Discussion Question – Coursework Example

Discussion Question Discussion Critics consider opinion polls to be the worst way of measuring public behavior, public opinion or predicting elections (Maher, 2008). The claim could be authentic as the results of the opinion polls often differ significantly, with the outcome of the elections. However, there are methods that statisticians and reporters can use to increase the chances of the results being accurate. The paper explains these methods.
First, the results may differ because of errors regarding sampling designs. The mode of selection of phone numbers during telephone surveys are erratic. Usually, there is bias in choosing individuals to participate in the survey. In correcting this mistake, the survey needs to employ non-biased methods of sampling. Here, they will select numbers without regard to political affiliations.
Second, the problem of non-availability where the individuals available differ on the variables measured than the unavailable individuals. The problems can be rectified through the formulation of variables that are compatible with all people or respondents available.
Third, the problem of refusal also crops up in the process of carrying out the polls. At times, people refuse to react on a variable that is under measurement. In correcting this problem, opinion poll officials must formulate variables that all respondents can respond to in the research process. The technique will ensure that they do not over emphasize on one variable.
Fourth, problems also exist in the structuring and wording of the survey questions. In most circumstances, the questions have confusing words. These words may lead the respondent to particular desired responses. The appropriate technique for use here is the formulation of neutral questions that do not favor any political mass.
Additionally, the weighting of data may be inadequate. Here, is where some data are given more emphasis during statistical manipulation. Proper statistical techniques can be used for analysis. It is appropriate to employee more than one method for data analysis to compare the outcome.
In addition, the respondents also deliberately or unconsciously lie. The reporters can also report correct results falsely. The error can be minimized through enlarging the sample size. However, false reporting can be minimized by instating a series of verification processes and involving independent bodies.
Finally, the order of the survey questions can also cause the error in the results to occur. At times, the questions may be clustered in a way that influence the respondents into giving particular responses. However, the poll officials can randomize the prompts to ensure that the respondent do not correlate the outcome of one question with another.

Reference
Maher, B. (2008). Poll results: look whos doping. Nature, 452(7188), pp.674-675.