Learning how to hire #2: Advice about hiring new lab members

I just went through the month-long process of hiring a first research assistant for my new lab. If navigating the posting of a position through HR was tough, what followed was even tougher. Partially thanks to Craigslist (see here), we ended up with almost 200 applications in 4 days, got them down to 10 viable candidates for phone interviews and 6 candidates for in person visits. This decision also involved dozens of phone calls to references and a lot of thinking about defining the responsibilities of the position and the required experience and how much training should be involved. Overall, it was very informative and we are really happy with our selection.

I tried to follow some of the advice I had received in the past from senior investigators (here and here) about taking my time and being mindful of my "gut feeling", but a recent series on LinkedIn from Joel Peterson, the chairman of Jet Blue, has a lot of great insight on the 10 biggest hiring mistakes.

Mistake #1: Hiring yourself. Diversity is good to bring balance to the workplace. You just want to make sure that all personalities gel together. You also cannot expect that everyone will be just like you and have to accept individual differences.

Mistake #2: Hiring too fast. Take your time hiring and finding the right person for the job. This has been repeated to me many many times by countless people. Do not rush to fill your empty lab or you will risk hemorrhaging cash and not being able to start on your best foot.

Mistake #3: Hiring the resume, not the person. Someone may look great on paper, but may not fit the culture or the energy of your lab. The first few people are incredibly important in defining the way you want your workplace to feel: vibrant and exciting? fun and relaxed? focused and ambitious? Peterson puts in terms of "brain and heart", you want someone smart and motivated, but also someone passionate about the job at hand. One more post I wrote about establishing your lab culture here.

Mistake #4: Interviewing on autopilot. Do not just go through candidates all in a row without a strategy.  I interviewed 8 people on the phone in one day and by the end of it I was completely exhausted. Because HR asked me to be consistent, I developed a script for what I was going to say about the position and for the questions I was going to ask and I took notes on all the answers, and that was my saving grace because by the end it all blended together. I took notes on qualifications, but also "gut feeling" on whether the person would be a good fit. The in-person visits were split in several different days so that we really had the time to talk to the candidates and better explore insight or inconsistencies coming from the previous chat.

Mistake #5: Lazy reference checking. Always always check multiple references ON THE PHONE and ask difficult questions if necessary. This may be by far one of the most important things. Firing people is difficult and emotional, but for some reason firing people in the lab seems sometimes impossible and is always messy. I have seen a lot of horrible situations where the fit was just not right and where information may have emerged from candidly talking to the references. During my own recent hiring, I received tons of honest feedback positive and negative which greatly informed my hiring decision.

Mistake #6: Freezing out your team. The entire team should be involved in the decision process and I may add, people should have power to veto a hiring decision. For a large (40 person) lab, my postdoc lab had an amazing combination of people who for the most part got along well, which is very rare for such a group. One of the most important features of the hiring process was that we all participated in it and talked to candidates, took them out for lunch and got to know them a bit. Everybody's opinion was considered and that was very important in making sure that only great people got in.

Mistake #7: Only hiring inside or outside. By definition in science you almost always hire from outside because people cycle through as they rise through the academic ranks. Moving around is favored and people are discouraged from remaining in their PhD lab for a postdoc. But continuity has its benefits and there are positions that may require more continuity and better knowledge of your institution.

Mistake #8. Blowing the first 90 days. This is another very important one! The way someone interacts with the team and absorbs the lab culture is dictated during their onboarding process. The culture you want them to absorb and the techniques you want them to master will become the focus of the training period and the way your new hire performs may depend on this.

Mistake #9. Focusing on money. If you wanted a high-paying job, you wouldn't be in academic science, but the lesson here is to provide rewards and positive feedback that go beyond compensation. In his great book about motivation, Drive, Dan Pink discusses Autonomy, Mastery and Purpose as important driving forces in employee motivation. See my post here for using these forces in the lab.

Mistake #10. Not firing a bad hire. Firing often seems impossible for scientists and a lot of "lab horror stories" revolve around people who should have been fired years ago (or not hired in the first place). I am not looking forward to this and I really hope it never happens, but it is so much easier for all parties involved to end things quickly and ease a transition to a different position.

Hope this is helpful. I will definitely try my best to avoid these mistakes.

A new paradigm for autism therapy from cancer research

I do not have much time to read for pleasure, but before a long flight I finally managed to pick up a copy of The Emperor of All Maladies, Siddartha Mukherjee's Pulitzer-prize winning "biography of cancer". The book is masterfully written and absolutely mesmerizing, but it is not my goal to review it here, as more poignant reviews were penned 3 years ago when it was first published.

As a scientist one of the most fascinating things is to follow the process of discovery: see what physicians and scientists inferred from what they saw and how they came to often very wrong conclusions with the very best intentions, how ideas became fixed in the common knowledge and how revolutionary approaches came to be. Like the organisms and molecules it studies, medical science is constantly evolving and ways for looking at scientific problems with new eyes are always necessary.

One thing that struck me of the evolution of the thinking about cancer, is the fact that at first we did not know that the same name encompasses multiple different disorders and the common approach was to find a cure-all drug that would work on everyone. After much often devastating trial and error, we now known specific drug regimens may only work on a specific form of cancer. Cancer is in fact becoming the poster-child for personalized medicine, an approach requiring a specific understanding of the individual patient's and tumor's history and genetic makeup. The rationale is that each type of cancer and maybe each individual cancer is different and the hope is that once we figure out what makes each tumor explode in millions of proliferating cells, we may be able to design a tailored regimen to stop it.

Reading about those initial struggles on how to deal with a very heterogeneous disorder that yet often looks similar in many patients, made me thing about the changing concept of autism nowadays. The herculean genetic efforts of the past few years have determined that autism is not caused by common genetic changes, but it is more likely due to a single faulty gene in each case. The rub is that the mutated gene could be one of dozens or hundreds, and as many clinicians and parents may tell you, each affected child may need to be considered as a unique presentation of the disease. This confusion is reflected in the recent changes to diagnostic criteria listed in the latest version of the Diagnostic and Statistical Manual of Mental Disorders or DSM-5, which groups all autistic-like symptoms of different severity under the term Autism Spectrum Disorders (ASD). So, is it one big disease or hundreds of different ones that we cannot diagnose individually? Is there a miracle drug which will work on everyone in the spectrum or do we need to figure out a way to group patients with defects in similar molecular processes, or even treat everyone differently?

Some food for thought came from the recent demise of arbaclofen, an experimental drug for autistic behavior in individuals affected by Fragile X Syndrome. Fragile X Syndrome is a common cause of intellectual disability and is often associated with autistic features, but as in all forms of ASD the symptoms and the behavioral problems are greatly variable. As described in a New York Times article last month, arbaclofen was deemed a failure after it showed no affect on a phase 3 clinical trial and funding was pulled. Yet a small number of patients showed remarkable improvement in their social interactions and parents are desperate that they will now lose access to the drug. It is possible that arbaclofen may only be effective in a specific subset of cases and that other patients may need different drugs. It is also possible that the children who responded well to arbaclofen may have improved on their own or responded to anything, but we will never know if we do not test these conflicting hypotheses. If there is a "type" of autism that responds well to a specific drug, it may possible to discover more types and their treatment, as is being done for cancer.