Researchers versus biostatistics

I can't see the video as it's blocked for New Zealand viewers, but judging from your text I know what you mean about people asking for data analysis assistance too late!

One of my pet gripes is people approaching âusâ (computational biologists / bioinformatics scientists, in my case) after the data analysis has been done, rather than while the grant application is being written.

Try here I think this is the same video but I'm in Australia so I can't see it either.

Blocked in Canada too, but Tort's link worked.

There seems to be a problem with imbedded youtube links. Everywhere that I looked that had one, it is blank. Though it is okay if you link to the original youtube page:
http://www.youtube.com/watch?v=PbODigCZqL8

Also blocked in England.

Oh, my! That was about the funniest cartoon I've seen in a long time (excepting certain Bugs Bunny cartoons)!

I'm not a biostatistician, but I work with them often enough to have heard that exact story many times. Even worse is when the "researcher" comes to the statisticians after the study is finished and asks them to find statistical significance in random noise - something like "find a publishable result in my worthless data", but not so honest.

I think something also needs to be said about "wannabe" statisticians (e.g. deSoto and Hitlan) who think they understand statistics (and may even teach statistics), but really only have a superficial understanding of how to type the numbers into SPSS. As I said above, I'm not a statistician, but I had to take statistics (from the Mathematics Dept.) in my graduate studies and was required to derive the equations used. There's nothing like seeing the mathematical underpinnings of - for instance - the Student's t-test to see what it can and cannot do.

I'll be sure to pass this on to my long-suffering biostatistician friends.

Prometheus

As a wanna-be statistician, I'm a great fan of this quote from Ronald Fisher:

To call in the statistician after the experiment is done may be no more than asking him to perform a post-mortem examination: he may be able to say what the experiment died of.

The cartoon experiment hasn't been done yet, but seeing as how the grant goes in tomorrow, it might as well have been done already.

There are several monitors that need cleaning in gene expression labs at the University of Michigan.

Why can't she just tell me if three iPhones is enough GBs?

About 10 years ago district biologists and technicians had radiotagged quite a few moose, took hair and blood samples, sexed, weighed them, and then tracked them for two years. At the end of that, they came over to the science and information branch, told us what they'd done and asked, "What ideas would you like us to test?". Apparently they didn't have any goals or ideas when they started the project...they just tagged moose and figured they'd come up with something to research afterward.

If you have to ask repeatedly if three people is enough of a sample size for a medical study, and have to be told repeatedly it is not, you have a bright future in homeopathy.

@Drose, well, it is true that under very certain circumstances, n=3 may be sufficient for a study. But that depends on what is being studied, and requires that it be such a rare or dangerous occurrence that asking for a greater n would be exceedingly difficult.

I'd have told the researcher in question to punch it into C-Power at 5% alpha. Then I'd have kicked him in the crotch.

Either way, the lot of us laughed at it over here. Y'all owe me a new keyboard here.

Blocked in Finland as well...

@2 Tort, thanks! :-)

@4 Chris, actually that doesn't work for me either! (Even if you cut'n'paste the URL into a new tab.)

Now that I've seen the video, beyond the point that you need to talk early, my take is that this cartoon nicely highlights the need for people who understand both or all âsidesâ of a bi- or multi-disciplinary area. I often find myself having to gently ask people to back up a bit and tell about the problem in biological terms. Too often people try tell me what analysis that they want done, as if I'm unfamiliar with biology. They mean well, but I'm usually (but admittedly not always) better placed to map the biological questions into an analysis strategy. (I am both biologist and computational analyst, as it were; itâs also why I prefer âcomputational biologistâ to âbioinformatics scientistâ.)

@7 spudbeach: Great quote. It reminds me of being approached with the results of a small-scale gene expression study (many years ago before gene arrays took off) which had no replicates or controls...

Rats. Forgot to add: thanks for showing this, âOracâ!

My favorite videos of this series are "Orthopedics vs. Anesthesia" and "Peds ER vs. Orthopedics". Comedy gold.

That's great. I know a pharmacology professor who'll likely (if possible) make good use of that in lectures. Forwarded.

I teach biostatistics to masters-level students. I think this will be highly entertaining to embed on the class website. Thanks!