Reliability classification:
1. Test-retest reliability: the consistency of the measurement results of the same scale at different times. If one weighs two kilos today, it will be one and a half kilos tomorrow, indicating that the reliability is not high.
2. Half reliability. Divide the measurement items into two halves by a certain method, measure them separately, and get the correlation of the results. Divide two kilograms into two kilograms, and if they are 1 kilogram and one and a half kilograms respectively, it means that there is something wrong with the reliability of folding in half.
3. Reliability among raters: the consistency of the results obtained by different raters using the same tool. Several people use the same scale to weigh them separately. If they are all two kilograms or close to two kilograms, it means that the reliability of the evaluator is high.
4. Replica reliability. Correlation between various measuring tools. If you weigh two kilograms with this brand and two kilograms with another brand of the same brand, it means that the copy is highly reliable.
Validity classification:
1, the coupon is valid. The results of empirical observation and measurement are the indicators to be measured. It looks like white sugar, which we subjectively think is white sugar. Of course, it may also be washing powder or salt.
2. Content effectiveness. Test results of related industries and specialties. Please ask the relevant authoritative experts to evaluate whether it is white sugar.
3. Structural validity. What are the measuring properties of measuring tools? From the taste, color, chemical structure and other aspects of sugar all-round measurement, evaluation and demonstration.
Is this ok?
Question 2: What are reliability and validity? Reliability refers to the consistency, stability and reliability of test results. Generally speaking, internal consistency is used to indicate the reliability of the test. The higher the reliability coefficient, the more consistent, stable and reliable the test results are. System error has little influence on reliability, because system error always affects the measured value in the same way, so it will not cause inconsistency. On the contrary, random errors may lead to uncertainty, thus reducing reliability. Reliability can be defined as the degree to which random error r affects the measured value. If R=0, it is considered that the measurement is completely reliable and the reliability is the highest.
Validity, that is, validity, refers to the degree to which a measuring tool or means can accurately measure what needs to be measured. Validity refers to the degree to which the measurement results reflect the contents of the survey. The more consistent the measurement results are with the contents of the survey, the higher the validity. On the contrary, the lower the validity. Validity can be divided into three types: content validity, standard validity and structural validity.
Question 3: Reliability coefficient R=0.85, P
If it is represented by R, it is more likely to test the reliability, because R stands for correlation, and the retest reliability is the correlation coefficient of the two tests.
Let's briefly explain the meaning of symbols.
R=0.85 means that the value of correlation coefficient is 0.85, which can also be understood as the value of reliability index is 0.85, and the reliability is usually above 0.8. And P Question 4: What are reliability analysis and validity analysis? 20-point questionnaire survey is a widely used survey method in educational research. The questionnaire designed according to the purpose of the survey is a tool to obtain information, and its quality plays a decisive role in the authenticity and applicability of the survey results. In order to ensure the high reliability and validity of the questionnaire, the questionnaire should be tested before the formal questionnaire is formed, and the reliability and validity of the test results should be analyzed. According to the analysis results, the questionnaire items are selected and the questionnaire structure is adjusted to improve the reliability and validity of the questionnaire. The methods of reliability and validity analysis include logical analysis and statistical analysis, and this paper mainly discusses the latter.
First, reliability analysis.
Reliability is reliability, which refers to the consistency of the results when the same object is measured repeatedly by the same method. Reliability indicators are mostly expressed by correlation coefficient, which can be roughly divided into three categories: stability coefficient (cross-time consistency), equivalence coefficient (cross-form consistency) and internal consistency coefficient (cross-project consistency). There are four main methods for reliability analysis:
1, retest reliability method
This method uses the same questionnaire to test the same group of respondents repeatedly at certain intervals, and calculates the correlation coefficient of the two test results. Obviously, the test-retest reliability belongs to the stability coefficient. The test-retest reliability method is especially suitable for factual questionnaires, such as gender and date of birth. There should be no difference between the two tests, and most of the respondents' hobbies and habits will not change significantly in a short time. This method is also applicable to the attitude and viewpoint questionnaire if the attitude and viewpoint of the respondents have not changed suddenly. Because the retest reliability method needs to test the same sample twice, the respondents are easily influenced by various events and activities, and the interval is limited, so it is difficult to implement.
2. Replication reliability method
Replica reliability method is to let the same group of respondents fill out two questionnaires at a time and calculate the correlation coefficient of the two questionnaires. Replica reliability belongs to equivalent coefficient. The copy reliability law requires that the two copies should be completely identical in content, format, difficulty and the direction of corresponding questions except for different expressions. However, in the actual investigation, it is difficult to make the questionnaire meet this requirement, so few people use this method.
3. Semi-reliability method
The semi-reliability method is to divide the survey items into two halves, calculate the correlation coefficient of the scores of the two halves, and then estimate the reliability of the whole scale. Semi-reliability belongs to the internal consistency coefficient, which measures the consistency of the scores of the two halves of the problem. This method is generally not suitable for fact-based questionnaires (such as age and gender can not be compared), and is often used to analyze the reliability of attitude and opinion questionnaires. In the questionnaire survey, the most common form of attitude measurement is the 5-level Likert scale. In semi-reliability analysis, if the scale contains ambiguous items, the scores of ambiguous items should be reversed to ensure the consistency of the scoring direction of each item, and then all items should be divided into equal halves as far as possible according to parity or before and after, and the correlation coefficient (rhh, the reliability coefficient of half of the scale) between them should be calculated. Finally, using Spearman-Brown formula:
The reliability coefficient (ru) of the whole scale was calculated.
4.α reliability coefficient method
Kronbach α reliability coefficient is the most commonly used reliability coefficient at present, and its formula is:
Among them, k is the total number of items in the scale, the intra-item variance of the score of the first question, and the variance of the total score of all items. As can be seen from the formula, the α coefficient evaluates the consistency between the scores of each item in the scale and belongs to the internal consistency coefficient. This method is suitable for the reliability analysis of attitude and opinion questionnaire (scale).
Second, the validity analysis
Validity, that is, validity, refers to the degree to which a measuring tool or means can accurately measure what needs to be measured. Validity can be divided into three types: content validity, standard validity and structural validity. There are many methods for validity analysis, and the measurement results reflect different aspects of validity. The methods commonly used to analyze the validity of questionnaires mainly include the following.
1, analysis of the correlation validity between single item and sum
This method is used to measure the content validity of the scale. Content validity, also called face validity or logical validity, refers to whether the designed project can represent the content or theme to be measured. The evaluation of content validity usually adopts the method of combining logical analysis with statistical analysis. Logic analysis is generally used by researchers or experts to judge whether the selected project "looks" in line with the purpose and requirements of measurement. Statistical analysis mainly adopts single item and sum correlation analysis to get the evaluation results, that is, according to whether the correlation is significant, calculate the correlation coefficient between the scores of each item and the total scores of each item. & gt
Question 5: What is reliability? The reliability of any measuring equipment represents the degree to which the result can be measured repeatedly. Simple and convenient, reliability refers to the consistency of measurement results under different conditions. A 20CM ruler, whether you use it today or tomorrow, in Beijing or Tokyo, or I use it, the result is the same.
Question 6: Popular explanation of validity and reliability. The validity of 20-point reliability can be simply summarized as "two Jin of sugar", two Jin is reliability and sugar is validity. If it is one and a half kilograms or two centimeters, the reliability is questionable. If the measured salt or rice is not the sugar we want, it is a question of effectiveness.
Reliability classification:
1. Test-retest reliability: the consistency of the measurement results of the same scale at different times. If one weighs two kilos today, it will be one and a half kilos tomorrow, indicating that the reliability is not high.
2. Half reliability. Divide the measurement items into two halves by a certain method, measure them separately, and get the correlation of the results. Divide two kilograms into two kilograms, and if they are 1 kilogram and one and a half kilograms respectively, it means that there is something wrong with the reliability of folding in half.
3. Reliability among raters: the consistency of the results obtained by different raters using the same tool. Several people use the same scale to weigh them separately. If they are all two kilograms or close to two kilograms, it means that the reliability of the evaluator is high.
4. Replica reliability. Correlation between various measuring tools. If you weigh two kilograms with this brand and two kilograms with another brand of the same brand, it means that the copy is highly reliable.
Validity classification:
1, the coupon is valid. The results of empirical observation and measurement are the indicators to be measured. It looks like white sugar, which we subjectively think is white sugar. Of course, it may also be washing powder or salt.
2. Content effectiveness. Test results of related industries and specialties. Please ask the relevant authoritative experts to evaluate whether it is white sugar.
3. Structural validity. What are the measuring properties of measuring tools? From the taste, color, chemical structure and other aspects of sugar all-round measurement, evaluation and demonstration.
Specifically, it is more direct to find professional papers and compile questionnaires by yourself.
Question 7: What does split-half reliability mean in the scale? The standard term is semi-reliability
It is one of the commonly used reliability testing methods. Reflect the degree of internal consistency of test items, that is, the degree to which tests measure the same content or characteristics. The specific analysis is to divide the test items into two equal groups (two halves) after the test. Usually, the odd-even grouping method is used, that is, the test items are divided into two halves according to the odd and even numbers of the serial numbers, and then the correlation between the two items is calculated. The higher the correlation, the higher the reliability or internal consistency.
When the variance of the two parts is equal (when the variance homogeneity test is homogeneous), it can be modified by Spearman-Brown formula: r = 2r'/( 1+r') r' is the correlation coefficient of the two parts, and r is the reliability value of the whole test.
If the variance homogeneity test is not uniform, the following formula can be used:
(1) Flanagan formula: r=2( 1-(S 1+S2)/S)
S 1 and S2 respectively represent the variance of the scores of all subjects in the half-test, and S represents the variance of the total scores of all subjects in the whole test.
(2) Lulun formula: SD/s.
R is the reliability value of the whole test, Sd is the variance of the difference between the scores of the same group of subjects on the bisection test, and S represents the variance of the total scores of all subjects on the whole test.