When I say polls are anecdotal. I mean I put them same category of evidence. Right at the bottom. Or the start. Depending on your frame of reference.
To me anecdotal evidence relies on self reported opinions and/or experience.
If you can ask 1000 people if they like cheese. You are just going to get back a yes or no opinion on their taste for cheese from a 1000 people. Which might not be relevant or useful to what you trying to measure. They could all thinking of different types of cheese. Some might not understand what cheese is and confuse it for something else. Some might lie. Maybe they told the poll taker to fuck off and the pollster filled the answer in for them. Maybe they are thinking of different way of using cheese than eating it.
I would rather actually hear their individual cheese stories without the prompting, to get a better indication of what kind cheese interactions they have had and gain better insight into what they are basing their opinions on.
I think that would be much more useful when finding a direction to take my cheesy research.
Thatās what poll design is for, finding questions that try to avoid misunderstandings of the question or use proxy questions to prevent preference falsification. And if any polling outfit is credible on this it would be Pew.
Then you arenāt looking for quantitative analysis and I have no idea why you would then try to criticize in quantitative terms like youāre reading out of an undergrad statistics textbook. Youāre looking for qualitative analysis e.g. narratives, interviews etc. Which doesnāt involve nās or p values or regressions at all. And is definitionally anecdotal.
Polling is definitionally quantitative analysis. A regression is only one type of quantitative analytic tool.
In the former you are attempting to quantitatively measure public opinion given a n without systematic error introduced or a specific n (e.g. Latino, paraplegic, Nurse Practitioners etc.) opinion. In the latter you are attempting to measure correlation between two data sets.
You would use a regression if you were attempting to see how polling on e.g. gay marriage correlated with polling on e.g. cannabis legalization over time; or even within that polling data to see how age, sex, race, or self identified political party correlated with support for gay marriage or cannabis legalization.
If what you are looking at is the data presented by the poll itself you wouldnāt do that because the data you are looking for is already presented.
It seems to me youāre trying to attack the evidence presented but arenāt sure how. First you say it is anecdotal, which it definitionally is not, then say you would prefer literal definitionally anecdotal evidence instead. Then when I describe an extremely basic concept of quantitative analytical tools you look like a startled NPC from Metal Gear with a big old ? hanging above your head.
All after coming in really fuckin hot accusing me of being a fraud (as if claiming to be a trained political scientist is some kind of valor anyone in their right mind would be looking to steal and that I am in my career distancing myself from anyway, I would trade it back for the time wasted anyway) and an idiot and you lack the extremely fundamental understanding of the field youāre trying to school me in.
If you just want to say you think my opinions are mean and Iām a doody head then just do that and donāt try to pretend youāre defeating me with āfacts and logicā like Ben Shapiro. If you want to present an argument about something then I donāt expect you to give me a works cited page but at least donāt be a cunt about it.
I donāt know how more plainly to make it. If you donāt understand a word I wrote just ask me what it means e.g. quantitative analysis vs qualitive analysis. I promise you Iām not trying to obfuscate anything. I donāt want to believe you are sub 90 IQ but youāre making it very difficult not to.
No, I didnāt point to that one thing and write āsee this PROVES itā. I provided that data set and noted public elite cues (e.g. Obergefell, Obama admin change on gay marriage) and wrote that the totality of this evidence points to elite driven opinion.
I then provided some links to quantitative studies asking the same question and finding the same thing only a few comments above.
Every single reply in any poll is just a simple anecdote. With the questions worded to remove any long worded reply to make the answers easier to quantify and somehow that makes system more objective. Because obviously math is always objective right. 1+1 is always two. So if you make yes=1 and No=-1 youāll always get the objective opinion of a group when you do the math.
This what you trying to sell me anyway. As is my understanding.
Clearly you do not. You literally claimed that a definitionally quantitative analytical tool is āanecdotalā and that you would rather a literal qualitative tool that literally collects anecdotes.
Bro. I provided peer reviewed literature to support the admittedly very basic and casual collection of evidence I presented. One of those is even on the same topic!
My face when this entire back and forth:
At this point just donāt even bother replying. I am only going to respond with āThatās nice honey.ā Thereās no other value in engaging seriously with you.
No, Iām saying that using a quantitative method to analyse qualitative inputs doesnāt get rid of problems that can be inherent in qualitative evidence.
Garbage in garbage out.
I would rather have my qualitative answers undistilled in their naturally raw occuring state. Then ran through the shit pipe that is a poll. Then sold the idea that all the shit was magically filtered out.