What makes a good formative assessment?
If teachers are going to solicit information from students to understand what a student knows and is able to do, then they need to understand what makes a good assessment. The National Research Council published "Knowing What Students Know: The Science and Design of Educational Assessment," which identifies three elements that underlie good assessments:
 Cognition: What do we know about how students learn?
 Observation: How do we create situations that allow us to observe student performance?
 Interpretation: How do we draw inferences from the performance?
Teachers must understand these elements to create good classroom assessments.
What do we know about how students learn?
Teachers may be able to identify some cognitive models for how students learn such as Bloom's Taxonomy or Dimensions of Learning. However, they may not have used it as a foundation for creating opportunities to evaluate what students know. Most teachers will need capacity building in the areas of creating opportunities for students to demonstrate what they know as well as in interpreting student performance. These opportunities for students should not be limited to tests or quizzes but should also include assignments, classroom activities, and projects.
How do we create situations that allow students to demonstrate what they have learned?
Creating good assessments is in large part dependent on teacher understanding of the content standard indicator/objective they are assessing and a clear idea of the characteristics of proficient work for students at a particular grade level. Richard Stiggins argues that "to assess student achievement accurately, teachers and administrators must understand the achievement targets their students are to master. They cannot assess (let alone teach) achievement that has not been defined." He believes that this is an essential condition for schools to be able to integrate assessment into the teaching and learning process.
For example, if an objective begins with evaluate and you ask the student to identify, then your question is not in alignment with the objective you're trying to assess. An ageold challenge for teachers is how to recognize the quality of the evidence, not just the number of pieces of evidence when determining proficiency. Often scoring rubrics used by teachers are quantitative—checklists and counts  rather than qualitative—evaluation of the depth of understanding students have of the material. This is one more area that teachers at the same grade level need to make sure they are on the same page.
There are many ways teachers can build in opportunities for students to demonstrate understanding in every instructional lesson. In Black and Wiliam describe a number of ways this can happen or can be inhibited in their article, "Inside the Black Box: Raising Standards Through Classroom Assessment."
"Discussions in which pupils are led to talk about their understanding in their own ways are important aids to increasing knowledge and improving understanding. Dialogue with the teacher provides the opportunity for the teacher to respond to and reorient a pupil's thinking. However, there are clearly recorded examples of such discussions in which teachers have, quite unconsciously, responded in ways that would inhibit the future learning of a pupil. What the examples have in common is that the teacher is looking for a particular response and lacks the flexibility or the confidence to deal with the unexpected. So the teacher tries to direct the pupil toward giving the expected answer. In manipulating the dialogue in this way, the teacher seals off any unusual, often thoughtful but unorthodox, attempts by pupils to work out their own answers. Over time the pupils get the message: they are not required to think out their own answers. The object of the exercise is to work out—or guess—what answer the teacher expects to see or hear."
How do we interpret student performance?
Understanding how to interpret student performance is another area that teachers may need to build capacity. Many teachers are still using holistic rubrics to score student responses and don't use them consistently as well as don't recognize hey are not formative assessments. If you are looking for diagnostic information to inform your instruction, then you can't use holistic rubrics. Though holistic rubrics are an excellent way to assess where a student is in their journey to proficiency, they are not a good way to determine what a student knows and still needs to learn because they use one score to represent the whole performance.
The collaborative discussions of the characteristics of proficient student work provide the basis for diagnosing and lead to invaluable insights not only about what students know but also about what they did or did not learn from the lesson. This information is a valuable resource for both addressing student needs and for adjusting instructional strategies.
