09 Spring Southwest Network University "Educational Evaluation and Measurement" the first answer

"Educational Evaluation and Measurement" answer

?

For the first time

I. Glossary title

?? 1 ,? norm:

??? norm is an average of a representative group of test subjects score groups. It represents the general situation of ordinary people. Differs from the standard norm.

?? 2, export scores

Export score is based on the conversion of raw scores, in accordance with certain rules, after statistical processing obtained a certain reference point and units, and can be compared with each other's scores.

??

?? 3, relative to the reference point

?? 4, the behavior of the sample

???? behavior refers to a group of samples are extracted, and as a direct measurement of the behavior of the object.

Second, the short answer questions

1, the meaning and source of measurement error

A: refers to a measurement error in the measurement process is a variation of those factors unrelated to the purpose of measuring the resulting measurements inaccurate or inconsistent effects. Here, we have to be understood in two ways: First, measurement errors due caused by those unrelated to the purpose of measuring: Second, the measurement error performance in inaccurate or inconsistent in two ways. There are three main sources of error, i.e. measurement tool, and the measured object Surveying process.

2, features percentile rank scores

answer:

3, two tests should be what kind of conditions are met can be called parallel test

answer:

4, the greater the difficulty of the test discrimination, the better, right?

A: Wrong. In fact, if all the projects tests are medium difficulty, inherent only in projects related to 0, the entire test scores only produce normal distribution. The actual tests, generally have some degree of correlation between each item, taking into account this point, we analyze the choice of questions, the difficulty of the project should be widely distributed some gradient larger, the difficulty of the whole test distribution and utilization projects a normal distribution, and to maintain an average level of 0.50. So can the various levels of people to distinguish, and differentiate too thin.

Third, the essay questions

1, in order to improve the measurement validity, what measures can be taken.

A: To improve the measurement must try to control the validity of civil random errors, systematic errors reduced, while particularly apt also select a good criterion, the validity coefficient calculating out accurately. Specifically, the validity improve the measurement methods are:

(A) well-prepared test scale, to avoid the system error.

This requires a good topic to be able to say on behalf of a sample to be measured content or structure, to avoid the subject bias.

At the same time, the degree of difficulty of the subject, but also to distinguish between proper reading, the number of items should also moderate. Too hard, too easily, too much, too little is detrimental to measure validity.

In addition, print a test paper, entitled to answer the requirements of the standard score scoring entitled meaning of the expression, etc., must be strictly checked to avoid all avoidable errors.

(Ii) test well-organized, random error control.

In the test process embodiment, the system error is less obvious, but there may be random error control. This requires testing the perpetrators must be in strict accordance with guidance language manual operation, to minimize interference unrelated factors.

(C) the creation of a standard examination situation, so that each test can play a normal level.

In various tests, some subjects often for various reasons should not play level (such as excessive levels of anxiety disorders and other causes), therefore, we should let the subjects change their attitude exam, get them from physical, psychological on the knowledge finest prepared as they should. Otherwise, anxiety and other factors unrelated to the impact is too large, it will inevitably reduce measurement validity, undetectable want to test the validity of the content or structure.

(D) selecting the proper criterion, the appropriate criterion will be a good measurement, correct use on formulas.

When evaluating a test is effective, efficient choice is an important aspect of the subject. If improperly selected criterion or selection criterion can not be quantified, it is difficult to correctly estimate the empirical validity of the measurement. If the criterion and criterion measurements are desirable, the selection formula also affect important aspects of validity estimated.

2, with practice instructions on how to report test scores to the test subject.

answer: