Big Data, Small Data in Personalized Learning Professional Development

• In Event: 68.049 - Data, Big and Small

Abstract

Personalized Learning is a growing trend in U.S. classrooms. Personalized Learning (PL) means various things to various people (Redding, 2013; Zmuda et al., 2015), but tends to include the use of digital technology in order to enable students to learn at their own pace.

In order to prepare teachers to work in PL classrooms, a graduate certificate in personalized learning has been developed by a large tier 1, southern university. This graduate certificate was created through conversations with various district administrators, NGOs, and also through collection of both ‘big data’ and ‘small data’ from teachers, principals, and district administrators. This paper analyzes the types of confirming, harmonious, and also contradictory data we received from collecting both ‘big’ and ‘small’ data.

While, in theory, these two types of data collection can work together to create a deeper and more accurate understanding of a problem or phenomenon, the actual combination of these techniques is often very difficult (Lazer et al. 2014). While the combination of these types of data is difficult, multiple scholars have also argued that using only Big Data also tends to create problems and a lack of accurate understanding of a phenomenon or question as well (Lazer et al, 2014; Harford et al, 2015). Thus, we collected both ‘big’ data in the form of surveys and meta-data on the surveys. We collected ‘small’ data in the form of interviews and observations.

There were some items where the survey data and the interviews agreed—they painted the same picture. For example, both the surveys and the interviews suggested that teachers needed help with organizing their time for better PL teaching. However, there were other items where the teacher responses on the surveys were not confirmed by the interview and observational data. For example, interviews and observations suggested that teachers were not engaging in data-driven instruction (DDI), even though the surveys suggested that teachers believed that they were engaging in DDI. This created a dilemma. Interview data provides a deeper, thicker, first-hand account of what is happening. The survey data, though, does a better job of getting to more people.

This paper addresses both the harmonies and contradictions of the data, as well as the power imbalances that come with these different types of data as they are situated within the cultural imaginary. There are different types of bias in both interview data and survey data. The survey data speaks with the voice of authority that comes from a society that invests power into the idea of quantitative numbers, mapping analysis, means, and standard deviations; and this pushes against the more narrative work of interview data that focuses on deep on-the-ground experiences of a few people. This created a moment of reckoning as we navigated the data in order to create the courses that are part of the graduate certificate. We were forced to ask: Which type of data, and whose voice, takes precedence?

Author

• Heather Greenhalgh-Spencer, Texas Tech University