Recapping: the purpose of the survey was to get an idea of how willing people are to throw good money after bad and that cost shouldn’t have mattered. It was a metaphor. If the survey failed to prove this it’s because it was flawed or data didn’t support the hypothesis. Truly, it was just something to play with. The ongoing comments from part two continue to intrigue me.
Several wanted the data set per respondent for analysis but I can only get 100 responses with a free account. I don’t do enough surveys to justify the expenditure and also, this wasn’t a scientific survey by any stretch. But, I appreciate the offers.
Again, keeping in mind my narrow dictates and results I wanted to measure (not the same thing as what was), comments to the second post are illuminating and fun to play with.
Some said they’d pay with a dollar bill thinking the coin drop wasn’t working, a very likely scenario. Your choice could depend on your orientation to the world. I wouldn’t try a dollar bill because I would interpret that as a system wide failure. If they can’t get the lowest hanging fruit to work, anticipating that a more sophisticated mechanism (bill sensor) would work is increasing the degree of risk by an order of magnitude. But that’s just me.
The topic of low cost was also mentioned as a contributing factor toward willingness to risk but again, cost is relative. Many of you couldn’t afford to risk the metaphorical 75 cents on a bad sewing contractor but others could. Whether you can or not should not matter. It’s a matter of value. Of course the vending machine is a crappy example because you could get a partial good solution with a contractor, enough to find enough common ground to work with and improve upon so an all or none situation wasn’t ideal (you couldn’t get a partial can of coke).
The issue of desperation also came up again, being very thirsty. No one mentioned a possible preventative -planning. Admittedly, that wasn’t an option you were given. Here in NM, it is very common that people carry a gallon of water in their cars at all times. Not everyone plans for negative outcomes be it with contractors or thirst but if a greater number were to, there would be less exposure to unanticipated circumstances and the need to take risks. That was kind of my point. A cold icy soda would be lovely on a hot day but the 90+ degree water you had would keep you alive until you could find a refreshing beverage elsewhere. I guess I’m trying to draw the conversation aligned to something Gini said:
It has been all too easy for me to think I am persevering in the face of adversity when in retrospect I just stupidly persisted rather than stopped to think.
Marian had this to say (snipped) about cutting your losses:
A vending machine is uncomplicated — it works or it doesn’t. People are a whole different matter. I’d need to know: can I trust this person? If someone has a history of reliability and everything fell apart in their shop last week because their mother suddenly died, I would try again if I could. But, if a mother dies one week, the kids get sick the next… then no, I would tell them I’m sorry things are so difficult for them and take my business elsewhere.
The only issue is, when do you cut your losses? Contractors may fail us or maybe our expectations are the problem but who can know the truth of it if you cannot be both parties? I think Ruth speaks (snipped) for some of us:
I would have put the coins in but not the dollar bill. [T]he first time you don’t pay me is the last because I don’t accept any more work from you. [A]fter imposing this rule, receipts stayed the same but work went down 50–60%. That is to say, my paying customers were no longer paying me to attend to someone who had no intention of paying for the work I did.
In the end, there are no answers but it remains something to think about. One should spend some time to try to define clear lines for when to cut their losses. And, planning can limit exposure to risk.
Now I have to go freshen up that gallon of water in my car.
Actually, a sample size of a hundred is still pretty good for these purposes. For anything medical, one would want more, but if this is just to get a sense of how the groups compare, there should be no problem so long as the responses are random (as in surveymonkey didn’t categorize them, then chop off the first hundred as that will skew the test because certain groups weren’t fully represented- I can check for this anyways). At you mentioned, there are already some other factors that would affect people’s answers, plus the fact that there is non-response bias (respondents are self-selected and must have internet access and read your site!).
But yeah, if you’re interested, email me at brit-anna(at)hotmail(dot)com, and I can at least try to analyze it and maybe make some pretty charts. :)
Just like any social science stats research – a whole lot of caveats and parameters, and the difficulty of assessing why an individual behaves during an experiment, or what the experiment’s effects are, since there are so many other possible influences from the individual’s life experiences.
I took media studies in school. The school I went to used to focus on stats research to determine TV/Media’s affect on populations, but the results were never strong enough to ever prove anything. By the time I went there, they snubbed that sort of research and focused on critical theory.
@ Denise:
Technically it’s not an experiment, so we don’t have to worry so much as we aren’t trying to find causation. But considering who exactly is taking this, I think the probably of a type one error is bound to be pretty high anyways. A better survey might ask questions geared to business at the end and multiple other scenario questions unrelated to business. And I do think that this survey shouldn’t be trying to leave out life experiences as those also affect business decisions, although, people do behave differently in different scenarios, hence my thoughts about the survey.
Talking about research in media’s effect being unable to prove something, even in such fields as economics, p-values like 0.40 are still considered pretty good!
I feel sad that they took out statistics. Statistics is everywhere, statistics is good, and no, I’m not biased. ;)