Good-day, Bias can occur in the most unexpected ways (without even knowing). you can have a sampling bias, instrument bias (how your questions are posed/phrased), response bias etc.

Its almost impossible to eliminate all bias but you can try your best that you can. hence i would say start from the instrument itself and ensure that the questions are neutral and objective. Random sampling or census method is usually good from a sampling perspective.

However if you have already collected your data, then just ensure it is coded properly and correctly and backed by reliability techniques. Depending on the type of data you have collected, i would also highly recommend 'triangulation' for increased validity and mitigating bias.

Creation of questionnaire is the most important part of the survey process for measuring the opinions, experiences and behavior of respondents. Accurate sampling and high response rates will be wasted if the information gathered is built on a shaky foundation of ambiguous or biased questions. Creating good measures involves both writing good questions and organizing them to form the questionnaire. Following links may be useful for designing effective questionnaire .

https://www.pewresearch.org/methods/u-s-survey-research/questionnaire-design/

https://www.b2binternational.com/2006/05/12/the-key-principles-of-effective-questionnaire-design/

https://www.b2binternational.com/b2b-blog/2006/01/17/questionnaire-design-given-5-stars-on-amazon https://www.b2binternational.com/b2b-blog/2006/01/03/our-latest-book-doing-well-on-amazon

https://ag.purdue.edu/extension/pdehs/Documents/Designing%20an%20Effective%20Questionnaire.pdf

Bias can occur at any phase of research, including study design or data collection, as well as in the process of data analysis and results. In quantitative studies having a well-designed research protocol explicitly outlining data collection and analysis can assist in reducing bias .Qualitative researchers must demonstrate rigour, associated with openness, relevance to practice and congruence of the methodological approach.

Many thanks for your responses, Gentlemen.. I appreciated!

First of all, describing the answers can help to find outliers, and then any statistical analysis proportional to main questions and hypothesis.

Survey construction is primary in avoiding biased responses or interpretations. Questions or sentence stems can be worded in ways that produce biased responses (as well as confusion on the part of the respondent).

Depending on the response set chosen, aggregating responses into larger categories can minimize bias in interpretations.

I am not sure if this relevant here, but I found when collecting qualitative survey responses that were recorded, that it was better to listen and transcribe the results myself (rather than have someone else do it). In doing this, you're more likely to spot emphasis, nuance etc than if you just read or analyse a cold transcript.

Again, if you can somehow factor in an ability to triangulate or affirm your initial findings, all the better from a validity perspective.

I hope this is helpful - apologies if I've misread your question!

Dear all,

I think that your responses are pretty helpful.. especially about the qualitative side which is always an issue .. For those who want to provide more guidance about this project, here i have a beta version www.talentandtalent.net It is still under construction but there is plenty of information for brainstorming ..

First, I would say that when composig a questionaire, it is important to understand clearly whether we wantas result of evaluation the opinion of the person (fenomenologic research - what people think about something) or if we want to show the reality itself (and being so, the opinion is data that should be interpreted). this point being clear, I think that the problem of bias could be reduced to lower problem as long as the researcher keeps in mind the need to avoid his opinion as filter of interpretation, what can only happen when the position is recognized and containned.

Good answer Dear Sachin Suknunan

I agree with Enéas Arrais Neto