Archive for: May, 2015

Young Scientists Caught between Scylla and Charybdis

May 29 2015 Published by under Uncategorized

Scylla and Charybdis has always been one of my favorite expressions.  The original rock and a hard place.

I am writing a grant with a young colleague. She brings one set of strengths to the table and I another. We've been talking about this project for about a year, and finally came up with a mechanism, design and Spec Aims that will, shall we say, blow something out of the water. We hope.

She is a glorified postdoc in another lab. She would like PI status, but the med school where she is has something of a problem: you have to demonstrate that you can get funding and be a PI before they will let you do it. Hard to do. Her main mentor/advisor there has agreed that the joint project she and I are working on is hers (even though I am the PI) and that if we get scored she can have the status.

Please note: I am doing much of the heavy lifting for the project, and abundantly aware of the ethical pitfalls in this collaboration. She is also saavy enough to understand what PI status without a TT position means. Finally, she has family/personal limitations and I do believe she is doing an excellent balancing job.

Anyway, we had hammered out a good, but not quite finished set of SA's.  She was sending me one last version to tweak, before moving on to finish up the rest of the proposal. I looked at them and for a minute thought she had sent a much earlier version, because one of the two aims was entirely different, and not something that NIH would be the least bit interested in. There were two new paras, and the writing was awkward. My thought: WTF?

Rather than editing, I said "lets talk about this". We did and it scnturned out that she showed the Specific Aims to her mentor at new place (who is not NIH-science, but NSF-science). He hated them. Said the writing was dreadful. Said the SA's didn't make sense. So she rewrote along the lines of his suggestion. What a horrible place for her. We talked for about 3 minutes and she got it. One of the reasons, beyond my excitement about the specific project, that I want to work with this woman, is that she is very very good. My first response to her was to make sure she understood that I cared about the work and that I was giving this the best I could. For getting funded, I felt the most important thing was to be honest and explain that I thought her other mentor was mistaken here, and why. I said if the mentor wants to write out specific comments, of course we can take those into account. We are now back on track. The grant will go in for the June 21rst deadline.

 

5 responses so far

Ray Bradbury quotes for the morning

May 28 2015 Published by under Uncategorized

This quote about writing could also be about science. Its not easy. There are lots of caveats, bumps and compromises. But...
Love. Fall in love and stay in love. Write only what you love, and love what you write. The key word is love. You have to get up in the morning and write something you love, something to live for.
This pairs nicely with another Bradbury quote:
“The first thing you learn in life is you're a fool. The last thing you learn in life is you're the same fool. Sometimes I think I understand everything. Then I regain consciousness”
Ray Bradbury, Dandelion Wine

No responses yet

Research Assistant Professors

May 27 2015 Published by under Uncategorized

I have a brilliant marvelous wonderful postdoc (oh hai. I know you're reading this). This postdoc has life balancing /personal issues that are entirely appropriate. Issues that got me thinking. The best thing for the is PD is to get A Real Job. I am working on that. Meantime, I am going to try and get this postdoc an appointment as a "research asst prof" of which there are several in my department at almost-MRU. This title means independent (largely), means PI status for writing their own grants. It does not mean their own lab, unless they score R-level NIH or equivalent funding. It might, given funding, mean a small amount of department support for research, < $50K.

I have all sorts of various feelings about this. On one hand, it can be a Good Thing. If young person can't move, for whatever reasons, then this is a way to be more than the pawn/gofer for a "real" faculty. If jobs are scarce, its one way to stay in the business. Needless to say, this implies the right mentor, with the right attitudes.

But, I see a lot, even in my marvelous new department, where its a ploy to tie someone tighter to the faculty mentor. There aren't really any BSD's here, in part because its not a BSD-kinda place, but also because the department has hired carefully over the years (there are more R01s here, in the hands of Gen-Xers, thank you very much, than at my old MRU department). At almost-MRU its hard to attract trainees as we are not a BSD-MRU. So there is motivation for keeping someone good.

But my postdoc, as is true for others here, would undoubtedly be better served by getting their own position. I have said so to the postdoc. We have talked about it. We will keep talking (as is our habit, oh hai, I know you're still reading).

It is a sign of maturity to let go of people when it is best for them. It is a sign of maturity to know when it is best for them. When I bring this up to colleagues, I hear a lot of hmming, and hawing, and toing and froing. Fuck that. Figure out what's best for your people. You have to get up with you in the morning.

9 responses so far

John Muir quote for the morning

May 26 2015 Published by under Uncategorized

Some morning, one just feels. A bit sappy, but here I am:

John Muir

“The sun shines not on us but in us. The rivers flow not past, but through us. Thrilling, tingling, vibrating every fiber and cell of the substance of our bodies, making them glide and sing. The trees wave and the flowers bloom in our bodies as well as our souls, and every bird song, wind song, and tremendous storm song of the rocks in the heart of the mountains is our song, our very own, and sings our love.”

John Muir

No responses yet

P-Hacking

May 22 2015 Published by under Uncategorized

During the journal club discussions about p-hacking, we went through the list of six things that Head et al talked about in their paper. The discussion was great (side note: the journal club is a set of three labs that work on different systems, and ask different questions yet we use similar methods, and to the larger world, yeah, well, we do the same thing. The other two PI's are part of the reason I came to almost-MRU. The relationships among the three of us are very good, and sharing a large lab, with dedicated spaces to specific activities, has prompted collaborations amongst us. Also, our post-docs share an office and our grad students, too. It works).

Our first conclusion was that yes, p-hacking happens. Yes, we have all probably done something close to this, not on purpose, but as a function of analysis. The second large conclusion came after we went through the list of six sins, one at a time and discussed them in the context of the data we were collecting and analyzing. I should add that the three of us are all fairly well trained in data analysis, and are the go-to people in the department for stats. What came out of our discussion was that yes, these are wrong, but there are subtleties there, and that blanket statements seldom encompass all the issues.

I'd like to talk about the sins problems here, maybe not one at a time. Some of the perceptions/insights we came to, or were coming to, at the end of our meeting are worthy of further discussion.

So... to start the first issue Head et al raise: (I am going to try and put the citations in here, even if it makes the post a bit longer):

conducting analyses midway through experiments to decide whether to continue collecting data [15,16];

  • 15. Gadbury GL, Allison DB (2014) Inappropriate fiddling with statistical analyses to obtain a desirable p-value: Tests to detect its presence in published literature. PLoS ONE 7: e46363. doi: 10.1371/journal.pone.0046363 View Article PubMed/NCBI Google Scholar
  • 16. John LK, Loewenstein G, Prelec D (2012) Measuring the prevalence of questionable research practices with incentives for truth telling. Psychol Sci 23: 524–532. doi: 10.1177/0956797611430953. pmid:22508865 View Article PubMed/NCBI Google Scholar

Obviously, if one has made a decision, done the power analyses and determined the number of subjects/trials/experiments, then one should stick with it. Stopping early may give false significance. So why would anyone do this? We came up with two reasons, one of which they mention at the end of the paper:

Amazingly, some animal ethics boards even encourage or mandate the termination of research if a significant result is obtained during the study, which is a particularly egregious form of p-hacking (Anonymous reviewer, personal communication).

This is not just animal ethics boards, but IRBs that oversee human subject research. Possibly one of the most famous examples of this was the first aspirin RCT called The Physician's Health Study. In this case the effect size (the value of aspirin in preventing cardiovascular mortality) was so remarkable, that in fact the study was terminated three years early. The abstract, from The Annals of Epidemiology (1(5):395-405):

The Physicians' Health Study is a randomized, double-blind, placebo-controlled prevention trial of 22,071 US physicians, using a factorial design to evaluate the role of aspirin in the prevention of cardiovascular mortality and beta carotene in the reduction of cancer incidence. After approximately 5 years of follow-up, the aspirin component was terminated, 3 years ahead of schedule. Several factors were considered in the decision to terminate, including a cardiovascular mortality rate markedly lower than expected in both aspirin and placebo subjects, precluding the evaluation of the primary aspirin hypothesis, and a highly significant (P < .00001) and impressive (44%) reduction in the risk of first myocardial infarction in the aspirin group. Issues in the decision to terminate are described in this report.

The question here that arises is one of ethics. If a researcher can show that something is valuable is it ethical to withhold this information or delay it? The important point is "can show". If you have p-hacked, have you shown? In the aspirin study, the significance level and effect size were so remarkable that there was no question of having shown. But how do you know that your result is remarkable without doing something wrong in terms of data analysis? Just breaking the blinding (which had to be done to show the effect) is wrong. Maybe there is some more sophisticated reasoning on this of which I am unaware. But the rationale for the animal studies is the concept of "reduction" which is one of the hallmarks of ethical use of animals- use no more than you need.

From a trainee or PI's view this is a tremendously tempting sin. Saving resources, be they animals, research costs and time, let alone animal or human lives, is a powerful motivating factor. What we agreed in the discussion is that acute and incisive pre-planning is one way to avoid this problem. Preliminary studies (which are not then folded into the larger study) to estimate sample size, or more importantly identify problematic covariates and control for variation in response make it likely that one will get the appropriate final sample size.

There is another factor that prompts people to do analysis half-way through data collection: timing of abstract submission for scientific meetings. Abstracts are usually due anywhere from 4 to 12 months prior to the meeting. One needs to have Something Important to say in the abstract. One of the (clinical) organizations of which I am (reluctantly) a part requires p-values and statements about significance in the abstracts (and they review and select, and less than 50% end up presenting at the meeting). Someone mentioned that some societies only accept titles in advance (for program organization) and that abstracts are submitted right before the meeting. Since the abstracts are for the most part "published" in one form or another, and get cited, or constitute part of a trainee's portfolio, there is pressure to get something significant rather than just exploring data.

If we wish to reduce the p-hacking, that either goes on in our labs (inadvertently, of course) or that we perceive in people we advise, there is more culture change than just being aware of the problem.

6 responses so far

For the statistical sins we commit when we....

May 21 2015 Published by under Uncategorized

There is a marvelous paper in PLOS Biology about p-hacking:

P-hacking is changing the analysis to get a significant p-value. P-hacking is defined in the paper as "selective reporting" or "inflation bias". The paper defines it very well:

It [p-hacking] occurs when researchers try out several statistical analyses and/or data eligibility specifications and then selectively report those that produce significant results [1215]. [citations from original paper:

  • 12. Brodeur A, Le M, Sangnier M, Zylberberg Y (2012) Star Wars: The empirics strike back. Paris School of Economics Working Paper 2012. ttp://ssrn.com/abstract=2089580.
  • 13. Cumming G (2014) The new statistics: Why and how. Psychol Sci 25: 7–29. doi: 10.1177/0956797613504966. pmid:24220629 View Article PubMed/NCBI Google Scholar
  • 14. Simmons JP, Nelson LD, Simonsohn U (2011) False-positive psychology: Undisclosed flexibility in data collection and analysis allows presenting anything as significant. Psychol Sci 22: 1359–1366. doi: 10.1177/0956797611417632. pmid:22006061 View Article PubMed/NCBI Google Scholar
  • 15. Gadbury GL, Allison DB (2014) Inappropriate fiddling with statistical analyses to obtain a desirable p-value: Tests to detect its presence in published literature. PLoS ONE 7: e46363. doi: 10.1371/journal.pone.0046363 View Article PubMed/NCBI Google Scholar

What I really really liked was the list of all the ways in which people commit sins of p-hacking. It reminds me of the prayers that get said at Rosh Hashanah and Yom Kippur  when one asks for forgiveness for the sins one has committed in the year. For the statistical sin I committed by

  • conducting analyses midway through experiments to decide whether to continue collecting data [15,16];
  • recording many response variables and deciding which to report postanalysis [16,17],
  • deciding whether to include or drop outliers postanalyses [16],
  • excluding, combining, or splitting treatment groups postanalysis [2],
  • including or excluding covariates postanalysis [14],
  • stopping data exploration if an analysis yields a significant p-value [18,19].

[see original paper for all the citations]

As someone originally trained in statistics and data analysis (before biology), I think the movement to reconsider how we use statistics and NHST (null hypothesis significance testing), as well as the lack of thinking about effect size are absolutely critical to Doing Good Science.

My lab group talked about these sins, er problems today during journal club. I think there's a lot to be said, but each deserves its own post.

 

One response so far

Musings on Alzheimer's Disease

May 21 2015 Published by under Uncategorized

Part of what is hard about Alzheimer's is the feeling that it is changing who one is. There was a long article this past Sunday in the NYTimes magazine about Sandy Bem, a psychologist with AD who decided to kill herself. She knew what was going on and had made the decision when she was cognitively intact. The story is powerful, in part because Bem was a powerful person, in part because of the story and in part because of the excellent writing of Robin Marantz Henig.  The story included swathes of her (relevant) life, including this passage which resonated with me:

As a parent in the 1970s, Sandy turned every interaction with her children into a political act. During story time, she would go through their picture books with a bottle of Wite-­Out and a Magic Marker, changing a hero’s name from male to female, revising plot lines, adding long hair or breasts to some of the drawings.

The story is not just about a woman making a hard decision. It is about this woman in the life she lived making a hard decision. Sandy Bem was a professor, and became a clinician late in life. She was a psychologist and knew about cognition and its changes over the lifespan. At times it seems the decision was easy, and then Henig gives us the consequences to family as Bem's disease progresses. One of the quotes from the article that  I found compelling concerned her ex-husband, who became one of her strongest sources of support as she got worse:

“If some devil had asked whether I would be willing to buy Daryl’s deeper self at the cost of my developing dementia,” she wrote, “I would say NO without hesitation. But if it comes free with my unstoppable decline into hell, I’m thankful for the gift.”

This quote is really about Bem, and her acceptance of the changes the disease brought to her. Nobody wants AD, but she was willing to look it in the face and make decisions about what happens next and how to navigate her changed life. Sandy Bem lived a powerful life, making hard choices <cue up quote about making hard choices from Anne Bancroft  in G.I. Jane. As my friend Maye says: there is a quote for every life situation from GI Jane>. The article, to me, was about living (and ending) life on your own terms, by your own choice. That is such a hard thing to do. It is so easy to find blame, but I found not a drop of self-pity in Bem. <cue up 2nd GI Jane quote about self-pity>.

Sandy Bem falls between me and my mother in age. My mother is end stage Alzheimer's and is well beyond the place where Sandy Bem decided to end her life. My mother never could have made the decision that Bem did. She was in absolute denial about things going wrong. My mother was smart and articulate and spent years pretending nothing was wrong. She never would or could have chosen to commit suicide. When intact she would have kept hoping for a cure until it was too late to make the choice that Bem did. And even now, with very little cognitive ability left and her dignity rapidly vanishing, I am not sure that my mother-then would want my mother-now to chose death.

Part of what is hard about Alzheimer's is the feeling that it is changing who one is. I write "my mother-then" with unease. One of my sibs has totally abandoned my mother. He has not seen her in years, does not communicate with me to find out how she is, what she needs, what she was like this week. To him, his mother is dead and gone and there is someone else in her place.

I try to think of this as a continuum. Just as any of us is not the person we were when were 2 let alone 12, my mother is not the person she was at 70 or even 80 when she could have made the choices that Sandy Bem did. There is still a person inside the husk that bears my mother's name. But I am uneasy. I do not know how you judge what this person in front of me wants. I am the one who is making the hard decisions, not the person who the decisions impact. I will try to chose, and to feel no self-pity as I watch my mother disintegrate.

 

5 responses so far

Greaseless Bearings, Popular Mechanics and My Nerd Self

May 20 2015 Published by under Uncategorized

I fully admit that I loved working on cars when I was younger. I cannot explain what pleasure it gave me, like listening to Bach. It was seeing something well-ordered and logical. I still read Popular Mechanics from time to time. Besides, just watching machines is watching beauty. I loved going to factories and seeing things being made. Back when you could get into the real manufacturing facility at Hershey's (not the theme park it is today), when you saw real machines wrapping chocolate kisses, now that was heaven for me.

But today's thrill is an article on greaseless ball bearings. I know this is not top of the queue for most people. It's still worth remembering we all use lots of stuff that we don't know how it works. Cars. Vacuum cleaners. Washing Machines. Mechanical things that make life easier for us. For a while I had a partner who worked in the steel industry. I knew more about the steel industry than I ever wanted. One of the things I learned is that just the making of steel which is there in a million ways in our lives, is complex. It involves gears and bearings and all sorts of other things I barely understood then, and still don't really understand now.

So, in tribute to that beauty, I give you: greaseless bearings. Ball bearings in general reduce friction when two rotating parts move BallBearingagainst each other. Here's a graphic from Wikipedia. If the inner ring just rotated against the outer ring, there would be a lot of friction. Because the balls themselves are rotating, there is much less friction. The first patent on roller bearings was awarded to one Philip Vaughan in 1794. These balls are usually greased up or lubed to further reduce friction and also extent the life of the elements in the system. Designing the bearings, and choosing the lubricant are critical steps in many machines. Getting it right (grease or oil for example) can make a significant difference in how well it operates and the lifetime of the bearings as well as the bearing races or cages (the outside rings). Sometimes the bearings are put into a "cage" to keep them from hitting each other, as such:

gallery-1431970116-ball-bearing-plain

The new ball bearings are based on putting small divots into the tracks the balls roll in, it makes them roll at very specific and changing speeds so that they don't touch and there is much less friction.

My geeky heart leaps for joy.

Coda: I hugely support things like Girl Geek Con (and you should too follow them at @GeekGirlCon ) for all the girls out there who need something like this.

 

 

 

 

 

No responses yet

Changing one's name at marriage

May 12 2015 Published by under Uncategorized

I have always been mystified why hetero-women want to change their name to their husband's when they get married. Within lesbian communities this has historically been less of a problem, if for no other reason than historically lesbians couldn't get married.

I had a hetero-friend in grad school (actually had more than one friend in grad school, and most were hetero, but that's another story), who had married right after college, and published her first papers then. She got divorced (ugly) during grad school, and published another few. Then after grad school got married again, and didn't change her name, despite being rather conservative about such things.  "Two different names on my CV is sufficient, I don't need a third. No one needs to know my whole marital history".

I know when anyone gets married (or anyone I know, which is indeed, a limited set), it is forever. Margaret Mead thought so, each of the four times she got married. I understand forever. Or the hope of forever. Been there, done that, wore out the t-shirt. I wish everyone to have forever. and ever. It's foolish to talk about statistics, because you and me and our friends are not those numbers. We are, of course, different.

Many of my friends from college/grad school didn't change their names. At least the ones who got married. I had a bunch of friends who didn't get married, and some of those changed their entire names, but that is also a different story. But I noticed that after a while more students I had seemed to be changing their names at marriage. Some of those marriages lasted (and have wonderful children who are getting married, and changing their names). Some didn't. I still believe in forever.

But apropos a discussion on the twits this morning about being called Dr. or Ms. or Mr. or Mrs.:

  Changing your name, similar to being Mrs. or Ms. or Miss, is a particular problem to women.  Everyone has the right to determine what they are called, name and title. But no matter what you do choose, you are making a statement about something.

8 responses so far

Jackie Robinson

May 11 2015 Published by under Uncategorized

When I grew up, Jackie Robinson was one of my mother's heroes (Joe DiMaggio was another).

jackie robinson

So when I found an article on the news site OZY about him as a businessman, I was interested.

It has an interesting quote:

Indeed, Robinson joined Chock Full o’Nuts on the condition that he would be more than a figurehead used to goose sales. (With all due respect to Jay Z, he saw greater power in being a businessman than in being a business, man.)

I'm not sure I see it the same way. It's akin to the question: do you want to do science, or be a scientist? What is it that you want. Figuring that out is the hardest thing.
Robinson wasn't a radical, but he radically changed all sorts of things along the way. I know lots of times we tend to think "I am being a radical" when we are in the midst of helping things change. But what constitutes effort towards change, change that lasts, is only something that gets judged in retrospect.

 

No responses yet

Older posts »