Study HN: The Blame Study Pre-Study (for sociological survey) | Hacker News Reader

Study HN: The Blame Study Pre-Study (for sociological survey)

spreadsheets.google.com

8 pointsmarkbao13 comments

All these questions are technically unanswerable because information to answer the questions is not given.

For example:

Q2, I feel that blame should be given to both parties, and possibly even the driver, but all that depends on information that was not given.

The car company should test the stuff that it buys better, the company that made the cruise control should test their stuff better to avoid errors, and it should 'fail safe', in other words it should fail in a way that minimizes the risk of accidents.

Also, an error is pretty ambiguous, clearly it is one that seems to have a failure mode that is disastrous, but any driver should normally be able to disengage the cruise control if it would malfunction, so maybe the driver should carry some blame too.

Similar stuff applies to the other two examples.

For instance in #1 it might have been the producer of the cake (not neccesarily the supermarket), it might have been the delay between the purchase and the use, it might have been the person transporting the cake from the baker to the supermarket and so on.

In Q3 the error was made by the employee, but it may have been due to bad training, working too many hours or one of many other reasons that would change the take on the situation.

To assign 'blame' with this little information and so many unanswered questions is un-doable.

There is a reason why accident investigations take a while, all those questions need to be answered. And the reason is not to assign 'blame', even though in a litigious society that is how they are used, the reason is to make sure that we learn from these errors and change our procedures to make sure that it will not happen again.

Stop seeing the world in color. Black and white is so much simpler.

Until you are asked what color you see.

I'm sure he'll have plenty of data without you...

If you take an inkblot test, do you say: "a blob of ink, a dab of ink, a splash of ink..."?

When after a robbery you're asked which one of three innocents did it will you say 'none' or point at one of the innocents ?

On another note, the author specifically asked for feedback, so apparently I answered at least one of this questions.

his harmless survey is more akin to another psychology test than wrongfully accusing someone you know to be innocent...no idea why you even threw that reply in there.

Let me help you out here then, if you are given a multiple choice question where 'all' or even 'more', or 'none' could be valid answers but the test does not allow that and you are asked for feedback on the test you can choose to ignore the test (not answer the question) and just provide feedback.

You used an analogy to make your point, I used an analogy to make mine.

I suspect that that's the point of the survey, though. Even though the questions can't be answered definitively, and even though they may really be the wrong questions to ask, we're faced with them on a daily basis.

Shakesville (http://shakespearessister.blogspot.com/) has a regular series where they point out how regularly various media outlets write stories about domestic abuse where the story is framed so that the victims (overwhelmingly women) are blamed for their assaults. This is an extreme case, but it's illustrative of how society at large (and each of us at small) doesn't approach blame in a reasoned and careful way that considers all the angles and second- or third-order causes.

I doubt the point of the survey is to actually determine blame for specific events, but to investigate how our faulty meat-brains tend to lay blame when we have limited information.

Anyway it should have included an option "Can't confidently blame anyone", otherwise it assumes that we have to blame someone.

that option is there: it's the "back" button.

Exactly, that's what I did, but my action was not recorded in survey results which became more biased because of lack of options. It's like asking what my favorite color is and only offering 3 options without the option "other".

If your goal is to assign blame, it can never be done on such a naive scale. The US has very well written liability laws, e.g. for Q2 every party involved assumes some liability. For every car crash, the individual assume some liability (that's why we all carry insurance), car companies assume some of it, parts suppliers do (remember the Ford/Firestone issue where the explorers were flipping over?), USDOT (e.g. CalTrans in California) does and the regulators at NHTSA do.

There is a reason why legal and insurances industries are so advanced. Blame is never black/white.

There should be a sliding scale of apportionment, and a statement of the perspective doing the evaluation (and there is no 'global' perspective).

For Q1, it's fairly easy to say 100% store -- there's an implied warranty their food is fit to eat, and the friend has no special insight or customary duty to double-check (via lab equipment?) the healthfulness of already-warranted food.

For Q2, it could approach more like 50/50 -- and depends on the perspective involved. From the perspective of the driver or victim, it's easy to say 100% car company -- again because the risk flowed through the car company, which had the expertise and commercial duty to sell a safe product.

But between the car company and the original feature developers it depends on their arrangement. If the developers are an outside firm that sold the system as fit for a particular purpose, much or all of the blame then flows to them. If the developers are the car company's in-house employees, the company probably agreed to take on the blame along with the profits, as the company is the coordinator and tester/seller of a multi-person process that failed.

For Q3, again it could vary and be a matter of perspective. From the perspective of the customer, it's 100% the shop, which has taken on responsibility for the benefits and costs generated by its employees in the performance of their duties. (From the customer's perspective, the barista and the shop are one and the same.) For the shop owner, responsibility may be thought to lie with the barista, especially if the rate of mistakes goes beyond typical/achievable levels. Still, as a practical matter of employee morale and labor law, a policy where the barista pays the costs of occasional errors is unlikely. The remedy for excessive errors is change of employees, not compensation for individual mistakes.

Reply on news.ycombinator.com