Happy to help, sorry rushed a little bit to finish the last reply. Effectively the point the original replier was trying to make is that the data set is polluted with bad data because the collection method is just terrible. So back to the analogy I started setting up earlier. If the goal is to get everyone food, you technically win… Job done… Good job. Food will arrive, some people will get the chicken they specifically requested, and maybe a few people who actually wanted hamburgers will be happy too…
But if the goal is to know what your coworkers actually wanted to eat and get it for them, then the only orders you will certainly get right would be for the people who actually wanted chicken, had the opportunity to reply, and took time out of their day to confirm there order. But you will also have people who maybe aren’t that keen on chicken but ordered it because they really didn’t want a hamburger.
Everyone else will now get a hamburger… That includes people who actually wanted hamburgers, people who didn’t have a preference, people with a preference but it’s something other than chicken or hamburgers, people who actually wanted chicken but didn’t get their order in on time, people who brought there lunch and planned on eating it instead, people who thought the message you sent was a scam and didn’t reply but would have said chicken if they had known it was actually legit, people who told you in person they wanted chicken and didn’t realise they still needed to email you, people you sent the email to but were actually on vacation or working from remote that day, etc. All of them, hamburgers… How exciting… LOOK HOW POPULAR HAMBURGERS ARE EVERYONE! I can’t believe hamburgers beat chicken! Can you believe that 67.3% of our office is such fans of hamburgers?!
Basically the results of a poll constructed like in the original post would be utterly trash, because the method being used is horseshit and not how any serious poller would/should ever conduct a poll.
Beautiful! Yeap that’s a very clean parallel and also extremely important in that field as well.
Since this has turned into a lovely exchange I want to offer one more point for your consideration.
Where the outcome of marketing data typically intends to position a product or service in to it’s most profitable position, and the quality of the data produced can be somewhat validated by future sales/market share/market depth/etc.
Polls like the one we are discussing aren’t constrained in the same manner and may be maliciously and purposefully designed to generate biased data. Humans are inherently vulnerable to hostile psychological manipulations. If this poll specifically isn’t just an outright scam intended to get its recipient to click on some link that load a payload of malware, it is certainly designed to purposefully create skewed empirical data.
Not everyone, but certainly a small minority of people who may have not necessarily felt certain about where they stood either direction could look at the results of a poll like this and might find a tinge of doubt in the back of their heads. This sort of tactic hopes that a person will feel a paranoia that everyone else seems to know something significant that they don’t and drive social anxieties up. Again, not always, but also not uncommonly, we can find ourself doubting even deeply felt personal resolve on a topic or position if it feels like the vast majority of people disagree with us. This sort of cognitive bias warfare isn’t intended to immediately flip a persons perspective, but rather it’s designed to soften a persons resolve and introduce enough doubt that they may become susceptible to being flipped later. It’s why we need to embrace healthy skepticism and be willing to be more stoic with how we consume numbers others prepped for us.
Much love friend, have a great day/evening!