In research methodology, there is a huge scope of errors; it could be while preparing the questionnaire or while performing the data analysis. It is very important to manage and maintain the error so as to conduct survey and obtain results accurately.
In order to maximize the accuracy, error needs to be minimized.
Total error has basically two components.
Total error = Sampling error + Non sampling error
Total error is measured in terms of Total error variance.
(Total error) 2 = (Sampling error) 2 + (Non sampling error) 2
Sampling error refers to the errors that occur from selection of elements from the population. Non sampling error refers to those which occur due to research project; prejudices or preferences of the researcher.
There are seven types of errors. They are
1.Population Specification error
This error occurs when the researcher chooses a wrong sample to obtain data. For example- Imagine a survey about breakfast cereal consumption. Who should one survey? It might be the entire family, the mother, or the children. The mother probably makes the purchase decision, but the children influence her choice.
2. Sampling error
This error occurs if probability sampling is used and that sample does not represent the appropriate sample. For example- A random sample of 500 people of age 10-20 years may not represent the child population.
3. Selection error
This error occurs when sample is selected by non-probability methods. For example- Often researchers collect responses from easily accessible friends, family members but they might not be the appropriate target to conduct surve
4. Frame error
This error occurs when wrong sub-population is used to select a sample. For example- Imagine you’re conducting survey for mobile operators and you collect responses from children who do not possess personal cell phone.
5. Non response error
This error occurs when the sample obtained differs from the sample selected. It can occur in two ways
(a) Non-Contact: It occurs when you were not able to reach respondents.
(b)Refusal Error: It occurs when respondents refuse to answer some questions in the questionnaire.
For example- In some surveys people hesitate to choose the age group, answer questions with don’t know and can’t say options or leave the questions unanswered.
6. Surrogate Information Error
This error occurs when the responses are collected from substitutes and not from the original ones. For example- While conducting survey, answers to the questionnaire was based on past data which was used to predict future data.
7. Measurement error
This error occurs when the measurement process itself and represents the difference between information generated and information wanted by the researcher. For example- If in a survey, there is a question about the economic status and options are given such as high class, upper middle class, lower middle class and low class. There is ambiguity in the options as the categorizing is quite not clear to the respondents. Each respondent will have his or her own take on each option and will fill accordingly.
8. Experimental error
This error occurs if an experiment is conducted, the researcher attempts to measure the impact of one or more manipulated independent variable on some dependent variable while controlling the impact of exogenous variable. For example- If trying to find out reasons for cancer may conclude that as all the patients under observation use to eat bread therefore eating bread is the cause of cancer.
For any research project it is important to identify these errors as well maintain them at the same time. There are two basic approaches for reducing error. First approach is to minimize errors through research design. In this process, effective use of research method and techniques are utilized to lessen the impact of both sampling and non-sampling errors. Second approach is to estimate and measure error. The researcher should be in a position to estimate error and estimate how accurately the research is. Though estimating error is difficult to estimate because of the nature of errors. Statistics helps us to reduce sampling errors to a large degree but reducing non-sampling errors is difficult since it depends on the researchers’ intuition.