Reliability and Validity

According to the author a test is reliable if...

• The test measures what is supposed to measure.

• the test is neither too long nor too difficult.

• the test does not have any content that wasn't taught specifically in the classroom.

• if students get similar scores on the test regardless of the day when they take it.

The reliability coefficient is...

• A number that indicates the degree of validity the test has.

• A number that indicate how much we can trust in a test.

• A number indicating the amount of ideal items in a test.

• A number that indicates we can't trust a test.

Test-retest method is...

• A method to get two set of scores of the same test to compare.

• A method to make students retake a test to finally pass the exam.

• A method to ensure the test is valid enough and can be applied.

• A method to make sure not many students fail a course.

In the split half method...

• The students take one half of the test one day, the other half another day and then are marked and marks are compared to measure reliability.

• The class group is divided in two and they take two different versions of the same test.

• Students decide what half of the test to do and are scored on it. Then results are compared to measure reliability.

• Subjects take the test in the usual way, but each subject is given two scores, one for each half of the exam. The two scores are compared.

One way of making tests more reliable is...

• Including lots of objective multiple choice items.

• Include as many items as possible on the test.

• Including a lot of items without right or wrong answers.

• Including the most important structures and contents.

To ensure the high level of reliability of a test, the additional items included should be...

• Independent of each other and of existing items.

• Similar to other items and the existing ones.

• Fairly well devised and complex in their nature.

• reliable and valid.

In general, the more important the decisions based on a test, the...

• better written the test should be.

• longer the test should be.

• higher its validity should be.

• higher its level of difficulty should be.

To ensure the reliability of a test candidates...

• should be given options and freedom to complete a task.

• shouldn't be given many options. Tests items should be controlled

• should be provided with the most ambiguous items possible.

• should be allowed to complete tasks in the way they feel like.

One recommendation to ensure reliability when providing spoken instructions is...

• to use conversational language that sounds natural to students.

• to use language approppriate to the age and cultural background os students.

• to read them from a prepared text in order to avoid confusion.

• to read them as slowly as possible to make sure people understand.

Another way of making sure a test is reliable consist of...

• providing different types of questions to make sure the results are consistent.

• providing open questions so students can think of the best way to respond a test

• providing items that permit scoring which is as objective as possible.

• providing items that are very well-written.