization systems and improve upon these as part of their continuous improvement processes.