Type of Errors in Research

Reasearch Design Methods We may often hear the statement that a good research is one that has the smallest error value. In spite of almost impossible that a research is completely free from any errors, we can still minimize potential errors. One of the way is by making a comprehensive research design. To find out more about the error in the research, this material is intentionally set up for you. Happy reading!

What is meant by error in research?

Let us recall again our memories on statistical learning in school or college. We certainly are familiar with the term of research error. In statistical science, the term of error refers to the difference between the true value of the population (or parameter values) with the values obtained from the alleged sample (statistical value), assumption used is in normal distribution condition. Total error in the research is a combination of random sampling error and non-sampling error as will be discussed below.

1.Random Sampling Error

Random sampling error occurs because the sample cannot represent the population or research objects. It is caused by we choose wrong samples. For example, the average income of the population is Rp 3,000,000, but the sample studied had average income of $ 3.2 million at a 5% confidence interval (CI). This sample cannot be used because the difference between the population and sample is 200,000 or more than the value of CI of ± Rp 150.000 (5% x Rp 3,000,000).

2. Nonsampling Error

This error occurs because any mistakes other than taking wrong sample. This error can occur at random or non-random. The cause of this error could be many things, such as an error in the definition of the problem, the error in choosing research approach, inappropriate questionnaire design and interview methods, as well as the wrong in preparing data and analyzing data collected. Non-sampling error generally is divided into two types: response error and non-response error. Non-response error happens because the sample did not answer the questions given. Usually, it is caused by the sample of respondents refused to answer or were in place.

Response error is an error that occurs because the respondents gave answers that do not fit with any questions - misunderstandings in reading the question. Response error may be caused by the researcher, the parties asked respondents (interviewer error), and the respondents themselves.

a. Response error cause by researcher (researcher errors) involves five type of errors.

• Surrogate information error. It is an error that emerges due to the information to be obtained is different from information searched. For example, a researcher wants to conduct research on the reasons for consumer switching to other brand of particular product, but he uses information about brand preference, which is not appropriate to explain his research problem.


• Measurement error. This type of error occurs because researcher use inappropriate measurement and scale, e.g. when we conduct research on consumers’ preference toward something we would be better using an interval or Likert or Semantic scale of measurement, but the researcher used a nominal scale, which is not suitable to explain the level of consumers preference toward particular object.


• Population definition error. It is an error caused by the researcher wrongly defines the population being measured. For example, a researcher is trying to figure out the teens behavior in using internet. He defines teens category as they whose age at 15-28 years old. However, according to existing law, they are who called teenagers are those only upto 18 years old.


• Sampling frame error. It is an error that occurs due to available sampling frame does not fit with the criterias of population being observed. Take an example, phone book issued by phone provider usually cannot represent accurately some categories of customers, such as : new entrance who just stopped subcribing, and potential customers who may be subscribe in the future. customers, customers


• Data analysis error . It happens because researcher uses inappropriate method of data analysis, or it could be caused by researcher used wrong statistics procedure. As a result the data lead to an incorrect conclusion.

b. Response error caused by interviewer (interviewer error) consists of:

• Responden selection error. This type of error occurs when the interviewer wrongly select respondent that doesn’t fit the criterias of sample required in research. For example, the interviewer should take a sample which is a loyal customer a store, but the interviewer is actually interviewing new customers, who only make purchase twice. This causes data obtained to be bias.


• Questioning error. Questionnaire design process is a stage in research that contains a lot of potential error. A researcher can go wrong in arranging questions, such as: use ambiguous statement or difficult term that can lead to respondents' misunderstanding. Another common mistake is that researchers use inappropriate measurements, and there is an error in printing questionnaire.


• Recording error. When we are conducting a research, it is common to happen that our recorder didn’t work well, or interviewer perhaps forgot to turn it on. As a result, data gained is unclear and it will disturb you in making transcript.


• Cheating error. This type of error can cause fatal effect. It is not only lead to incorrect result, but also violating research ethics. As a researcher, we must be objective and honest because our research result will be applied to answer any problem related to human life.

c. Response error caused by respondents (respondent errors) :

• Inability error. It is type of error caused by respondents do not answer the questions given. Commonly, there are two reason why they refuse to answer certain quesntions. First, it is because they really don’t know the answer, as they have not experienced that. Second, it can be because they don’t understand the meaning of the question given.


• Unwillingness error. Respondents refuse to answer certain questions is not only because they don’t know the answer, but some respondents choose to not answer intentionnally as the questions are something sensitive topic for them. Question about take home payment is one of sensitive question for most people.

Now we already know the kinds of error in research, and we also know the causes of error in research. Understanding any potential errors in research can help us prevent the error before it happens. One of solution to overcome that error is by preparing your research design as the best you can. Be careful in formulating sampling strategy. Look back to your research objective and foundation theories. For example, to reduce errors caused by interviewer and researcher, we can implement an evaluation system or checking the data before it is inputted. Or, we can also train our interviewer before he/she conducted a fieldwork job.