Malcom Gladwell Knows Everything!

In this rather lengthy video which ultimately may be considered a preview of his new book, Malcolm Gladwell regurgitates a thread from Moneyball and one which is the core topic of concern to any selection specialist. The focus is on the problem of identifying effective predictors of job performance. As in Moneyball, MG discusses the use of field tests as predictors of future performance in sport. He also uses examples of cognitive test in the National Football League, physical abilities tests and subjective tests as predictors of performance in hockey, football, and basketball.

The focus is on what MG has termed 'mismatch'. Mismatch means identifying the wrong person for the job and fundamentally rests on choosing the wrong predictors (as was the story in Moneyball). To generalize the problem, MG goes on to provide examples of the difficulty of predicting effective performance in teaching and in the legal profession.

In teaching, the question is what predicts teacher quality? This is clearly important as it has an impact on student learning. Traditional predictors include various dimensions of education and experience. To teach in the US you must have a first degree, certification, state licensing, and academic work specific to the specialty. As reported by MG apparently there is no relationship apparently between these predictors and teacher quality.

What is missing from this discussion is a definition of what being a good teacher is. It is true, that in part a good teacher is one who contributes to improvements in test scores of students. However, many would argue that simply seeing standardized test scores improve cannot be judged the only indicator. What if you are lucky enough to work in a school where you have bright students. They will learn despite your efforts to the contrary! What if you are unlucky enough to work in an environment where students have other serious distractions such as family troubles, poverty, gang violence? Can we judge a teacher as poor even if her achievement is simply to keep kids in school? How will we measure this criterion (similar issues for Doctors and Lawyers by the way)?

Unfortunately, in his rant, MG has made a fundamental error. He never considers the complexity of measuring the criterion - what is an excellent athlete, what is an excellent teacher or lawyer? By obfuscating this half of the equation, it is unlikely that any satisfactory analysis will be achieved.

However, Gladwell does pose a challenge to selection specialists everywhere. In a world that is complex and dynamic, where the required behaviors are uncertain, should we continue to search for predictors of performance? Here in my opinion, are some of the flaws in the logic of Gladwells argument:

1. He ignores the fact that even a weak predictor is actually likely to improve your odds over no predictor at all (chance) or a bad predictor (stereotypical ideas of a good performer). If you combine several modestly predictive measures you can significantly impact job performance of new hires.

2. Performance is complex and must be defined clearly. Sometimes we can predict parts of a job performance rather than overall job performance.

3. Using standardized predictors increases actual and perceived fairness. It helps us to ensure that blacks, whites, women, men, old and young etc. all receive consistent treatment. Furthermore, it enables us to hold companies accountable for how the decision is made.

4. Despite the examples of cases where selection systems do not work, MG ignores the fact that when done well, careful selection improves the average performance of the employees hired and that has a positive impact on organizational performance. This was the other side of the story in Moneyball!

I am looking forward to Mr Gladwell's book. He chooses interesting topics and writes beautifully. What is your opinion, is selection more or less important as the world increases in complexity?

Waive goodbye to your employment rights

An article from the Detroit Free press suggests that it is rather easy for an employer in the US to overcome statutory limitations on filing sexual harassment and other discrimination charges. While the EEOC allows up to 300 days, and other state laws allow up to 6 years for a statute of limitations, employers may simply build shorter limitations right into their employment contract. This is apparently what Chrysler (alone among the major US auto manaufacturers) is doing. Specifically, it is limiting employee claims to 6 months. Any occurances older than six months are disallowed under the employment contract - and this has been upheld by US courts.

Consider the situation - you are denied opportunities or benefits based upon your age, sex, or national origin etc. The first time this happens you are unaware of the issue, over several months or years the pattern becomes clear. Eventually you become so concerned that you wish to act. You file a complaint. Only the specific events that have occured within the six months of your complain may be considered as relevant. How would you feel? How would you feel when you realize that the intent of the labor law of your country was quite simply undermined by the fact that employers are allowed to ask employees to waive their legal rights in exchange for a job?

Now consider that unemployment is rising throughout the developing world. The pressure to accept a job when offered, no matter how unfair or unreasonable the terms, is rising. At a time when employees only defense - the labor union - is at it's all time weakest worldwide, who or how will employees redress the balance in this very unfair bargain.

It will be interesting to see how smart employers deal with this weakness in the labor union and the increased power that they have for doing bad (evil may be too strong a word here) as well as doing good. What happens when you build a reputation for such shenanigans and then the labor market gets tight once again as we expect it will? Who will want to work for you? Will you labor costs be higher than your competitors? Will you employees be less committed?

We call it regression to the mean

When you see outstanding performance on a regular basis, it is quite possible that you will sooner or later see performance that is closer to the average. This is a statistical necessity called regression to the mean. In the case of a company with an outstanding record, it is quite likely that we will sooner or later see a drift towards the average.

Google seem to be experiencing issues like this recently in terms of their people management practices. Not surprisingly, as the firm grows and grows it needs to control costs and promote the return on its people practices. Now, Google is receiving flack on a number of fronts. Just this week, the New York Times reports that Google has announced an increase of 75% in the price of child care for employees:

"Parents who had been paying $1,425 a month for infant care would see their costs rise to nearly $2,500 — well above the market rate. For parents with toddlers and preschoolers, who were charged less, the price increases were equally eye-popping. Under the new plan, parents with two kids in Google day care would most likely see their annual day care bill grow to more than $57,000 from around $33,000."

According to Silicon Valley Insider:

"the Google co-founder Sergey Brin said he had no sympathy for the parents, and that he was tired of “Googlers” who felt entitled to perks like “bottled water and M&Ms,” according to several people in the meeting. (A Google spokesman denies that Mr. Brin made that comment.)"

Is this also a case where a company has built a reputation for its remarkable HR practices only to suffer when reality creeps in? Could it be that HR must be careful in terms of long term planning - how fun and attractive can we reasonably make this company if we expect to be held accountable for costs and returns in the long run. Is it really true that the 'Googlers' feel a sense of entitlement, or is this backlash to be expected when you seem to be breaking a psychological contract. Google promised explicitly not to be evil. They also promised, implicitly, to take great care of all of their employees needs. If they are seen as breaking these promises by insiders and outsiders, it will no doubt impact their employment brand and ultimately their corporate reputation.

There are a number of clouds on the horizon. In a recent blog, one insider explained why he was returning to work at Micrsosoft:
"There are many things about Google that are not great, and merit improvement. There are plenty of silly politics, underperformance, inefficiencies and ineffectiveness, and things that are plain stupid. "

It will be interesting to see how Google handle this next phase in their life - as they move beyond the 'everyone loves us' phase. Are they going to regress to the mean - at least an 'industry' mean if not the 84.hour.week.working.at.the.steel.mill mean.