3 ms·
I think it shows that their model had too few features, and was an excellent example of the streetlight effect of genetic testing. Because genetic results are s
by rz2k 10y ago
I think it shows that their model had too few features, and was an excellent example of the streetlight effect of genetic testing. Because genetic results are so easy to obtain the signal is readily overweighted by people for decision making.
I would far rather have Ethan Hawke's character as a coast guard swimmer rescuing me in a storm than his brother, even if his brother had no similar heart defect. I would do that because I saw two instances in the movie of him out-performing his brother in feats of swimming endurance when under pressure.
His character is statistically more likely to die on the mission. Just to get a handle on the numbers let's say that with their filtering a 30 year old has about 1/20,000 chance of dying in a given month, and he might have a 1/1000 chance. We haven't seen anything about how he outperformed other candidates, but we can consider the respect of his peers and that he has succeeded in spite of appreciable time spent on establishing an identity and tracking down his DNA.
It probably doesn't matter if he dies on a routine mission. It is the rare mission with an anomaly where someone who is truly exceptional matters. If he is only 20 times more likely to provide the unusual solution that saves the mission, then it is a wash.
Anyway, my point is that one perspective on filling positions is to develop a long list of dis-qualifiers like health forecasts (and minimum competence). That is different than choosing the most extraordinary people possible (where negatives play a role in that determination). I know there is a tendency toward the former when choosing people for very selective and highly sought after roles, but it probably only appropriate when trying to fill out a large headcount with minimal resources for selecting the people. Those who can excel during crunch time are probably far more important for high level positions that are routine but punctuated requirements for extraordinary performance, than the people who look good to a hiring manager playing it safe and relying on clear but potentially less important signals.
- mikeash 10y agoDoes this just come down to a difference in the assessment of the protagonists chances? I came away from the movie thinking there was a ~75% chance he'd die en route.
- rz2k 10y agoIt must, because I'd feel different if the actual likelihood of him dying were that high. I haven't watched the movie in a while, but was it hyperbole and were they using statistics as a figure of speech? With intellectual laziness being the defining characteristic of the movie's society, exaggerating risks as post hoc justifications of biases against "in-valids" seems possible. DNA makes more sense than, let's say, blood type or astrological sign for hiring decisions, since there is a conceivable mechanism for determining future job performance. However, I wonder if the permission society that they lived in was capable of doing good analysis. How would they cope with the relative weight of environment or even gut flora when DNA practically part of a civic religion with people being randomly tested on the street? Given a situation like the declining accuracy of Google Flu Trends over time, would they insist that the predictions were right, and the indicators of inaccuracy were wrong?