The general public opinion polls being analyzed and compared include the following two: a poll by the Pew organization on the topic of outside campaign spending and a poll by the Rasmussen organization on the Electoral College. Before proceeding to a comparison or analysis, the property in question for both must be identified because it will allow for analysis of aspects such as proper sample sizes and methodologies in regards to the question, which will allow for any discrepancies to appear. These discrepancies will be the evidence in support of the conclusion regarding the strength of one poll over the other.
The property in question for the Pew poll measures the public’s knowledge about a growth in the amount of money being spent by outside campaign groups (Super PAC’s) on elections. In the Rasmussen poll, the property in question gauges the sentiment towards the idea of eliminating the Electoral College in order to make the popular vote the only method to select a U.S. president. Both are political topics, but the approach each took in collecting the data was distinct.
Here are some of the data facts provided by the organizations. The sample size for the Pew poll was a group of 1,010 adults residing within the United States, with the sample population coming from people who have a landline or cellphone number within the 50 U.S. states and the District of Columbia. The sample size for the Rasmussen poll was a collection of 1,000 “Likely Voters”, with the sample population taken from pool of landline telephone numbers and online surveys. The sample size used by both of the organizations is large enough to avoid any hastiness according to Table 10-1(354), where a sample of 1,000 yields an error of margin of about plus or minus 3%. The corresponding rate for both is also within the range of 6 percentage points, enough to hold the previous statement true even if the population size is that of the U.S.
The methodology for the Pew poll data collection was very complex, but the whole process seems fair and allows for random selection. They collect a random sample of landline and cellphone numbers but in a way and using statistics that ensures a diverse demographic can be represented, with 40% being landlines and 60% being cellphones. For landlines they select a batch of numbers from a single county proportional to the amount the county has in total, and the cellphone numbers come from a wireless and shared bank of numbers. When they call and request a specific gender, they ask for the youngest member of the household that is above the age of 18, this way they can ensure that the younger members of the population are reached. They estimate their average sampling error at no more than plus or minus 2.9% with a 95%...