We're huge fans of evidence and data. So we decided to create a local politics poll, essentially an online "sample ballot" for people to fill out. This is our first poll ever and we have a limited reach on social media - as such our poll is not perfect. Please read over this entire post before analyzing the results. We're looking forward to comparing our results to the actual election in November. For now here's a snapshot of who (some) Burke County voters are thinking of voting for!
- Due to anonymity we do not have 100% certainty that the respondents are residents in Burke or are registered to vote in Burke. The first question "Do you live in Burke County?" was answered in the affirmative by the majority of respondents, however this does not ensure that they are or will be registered to vote. Additionally those that checked "no" on this question could be absentee voters so we still counted those in the final results.
- Unlike real voting, we required a response for the entire ballot. Therefore those that only knew some candidates and not others may have chose not to complete the poll at all. To be honest there isn't a strong justification for why we did this, it was simply not thought about thoroughly until after the poll was public. Our bad.
- There is nothing to prevent someone from voting multiple times, which would obviously skew the results.
- For the state positions, we find it unlikely that we sampled beyond Burke County proper. The districts those positions serve are not explicit to the county. For example, Senate District 46 covers much of Cleveland County as well, where it is unlikely we have many listeners (if we are wrong on this and you feel like we reach beyond the county please let us know!).
- We chose not to included demographic questions because we wanted the poll to be as quick and easy as possible in order to increase the amount of respondents. This means we know nothing certain of political affiliation, age, gender, economic status, etc.
- Due to lack of demographics, we are unable to know if our results are reflective of Burke County as a whole. To say this we would have to confirm that the participants were of a similiar ratio as the county demographics in terms of political affiliation, race, gender, etc.
While we can't say that our results are a 100% accurate representation of the county, we do feel confident that it is reflective of more than just our followers on social media (both on our podcast page and personals).
The post was shared 42 times for a total reach of 5,163 people according to our Facebook statistics. Each share widens the scope of the respondents to the friends of those that shared, expanding the reach beyond just our immediate followers.
Additionally, it was shared by a few popular neutral pages, namely Downtown Morganton and the Board of Elections. We reached out to a few other local political pages and requested sharing without any success.
Lastly, we paid for a Facebook promotion that specifically targeted people that were not our friends and lived in the various cities, towns, and townships in Burke County. This expanded the reach by 1,700 people or 34% of the total respondents.
As stated above, the results could be skewed by multiple votes from the same person. Given the wide reach on social media and our general optimism about our followers and community, we feel that many of the votes were genuine single votes. Unfortunately there is no way to know for sure.
Due to these reasons, we feel like we can say with a good degree of confidence that these results can be interpreted as accurately representative of some mixture of "Burke County as a whole" and "our immediate followers on social media". Just don't ask us to put a number on it.
Total Respondents: 400
Raw Data: Burke_Pre-Election_Survey_9-11-16.csv (Right click, save as...)