Politics
Out of context: Reply #30864
- Started
- Last post
- 33,768 Responses
- ********-3
so ukit,
to what we discussed and your questions. I do not think there is any bias in the poll. the poll is a pretty standard measure over and above as you point out on polling figure. However its scary when people think they can have a bead on society through poll with about 16 popular metrics. Chasing lowest common denominators and only measuring the basics make you miss out on the larger spectrum of society. But the big hope for central planners in making "correct" decisions...
As far as bias in the graph... without a doubt its biased.... i could't see the article due to paywall... maybe there was more than I have seen... but the single graph you posted was directly related to showing a rep/dem bias on war. Which is a single metric point. And all the other metrics in the poll being ignored in the graph dictate that was the end game point. If it wasn't why not share all the data and let people decide what they take away. Why go through effort and labor to create a graph? Which designer also got wrong graphing a dataset not in the data. Which also calls into question interpretation issues. Which is similar to the 5am algo buy on trade talks cnbc did this morning. Sometimes when framing info we fuck up graphicly and can completely distort message and perceived reality.
In conclusion my little point calling out the strange "us adult" stat in the graphic (which i was right on as an error) was the main point. The rest is just commentary on the nature of polling being the soft science and highly biased commodity it is. You can agree or disagree and error of margin is anyones guess daily. I think you have mentioned the bias yourself, but not sure you see the big market for it. Not sure you even care. But hey no hard feelings. I wish I could explain better the dangers of chasing LCD poll data and holding certain measures up out of context for a sell
- For the record I wasn't claiming this or any other poll is perfect...but I think your criticisms in this case are based on pretty much nothing.yuekit
- The reason it says "US adults" is because they aren't going to poll a 5 year old on a question about war. So they have to draw the line somewhere, probably atyuekit
- 18 and older. Pretty standard stuff. Not sure why you see that as suspicious.yuekit
- It's also not unusual at all to have a poll that asks a number of questions and then highlight one of them in a graph or article.yuekit
- Doesn't indicate bias, just interest in a particular topic or question.yuekit
- Anyway this is way more discussion than a single poll question deserved, but if those are the only issues, I'd say maybe the poll is not so bad :)yuekit
- the us adult is a percentage of the polled related to the question. but the whole poll is 100% us adults (18+) there is no need to graph it.********
- Which begs the question where the arbitrary data came from. Considering I didn't see poll geography info was i wonder if it was outside US.********
- They piggybacked on that other data source I said I didn't look at. If that is the case and sample was outside US than those political terms do not matter.********
- Unless a GD fucked it up and assign values to the graphs that where attributed to something else. doesn't add up.********
- and yes its not unusual for polling to collect a variety of info. The variety makes it more valuable to sell/use to shape for bias.********
- Choosing one dataset vs others is bias. Like choosing ingredients in a meal you do it with an endgame in mind. Great to have stocked kitchen********
- But without a doubt each meal derived from the ingredients is purposefully designed. Why its not common for so many people to ignore "polls". they get it********
- still surprised this deep in asking same question but my fault for my extra commentary of the obvious********
- So your problem with it is that infographics in general are biased -- got it lol.yuekit
- Well yes but that was the secondary point. My primary objection is that us adult data being graphed. Where it came from and why? Really that bugs me more than********
- anything. Still thinking a young designer, but where would he have pulled the data to graph the percentage?********
- It says they surveyed 1,500 US adults. It is really not complicated, in spite of you trying to make it that way. I think you have too much time on your hands.yuekit
- OMG dude! than why graph 47, 31, 22% when clearly the study is 1500 adults = 100%. And no need to graph. Now here did the designer pull those numbers from?********
- Not sure how to make myself any more clear... Have you just been misunderstanding me this whole time? Maybe still...?********
- haha no idea if you are trolling or not but I can’t waste time on this anymore sorryyuekit
- There were 1500 US adults polled. Out of that 47% answered yes. 31% of those adults who are Democrats answered yes. Etc...yuekit
- It’s super easy to understand and no offense but in 100+ notes you haven’t been able to raise a single legitimate point. Time to call it a day dude...yuekit
- im not trolling. the percentage of adults is its own stat. of 1500 (all which are adults) the breakdown is dem/ind/rep = 31/40/77 (100% may not add up due to********
- rounding... 31-40-77 = 148% not a 1-2% for rounding. Plus the added US adult comparison. If 47 us adults strongly supported who was the 53% non us adults?********
- the graph you linked is all kinds of fucked up. you have to be trollin at this point********
- 31% of Democrats answered yes. Not 31% of all the respondents. Let's say a third of the people polled were Democrats, 31% of that third answered yes.yuekit
- HA why would some one do it that way? Now we don't even have a figure of dems vs reps etc polled by the graph. Basically if 80% polled were reps.********
- Did you seriously think the pollster made a mistake where the numbers added up to 148%?yuekit
- and still doesnt explain my pet peeve. why graph us adults if all are adults? can you explain that one? are their kids or non us adults? how do you get that********
- US adults just means "everyone." There's no mistake or trick going on there, it's just showing the overall percentage that answered yes, no or don't know. Andyuekit
- then the percentage that answered this for each subgroup.yuekit
- The breakdown of respondents in terms of Dems, Repubs and independents is listed in the detailed poll results I linked earlier (it's about a third for each).yuekit
- Ooh I think I get it. you're sayign they are creatign a new data set outside poll info of the graph to display SUPPORT, OPPOSE, NOT SURE of sample********
- but that is bad very bad design. first theyre combining percentages which distorts percentages. than graphing a percentage next to it which has % is not connect********
- ed at all.... its jsut a very bad infographic********
- I dunno if it's fair to blame the pollster or designer just because YOU personally couldn't understand it. I think the average person would not have an issueyuekit
- Would have left off whole as % and just graphed the 5 data sets by party ID. Less confusing and better understanding of data********
- with it, considering I've seen many polls presented the exact same way. Bottom line, there was no mistake or bias here. You were just wrong.yuekit
- perhaps. i did mistake it. as far as data go I'd now be curious of actual number of said variable vs percentages.********
- and no mistake. i see how it was done and can speculate reasons why. but any reason is bias. wether a person cares about info display vs or 3 fits better than 5********
- and is less work. but i will say an honest thank you for sticking with me on it to see it through. i would never do data this way, so my own bias blinded me********
- does beg the question in how such basic communication took so long though. and could be an example of problems of larger communication********
- why did my question take so long. Did i sideline it out of just basic trash talk? Did I hope my lazyiness in understanding data would result in someone giving m********
- me answers based on basics i saw that didnt add up. curious why it took you so long ukit to identify primary goal? Help explain it?********
- were we talking at and not with. What changed?********