• The Spark

The Data Promise: Turning Data into Actionable Insight, Part I

Author:  Stevin Smith, Senior Vice President, Technical Services, Houghton Mifflin Harcourt  | 09/28/2017

Among the most powerful benefits that digital technology brings to education is the ability to generate and enable the utilization of data to improve teaching and learning outcomes. Yet despite all the attention to data as a transformative enabler in education, in practice the meaningful usage of data remains elusive (Papamitsiou & Economides, 2014). From big ideas like those being advanced by Max Ventilla at Alt-Schools, to more pragmatic calls from the Center for Digital Education to make data easier to access for teachers, efforts to utilize data more effectively are increasingly gaining the focus of the educational industry.

With more than 55 million K–12 students in the United States, the amount of teaching and learning data that is generated each day is mind-boggling. Yet this data alone has little value, and the sheer amount of it, even for a single mid-sized school district, makes it difficult to extract meaning and use it effectively. It is only through the application of analytics that data becomes actionable.

From Data to Decisions

Addressing the opportunity to utilize data to make better decisions about teaching and learning requires an understanding of the types of analyses that are possible, the specific data that is required to support these analyses, and the people who ultimately make use of the data. The intersection of analyses types, information types, and use cases defines the value of the analytics that can be performed. But regardless of the situation, there are five important steps to plan when launching a new data analytics project:

  1. Define the Question
  2. Identify the Data
  3. Gather the Data
  4. Conduct the Analysis
  5. Take Action

Intentionally planning around these five steps will help avoid common problems with data projects and help ensure a successful analytics implementation. This first post of my series on data analytics will focus on the first step:

Step One - Defining the Question

Figure 1

In the data world there are many different kinds of information. There is information about buying patterns, weather patterns, Internet usages and television watching. In the education world, there are essentially five types of information that matter most to edtech providers, represented in the five boxes in Figure 2.

These types of information range from basic information about what learners (and teachers) are doing in class to measures of growth, proficiency, and return on instructional investment. There is also an important category of information that addresses student emotions, empathy and feelings which, while less quantifiable, is an important part of the learning process that has been shown to have a positive impact on learning outcomes (Durlak, 2011). Additional information about each of the types of teaching and learning data are described, with examples, in the table below.

Figure 2: Framework for Using Teaching and Learning Data

Type of Information

Description and Example(s) from Education

Engagement

Tracks the usage of resources, such as digital programs and applications: Number of logins, time on task, activities completed

Fidelity

Identifies how well a learning program is implemented, whether the prescribed and/or recommended pedagogy is employed in the teaching and learning process: examples include activity sequencing, teacher behaviors and actions, formative assessment completion

Growth and Proficiency

Comparative evaluation of learner performance: examples include summative assessment score compared to benchmarks (proficiency), summative assessment changes over time (growth), and records of standards mastery progression

Return on Instructional Investment

Compares efforts to improve learner performance with actual changes in learner performance, in order to determine if the efforts expended (in time and/or dollars) are worthwhile investments, given a discrete set of resources; examples include comparisons of program utilization and in-program assessment results with scores on state standardized assessments

Social-Emotional

This information supports social-emotional learning, a process by which students acquire the knowledge and skills necessary to deal with emotional aspects of learning: examples include goal setting, building relationships and making good decisions

Using Data to Answer Questions and Make Decisions for Users

Information becomes actionable when the right teaching and learning data is analyzed appropriately and made available to the right people. Data users can be identified based on the questions that are being asked and answered, such as:

  • How are specific students performing in the classroom environment?
    • teachers, students, parents, school administrators
  • How are teachers and schools performing overall?
    • teachers, parents, school and district administrators
  • What strategies are having the greatest impact on learning outcomes?
    • teachers, school and district administrators, state and federal officials, researchers

While data can help users answer questions, its real power exists when advanced analytics are used to help make decisions that enable people to perform better and achieve more.

Watch this space next month for Step Two: Identify the Data.

  • Add your comment

    Your name will not be published to our website. Please enter comment name. We’re sorry—there are some invalid characters.

     

    Your nickname will appear with your comment. Please enter comment nickname. We’re sorry—there are some invalid characters.

     

    Please enter comment email.

     

    Please enter comment. We’re sorry—there are some invalid characters.