Discussion: View Thread

I may end up giving a talk to researchers about the statistical aspects of writing grants, and part of that would be how to find a good statistician, how to recognize when you have someone good, and how to best interact with him/her.

I have some ideas about this topic, but I'm sure that others have a lot to offer on this as well. Thanks in advance for your help.

-------------------------------------------
Stephen Simon
Independent Statistical Consultant
P. Mean Consulting
-------------------------------------------

I have to laugh about this one, as it comes up very often!  So here are my top 5 tips that I tend to give:


1.  Find your statistician at least one week before your submission date, instead of the night before.
2. Engage your statistician in all aspects of the grant, and not just a portion that you think they may be interested in.
3. Does your potential statistician talk more than they listen.  If so, walk away.
4. Does your potential statistician discuss money for their services first.  If so, walk away.
5. Has your statistician ever applied for a grant as PI before, or have been part of that process?

I have my top 100 of course as well, but I tend to stick to these 5 first!
Enjoy,
Nicole


-------------------------------------------
Nicole Close
President and Principal Biostatistican
EmpiriStat, Inc
www.empiristat.com
-------------------------------------------

Nicole's comments are in a humorous vein but make some serious points.  I will provides some suggestions but without attempting to be humorous.  Since this involves grant writing makes sure that the statistician has good writing skills.  Ask to look at various journal publications and find out if he or she has submitted grants and whether or not they have been successful and what they attribute their failures to.  In a highly competitive world for grant money we probably fail as much or more often than we succeed.  But it doesn't mean we made a bad attempt.  I think you should want somebody with a lot of real-world experience.  Maybe at least five years of post-doctoral research would be sufficient.  Being able to give good oral presentations is important too.  Nicole assumed that you were looking for someone who is a good professional consultant as well as a good statistician.  With consulting I do not take it as a red flag if the consultant  discusses money first or near the beginning.  The consultant is a professional and expects fair pay for his/her services.  Too often people will take a friendly approach trying to get free service.  So the smart professional consultant is rightfully on guard and you should not hold his interest in money against.  But of course you want to be sure that he/she is not just interested in the money.  The consultant should be very attentive to the clients problem and you want someone who is both a good listener but who also will ask a lot of good probing questions.  For grants name recognition is something to consider.  If you are going to NIH, NSF or other government funding agencies it would be nice to have someone with a good national or international reputation so that the reviewers will take your proposal seriously.  In most cases the reviewers will consider the proposal on its merits but psychologically a name they recognize will add to their confidence.
-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------







Original Message:
Sent: 09-26-2011 16:55
From: Nicole Close
Subject: What should clients look for in a statistical consultant?

I have to laugh about this one, as it comes up very often!  So here are my top 5 tips that I tend to give:


1.  Find your statistician at least one week before your submission date, instead of the night before.
2. Engage your statistician in all aspects of the grant, and not just a portion that you think they may be interested in.
3. Does your potential statistician talk more than they listen.  If so, walk away.
4. Does your potential statistician discuss money for their services first.  If so, walk away.
5. Has your statistician ever applied for a grant as PI before, or have been part of that process?

I have my top 100 of course as well, but I tend to stick to these 5 first!
Enjoy,
Nicole


-------------------------------------------
Nicole Close
President and Principal Biostatistican
EmpiriStat, Inc
www.empiristat.com
-------------------------------------------

Below is the URL for a 2010 piece in Chance titled "Developing Grant Proposals: Guidelines for Statisticians
Collaborating Under Limited Resources." It comes from:

Glickman, M., Ittenbach, R., Nick, T., O'Brien, R., Ratcliffe, S., and Shults, J. (2010).

Statistical consulting with limited resources: Applications to practice. Chance, 23:35-42.

http://hal.case.edu/~robrien/NickOBrienDevelGrantProposals_Chance10.pdf

Full article:
http://hal.case.edu/~robrien/Glickman10Statistical%20consulting%20with%20limited%20resources.pdf

-------------------------------------------
Ralph O'Brien
Case Western Reserve University
-------------------------------------------

I was involved in writing grants for a decade, while working at NDRI, and I've helped write some since then as well.

How to find a good statistician: My advice would just be word of mouth - ask colleagues who write grants similar to yours; or search the places statisticians hang out online (like here, for instance!)

How to know when you've found someone good: A good statistical consultant will ask you questions that you have a hard time answering and tell you things you don't like hearing. He or she will answer your questions and question your answers.

How to interact best: 1) Consult EARLY.  No, earlier than that!  When you have a glimmer of an idea of something you'd like to maybe submit a grant application for, find a stats person to talk to. 2) Involve the stats person in your proposal writing process. If he/she doesn't want to be involved, go find someone else. 3) Talk to your stats person about the SUBSTANCE of your grant, not the stat methods.  The stat methods are the consultant's forte. YOU know the substance (well, we hope you  do!).  4) Don't assume the stats person knows all about your subject. Discuss the substance as if you were telling someone what you do. 5) If you don't understand what your stats person is saying ASK.

-------------------------------------------
Peter Flom
-------------------------------------------

Of course, the flip side to this is How does a stat consultant decide whether to become involved in creating a new grant application? As per the classic 1984 essay by Lincoln Moses and Tom Louis, stat consulting is a two-way street. (http://hal.case.edu/~robrien/MosesLousiStatMed84TwoWayStreet.pdf)

One of the best statisticians I know is the VERY BEST I know at declining to be involved with projects that do not suit her. She can "Sorry, but I don't have the time needed to do this well. But thanks for asking." in a way that does not seem to anger people. In academic biostatistics, this is a most valuable skill.

Word of mouth here is valuable, too. Long ago, as the just-hired faculty member of a biostat group, I was assigned to work with a grant PI who had had run-ins with every other faculty member in the group. Thinking then that I could work with anyone, I dove right in. I hung in there for 3+ years but it became a nightmare.

Look hard before you leap. Learn to say "No, thanks" in just the right way.

If you are good at what you do, then you can fill your professional "dance card" with investigators and projects that make being a statistician rewarding and enjoyable.

-------------------------------------------
Ralph O'Brien
Case Western Reserve University
-------------------------------------------







Original Message:
Sent: 09-27-2011 06:25
From: Peter Flom
Subject: What should clients look for in a statistical consultant?


I was involved in writing grants for a decade, while working at NDRI, and I've helped write some since then as well.

How to find a good statistician: My advice would just be word of mouth - ask colleagues who write grants similar to yours; or search the places statisticians hang out online (like here, for instance!)

How to know when you've found someone good: A good statistical consultant will ask you questions that you have a hard time answering and tell you things you don't like hearing. He or she will answer your questions and question your answers.

How to interact best: 1) Consult EARLY.  No, earlier than that!  When you have a glimmer of an idea of something you'd like to maybe submit a grant application for, find a stats person to talk to. 2) Involve the stats person in your proposal writing process. If he/she doesn't want to be involved, go find someone else. 3) Talk to your stats person about the SUBSTANCE of your grant, not the stat methods.  The stat methods are the consultant's forte. YOU know the substance (well, we hope you  do!).  4) Don't assume the stats person knows all about your subject. Discuss the substance as if you were telling someone what you do. 5) If you don't understand what your stats person is saying ASK.

-------------------------------------------
Peter Flom
-------------------------------------------

The link doesn't work because Ralph made a typo.  This will work:
http://hal.case.edu/~robrien/MosesLouisStatMed84TwoWayStreet.pdf   It was just that "Lousi" should have been "Louis".
-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------







Original Message:
Sent: 09-27-2011 09:03
From: Ralph O'Brien
Subject: What should clients look for in a statistical consultant?

Of course, the flip side to this is How does a stat consultant decide whether to become involved in creating a new grant application? As per the classic 1984 essay by Lincoln Moses and Tom Louis, stat consulting is a two-way street. (http://hal.case.edu/~robrien/MosesLousiStatMed84TwoWayStreet.pdf)

One of the best statisticians I know is the VERY BEST I know at declining to be involved with projects that do not suit her. She can "Sorry, but I don't have the time needed to do this well. But thanks for asking." in a way that does not seem to anger people. In academic biostatistics, this is a most valuable skill.

Word of mouth here is valuable, too. Long ago, as the just-hired faculty member of a biostat group, I was assigned to work with a grant PI who had had run-ins with every other faculty member in the group. Thinking then that I could work with anyone, I dove right in. I hung in there for 3+ years but it became a nightmare.

Look hard before you leap. Learn to say "No, thanks" in just the right way.

If you are good at what you do, then you can fill your professional "dance card" with investigators and projects that make being a statistician rewarding and enjoyable.

-------------------------------------------
Ralph O'Brien
Case Western Reserve University
-------------------------------------------

For some strange reason if I use this link directly it doesn't work. But if I copy and paste it into my Internet Explorer search window it does.

-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------







Original Message:
Sent: 09-27-2011 10:11
From: Michael Chernick
Subject: What should clients look for in a statistical consultant?


The link doesn't work because Ralph made a typo.  This will work:
http://hal.case.edu/~robrien/MosesLouisStatMed84TwoWayStreet.pdf   It was just that "Lousi" should have been "Louis".
-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------







Original Message:
Sent: 09-27-2011 09:03
From: Ralph O'Brien
Subject: What should clients look for in a statistical consultant?

Of course, the flip side to this is How does a stat consultant decide whether to become involved in creating a new grant application? As per the classic 1984 essay by Lincoln Moses and Tom Louis, stat consulting is a two-way street. (http://hal.case.edu/~robrien/MosesLousiStatMed84TwoWayStreet.pdf)

One of the best statisticians I know is the VERY BEST I know at declining to be involved with projects that do not suit her. She can "Sorry, but I don't have the time needed to do this well. But thanks for asking." in a way that does not seem to anger people. In academic biostatistics, this is a most valuable skill.

Word of mouth here is valuable, too. Long ago, as the just-hired faculty member of a biostat group, I was assigned to work with a grant PI who had had run-ins with every other faculty member in the group. Thinking then that I could work with anyone, I dove right in. I hung in there for 3+ years but it became a nightmare.

Look hard before you leap. Learn to say "No, thanks" in just the right way.

If you are good at what you do, then you can fill your professional "dance card" with investigators and projects that make being a statistician rewarding and enjoyable.

-------------------------------------------
Ralph O'Brien
Case Western Reserve University
-------------------------------------------

I did not type anything, just copy and pasted right from the site that holds it. The problematic "point-and-click" URL must have supplied by this ASA system.

-------------------------------------------
Ralph O'Brien
Case Western Reserve University
-------------------------------------------







Original Message:
Sent: 09-27-2011 10:26
From: Michael Chernick
Subject: What should clients look for in a statistical consultant?

For some strange reason if I use this link directly it doesn't work. But if I copy and paste it into my Internet Explorer search window it does.

-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------







Original Message:
Sent: 09-27-2011 10:11
From: Michael Chernick
Subject: What should clients look for in a statistical consultant?


The link doesn't work because Ralph made a typo.  This will work:
http://hal.case.edu/~robrien/MosesLouisStatMed84TwoWayStreet.pdf   It was just that "Lousi" should have been "Louis".
-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------

The typo wouldn't be a point and click issue.  Somebody made it.  But I guess Ralph's explanation would apply to my point and click of the edited version of the link.

-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------







Original Message:
Sent: 09-27-2011 11:28
From: Ralph O'Brien
Subject: What should clients look for in a statistical consultant?



I did not type anything, just copy and pasted right from the site that holds it. The problematic "point-and-click" URL must have supplied by this ASA system.

-------------------------------------------
Ralph O'Brien
Case Western Reserve University
-------------------------------------------







Original Message:
Sent: 09-27-2011 10:26
From: Michael Chernick
Subject: What should clients look for in a statistical consultant?

For some strange reason if I use this link directly it doesn't work. But if I copy and paste it into my Internet Explorer search window it does.

-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------







Original Message:
Sent: 09-27-2011 10:11
From: Michael Chernick
Subject: What should clients look for in a statistical consultant?


The link doesn't work because Ralph made a typo.  This will work:
http://hal.case.edu/~robrien/MosesLouisStatMed84TwoWayStreet.pdf   It was just that "Lousi" should have been "Louis".
-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------

In my career as a clinical trials statistician I have only helped a few times in writing a grant proposal.  I never enjoyed it.  I think I did it mostly for some friends at the local university.  However, I have worked on writing many protocols within the pharmaceutical industry and I've noticed that your points are also valid here.  These points hold for grants, protocols, study design and sample size estimates, SAPs, helping write the final report, and helping with abstracts and papers.  I think these are all values good statisticians need to have to be successful.

The only difference I see is that perhaps gaining experience in grant proposals would be nice so the statistician has knowledge as to what the reviewers of a grant proposal arelooking for. 

-------------------------------------------
Rocco Brunelle
Senior Statistician
Bowsher Brunelle Smith LLC
-------------------------------------------







Original Message:
Sent: 09-27-2011 06:25
From: Peter Flom
Subject: What should clients look for in a statistical consultant?


I was involved in writing grants for a decade, while working at NDRI, and I've helped write some since then as well.

How to find a good statistician: My advice would just be word of mouth - ask colleagues who write grants similar to yours; or search the places statisticians hang out online (like here, for instance!)

How to know when you've found someone good: A good statistical consultant will ask you questions that you have a hard time answering and tell you things you don't like hearing. He or she will answer your questions and question your answers.

How to interact best: 1) Consult EARLY.  No, earlier than that!  When you have a glimmer of an idea of something you'd like to maybe submit a grant application for, find a stats person to talk to. 2) Involve the stats person in your proposal writing process. If he/she doesn't want to be involved, go find someone else. 3) Talk to your stats person about the SUBSTANCE of your grant, not the stat methods.  The stat methods are the consultant's forte. YOU know the substance (well, we hope you  do!).  4) Don't assume the stats person knows all about your subject. Discuss the substance as if you were telling someone what you do. 5) If you don't understand what your stats person is saying ASK.

-------------------------------------------
Peter Flom
-------------------------------------------

I liked Ralph's story and will go to his site to reader Lincoln Moses' article.  Lincoln was a wise gentleman with with the look of a Santa Clause when i knew him both as a graduate student at Stanford and later when he headed up the EIA.  To Rocco I would say the big difference between a invitation to a grant application and writng protocols and statistical analysis plans is that the latter is your job.  You can't turn it down.  What you need to do is the best job you can and be assertive in your position to keep the statistical approach sound.  Your employer knows that if they don't follow your advice they may have to face the wrath of the FDA down the road.  So that is your trump card.  I have always enjoyed the job of a clinical trials biostatistician.  Sometimes the FDA is there to agree with me and other times to frustrate me.

-------------------------------------------
Michael Chernick
Director of Biostatistical Services
Lankenau Institute for Medical Research
-------------------------------------------







Original Message:
Sent: 09-27-2011 09:05
From: Rocco Brunelle
Subject: What should clients look for in a statistical consultant?


In my career as a clinical trials statistician I have only helped a few times in writing a grant proposal.  I never enjoyed it.  I think I did it mostly for some friends at the local university.  However, I have worked on writing many protocols within the pharmaceutical industry and I've noticed that your points are also valid here.  These points hold for grants, protocols, study design and sample size estimates, SAPs, helping write the final report, and helping with abstracts and papers.  I think these are all values good statisticians need to have to be successful.

The only difference I see is that perhaps gaining experience in grant proposals would be nice so the statistician has knowledge as to what the reviewers of a grant proposal arelooking for. 

-------------------------------------------
Rocco Brunelle
Senior Statistician
Bowsher Brunelle Smith LLC
-------------------------------------------

Here are some "bewares" to add to Nicole's list:

1.  Beware of a statistician who does not ask questions that make you think.
2.  Beware of a statistician who simply does what you tell him to do.

-------------------------------------------
Emil M Friedman
Principal Scientist/Statistician at MannKind Corporation
Part-time independent consultant
The views expressed on this Web site are mine alone and do not necessarily reflect the views of my employer.