Current location - Training Enrollment Network - Mathematics courses - Understanding and Cases of Reliability and Validity-Partial Supplement
Understanding and Cases of Reliability and Validity-Partial Supplement
Let's take a popular example first: If you weigh it three times in a row on the weighing scale, the readings are different, which means the scale is unreliable.

Second, learning-related examples: a questionnaire, which the subjects did not fill in carefully, is said to have no reliability.

1. Popular example: If you weigh it three times in a row on another scale and the readings are the same, it means that this scale is reliable. However, if the three consistent readings are not accurate, it means that the scale is not effective enough or not. At this time, another weighing scale was used, and three consecutive readings were consistent and accurate, indicating that this scale is reliable and effective.

Second, learning-related examples: a very simple math test, many subjects answered correctly, indicating that the math test is reliable, but it can't test the math ability of the subjects. If we change a set of data test questions with moderate difficulty, the number of correctly answered questions in many subjects may be different. If the scores of these subjects can be divided into good, medium and poor, we can think that the test questions have good validity.

It refers to using the same measuring tool to measure the same group of subjects repeatedly at a certain interval to investigate the correlation between the two measurement results. Correlation analysis can be directly used, and the obtained correlation coefficient is the retest reliability coefficient.

It means that the same group of subjects fill in two parallel questionnaires at a time and calculate the correlation coefficient of the two data. The reliability of duplicate requires that the two questionnaires should be identical except for different questions, which is difficult to operate in practice.

It refers to dividing a questionnaire into two parts and calculating the correlation coefficient between the two parts, that is, the semi-reliability coefficient, so as to measure the reliability of the whole questionnaire.

It is the most commonly used method to measure the reliability of internal consistency. The calculated Cronbach α coefficient is the average of all possible semi-reliabilities, and the value is between 0- 1. The higher the coefficient, the better the consistency, indicating the higher the authenticity of the data. α coefficient is the most important index to test the authenticity of questionnaire data.

It is a qualitative evaluation standard, mainly judged by experience, mainly by experts and experienced insiders. You can also use the questionnaire in a small scale before it is officially used, and modify the items according to the results to illustrate the effectiveness of the questionnaire. Generally speaking, content validity does not need to be analyzed by SPSS, but needs the guidance of experts and authoritative teachers, pre-test the revision process, and finally explain the content validity of the questionnaire.

Refers to the corresponding relationship between measurement items and measurement direction.

The measurement result is artificially designated as the "accuracy standard", and whether other results to be measured are consistent with it is investigated. For example, the multiple-choice questions in the test paper will have accurate answers, so the correct rate of each question is the effectiveness of this lesson on this question.