How to control the error in the survey, and whether the error belongs to the error of the respondent

1. Establish a standardized system. Standardizing survey activities is an important means of controlling non-sampling error, and it is necessary to establish a standardized system first, including standardized terminology (survey terminology and project terminology) and standardized tools (sample sheets, questionnaires, cards, etc.).

The significance of standardization work is to enable all aspects of the investigation to operate under unified requirements, and the work exchange and information exchange of each link can be accurate and effective. Only by establishing a standardized operating system can it be possible to ensure the stability of the quality of investigation.

2. Strengthen document management. The document has the function of information transmission, document management belongs to information management activities, the data obtained by the investigation are information products, the document is the carrier of these information products, and the investigation activities involve many documents, such as various design schemes, sampling data, questionnaires, interview records, data processing records, etc. Strengthening document management is an important means of clarifying responsibilities.

3. Improve internal measurement work. In the process of survey implementation, there is a lot of internal measurement work to be completed, such as some common measurement indicators include average access time, rejection rate, etc., these measurement indicators are the standard for evaluating the quality of interview work on the one hand, and the basic data for survey design on the other hand.

4. Establish a quality responsibility system. The establishment of a quality responsibility system is the premise of quality control, and the operation specifications of each link of the investigation activities can only be guaranteed to ensure that the actual operation is carried out in accordance with the control requirements and specifications only under the premise that the corresponding personnel understand their own responsibilities.

5. Do a good job in quality education. Quality education is a positive means of quality control, on the one hand, through the education of quality awareness, improve the investigators' understanding of the significance and requirements of controlling the quality of the investigation, deepen the understanding of the requirements of the norms, and consciously improve the sense of responsibility of the work; On the other hand, strengthening the learning of relevant business knowledge can further improve the ability and level of work.

The right design questions, and more importantly, choosing the right group of participants!

Only in this way can the question be accurately analyzed by the person being asked.

The questions should be scientifically designed, with clear meanings, appropriate wording, conciseness, and logic.

1. Human factors.

Errors caused by human factors, including misreading, miscalculation, and parallax. Miscalculations often occur when a calculation is incorrectly made or when incorrect data is entered. Parallax often occurs when the direction of reading the measured value is different or the scale planes are not in the same plane, and the difference between the two scale planes is about mm, and if the reading size is not perpendicular to the scale plane, it will produce the amount of error.

2. Measuring tool factors.

Errors caused by gage factors, including scale errors, wear errors, and uncorrected before use. Whether the scale division is accurate or not must be corrected and traced by more sophisticated instruments. After a period of use, the measuring tool will produce a considerable degree of wear, so it must be calibrated or sent for repair before it can be used again.

3. The strength factor.

Errors due to the contact force used in the measurement or the deflection caused by the contact. According to Hooke's law, when measuring the size, if the measuring shaft is in contact with the machine with a certain measuring force, the shaft and the machine will be partially or comprehensively elastically deformed, in order to prevent this elastic deformation, the shaft and the machine should be made of the same material. Secondly, according to Hertz's law, if the axle and the machine are made of steel, the amount of error caused by its elastic deformation.

4. Measurement factors.

During the measurement, errors caused by poor instrument design or placement, including cosine error, Abbe error, etc. The cosine error occurs when the measurement axis is tilted at a certain angle to the surface to be measured. In general, cosine errors occur in both measurement directions and special care must be taken.

5. Environmental factors.

The measurement is affected by the difference of environment or site, and the possible errors are thermal deformation error and random error are the most significant. The thermal deformation error usually occurs due to room temperature, human contact and the temperature of the workpiece after processing, so it must be under the control of temperature and humidity, and the workpiece and measuring tools cannot be touched by hand, and the workpiece is measured after cooling after processing. However, in order to shorten the processing time, real-time measurement is required during processing, so the thermal expansion coefficient of various materials must be considered as compensation to cope with the error caused by the different thermal expansion coefficients of the materials at different temperatures.

1.Error**.

Device error. Standard error: A standard is an instrument that provides a standard amount, such as a standard battery, a standard resistor, a standard clock, etc. They all have errors in the quantities they embody.

Instrument error: such as the error of the meter, balance, vernier itself, etc.

Accessory error: Error caused by auxiliary accessories used in measurement, such as switches, power supplies, and connecting wires.

Environmental errors. Due to the inconsistency of various environmental factors (such as temperature, humidity, air pressure, vibration, lighting, electromagnetic field, etc.) with the required standard state, and its spatial gradient and change over time, the error caused by the change of the measuring device and the measured itself.

Personnel error. Measure the minimum resolution of the person's physiology, the physiological changes in the senses, the speed of reaction, and the errors caused by inherent habits.

Method errors. Empirical formulas, the approximation of the selection of function types, and the errors caused by the approximation values determined by the coefficients in the formula.

Errors caused by factors that are not reflected in the derivation of the expression of the measurement results but actually play a role in the measurement process, such as errors caused by some factors such as leakage, thermoelectric potential, lead resistance, etc.

Methodological errors due to insufficient knowledge or insufficient research.

2.Classification of errors.

Systematic error. Systematic error refers to the error that the absolute value and symbol of the error remain constant when the same quantity is measured many times under the same conditions or change according to a certain law when the conditions change, which is characterized by its certainty. Once the experimental conditions are determined, the systematic error acquires an objectively constant value. The average of multiple measurements does not weaken its effect either, and changing the experimental conditions or changing the measurement method can reveal systematic errors that can be eliminated by correction.

Accidental errors. Accidental error refers to the random change of the absolute value and sign of the error when the same quantity is measured many times under the same conditions, which is characterized by randomness, no certain law, sometimes large and sometimes small, sometimes positive and sometimes negative, and cannot be determined. Because the accidental error has the nature of chance, it can not be known in advance, so it can not be corrected or eliminated from the measurement process, but the accidental error, obeying the statistical law in multiple repeated measurements, can be controlled by increasing the number of measurements under certain conditions, so as to reduce its influence on the measurement results.

Gross error (gross error).

A negligent error is an error that significantly distorts the measurement results. This is due to the error caused by the surveyor's unreasonable methods, carelessness, and misremembering data in measurement and calculation. It can be avoided as long as the experimenter takes it seriously.

Summary. Hello belongs, the error belongs to the investigator error.

Hello belongs, the error belongs to the investigator error.

Non-sampling error: mainly including: sampling frame error, error, no error, investigator error; It is an error caused by the human factors of the investigator or the respondent in the process of investigation.

The errors caused by the investigators mainly include: errors in filling in the survey due to unclear provisions or interpretations in the survey plan, errors in transcription, and errors in summarizing and shooting; The errors caused by the respondents mainly include: intentional misreporting or concealment of survey data due to human interference.

Non-sampling error can theoretically be eliminated.

Theoretical errorTheoretical error, also known as method error, is an error caused by the measurement method itself or due to the imperfection of the principle on which the measurement is based. Example 1: The heat loss caused by heat dissipation is not considered in the thermal experiment, and the influence of the internal resistance of the meter on the experimental results is not considered when the voltammetry measurement resistance.