Introduction to SpiralMath

SpiralMath is an EdTech startup based in the greater Baltimore region and a member of the Bytelion Analytics’ Business Intelligence Pilot Program cohort. SpiralMath is working on creating an instructional protocol and software package to help teachers present Common Core Mathematics in a way that improves student retention. With a year of pilot testing complete, the founding team from SpiralMath is working with Bytelion Analytics to validate their product is working effectively before making the necessary capital expenditures to expand.

The SpiralMath instructional protocol was developed by Joseph Mills, a career instructor and State Supervisor of Mathematics, utilizing five principles of learning psychology developed at Washington University, St Louis, where researchers have identified methods that improve student retention of new knowledge. To administer the protocol, students receive a daily five-question quiz that is completed, scored, and reviewed within a window of 45 minutes at the start of the day. Through intentional spaced repetition students build and affirm their knowledge of each concept over the course of the school year. 

The Data

The Common Core State Standards is a list of skills that students are expected to master. Typically, each year of schooling evaluates students on 20 to 30 standards. For each standard, Joe Mills prepared some 20 questions (similar, not identical) that are used to assess the student’s mastery over time. When equivalent questions are repeated at increasing intervals, students’ long-term retention improves. This approach is known as Assessment for Learning.

Throughout the 2018-2019 school year, SpiralMath worked with a third-grade teacher at a Baltimore City Charter school to pilot the program and collect initial data for analysis. Over the course of the year, SpiralMath collected over 9,000 data points from a class of 25 students. For the purpose of this statistical examination, Bytelion Analytics utilized a data set that showed the daily average results of the class population sorted by Common Core standards. Since the number of questions varied from standard to standard, our team took a sample of the standards that have been assessed at least 10 times. This provided us a sample data set that included 110 total data points spanning the results of 11 standards.

Statistical Testing

For our statistical testing, we began by identifying our research question: Will the SpiralMath Protocol lead to an improvement in test scores? To answer this question, we identified our null hypothesis: the SpiralMath Protocol does not lead to an improvement in the test scores of the class population. A key guiding principle for our testing was that we were looking to see whether the difference between each assessment instance and the first assessment was significant, i.e. was the difference between the 2nd assessment and the 1st an improvement and statistically significant, etc. A t-test is the testing procedure for significance between two populations was selected; it compares the difference between means and dictates whether the null hypothesis should be accepted or rejected. 

Overview of Results

At surface level, there was an improvement in the testing population from the 2nd assessment. This trend continued to show positive results as the number of assessments continued; however, results from the protocol became significant from the 7th assessment and onwards. We are able to conclude, with a 95% level of confidence, that the SpiralMath’s protocol helps students retain information in 7 administrations of their assessment, agnostic to the testing interval.

Graphic One: The class averages are shown over the course of the 10 assessments. 

Graphic Two: The results of the t-test display that from the seventh assessment onwards the improvements were significant to the testing population.

What is next for SpiralMath?

The results of this product validation are just the beginning for SpiralMath. Although our quick and simple method, by definition, provides us 95% confidence in our results horizon, there are many other variables to test in order to better understand what actually drives the retention of information for the members of the classroom. SpiralMath will begin phase two of pilot testing with an emphasis on validating the success horizon and understanding the relationship between variables, such as question formatting, length of the spacing interval, and the number of test questions, in their protocol to create an algorithmic administrative process.

A special thanks to the leadership team at SpiralMath – Joe Mills and David Robson, for joining the Business Intelligence Pilot Program cohort! To learn more about SpiralMath and enrollment information for their ongoing pilot program please visit their website: www.spiralmath.net. If you would like to explore how we can leverage analytics and business intelligence to help you achieve the goals of your business, please reach out to Sidd Chhabra, Business Intelligence Lead, at sidd.chhabra@bytelion.com or read more at www.bytelion.com.