Definitely, the interviewer’s physical and mental conditions, the place of interview and the questions used are important factors in data collection. The development of valid questions is a very important factor. It must be decided as to what form of questioning must be done, whether short answers, like yes or no, or detailed answers should be accepted. There can be biases in both forms of questioning and a proper balance must be maintained to get the required results. Moreover, the answers to these questions should not carry any personal views of the interviewer. Other factors are the number of questions that must be asked and the time duration that must be devoted to each interviewee. Questioning for long periods may not yield good results and may give wrong impression to future interviewees.
In case of survey samples that are complete in all respect, rejection of sample questionnaires that seem adequately complete should have very strong reasons. These reasons may include relevance of the data with the question asked and the time in which the questionnaire has been submitted etc. When seen clearly, tossing a questionnaire sample becomes valid only when the answer is irrelevant with the question asked. If samples are dropped for any other reasons, personal or professional, the overall response to the question becomes useless. For example, the person who is supposed to select and sort out data in different categories may not like the response of some participants and may drop those samples altogether. This would be a matter of personal and professional ethics and would bias the whole research. It is the editor’s duty to entertain all the data first and toss samples that are incomplete or irrelevant. Therefore, it is important to at least go through the data once before any of the data is rejected. However, if the answers to questions do not solve the purpose of the research, are irrelevant or demean the research, the sample must be tossed out of research.
My Response to John Grace
Tracking down and locating individuals who did not complete their questionnaires would be a time consuming process and would require additional efforts. Firstly, time is of great importance in producing results in a research and if locating individuals is carried out, precious time would be lost that can be utilized in data analysis.
Secondly, locating individuals and asking them to refill or complete their form would make them feel uncomfortable because most of these researchers are carried out on the basis of anonymity. If tracked and located individuals may feel that their personal information and their activities are being monitored. This can become very uncomfortable for the respondent and there is no guarantee that the individual will fill out the form in the same way they did during the survey was conducted. The setting will be different and the answer will differ from what it originally may have been during the survey. Hence, locating respondents may have psychological issues related to it. Also, tossing out data that is incomplete may only be right if the data cannot be used at all. If the data has something relevant to the research it can be used to satisfy the results that we get using analysis methods. Incomplete but relevant data can be used to prove the correctness of the results of probability analysis used to predict the data.