Interim Assessments: A User's Guide

0809_SEP_1.qxp 8/5/08 8:20 AM Page 64 Interim Assessments: A User’s Guide Interim assessments are an important tool for school improvement, but they are easy to use poorly. Mr. Marshall provides 10 guidelines for using these tests effectively. BY KIM MARSHALL NTERIM assessments are hot in American and their intuitive sense of how to bring out the best I schools. Also called benchmark or periodic in children is confirmed by three strands of research: tests, these assessments are given every four • Benjamin Bloom’s work on mastery learning to nine weeks to check on students’ progress. (which found that when teachers look at unit assess- Small wonder they are popular, since they ment results and work to get all students to 80%-85% embody three powerful insights: first, that mastery before moving on to the next unit, year-end initial teaching, no matter how good, can’t achievement improves dramatically); bring all students to proficiency because of • the “effective schools” research (which found that differences in their prior knowledge, attention, and motivation; second, that we shouldn’t wait till the ■ KIM MARSHALL was a teacher, central office curriculum ad- end of the year to find out who’s confused; and third, ministrator, and principal in the Boston Public Schools for 32 years. that if we put our minds to it, we can fix many learn- He now works as a leadership coach and faculty member for New Leaders for New Schools, a nonprofit that recruits, trains, and sup- ing problems before they snowball. ports new principals in nine cities. He also writes the Marshall Great teachers, athletic coaches, and drama and Memo, a weekly e-newsletter that summarizes useful research music instructors have always applied these insights, (www.marshallmemo.com). 64 PHI DELTA KAPPAN Photo: PhotoSpin 0809_SEP_1.qxp 8/5/08 8:20 AM Page 65 beat-the-odds schools almost always made good use give teachers enough information to have useful con- of data from ongoing assessments); and versations with their colleagues on ways to improve • Total Quality Management (which showed that instruction. factories can produce higher-quality products if they • When interim tests are scored externally, teachers pay attention to input from teams of workers and em- have less ownership and interest and may shrug off the power them to stop production lines and fix problems test reports. When teachers have to go online to get on the spot). their results, navigating through complex websites, What happens when teachers don’t use interim as- few are likely to persist and extract the data they need sessment data? The achievement gap widens. As to improve their teaching. Grant Wiggins puts it, “The more you teach without • When turnaround time after interim assessments finding out who understands the information and is long (as much as three weeks in some districts), the who doesn’t, the greater the likelihood that only al- results are stale and outdated by the time teachers sit ready-proficient students will succeed.”1 Unfortu- down to discuss them. nately, this is a very common state of affairs, which is • When the “grain size” of interim test reports is too why most schools are engines of inequality. The stu- fine, teachers can get lost in the data and fail to focus dents who enter with disadvantages tend to be the on a few manageable challenges. When the grain size same ones who don’t understand after initial teach- is too large (e.g., data are reported only on “number ing, and they are also the ones who are harmed most sense” or “comprehension”), follow-up conversations when teachers move on without checking for under- become superficial and unhelpful. standing and following up. The rich get richer and the • Union or scheduling issues sometimes prevent poor get poorer. same-grade or same-subject teams from meeting to This is not the way we want our schools to be — discuss the data. This deprives teachers of one of the hence the popularity of interim assessments, which best forums to share best practices. While cross-grade claim to counteract such gap-widening tendencies. teacher meetings are good for many purposes, the most powerful conversations about data occur when PROBLEMS USING INTERIM ASSESSMENTS teams of teachers give common assessments to the same level of students on the same schedule — and In my work coaching principals in a number of dis- have enough time to pore over the results. tricts, I’m seeing problems. Here are some common • Many meetings designed to look at student data glitches in the use of interim assessments: fall victim to the “culture of nice” — teachers chat • Administrators sometimes fail to explain the ra- amiably and don’t confront ineffective practices or tionale behind interim assessments, and so teachers push one another to higher levels of performance. see them as “one more thing” from the clueless cen- • Astonishingly, some schools give interim assess- tral office. (All we do is test, test, test. Why don’t they ments and then don’t follow up with re-teaching and just let us teach?) When this happens, teachers tend help for struggling students. This is the biggest reason to communicate their negativity to students, thereby that critics of interim assessments charge that they are souring the whole process. nothing more than summative tests scheduled during • Teachers often fear that interim tests will be used the school year. to blame them for student failure. This makes them • Some schools use the results of interim tests to fo- tighten up and not engage in the kind of free-flowing cus only on the “bubble” students — those on the discussions of assessment data that can drive improve- cusp of proficiency who might, with just a little im- ments in teaching and learning. provement, help a school make AYP (adequate yearly • Commercial interim tests are often poorly aligned progress). This amounts to educational triage and with standards, state tests, and pacing calendars. does a huge disservice to other students who need When students are required to take tests on material help. they haven’t been taught, they get discouraged, and As I’ve watched well-intentioned, hard-working their teachers get mad. educators make these mistakes, I’ve realized that in- • When interim tests are given only two or three terim assessments are a lot harder to implement well times a year, teachers can’t fix learning problems in a than a lot of us thought. My colleagues and I in New timely manner. February is too late to find out about Leaders for New Schools have gradually honed a list serious gaps in understanding. of the most important steps in implementing interim • Interim tests that are short and superficial don’t assessments so that they really make a difference. SEPTEMBER 2008 65 0809_SEP_1.qxp 8/5/08 8:20 AM Page 66 INTERIM VS. FORMATIVE ASSESSMENTS guidelines to help schools Many educators are using the terms “interim” and “formative” interchangeably and are fuzzy about the difference between these two quite different types of 10 exploit the full potential of interim assessments. If you are using interim assessments, you during-the-year assessment. Let me clear up this am- biguity and spell out the very different possibilities may want to rate your school on a 4-3-2-1 scale and challenges that each presents. to see which of these areas are going well and In their influential 1998 study, “Inside the Black which need work. Box,”2 British researchers Paul Black and Dylan Wil- iam focused almost entirely on what I’ll call in-the- 1. Build understanding and trust. moment assessments — not the more formal, every The principal needs to explain interim assessments to four- to nine-week interim assessments that Mike the leadership team and the full staff and make sure Schmoker, Douglas Reeves, Richard DuFour, Jeff everyone understands the powerful role these assess- Howard, and other U.S. educators have been touting. ments can play in closing the achievement gap. Teach- The in-the-moment toolbox includes: students writ- ers need repeated assurances that interim assess- ing answers on small whiteboards and holding them ments are low-stakes tests and will not be used as part up so the teacher can gauge the overall level of under- of performance evaluations. Principals can build trust standing; “clickers” that gather instant electronic da- by distributing copies of the tests in advance and in- ta on in-class quizzes; journal writing and exit cards volving teachers in tweaking and improving them. to give teachers quick feedback on what students un- >> The outcome should be a climate in which con- derstand; more effective teacher questioning; and tinuous adult learning can take place. Data without methods that randomize which students are called on blame. (e.g., pulling popsicle sticks out of a can, each one with a student’s name written on it). 2. Clarify learning outcomes. It makes perfect sense that these in-the-moment as- All teachers need clear, manageable standards that sessments improve teaching and learning. If teachers spell out what their students should know and be able find out immediately which students don’t under- to do by the end of the year. And they don’t need them stand and which concepts aren’t getting through, they on web sites or in hulking three-ring binders, but in slim can clarify and re-teach before misconceptions and booklets right on their desks. misunderstandings widen the achievement gap, and >> Standards should be visible to students and par- they can use the insights to teach the concept more ents and backed up by exemplars of proficient stu- effectively the next time around. In addition, when dent work. No surprises, no excuses. students know they might be quizzed on their under- standing at any moment, they are more engaged and 3. Set a multi-year target and annual active learners. Expert teachers have always known SMART goals. this, and now Black and Wiliam and their colleagues Major gains in student achievement don’t happen are helping thousands more educators add these pow- overnight, so it’s very helpful for the leadership team erful methods to their repertoires. and the teachers to agree on an ambitious yet attain- There’s another way that in-the-moment assess- able long-range goal. For example, 85% of fifth graders ments help improve learning: recent research at Wash- will be reading at Fountas-Pinnell Level W (instruction- ington University in St. Louis (reviving research from al level) by June 2012. Grade-level teams can then set the 1930s) is showing that quick assessments (within annual SMART (Specific, Measurable, Attainable, Re- 24 hours of initial teaching) significantly improve sults-oriented, Time-bound) goals. For example, 85% long-term memory by helping students retrieve, re- of first graders will be reading at Fountas-Pinnell Level view, and make better brain connections with new in- I by the end of this year. formation.3 In-the-moment assessments are powerful. When >> SMART goals should be ratcheted up year by year, as high-achieving students progress through teachers use them well, more of their students will the grades. reach high levels of achievement. But the other kind of assessment for learning — interim assessment — 4. Get good interim assessments. checks for understanding several weeks after initial in- Whether it acquires them from the outside or writes 66 PHI DELTA KAPPAN 0809_SEP_1.qxp 8/5/08 8:20 AM Page 67 them in-house, each school should have reading, writ- to teachers, administrators, students, and parents. Dis- ing, and math tests that define on-the-way-to-college playing data in such a way can answer these questions: standards; are aligned with the sequence of school- How did students do on each test item? How did stu- based curriculum materials; reassess previous stan- dents do on each standard? What’s the big picture of dards to provide ongoing, cumulative review and a achievement at this point (i.e., what proportion of stu- sense of progress; evaluate higher-order thinking skills; dents is proficient and above)? and include both open-response and multiple-choice >> Robert Marzano has found such graphic display questions, as well as writing prompts with user-friend- of data to be the second most powerful factor in ly scoring rubrics. Interim assessments present the boosting achievement,4 and it is especially effective classic Goldilocks dilemma: they need to be long when teachers and administrators see the names of enough so teachers can have substantive conversa- individual students and how each of them is doing. tions about the results, but not so long that they are overwhelming to administer, score, and analyze. 8. Hold candid data meetings and >> Interim assessments are, by their nature, low- planning for action. stakes and don’t have to be psychometrically per- Discussions of interim assessment data need to take fect. However, they must be good enough and long place as soon as possible after each round of tests in enough to provide teachers with real insights for same-grade/same-subject teacher teams (or, in very classroom follow-up. small schools, in one-on-one meetings between teach- 5. Schedule assessments and time for ers and principals or instructional coaches). To be ef- immediate follow-up. fective, these meetings need to be hard-hitting, honest, test-in-hand, and low-stakes. They should celebrate Principals should block out time in the calendar for in- successes and then examine what students got wrong terim tests every four to nine weeks — along with times and figure out why they got it wrong. Focused data con- for rapid scoring, analysis, and data meetings (ideally versations rarely happen without a guiding hand, and within 48 hours). It’s also important that there be sever- many schools have found it helpful to have these data al days for re-teaching after each interim assessment meetings facilitated by someone from outside the team. cycle. >> Teachers should leave each meeting with specific >> Unless these dates are in everyone’s calendars, plans for next steps, such as a battle plan for whole- interim assessments will be constantly pushed aside class re-teaching, small-group explanations, tutori- by other events, and teacher meetings to look at da- als, and after-school work; teaching points for high- ta won’t happen with regularity. risk students; and distributed before-class work, mini- lesson, and homework topics. 6. Involve teachers in making sense of the assessments. 9. Involve students in the process. Teachers may complain about the work of scoring and analyzing interim assessments, but if time is set aside Curriculum goals and interim assessment data have (without taking them away from students), teachers will even more impact when they are shared with students end up appreciating and learning a great deal from so that each knows the answers to these questions: working on their own students’ tests. Of course, Where am I going? Where am I now? How am I going schools should take advantage of scanners to score to close the gap?5 multiple-choice items, but teachers must score the writ- >> Student investment in the improvement process ten responses of their students and stay close to the is one of the great untapped resources in American item analysis so they can celebrate their students’ suc- schools. cesses and form initial hypotheses about why students did poorly in some areas. 10. Follow up relentlessly. >> The heart of the process of interim assessment is Interim assessments are a waste of time if teachers teachers making new instructional decisions based don’t implement their action plans and check to see if on timely information. students improve. Richard DuFour and his colleagues have done some of the best work in this area, describ- 7. Display data effectively. ing schools that refuse to let students fail.6 Succinct spreadsheets and wall charts should make >> Principals need to monitor the teachers’ follow- students’ current status and progress graphically clear up efforts and provide as much support as needed. SEPTEMBER 2008 67 0809_SEP_1.qxp 8/5/08 8:20 AM Page 68 struction and takes advantage of a more formal struc- help them feel less threatened and more confident ture to provide an additional boost that can take about taking any kind of test. But interim assessments teaching and learning to even higher levels. Here’s can also be performance tasks and essays scored by how: rubrics, which can tap into the kinds of understand- • Interim assessments check to see if students re- ing and knowledge that state tests cannot. member material from one to nine weeks ago, some- So which are more helpful, in-the-moment or in- thing in-the-moment assessments can’t do. terim assessments? This is not an either/or question: • Interim tests can be more wide ranging and rig- we need both! But here’s a theory. Interim assess- orous, tapping systematically into students’ knowl- ments, done right, can have a ripple effect: they can edge and understanding of what’s been taught over fuel improvements in every other stage of the teach- several weeks and requiring students to apply what ing/learning process. This is why I believe that inter- they have learned in novel situations. im assessments, if handled well, constitute the most • The results of interim tests can be made visible to effective single initiative that a principal can imple- teachers, administrators, and students in spreadsheets ment. They can help teachers plan better, teach bet- and wall charts, which means they can be analyzed ter, use in-the-moment assessments better, and make and discussed more thoughtfully. powerful use of interim data to help close achieve- • Interim assessments, if they are cumulative, allow ment gaps during each year. teachers and administrators to track students’ progress as the year unfolds. While the idea of using during-the-year assessment • Data displays make it possible for same- data and sharing the outcomes is not new, it runs pro- grade/same-subject teams of teachers to discuss col- foundly against the culture of most American schools. laboratively what students misunderstand, why they Our tradition is for teachers to work in isolation and misunderstand it, and how the material can be taught be swept along by the pressure to “cover” the curricu- more effectively. Whereas in-the-moment assessment lum — and for principals to supervise and evaluate data are usually seen only by one teacher, team discus- the process of teaching rather than discuss the results. sions take assessment data out of the privacy of the The challenge for principals and other school leaders classroom and make possible a synergistic discussion is to get teachers to slow down, reflect on what’s work- of best practices across several classrooms. When ing and not working, and orchestrate a continuous teachers confront specific data on their students’ process of self-improvement, driven by insights from short-term errors and confusions, admit that certain real-time assessments. This ongoing conversation is teaching practices aren’t working, and listen to the vital because changes in teachers’ practices are deeper ideas of their colleagues, teaching improves dramati- and more lasting when they come from within, as part cally. of an ongoing, low-stakes, collegial dialogue about • Interim assessment data allow principals, other ad- the best ways to get all students to high levels of ministrators, and instructional coaches to get involved achievement. with the teams as they look at interim test results. • When administrators who are familiar with data 1. Grant Wiggins, “Healthier Testing Made Easy: The Idea of Authen- tic Assessment,” Edutopia, April 2006, available at www.edutopia.org/ discussions visit classrooms, it’s “as if they have healthier-testing-made-easy. donned 3-D glasses,” says Paul Bambrick-Santoyo, a 2. Paul Black and Dylan Wiliam, “Inside the Black Box: Raising Stan- school leader in Newark, New Jersey. They have much dards Through Classroom Assessment,” British Education Research As- sociation, short final draft, 6 November 2001. See also a version of the better insights into what’s going on as the curriculum same document, Paul Black and Dylan Wiliam, “Inside the Black Box: unfolds, and they can shift the conversation to results Raising Standards Through Classroom Assessment,” Phi Delta Kappan, when they give feedback to teachers. October 1998, pp. 139-44, 146-48. • Because reports of interim assessment data con- 3. David Glenn, “You Will Be Tested on This,” Chronicle of Higher Ed- ucation, 8 June 2007, p. A14. tain the names of struggling students and the specif- 4. Robert Marzano Classroom Assessment and Grading That Work ic areas in which they are having difficulty, they can (Alexandria, Va.: Association for Supervision and Curriculum Develop- be used to identify students for systematic follow-up, ment, 2006). including small-group tutoring and focused interven- 5. Rick Stiggins and Jan Chappuis, “Using Student-Involved Classroom tions with students of major concern. Assessment to Close Achievement Gaps,” Theory Into Practice, Winter 2005, pp. 11-18. • Interim assessments can simulate the content, 6. Richard Dufour, et al., Whatever It Takes: How Professional Learning format, and rigor of state tests, which can help reduce Communities Respond When Kids Don’t Learn (Bloomington, Ind.: Nation- students’ stress when they take formal state tests and al Educational Service, 2004). K 68 PHI DELTA KAPPAN