What to Look for in a Statistician

If you’re looking for a statistician to hire or subcontract to, here are some things you might consider. Before Data Mania and before statistics became the sexiest job in the 21st century, statisticians mostly had no image at all or were perceived to be dweeby types hiding in the “back room”, performing calculations and never talking to anyone.  

They were viewed as perfectionists preoccupied with mathematical theory and unconcerned with the real world. Statisticians have also been confused with programmers and data processing specialists because these functions are headed by statisticians in some organizations.

Successful digital transformation is a matter of know how and access to the best talent. We connect you to both.Click for more.

There is usually at least some truth to any stereotype. Besides, is being an uncommunicative nerd really that awful? Not everyone can be the life of the party, and introverts and theoretically-oriented folks do contribute to society.

Certainly, more than the now abundant used car salesman (these days typically selling software). Susan Cain has much to say about these points in her quiet, delightful way.

All this said, statistics is a multifaceted occupation and it is very difficult to generalize about statisticians. Different organizations have different requirements and these change over time.

Here are some of the things you might consider, bearing in mind that no one possesses all of these characteristics and that some may not be relevant in your case.

Personality and Intellectual Traits

Honesty can never be undervalued – even gangsters apparently respect it – and statisticians often have access to data that are personal or sensitive to an organization.

Humans by nature are hierarchical and seek heroes. It is important to understand that IQ, at least as far as I can tell, is not correlated with personal or professional ethics. The brilliant person with questionable integrity is among the most dangerous of our species.

Curiosity has probably killed few cats – felines have been in existence for millions of years after all – and most likely has caused the demise of even fewer statisticians. We live in a constantly and rapidly changing world, and many of the methods I now use did not exist, or were on the drawing board, when I was in grad school.

“It is not the strongest of the species that survives, nor the most intelligent that survives. It is the one that is most adaptable to change” has been attributed to Charles Darwin. Though this quote may be apocryphal, I suspect he would have agreed with it in substance.

A hunger to understand the Why is another important asset for a statistician, especially once they have a few years’ experience under their belts and have begun to feel confident that they know which techniques to apply in which situations.

At the junior level, perhaps all you really need is a number-cruncher but for more senior positions that will not be enough.

Imagination. Statistics is seldom just plugging numbers into formulas. Statisticians and the Real World and What Makes a Good Analyst dig a little deeper into this.

Along these lines, I feel some coursework in “softer” subjects such as history can keep our heads from getting to hard. Broadening Our Perspectives elaborates a bit more on this.

Would you want your lawyer or doctor to tell you only what you want to hear? Objectivity is also crucial to being a competent statistician, and if you want a yes man, hire a yes man, not a statistician.

Critical thinking is a must, as is statistical thinking. One can be a highly skilled programmer and a math virtuoso and still be a lousy analyst. This is not unusual, in my experience.

Successful digital transformation is a matter of knowledge and access to the best talent. We connect you to both.Click for more.

Technical Knowledge

Competence in a range of analytic methods is a major asset for most statisticians. Like baseball, statistics has its one-pitch pitchers. They may be very good at one thing and try to use it in every situation.

This can lead to serious mistakes and bad decisions. See www.cannongray.com/methods for methods many experienced statisticians and data scientists utilize.

In college, we need to declare a major at some point and some people who like math, or at least are not intimidated by it, major in statistics or a related discipline. They may have no real interest in statistics and do not belong to professional associations such as the ASA or RSS.

They may do little outside reading. These sorts of people may have other talents but will never develop very far as statisticians. It helps to ask about things such as professional memberships and reading habits.

Be on the lookout for point-and-clickers and coders. Just a few years ago, data analysis was uncool but now, perhaps following head hunters’ advice, it seems almost obligatory to claim statistical competence. Many people have now had experience with user-friendly stats software without really knowing what they were pointing at and clicking on.

Others are junior programmers – coders, really – who have learned to copy/paste R or Python code and make slight modifications as needed. Neither is true statistical analysis, and there are obvious risks in hiring or subcontracting to these sorts of people. To be fair, many of them know no other way to analyze data and it is not a case of them being dishonest.

Over the years I have used at least two-dozen statistical software products, both commercial and open-source. Competence with several statistical and data mining software packages helps round out one’s skills. One reason is that there are many ways to implement a statistical routine, and most have several options.

The same method may be programmed in various ways and provide different options, in other words. The software documentation, often useful as a reference, may conflict in the recommendations offered as well as how a method is described.

A marked difference between a statistician and the typical data scientist I encounter is that statisticians normally have had training in how to design primary researchAbility to design primary studies, not just mine existing data, is a major plus for data analysts – you will be a better analyst if you also know how to design research. A good statistician is also a good researcher, in my experience.

Good statisticians are more focused on problem solving than on the theory and mechanics of stats. This is not to say technical knowledge doesn’t matter – jazz trumpeter Freddy Hubbard possessed brilliant technique and extensive knowledge of music theory, but he also knew how to play.

Data

All data are not the same! Subject matter expertise relevant to your organization or field is a plus, and the person or persons you are considering should at least be eager to learn about your organization and discipline. Without knowledge of what they are analyzing or why, statisticians have one arm tied behind their backs.

Understanding how data are actually used in decision-making – as opposed to the utopian world of the blogosphere and sales pitches – is very, very important. Many enthusiastic and talented young people have been shocked when the results of their analysis are simply ignored or twisted for political purposes. More data and bigger data will not change the world, I’m afraid.

Understanding of business is critical in many organizations but not all. Obviously, in marketing research or sales and operations planning (S&OP) positions it’s a must.

Some understanding of data management – at least being able to communicate with those in your organization carrying out this function – is essential. Statisticians must also be able to create the data files they will need for their analyses from the bits and pieces available in your organization’s data bases or external sources.

However, it is not necessary except in rare cases that they be skilled at Hadoop or have other IT-level competencies. This is a separate profession, despite some of the blogosphere baloney I’ve heard.

People and Communication Skills

These are relevant to most jobs, so I won’t say much about this topic here. I will only add that, over the years, I’ve noticed that quant types often take statements very literally, which can lead to miscommunications and, occasionally, bruised feelings.

It’s not that we lack a sense of humor, it’s just that we don’t have a sense of humor. We also tend to zero in on a particular sentence…such as this one…and ignore everything else that is said or written. Smiley face.

Successful digital transformation is a matter of know how and access to the best talent. We connect you to both.Click for more.

Summary

As I mentioned earlier, you won’t get all the above in one unicorn, but you probably won’t need to. These are just some things I would consider. Of course, there are other important skills, aptitudes and personality traits besides the ones I’ve listed, but I hope you’ve found this short article interesting and helpful.

Arrange a Conversation 

Browse

Article by channel:

Read more articles tagged: Analytics, Featured

Data & Analytics