achieved when the number of observations is large enough for a teacher to exhibit a full range of teaching knowledge and skill. Consistency of performance is also established through repeated observations.

Data-collection methods can take many forms. Each has advantages and disadvantages. The choice among them is usually

constrained by tradeoffs between the type, quality, and amount of information gained, and the time and resources each requires. However, to serve the intended purpose, the choice of assessment form should be consistent with what one wants to measure and to infer. It is critical that the data and their method of collection yield information with confidence levels consistent with the consequences of its use. Public confidence in educational data and their use is related to technical quality. This public confidence is influenced by the extent to which technical quality has been considered by educators and policy makers and the skill with which they communicate with the public about it.

Assessment practices must be fair.

• Assessment tasks must be reviewed for the use of stereotypes, for assumptions that reflect the perspectives or experiences of a particular group, for language that might be offensive to a particular group, and for other features that might distract students from the intended task.

• Large-scale assessments must use statistical techniques to identify potential bias among subgroups.

• Assessment tasks must be appropriately modified to accommodate the needs of students with physical disabilities, learning disabilities, or limited English proficiency.

• Assessment tasks must be set in a variety of contexts, be engaging to students with different interests and experiences, and must not assume the perspective or experience of a particular gender, racial, or ethnic group.

[See Program Standard E and System Standard E]

A premise of the National Science Education Standards is that all students should have access to quality science education and should be expected to achieve scientific literacy as defined by the content standards. It follows that the processes used to assess student achievement must be fair to all students. This is not only an ethical requirement but also a measurement requirement. If assessment results are more closely related to gender or ethnicity than to the preparation received or the science understanding and ability being assessed, the validity of the assessment process is questionable.

ASSESSMENT TASKS MUST BE REVIEWED FOR THE USE OF STEREOTYPES, FOR ASSUMPTIONS THAT REFLECT THE PERSPECTIVES OR EXPERIENCES OF A PARTICULAR GROUP, FOR LANGUAGE THAT MIGHT BE OFFENSIVE TO A PARTICULAR GROUP, AND FOR OTHER FEATURES THAT MIGHT DISTRACT STUDENTS FROM THE INTENDED TASK. Those who plan and implement science assessments must pay deliberate attention to issues of fairness.