Big Data, Trying to Build Better Workers | Hacker News Reader

Big Data, Trying to Build Better Workers

nytimes.com

38 pointsrmah7 comments

A couple of concerns:

1. At what point (if any) does high-resolution employee data and its analysis become subject to regulations on health information privacy?

2. How do we control the risk that data-driven employee management may introduce some practices with unforeseen and systematically bad consequences? Can we expect that businesses will use the best available designers and interpreters of psychological measurement instruments? How relevant is the current state of that art to this purpose?

It doesn't seem that they're trying to actually build better workers, just that they want to select against poorer workers. I'd be more interested if someone tried to teach people the traits that make more effective workers rather than just using this data as a selection tool

Id say selecting better workers is a challenge in individual character development (honesty, work ethic, trust, helpfulness) while making more edfective workers is about skills and environment.

As the old adage goes, "good help is hard to find."

By selecting the right candidates for the right traits (traits previously unquantifiable and untenable before Big Data techniques) these companies might eventually, in the long term, build better workers. For example, if the selection criteria for a career as a developer favours people capable of getting a Computer Science degree, then more people who want to be developers will get CS degrees. But in reality, getting a CS degree does not always equal good developer. Assume there are some people without CS degrees, but with the skills and behaviours and personality more suited to developer work. These people would be better hires. Selecting for those criteria would not only allow your company to hire better developers, but would (like selection for CS degrees) encourage emerging professionals to develop the desirable traits and skills defined by the selection criteria. (In some weird social darwinist way, that I myself am not entirely convinced of.)

I'm reminded of this exchange

https://news.ycombinator.com/item?id=4848454

First of all, this sounds horrible.

Second... this: It adds a large dose of data analysis, a k a Big Data, to the field of human resource management

Genuine, actual data scientist here. We need to talk.

FOR FUCK'S SAKE. That is not what Big Data is! Big Data does not mean "data is important". Of course it's important! No, Big Data means that the size of the dataset is part of the problem, requiring approaches to computation that were previously on the fringe, such as non-relational databases and advanced distributed computing techniques, with all the concurrency fun that these needs provide.

Also, if I hear one more person call him- or herself a "data junkie" I will fucking lose it. Data is a pain in the ass (ever clean gigabytes of the stuff?) Sure, it's great to have it and you can do some really fun stuff once it's clean, but man...

good interfaces on data are rare

Reply on news.ycombinator.com