ISSUE 1: Slight Disapproval in Confirmation of Baobab

Sincluda

making my way downtown
Cabinet
Deputy Minister
Citizen
Pronouns
He/Him
Screenshot 2021-09-22 7.27.06 PM.png

ISSUE 1​

SLIGHT DISAPPROVAL IN CONFIRMATION OF BAOBAB

by Sincluda​



The Not-So-Daily Bugle recently ran a flash poll, and you know that it is totally accurate because it is a flash poll done by independent media. Regardless of the true accuracy of the poll, there is still something we can learn from it - particularly the surprisingly high amount of disapproval at a cabinet nomination. Ok, it hasn't been THAT long since a poorly received nomination *cough*whatermelons*cough*, but what's interesting is that this nominee - Baobab - has been a popular choice for his now-confirmed position (Minister of Culture) if you look at the EBC's executive satisfaction polls. Just this last one, Baobab received four nominations for MinCult - which was also more nominations than for any other position (Minister of World Assembly Affairs had 3, split between Sincluda (me) and Pland Adanna). Enough talking, I'll give you what you came here for - flash poll results!

Forms response chart. Question title: If you were a senator, would you vote AYE or NAY on the confirmation of Baobab?. Number of responses: 16 responses.

Here you can see what the title mentioned - a slight disapproval for Baobab's confirmation. For those not willing to figure out what 56.3% of 16 is, 9 people said that they would vote NAY on the confirmation of Baobab. 37.5%, or 6 people, said that they would vote AYE, and apparently Baobab answered to say that he, like always, would abstain (I had to take that jab, no offense, Bao). What we can gather here is that not only is there some tangible disapproval for a cabinet nominee, but there is a somewhat tight (albeit not intense and already resolved) debate.

1632577979819.png

This is the only comment section that got any actual responses. I could poke holes in either argument - hasn't Bao proven himself on a smaller level? Isn't that putting a lot of trust into a relatively basic and unconvincing promise? I won't go too in depth here, so make of it what you will. Also, as I said earlier, neither of the other comment sections got responses, so I won't look at them.

What did we learn here, today? We learned that more people than normal dislike a cabinet appointment. Big whoop, really needed a poll to tell us this. But, on a less accurate level, we could assume that a majority of Europeians are dissatisfied with Baobab's confirmation. However, like it or not, Baobab is the latest Minister of Culture, and we'll see who made the right prediction.

xXx​
 
I feel like 16 respondents (~27% of first round presidential election voters (60)) can't really provide any sort of reasonable estimate on how the region as a whole feels about it. Baobab was confirmed, so I feel it's pretty reasonable to assume that more dissidents are going to feel like responding and commenting on this poll than those whose preferred choice was actually enacted. It's been shown that under such kind of circumstances (where a factor directly bound to the survey or poll is likely to influence respondent rate) you very often need a convincing respondent rate in order to draw workable conclusions or inclinations from it speaking about the population as a whole (Groves, 2006). If there was some kind of unbiased sample selection going on then you could probably prevent that to some degree, but that's obviously not the case, so I feel like the proportion Aye vs Nay is basically discardable given the amount of sample bias this is prone to. I would personally hesitate to even publish something with such a low respondent rate, and if I did, I would strongly caution against using it to draw any kind of statement or conclusion from at all.


Reference for those interested:
Groves, R. M. (2006). Nonresponse Rates and Nonresponse Bias in Household Surveys. The Public Opinion Quarterly, 70(5), 646–675. http://www.jstor.org/stable/4124220
 
I feel like 16 respondents (~27% of first round presidential election voters (60)) can't really provide any sort of reasonable estimate on how the region as a whole feels about it. Baobab was confirmed, so I feel it's pretty reasonable to assume that more dissidents are going to feel like responding and commenting on this poll than those whose preferred choice was actually enacted. It's been shown that under such kind of circumstances (where a factor directly bound to the survey or poll is likely to influence respondent rate) you very often need a convincing respondent rate in order to draw workable conclusions or inclinations from it speaking about the population as a whole (Groves, 2006). If there was some kind of unbiased sample selection going on then you could probably prevent that to some degree, but that's obviously not the case, so I feel like the proportion Aye vs Nay is basically discardable given the amount of sample bias this is prone to. I would personally hesitate to even publish something with such a low respondent rate, and if I did, I would strongly caution against using it to draw any kind of statement or conclusion from at all.


Reference for those interested:
Groves, R. M. (2006). Nonresponse Rates and Nonresponse Bias in Household Surveys. The Public Opinion Quarterly, 70(5), 646–675. http://www.jstor.org/stable/4124220
Agreed, 16 is very low. I definitely would have made it into a normal 48 hour poll and re-advertised. Normal rules of polling and statistical inference don't tend to work well with very small populations like ours though. When we do a normal election poll, we're usually capturing 50-75% of the actual electorate, not just a small sample size.

But yes, given the last election total of 60 votes, and a disapproval count of 9 from this poll, it's likely that the disapproval rate is not actually over 50%.

This is one of the major issues with Flash polls not tied to elections, very low response.

If there was some kind of unbiased sample selection
This is not possible in Europeia or any community this small. Our polls will always be self-selected samples. We can't draw strictly statistical conclusions, we mostly use the polls as a sociological snapshot or referendum on how the most active members feel.
 
Just a reminder: I absolutely know that this poll can't possibly be that accurate, and I acknowledge that in the article.
 
So, I've been looking for best practices on accuracy in very small population sizes (~50-70, which is our electorate size) and the advice is that you basically have to poll almost the entire electorate to get accurate results, so, something like 60% or more. So, if we have 60 people voting in the election, we'd likely have to get around 35 responses to be able to rely on that data (and we'd still have pretty large margins of error, -+(8-10%).

All of this aside -- don't underestimate the power of a few highly motivated opposition figures to change the opinion of the region. We're all highly correlated opinion-holders in Euro, we're not tribalistic or entrenched in positions, we tend to be flexible on positions when good arguments are presented and move with the crowd in many ways.

I know I wouldn't want to go up against an Op-Ed (or a GH opposition thread) with 9 vocal supporters backing it.
 
Last edited:
Back
Top