This seems odd

From Macworld’s article on the Fair Labor Association’s preliminary report of working conditions at Apple supplier Foxconn:

Foxconn told the FLA it plans to reconcile most of these problems, through a variety of means: The company intends to meet overtime standards from both the Chinese government and the FLA itself, improve compensation packages, build new housing and canteen facilities, and more. Foxconn has already put in place a regulation that supervisors and workers are required to report all accidents that involve an injury.

From the report:

Additionally, Foxconn has agreed to change the system by which accidents are recorded. In the past, only those accidents that caused work stoppage were recorded as accidents. Moving forward, all accidents that result in an injury will be recorded and addressed.

It’s the last sentence that’s puzzling. Apple has publicly stated that a major reason for ramping manufacturing in China is that they need lots and lots of industrial engineers. Yet industrial engineers should want all accidents reported.

Now, these reasons are going to sound callous because engineers work with data, not emotions. The zeroth reason (above first) for avoiding accidents is to avoid hurting people. We’re all aware there have been companies and nations that have cared less about their workers’ or members’ health than they should have, and you start getting into that area when you remove the humanity and only discuss workers as “resources.” Yet the only real way we have to analyze problems is with data, and so it’s important that these analysis then be combined with a higher-level, unshakable concern for employee health, well-being, and success. Relying too much on data allows those harder-to-quantify concerns to suffer, and we’re not advocating or talking about that. We’re merely dealing with the data side of things, OK?

Industrial engineers need data to design better processes. Accidents that cause injury obviously have the highest priority not only because of employee concerns, but because they’re bad business (and this is the callous data side): even with an endless supply of cheap replacement labor, it’s better to keep your current skilled workers than to train new replacements. On top of that, injury accidents stop work, slow down everyone else (would you be as efficient just after watching a co-worker get injured?), and depending on the accident, destroy parts or machines that you need for business. These are bad news, and they have to be tracked.

But every accident is bad news for an industrial engineer. An accident that’s not 100% the worker’s fault (carelessness, intoxication, failure to follow specified procedures for personal reasons) means there’s a flaw in the process. If there are three non-injury accidents per day on the iPad assembly line, and each of them destroys one retina display, any decent industrial engineer should want to examine the process and try to find a way to make it so that kind of accident can’t happen.

The 20th century saw massive improvements in manufacturing thanks to industrial engineering. Work studies would find that workers, for example, were sharing expensive hand tools between two stations and therefore reaching in an unnatural direction hundreds of times per day to grab the shared tool, even if assembly was in perfect sync such that no two adjacent workers needed the tool at the same time. By purchasing more tools so each worker had one, it eliminated so many motions per day that production would increase by double digit percentages, allowing the tools to pay for themselves in months. Keeping tools on retractable cords within easy reach rather than cluttering a workbench could turn a 2-minute assembly into a 90-second assembly, allowing for 33% more assembly in the same time with greater comfort. The refinements are real, important, quantifiable, and benefit employer and worker alike.

Accidents mean these processes are failing at some points. Apple has spoken of having “thousands” of industrial engineers to help make these processes. The first thing those engineers should want is to design a process that allows reporting all accidents electronically in no more than 2 minutes. The data could allow redesigning manufacturing to help workers, save money, and improve productivity.

So volunteering to report all “injury” accidents to management? They should want data on every accident. It’s kind of puzzling that they don’t already do this.