What are the main ways to improve the validity of psychometrics

1. The design of the questionnaire should be scientific and reasonable: the content of the questionnaire has a very important relationship with the validity, if the content is set scientifically and moderately, the validity will be high, otherwise, the validity will be affected. Such as:

The setting of each topic should have a clear and definite concept, the difficulty of the question should always be consistent, and the arrangement of the question should be reasonable.

2. Careful selection of samples: the more representative the sample, the more it will reflect the main characteristics of the questionnaire, and the larger the full distance of the sample score, the higher the validity coefficient will be.

3. Pay attention to the ** rate of the questionnaire: if the ** rate is low, it may not be representative and cannot correctly reflect the characteristics of the measured subject.

4. The rationality of the effect standard: the effect standard is a prerequisite for validity, and the appropriate operation should be constructed according to the content.

5. Pay attention to the status of the respondent: In the process of filling in the questionnaire, the status of the respondent is an important factor affecting the validity, if the respondent is absent-minded or reluctant to answer, or even has a negative attitude, it will affect the accuracy of the questionnaire results.

6. Ensure that the test environment is free of interference: if the respondent is in a noisy and noisy external environment or the external environment cannot be controlled by the researcher, the respondent will be disturbed, and it is necessary to eliminate irrelevant interference to avoid affecting the correctness of the questionnaire.

Test validity is a measure of how well the content of a test matches the characteristics of the object it is measuring. A test is valid if it measures exactly what it is trying to measure.

2. Carefully compile test questions and test scales: First of all, the content of the test questions should be suitable for the purpose of the test, such as knowledge test questions can not fully reflect the intelligence level of the subject, it mainly measures its knowledge level. Secondly, the test questions should be clear and clear, the language should be understandable by the participants, and the order should be from easy to difficult.

Third, the difficulty and scale of the test should be appropriate.

3. Measure in strict accordance with the test procedure to prevent measurement errors: measure in strict accordance with the test manual, do not explain too much, score according to the standard, and the interval between the two tests should be appropriate.

4. The sample size should be appropriate: when the sample size increases, the representativeness of the sample to the population increases, the sample size is large, and the internal differences of the participants increase, which expands the variance of the true score and improves the validity. The sample size should generally not be less than 30.

In addition, the sampling method is also very important, generally using random sampling, when the population is large, stratified sampling, when the sample size is expanded, its representativeness increases.

Ways to improve test reliability:

1. The difficulty of the test should be moderate: the difficulty is the difficulty of the test, when the test is too difficult, the score of the subject is generally too low, showing a negative skew distribution; When the difficulty of the test is too small, the scores of the participants are generally high, showing a positive skewed distribution. Tests that are too difficult and too easy will reduce the difference in the scores of the participants and the variance of the take-home scores, thus reducing the reliability of the test.

2. The content of the test should be as homogeneous as possible: the test content of the same nature also requires the same ability, knowledge and skills for the participants; Tests with different content require different abilities, knowledge and skills. Therefore, in order to improve the reliability of the test, the test content should be as homogeneous as possible.

3. The time for the test should be sufficient: for a certain test, the vast majority of participants should be guaranteed to complete the test within the specified time; Otherwise, if the participant cannot answer all the questions calmly, it will not reflect the true level of the participant.

1. Create a standard test-taking situation, so that each participant can perform at a normal level;

2. Carefully compile the test scale (pay attention to the sampling, quantity, difficulty and differentiation of the questions) to avoid large systematic errors;

4. Select the correct standard of effect, set the appropriate standard measurement, and use the relevant formulas correctly;

Connections (1) Educational measurement is an important means to obtain data for educational evaluation;

2) Evaluation can only give correct judgment on the basis of measurement, and the evaluation made through measurement can be reliable;

3) Measurement is also inseparable from evaluation, and the results of measurement can only reveal their practical significance through evaluation;

4) Measurement is the basis of evaluation, and evaluation is the specific embodiment of measurement.

Difference: Measurement is concerned with the quantity, and evaluation is concerned with the value; Measurement is a purely objective process, and evaluation is the unity of subjective estimation and objective measurement. Measurement is a singular activity, and evaluation is an integrated activity.

The validity of the study includes intrinsic validity and extrinsic validity, and intrinsic validity refers to the correctness and authenticity of the research narrative; Extrinsic validity studies the correctness of inferences.

Ways to improve intrinsic validity:1Correct understanding of the theory:

The concept should be clear, and the explanation should be credible and reprognizable; 2.Operational fidelity to reduce errors: Concepts and variables can be defined according to their theoretical construction or specific content, and then effective assessment tools or tests can be designed; 3.

Appropriate sample and prevention of loss: Sample sampling should pay attention to the equality of the number of people in different groups, and pay attention to the appropriateness of the research situation and the ** rate of the questionnaire; 4.Eliminate irrelevant variables

Identify and eliminate extraneous interfering variables that can confuse or threaten conclusions, and reduce them as much as possible.

Ways to improve extrinsic validity:1Interpretation and analysis should be universal, objective, neutral, reasonable and authentic. 2.

Operational definitions represent conceptual meanings. The sample should be sufficiently representative, and the context of the study should be appropriate, preferably similar to the context in which it is to be applied or inferred in the future; 3.Observation is universal, and data collection should be pluralistic and objective; 4.

Extraneous interfering variables should be excluded as much as possible, and experimenter effects should be prevented.

Excerpted from "Questionnaire Statistical Analysis Practice - SPSS Operation and Application").

Do you mean a pedagogical approach to test validity for students?

There are four ways to improve the validity of a test:

1) In order to avoid large systematic errors, the test scale was carefully compiled;

2) In order to control random errors, properly organize the test;

3) Create a standard test-taking situation in order to make each participant perform at the policy level;

4) The calibration standard is selected correctly, the calibration measurement is appropriate, and the relevant formulas are used correctly.

The degree to which a test or scale can actually measure the psychological trait it is trying to measure is validity. It is relative to a certain purpose of measurement;

At the same time, psychometrics cannot be 100% accurate, only to some extent;

Validity is a combined reflection of measured random and systematic errors. To judge whether a measurement is valid, it is necessary to collect evidence from various aspects.

Validity refers to the degree to which a test or scale can actually measure the psychological trait it is supposed to measure.

Of all the variations, only the part of the variation caused by the observed psychological characteristics is truly measurable, and its proportion in the variation is the magnitude of validity.

The basic question of validity is: what the test is trying to measure. Reliability does not take into account systematic error, validity is considered.

2. Nature of validity: Validity is specific to the test results. It's for a specific measurement purpose, so it's not universal; There is only a difference in degree.

Psychological traits are more subtle and can only be inferred from their behavior to a certain degree of accuracy. So validity is a difference in degree, not a difference in total or none. (Additional note:.)

Relativity: The purpose and function of a test must be considered when evaluating the validity of the test).

3. Validity verification: the process of collecting a large amount of data and evidence to test the validity of the test.

Commonly used methods for generalizing validation of validity: cross-validity assessment Meta-analysis.

4. The relationship between validity and reliability: High reliability is a necessary and non-essential condition for high validity. So, 3 combinations:

High trust is efficient, high trust is inefficient, and low trust is inefficient. It can be seen that the reliability is not necessarily high, but for a test to be valid, it must be reliable. Test validity is subject to reliability.

6. Other validity:

Apparent validity: (is the interaction between what the test asks the participant to do and the participant's comparative understanding) in a technical sense, it cannot be considered a validity; It does not refer to the validity of what the test actually measurs, but to what the test looks to measure on its surface. The performance affects the test subject's motivation.

Appropriate performance is okay, and the test questions arouse the motivation and interest of the participants, but they are also well hidden.

Efficacy and discriminatory validity: two new criterion correlation validities developed by occupational psychologists.

Internal Validity A vs. External Validity B: A = Internal Consistency Validity, reflecting the construct validity of the test. b: Refers to the degree to which the results of the study are generalized to other contexts and populations.

8. Factors affecting validity: factors of the test itself: 1, test, calibration factors, representativeness of the sample, and interference variables.

9. Application of validity: the important role of validity in test development; Important influence in the selection of the former tomb An important influence in decision-making; combined with information; It has an important impact on the classification and placement of Hui manuscript relatives.

Ways to improve the validity of the measurement:1Control system error:

The content of the test should be suitable for the purpose of Lilafeng, clear and clear, arranged from easy to difficult, and the difficulty and curvature should be appropriate. 3.Strictly follow the test procedure to measure the amount of answers:

Don't explain too much, mark according to the standard, and the interval between the two tests should be appropriate. 4.Sample size should be appropriate:

Validity refers to the accuracy of a psychometric test.

Validity, or validity, refers to the degree to which a measurement tool or means can accurately measure what needs to be measured. Validity refers to the degree to which the measured results reflect the content to be examined, and the more consistent the measurement results are with the content to be examined, the higher the validity; Conversely, the lower the validity. There are three types of validity:

Content validity, criterion validity, and construct validity.

Validity is the degree to which a measure is valid, i.e., the degree to which the measurement tool can actually measure the trait it is intended to measure, or simply the accuracy and usefulness of a test. Validity is the most important condition that a scientific measurement tool must have. In social measures, there is a high requirement for the validity of questionnaires or scales as measurement tools.

To identify validity, it is necessary to clarify the purpose and scope of the measurement, consider the content to be measured, analyze its nature and characteristics, check whether the content of the measurement is consistent with the purpose of the missed measurement, and then judge whether the measurement results reflect the degree of the characteristics to be measured.

External Validity:

The external validity of a study refers to the degree to which the results of a study can be generalized and universally applied to the population from which the sample came and to other populations, i.e., the representativeness and general applicability of the results and variables to the conditions, time and context. External validity can be subdivided into two categories: overall validity and eco-validity.

Overall validity refers to the degree and ability of the results of a study to be adapted to the population from which the study sample came, or the general significance to the population. For the results of the study to be applicable to the population, a random sample must be selected from the population so that the sample is representative of the population. If the sample is biased or too small to be representative of the population, it is difficult to generalize the characteristics of the population.

Eco-validity refers to the extent and ability of research results to be generalized and adapted to other research conditions and scenarios. In order for the results of the study to be applicable to other research conditions and scenarios (e.g., differences in independent and dependent variables, research procedures, research background, study time, and investigators), it is necessary to design the research conditions and scenarios specifically to ensure that they are representative of other conditions and scenarios.

Answer]: B In psychometric testing, validity refers to the degree of conformity between what is measured and the psychological characteristics to be measured, or simply the accuracy of a psychometric test. Reliability refers to the degree of consistency of the same test being repeated by the same test (or another set of equal tests) in different lapitical times.

Correct measurement tools, ensuring the reliability of measurements, selecting appropriate norms (scales), avoiding subjective bias of observers, etc.

Carefully develop test scales to avoid large systematic errors. It is necessary to avoid bias in the questions, the difficulty of the questions, the degree of differentiation should be appropriate, and the number of questions should be appropriate.

Organize the test properly and control for random errors. The test taker should operate in strict accordance with the instructions of the manual to reduce the interference of irrelevant factors.

Create a standard test-taking scenario so that each participant can perform at a normal level.

Choose the right scale, make the right scale, measure the right scale, and use the relevant formulas correctly.