Question

To assess consumer acceptance of a new series of ads for a soft drink, suppose a certain company conducted a nationwide poll of 357 adults who had seen the ads. The following responses are based on that survey. How might nonsampling error bias the results of such a survey?

Answer 1

The most important non sampling errors are non-response errors, measurement errors, interviewer errors, adjustment error and processing error. The non sampling errors is the deviation between results of a survey and truth which not caused by random selection of the observation. Non sampling error occurred due to the poor sampling method. Here the soft drink company only take a less number of samples to find the acceptance of the product in the market. The samples were mainly the adults who seen the ads. At the same time, there are several people who know about the product who did not watch the ads. This may affect the final results of the study. There is a selection bias occurred here in this research. The researcher is only selected the respondents who seen the ads. If there is a bias occurred in the selection of the samples, there is a chance for non sampling errors.
Here the investigator assumes that the research is free from errors. But this is not possible when it become practical. These errors will affect the tabulation and it leads to invalidate results. One of the most important errors which the investigator made here is the less number of selections of samples. There is also an inadequate data also. If the data which is free from sampling errors have a chance for non sampling errors also.