Please enable JavaScript.
Coggle requires JavaScript to display documents.
Twenty common testing mistakes for EFL teachers to avoid, Ana Cristina…
Twenty common testing mistakes for EFL teachers to avoid
General examination characteristics
Redundancy of test type
Increasing the number of subtests adds no significant variance-explanatory information to the test battery beyond that which may be obtained from the three or four best, reliable subtests.
Nothing is added beyond the existing components of the test in terms of ability
Lack of confidence measures
The manual provides us with information about the reliability and validity of the tests both what they are and how they were ascertained.
Need to ensure that the persons on whom the test was tried out in its evaluation stage are from the same general population as those with whom the test is ultimately used.
An insufficient number of items
Test reliability is directly related to the number of items occurring on the test.
Negative washback through non-occurrent forms
Through use of inappropriate structures of the language it is possible to teach errors to the students.
It is necessary that options include incorrect forms as distractors, it is best if these forms have some possible appropriate environment in the
language.
Tests which are too difficult or too easy
Boundary effects (Accumulation of scores at the lower or higher ends of the scoring range
Reduced capacity of the test to discriminate among students in their ability.
Unreliable and unsuitable test for evaluation
Care should be taken to prepare tests and items that have about a fifty percent average rate of student success.
Item characteristics
Divergence cues
Not to provide cues regarding the choice of the correct option.
Convergence cues
Test-wise students can identify the correct option because of content overlap.
Redundant wording
Needless repetition
Reduce the amount of information available from a given period of time available for testing.
Option number
Irregularity in the numbers of options
It is best to be consistent in the numbers of options used for items within a test
Trick questions
Must be avoided
Such items impair the motivation of the students, the credibility of the teacher, and the quality of the test.
Test validity concerns
Common knowledge
Items that require common-knowledge responses should also be avoided.
Syllabus mismatch
Failure of a test to measure adequately either instructional objectives or course content.
Wrong medium
Care must be taken that the response medium be representative
of the skill being tested.
Mixed content
Sometimes tests have been claimed to measure something different from what many of their items are actually measuring.
Content matching
Mere matching of a word or phrase in a test item with the exact counterpart in a comprehension passage does not necessarily entail comprehension.
Tests involving such content-matching tasks are usually invalid as measures of comprehension.
Administrative and scoring issues
Administrative inequities
Other factors as well may impair the reliability of the test.
Lack of piloting
It is important to try out the test on a restricted sample
from the target population before it is put into general use.
Inadequate instructions
If the students fail to understand the task, their responses may be invalid, in the sense that the students would have been able to supply the correct answers if they had understood the procedure.
Procedures should be carefully standardized even if this requires special training sessions for test administrators.
Subjectivity of scoring
Instructors give subjective, opinionated judgments of student performance.
If subjective judgment must be relied on, several mitigating procedures should be employed.
More than one judge should be consulted on marks assigned by other judges. The total of all judges’ ratings should determine the student’s mark.
Judges should make use of some precise rating schedule. Judges will be giving equal weight to the same kinds of performance.
Sufficient samples of language should be elicited from the students.
Lack of cheating controls
The teacher should take care to separate students, and where possible use alternate forms of the test.
When students obtain higher scores through cheating, tests are neither reliable nor valid.
Ana Cristina Reyes Vera