Module 6
Primary data collection through surveys involves numerous issues that can prove to be a serious impediment to achieve the desired outcome while analyzing the data. It is a common issue in the world of data analysis as a researcher is always in the dilemma whether he should use partial responses and few incorrect entries collected during the data collection process.
While every researcher has its own method of treating those partial answers, but in my opinion, partial non-response errors, which generally arises because of inadequate memory or reluctance to answer some questions, should be minimized rather than eliminating those responses completely from the analysis process .
Therefore, in order to improve such data, the researcher should send reminder emails to the respondents as this method has been effective in gathering more completed response. Doing so, it is important that respondents who have already submitted completed responses, should not be harassed and reminders should only be sent to respondents who have submitted partial answers. Additionally, if the respondents have not answered some questions with citing some sensitive opinions, the researcher should assure them of the confidentiality and educate them that their responses will be used as a whole in the form of a sample and not on an individual basis. Important to note, some researchers prefer to use a weighting method for adjusting unit non-response and imputation method for item non-response. However, these methods are valid for large scale research where researchers are sure of assumptions and estimation procedures. Accordingly, since we are performing a small-scale survey, we suppose theoretical based methods will be easy to follow here.
On the other hand, if confronted with incorrect answers, the researchers correct those entires using the data cleansing tools such as batch processing. This method will not only save data processing time but also costs of fixing data errors manually.
References
How to Avoid Nonresponse Error. (2013, August 19). Retrieved February 14, 2016, from http://fluidsurveys.com/university/how-to-avoid-nonresponse-error/
Rouse, M. (n.d.). Data scrubbing (data cleansing). Retrieved February 14, 2016, from http://searchdatamanagement.techtarget.com/definition/data-scrubbing
Reply#1:
Dear April,
I agree that the methods proposed in our textbook relating to correction of missing data such as casewise deletion and pairwise deletions, are widely accepted by the statisticians. However, in my perception, adopting a theoretical approach and arranging data of the respondents who came up with incomplete answers, and then contacting them again in a professional manner will be a more reliable way to approach and rectify the shortcoming. In addition, utilizing neutral values for rectifying the incorrect data will again invite the limitation of a mean value such as being affected by outliers, to the data and thus undermine the usability of the data.(Malhotra, 2013)
Regards,
Jay
References
How to Avoid Nonresponse Error. (2013, August 19). Retrieved February 14, 2016, from http://fluidsurveys.com/university/how-to-avoid-nonresponse-error/
Malhotra, N. (2013). Marketing research: An Applied Orientation. 6th ed. (Sixth ed.). Prentice Hall
Reply#2:
Dear Shyne,
Working on basics of data collection, such as avoiding incorrect answers and ensuring that the respondents are fully involved in the survey process, are essential to gather a correct survey outcome. However, in the given scenario, rather than focusing on trial and error methods, tested methods such as batch processing would more effective in terms of correction of the incorrect data. Similary, for rectifying the issue related to missing data, we should first contact the respondents and then use the methods proposed in our textbooks namely, casewise deletion and pairwise deletions.(Malhotra, 2013)
Regards,
Jay
References
Malhotra, N. (2013). Marketing research: An Applied Orientation. 6th ed. (Sixth ed.). Prentice Hall
Rouse, M. (n.d.). Data scrubbing (data cleansing). Retrieved February 14, 2016, from http://searchdatamanagement.techtarget.com/definition/data-scrubbing