Summative assessments alone do not provide enough information to find out what is working, what’s not and what we need to do differently, Victoria Bernhardt told attendees at the fourth annual Education Leadership Conference (ELC). 

The ELC is an event sponsored by Dover-based Measured Progress as a way to encourage dialogue between education leaders of all kinds. Superintendents, teachers, principals, directors of professional development, testing and assessment directors, curriculum coordinators, directors of special education and other education leaders gathered on May 11 and 12 in Wakefield, Massachusetts, to address the conference’s central theme: “Taking the Lead: Assessment Literacy for School Improvement.” 

Bernhardt is the executive director of the Education for the Future Initiative and a professor of professional studies in education at California State University, Chico. She has published many books based on her research, including the series, Using Data to Improve Student Learning. 

During her keynote presentation, Bernhardt advocated combining multiple measures of data to get a complete picture of a school or district. “We collect data every day, but we seldom pool it together and look at it,” she noted. Bernhardt outlined these multiple measures of data—demographics, perceptions, student learning and school processes. 

In Bernhardt’s model, demographic information includes enrollment, attendance, drop-out rates, ethnicity and gender. She emphasized that demographic data are much more than basic information. “Just by looking at student demographic data,” she told the audience, “I can predict, with 92 percent accuracy, how long a principal has been at a school I have never visited.” 

Perceptions provide equally critical information. Bernhardt described a survey in which 10,000 students were asked the question: What has to be in place in order for you to learn? The overwhelming response was, “I have to know that my teacher cares about me.” Students’ ideas about teacher attitudes are among many measures of perceptual data, including values, beliefs, attitudes and observations. Bernhardt presented other survey results showing that students’ perceptions for the entire school year are shaped during the first few months. In her words, “You can’t start caring in January.” 

Data on student learning include standardized tests, such as norm- and criterion-referenced tests, and teacher observations. Bernhardt said that in her “perfect world,” teachers, administrators and other stakeholders would always be entirely clear on what they expect students to know and to be able to do; they would always agree about data on student learning. 

The final measure, school processes, includes information on instructional strategies, curricula, programs and assessments. “If we are not monitoring and measuring these processes and programs,” Bernhardt said, “they probably don’t exist.” When discussing school processes, Bernhardt emphasized the importance of comparing the goals of a program and who it was intended for with what is actually happening and if it is working. 

These four measures of data can be used to discern different information, such as who schools have as students, how those students want to learn, the results schools are achieving and the processes that get those results. Understanding this information is critical to school improvement. “If we want different results,” Bernhardt noted, “we probably have to alter the processes by which we get the results.” 

After describing multiple measures of data and their uses, Bernhardt led attendees in an activity. Participants were given cards with terms such as “Student Background,” “Core Values and Beliefs,” “Student Achievement Results” and “Curriculum” and asked to sort the cards into three groups—input, process and outcome. “Input” includes data elements that are “givens” and usually beyond immediate control. “Process” describes the actions schools and districts implement to get the outcomes they are trying to achieve. “Outcome” is the results of these processes. This activity fostered considerable debate among participants, and many attendees admitted to surprise when Bernhardt revealed her own classifications. She particularly stressed that very few elements should be classified as “input.” “The chilling thing,” Bernhardt said, “is that some educators believe that many outcomes are givens.” 

Bernhardt then turned her focus to school improvement. She described three “preconditions for school improvement.” These are: instructional coherence, a shared vision and data-driven decision making. A shared vision, she said, is absolutely essential in implementing data-driven decision making. Bernhardt said, “Anybody who talks to you about data-driven decision making without also talking about a vision hasn’t done it.” 

To achieve this shared vision, schools must have effective leadership. Bernhardt described effective leaders as those who challenge current processes with data, who inspire a shared vision, who enable others to act, who model the way, and finally, who encourage the heart.