Is There a Correlation between Homework Completion and Learning?

I've just started a weekend workshop on data and assessment to drive student learning. We are being facilitated by Jennifer Sparrow, who is the assessment and data instructor at Singapore American School. As a member of the data committee of our school, we got the chance to meet with Mrs. Sparrow this afternoon for ideas for how our school could use data before the conference starts tomorrow and I left thinking about the concept of homework and its assessment as an means of student learning. I went home on Friday afternoon and spent the evening collecting and analyzing data from my first semester's math classes on a Friday night. Can you say "nerd"?

Does Doing Homework Lead to Learning?
Alfie Kohn argues no. He cites a number of case studies that say homework is no indicator of student learning in his book "The Homework Myth". I have met a number of educators that have accepted his conclusions as fact, but the thing that Mr. Kohn fails to mention in his book is the variablity and breadth of what homework is. For example, he does not mention what forms homework can take, nor does he mention the wide range of how homework is assessed. For his argument, homework is ultimately dreary, a chore and unrelated to learning outcomes.

I kindly disagree. Some teachers have insightful discussions about a homework assignment. Some merely give it a cursory glance and a rubber stamp. Some grade it, some don't. Such a variable must give the teacher the onus to assess whether or not learning outcomes are supported by its usage, and not merely use it as a means of grade inflation or deflation. It's an entry level formative assessment. However, homework, if assigned, must be meaningful, interesting and open-ended.

Getting back to data, I took it upon myself to answer the question: "Is there a correlation between homework completion and learning?" For this case study, I took a spreadsheet and graphed the number of missing or late assignments and what a student's semester grade was as a percentage by the end of first semester. I use a flipped classroom model which has very open ended responses but if assignments were not completed, students were asked to resubmit.

Analysis of C Block
For my "C" block class above, there is is a obvious but subtle negative correlation. Generally, the more assignments a student did not complete on time, the lower their semester grade was as a percentage. Students that completed all their assignments and had "0" marked as late had an average semester grade of 90.2% or an "A-". There were some outliers though. I had a student that had 4 late assignments who achieved a 91%. Also, I had a student that didn't complete four assignments (as pictured above) who had the lowest grade in the class at 70%. The lowest end of the range was a student that didn't complete 12 assignments on time or were incomplete who achieved 72%.

Analysis of H Block 
In the analysis of H block, there is also a negative correlation with noticeable outliers. One student had 12 late assignments and achieved an 85%. Also, one student that had 4 missing assignments which is well above the line of best fit. Still, group averages continue to support the hypothesis that more independent practice will raise mean understanding.

Sources of Error
This measure does not take into account the work that students did in class following the flipped learning model. Generally, students that had homework completed (notes, front loading of concepts ahead of time) came better prepared and used time more efficiently for in-class practice. Students that did not required more modelling, which gave less time for independent practice. The data points also represent individuals or groups of individuals which may affect sample size. There is also subjectivity regarding what it means for homework to be "well done". As the classroom teacher, if assignments were well written with explanations, definitions, and conjectures supported by examples, an assignment was reported as "complete".

In Malcom Gladwell's book "Outliers" he advocates that the more practice one does will increase mastery; he uses a benchmark of 10,000 hours of practice at a task will lead to mastery. The data above supports this claim as when students put in more time outside of class to review or prepare for concepts, they generally show higher levels of understanding on summative assessments. Another author, Paul Tough in his book "How Children Succeed" says that the basis for academic achievment is non-academic skills such as time management, grit, and the ability to ask for and seek help are the precursor to academic achievement. Because of this, another area of inquiry is the question: "Do students not understand because they don't do the homework or because they are unmotivated?"
It's no wonder that students who are detached, uninterested and apathetic tend generally show lower levels of understanding, although this is a statistical average, not a foregone conclusion.

We as educators must find ways to connect to these students. We must find ways to give them the practice and pedagogical approaches that adapt to their learning style. Every educator has a nagging feeling that when students don't put forth effort, they are setting themselves up for failure, so the data came as no surprise to me.

What's interesting about my data workshop has been this: everyone can collect data. Often there is a mountain of statistics that fall through the hands of teaching professionals every year. We as professionals must take the time to wade through this data of what we are doing and ask ourselves: "Is this an effective high-quality assessment that leads to learning?" We need not be published professionals, but rather humble teachers that are looking to improve our professional trade.

By the way, if you're reading this Mr. Kohn, I'm still a big fan. 

Alfie Kohn "The Homework Myth"
Paul Tough "How Children Succeed"
Malcom Gladwell "Outliers"

Related Posts
Formative Assessments in the Math Classroom
Turning Student Failure into Information
Using Data to Support Instructional Practice
Making the Most from Quizzes
Using Homework Effectively
Making Flipped Lessons Meaningful

Popular Posts