Independent topic

Before you get to know the key statistical components to doing a professional poll you will at first need to understand that the original purpose of conducting such polls is to magnify the voice of the public with regards to certain issues. This means that the ultimate goal is that the same public must have confidence in the results obtained from such polls. To maintain public confidence, therefore, the following four broad statistical components are very important: the sample, the method of interviewing, the questions and method of interpreting results.
In professional polling the importance of sampling validity is immeasurable. The first step is to fully identify the sampling frame from the population that the poll seeks to represent. This is followed by a random selection of a small percentage from within the sampling frame to represent the entire population targeted by that poll. Random here implies that everyone has an equal chance to be selected and this can only be achieved if the method chosen to identify individuals to be interviewed is free of bias. Even though Statistics postulates that the actual number of people interviewed for a given sample is less important than the soundness of ‘ randomness’ employed in the process, you still need to balance the economic cost of sample sizes with the desired level of accuracy required. Accuracy of plus or minus three percentage points margin of error is the acceptable level for professional polls.
Once the individual to be interviewed has been identified it is important not to change the selected person so as to eliminate possibility of bias. In the case of telephone sampling this would involve performing call-backs until such a time that selected individual is available. However, more importantly for the interview is the wording of the questions and the order in which they are set to be answered. Crafting fair and objective questions requires extensive knowledge of public opinion, care and discipline. To achieve this it is good practice to have exact wording of a question held constant from year to year to analyze trends and also to have multiple questions to help put the research within a certain context. These two practices come in handy when interpreting the poll results. Finally, for accountability purposes the polling company needs to make its data available to the public domain for any interested parties to draw their own conclusions about what the survey data means and to assess whether the poll was conducted in a professional manner.
The New York Times poll on the proposed immigration law used a sample size of 1, 079 adults which statistically would give us the acceptable margin of error of plus or minus three percentage points. The New York Times tells us that respondents were selected randomly from the adults within households that were selected from a list that exceeded 69, 000 active residential exchanges across the United States (para 2), that included listed and unlisted numbers (para 3) and was supplemented by cell phone numbers (para 4). It is clear that the pollsters took great effort in seeking to maintain a random sampling technique. We are also informed how they sought to ensure regional representation of their target population in the survey through the selection of the residential exchanges used. We therefore find the technique used to come up with the respondents to be in tandem with the industry standards. We also find the sample size of more than 1, 000 to be appropriate because it gives us a margin of error that is appropriate for such surveys.
However, we are not able to give a complete critical analysis of the poll with regards to the interviewing technique, wording of the questions and question order because we are not provided with that information. This would be vital for the critique because question wording and ordering are great sources of bias and error in surveys nowadays. Therefore, we cannot unequivocally give the New York Times poll a clean bill of health.
Works Cited
“ How the Poll Was Conducted”. The New York Times. 3 May 2010. Web. 16 May 2010.