Bias Hunter
  • Home
  • Blog
  • Resources
  • Contact

Pure Importance Says Nothing

26/5/2015

0 Comments

 
”Content of the job is more important than wage”

“Debt reduction trumps financial growth”

“Life is more important than money”

All three above statements are sentenced I could well imagine someone utter in an intelligent discussion. All the statements have one other thing in common: they’re pretty much meaningless.

It’s clear that we can – and often do – make such statements. In itself, there’s nothing wrong about saying that A is more important than B. For example, “the math exam is more important than the history exam” is a perfectly legit way of relating your lack of interest in what happened in the 30 Years’ War. But when it comes to talking about what you want, and how you should distribute your resources, importance statements are meaningless without numbers.

The third case is perhaps the most common one. Presumably the idea is to say that we should never sacrifice human life to gain financially. Of course, that’s flat out wrong. Even if you agreed with that in principle, in practice you’re trading off human life for welfare all the time. When you go to work, you risk getting killed in an accident on the way, but have a chance of getting paid. Buying things from the grocery means someone has risked themselves picking, packing and producing the items – if you really valued their health, you’d grow your own potatoes. In healthcare, we recognize that some treatments are too expensive to offer – the money is better used for other welfare-increasing things, like building roads for instance. Life can be traded for welfare, ie. money.
Picture
The first case also seems clear in intent: you want to have a meaningful job, instead of a becoming a cubicle hamster for a faceless corporation, no matter the wage differential. However, it’s surely not true that meaning of the job is infinitely more important. Would you rather help the homeless for free, or be a hamster for 10 million per hour?

The problem with importance without numbers is that they are hinting at tradeoffs, but grossly misrepresenting what we’re willing to accept. The choice examples involve tradeoffs, and tradeoffs are impossible if one goal is always more important than another. This causes an infinite tradeoff rate, causing you to favor a teeny-tiny probability of loss of life over the GDP of the whole world. Doesn’t sound too reasonable, does it? In fact, Keeney (1992, p.147) calls the lack of attention to range “the most common critical mistake”.

Naturally, we can always say that the examples are ridiculous, surely no one is thinking about such tradeoffs when they say life is more important than money, surely they’re thinking in terms of “sensible situations”. In a sense, I agree. Unfortunately, one’s ridiculous example is another’s plausible one. If you don’t say anything about the range of life and money that you’re talking about, I can’t know what you’re trying to say. It’s just much easier to say it explicitly: life is more important than money, for amounts smaller than 1000 euros, say.

Even this gets us into problems, because now if I originally have a choice problem involving 3000 euros and chance of death, you’d be willing to make some kind of tradeoff. But if I subdivide the issue into three problems, now suddenly human life always wins. If you think about utility functions, you can see how this can quickly become a problem. But the situation is still better than not having any ranges at all. Even better would be to assign a tradeoff ration that’s high but not infinite. 
0 Comments

Decisions people Face: Results of the Survey

17/2/2015

1 Comment

 
So, last week we did a survey with Tuomas Lahtinen about the difficult decisions people face in the coming year. We got a total of 22 responses. What now follows is some analysis of those responses. Tuomas has yesterday already done a fantastic job of looking at the open question responses, and also categorizing the results. You can see from the figure below (copied from Tuomas’s analysis) that most decisions are related to one’s career. Given that we promoted the questionnaire on Facebook, and like our friends, we’re just the age of finishing up our studies and entering or having just entered work life – it’s hardly a surprise.

Since Tuomas already took a good look at the responses, I’m not going to repeat that. Instead, I’ll do what any analyst always does: wrangle the data for any other useful nuggets of information. So if you’re interested in the responses to the open questions, I direct you to Tuomas’s analysis.
Picture
Okay, so with 22 responses the data is of course not exactly scientific quality, but let’s look at some averages nevertheless. Below are averages to the questions “How well have you figured out the objectives/alternatives/consequence/when to make the decision” by category of the problem. To make the data a little more robust, I’ve combined Education under Career, and took one of the three questions under Other and also put it under Career.
Picture
The finding here seems to be that alternatives are better known than objectives, consequences or time. I guess this makes sense, since often alternatives are mostly a matter of browsing the internet and finding out what’s available. On the other hand, shouldn’t objectives be even easier? After all, to find your objectives, you just need to take a look inside your own thinking and find out what you value. Well, it seems that is not the easiest part of the problem.

What I find interesting is the comparison of career and family. Career alternatives and decision time seem to be relatively well known, but objectives not so much. With family, the situation is the exact opposite: objectives are clear, but consequences and decision time are very vague. This is probably due to the fact that many family decisions we can still afford to put off for several years, whereas many career choices at this point demand our choice by a certain date.

What other interesting patterns can we find? Well, here’s one:
Picture
So, what the correlation matrix shows, is that knowing when the decision should be made seems to go together with better knowledge of alternatives and consequences. Of course, with the data I can’t know whether the causality runs this way – but relying on the fact that people are prone to procrastination – I believe that it’s at least plausible on the face. Deadlines make us focus on the problem, which quite likely helps us to find out better what kinds of alternatives and consequences there are.

All in all, I’m surprised how low the values  in Table 1 are, especially since the measure was a self-report one. It looks to me as if respondents are comparing themselves to a perfect world, which is a tad unfair. We will never have perfect information, and uncertainty is something we’ll just have to tolerate to a degree. Of course, when there are information-laden variables that can help you, you ought to measure. But even after all the measurements you could do, there’s still going to be uncertainty. So what’s one to do? Well, I recommend heeding the advice of Reid Hastie and Robyn Dawes:

“[--] our advice is to strive for systematic external representations of the judgment and decision situations you encounter: Think graphically, symbolically, and distributionally. If we can make ourselves think analytically, and take the time to acquire the correct intellectual tools, we have the capability to think rationally.” – Rational Choice in an Uncertain World, p. 334

So, in short. Recognize that you can’t have it all. Decide what it is that you want, and then apply focused, analytical thinking to reach that. 
1 Comment

    RSS Feed

    Archives

    December 2016
    November 2016
    April 2016
    March 2016
    February 2016
    November 2015
    October 2015
    September 2015
    June 2015
    May 2015
    April 2015
    March 2015
    February 2015
    January 2015
    December 2014
    November 2014
    October 2014
    September 2014
    August 2014

    Categories

    All
    Alternatives
    Availability
    Basics
    Books
    Cognitive Reflection Test
    Conferences
    Criteria
    Culture
    Data Presentation
    Decision Analysis
    Decision Architecture
    Defaults
    Emotions
    Framing
    Hindsight Bias
    Improving Decisions
    Intelligence
    Marketing
    Mindware
    Modeling
    Norms
    Nudge
    Organizations
    Outside View
    Phd
    Planning Fallacy
    Post Hoc Fallacy
    Prediction
    Preferences
    Public Policy
    Rationality
    Regression To The Mean
    Sarcasm
    Software
    Status Quo Bias
    TED Talks
    Uncertainty
    Value Of Information
    Wellbeing
    Willpower

Powered by Create your own unique website with customizable templates.