Pre-Order Now: Introduction to the New Statistics

Can a statistics textbook change the world? Maybe yes! At least that’s the aspiration behind Introduction to the New Statistics: Estimation, Open Science and Beyond, a new textbook by Geoff Cumming and yours truly (978-1138825529)ITNS

How can a statistics textbook change the world? By teaching the estimation approach to data analysis–one that emphasizes confidence intervals, replication, and meta-analysis. The estimation approach is vast improvement over Null Hypothesis Significance Testing. Students find estimation easier to learn and it supports better inference. Hopefully, this book will be an important step in the ongoing battle to abolish p values.

Switching to estimation is not enough, though. Students of research also need to learn the new Open Science practices that are evolving to enhance research rigor: pre-registration, Open Data, and Open Materials. So the textbook is also the first to teach these essential practices from the start.

To sum it up, Geoff and I believe that this is a unique book in a large sea of statistics texts–one which we hope will inspire the next generation of researchers who will be raised from the start on good statistical and methodological practices. Perhaps the replication crisis will fade into the

The book is now available for pre-order on Amazon. It should be published by August, 2016–which is running just a bit late for fall adoptions. If you’d like a desk copy, shoot me an email or leave a comment.

Bibliography

Rest easy — organic food probably does not make you into a jerk…

My student Eileen Moery and I have a new paper out today in Social Psychology and Personality Science. It’s a replication paper that I’m quite proud of (10.1177/1948550616639649). It represents some evolution in how I’m supervising replication projects.

The new paper replicates a study purporting to show that being exposed to images of organic food produces a strong decrease in prosocial behavior and a strong up-tick in being morally judgmental (10.1177/1948550612447114). This is a potentially fascinating phenomenon–something like ‘moral licensing’, the ironic effect of good behavior fostering subsequent bad behavior.

The original paper caught fire and the media covered these findings extensively. Rush Limbaugh even crowed about them as evidence of liberal hypocrisy. I noticed the media coverage, and this is how the original study made it onto my ‘possible replication’ list. Eileen found it there, read the paper, and developed a fantastic honors project to put the initial study to the test.

For her project, Eileen contacted the original author to obtain the original materials. She planned and executed a large pre-registered replication attempt. She included a positive control (Retrospective Gambler’s task) so that if the main study ‘failed’ we would have a way to check if it was somehow her fault. She also devised a nice memory manipulation check to be sure that participants were attending to the study materials. She conducted the study and found little to no impact of organic food exposure on moral reasoning and little to no impact on prosocial behavior. She did find the expected outcome on the positive control, though–so sorry, doubters, this was not an example of researcher incompetence.

One of the things I don’t like about the current replication craze is the obsessive emphasis on sample size (this paper is not helping: (10.1177/0956797614567341)). Sure, it’s important to have good power to detect the effect of interest. But power is not the only reason a study can fail. And meta-analysis allows multiple low-power studies to be combined. So why be so darned focused on the informativeness of a single study? The key, it seems to me, is not to put all your eggs in one basket but rather to conduct a series of replications–trying different conditions, participant pools, etc. The pattern of effects across multiple smaller studies is, to my mind, far more informative than the effect found in a single but much larger study. I’m talking about you, verbal overshadowing (10.1177/1745691614545653)

Anyways, based on this philsophy, Eileen didn’t stop with 1 study. She conducted another larger study using Mechanical Turk. There are lots of legitimate concerns about MTurk, so we used the quality controls developed in Meg Cusack’s project (10.1371/journal.pone.0140806 )–screening out participants who don’t speak English natively, who take way too long or too short of a time to complete the study, etc. Despite all this care (and another successful positive control), Eileen still found that organic food produced about 0 change in moral judgments and prosocial behavior.

Still not finished, Eileen obtained permission to conduct her study at an organic food market in Oak Park. Her and I spent two very hot Saturday mornings measuring moral judgments in those arriving at or leaving from the market. We reasoned those leaving from had just bought organic food and should feel much more smug than those merely arriving or passing by. Yes, there are some problems of making this assumption–but again, it was the overall pattern across multiple studies we cared about. And the pattern was once again consistent but disappointing–only a very small difference in the expected direction.

Although Eileen and I were ready to call it quits at this point, our reviewers did not agree. They asked for one additional study with a regular participant pool. Eileen had graduated already, but I rolled up my sleeves and got it done. Fourth time, though, was not the charm–again there was little to no effect of organic food exposure.

With all that said and done, Eileen and I conducted a final meta-anlysis integrating our results. The journal would not actually allow us to report on the field study (too different!?), but across the other three studies we found that organic food exposure has little to no effect on moral judgments (d = 0.06, 95% CI [0.14, 0.26],N=377) and prosocial behavior (d=0.03, 95% CI [?0.17, 0.23],N=377).

So–what’s our major contribution to science? Well, I suppose we have now dispelled what in retrospect is a somewhat silly notion that organic food exposure could have a substantial impact on moral behavior. We are also contributing to the ongoing meta-science examining the reliability of our published research literature–it gives me no joy to say that this ongoing work is largely painting a relatively bleak picture. Finally, I hope that we have now gained enough experience with replication work to be (modestly) showing the way a bit. I hope the practices that are now becoming routine for my honors students (pre-registration, multiple studies, positive controls, careful quality controls, and synthesis through meta-analysis) will become routine in the rest of replication land. No, strike that–these are practices that should really be routine in psychology. Holding my breath.

Oh – an one other important thing about this paper–it was published in the same journal that published the original study. I think that’s exactly as it should be (journals should have to eat their own dog food). Obviously, though, this is exceptionally rare. I think it was quite daring for the journal to have published this replication, and I hope the good behavior of its editors are a model for others and a sign that things really are changing for the better.

Bibliography

Grant Awarded to study the mechanisms of sensitization maintenance and decay

Woot! The Slug Lab has just been awarded a 3-year R15 grant from NIH to study the transcriptional mechanisms of sensitization memory and decay. What does that mean? It means that Irina and I will continue to be working our a**’ off trying to understand what genes are activated as an animal stores a long term memory, and even more importantly, as a long-term memory is forgotten.

Here’s the screen grab from ERA commons:
grant*

Big thanks to our dedicated and amazing students, and to the incredibly supportive colleagues and administrators we have here at Dominican University. We’re looking forward to crushing it with this project.

*Technically, that’s not the actual notice of the award, but of our priority score from a peer review of our grant proposal by a panel of esteemed scientists in the field. We got the award letter via email last week.

What’s the best way to teach methods and statistics?

No easy answer, but I’m co-author on a new paper that has some hints (10.1177/0098628315573139). Well, actually, it’s just a summary of some assessment data my department has been collecting to help evaluate an integrated and inquiry-based approach to teaching methods and stats. We provide lots of opportunities for apprenticeship, with student completing independent correlational and experimental projects across a 2-semester sequence. Currently, we’re using Nolan & Heinzen as a stats text, but I’m hoping that we’ll be using Cumming & Calin-Jageman by 2017.

Pliske, R. M., Caldwell, T. L., Calin-jageman, R. J., & Taylor-ritzler, T. (2015). Demonstrating the Effectiveness of an Integrated and Intensive Research Methods and Statistics Course Sequence. doi:10.1177/0098628315573139

Bibliography

New Publication on ERIN, Educational Resources in Neuroscience

Just after I started at Dominican in 2007 I had the good fortune to team up with Richard Olivo to help in the development of ERIN, an online curated database of educational resources for neuroscience. It was a long process, but ERIN has now been online for the past three years serving up great resources to faculty prepping their activities for neuroscience courses. You can take a spin for yourself at http://erin.sfn.org/.

To cap off our efforts with ERIN, Richard took the leading in writing up an article for JUNE, the Journal of Neuroscience Education. You can find it online here: (26240519).

Bibliography

Qualtrics Tips – A HTML5 and JavaScript Word-Search Task

Here’s a second experimental task I wrote for use with online social psychology experiments. This one is a word search. Again, the code is a series of kludges cobbled together from examples I could findwith Google. But it works out pretty well as a task you can embed in a Qualtrics survey. You can define the grid and the word list as you’d like, and you can have Qualtrics pass parameters that specify which grid and words list to use for a specific participant. It looks nice, and I’ve had good success using this online.

This particular grid is a control grid is a sample I adopted from a book of word searches. It’s a really tough one due to the size of the grid. If you want to see what finding a word looks like without, you know, actually finding a word: “Tradition” starts in the 8th letter of the bottom row.

I have a couple of papers I’m working on that use this task. When either goes to press, I’ll post the reference here for citation. The first one is finally in press, (10.1371/journal.pone.0140806) with three student co-authors:

Cusack, M., Vezenkova, N., Gottschalk, C., & Calin-Jageman, R. J. (2015). Direct and Conceptual Replications of Burgmer & Englich (2012): Power May Have Little to No Effect on Motor Performance. PLOS ONE, 10(11), e0140806. doi:10.1371/journal.pone.0140806, http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0140806

Email me if you’d like the source code, or scrape it directly from this page. The code is a hot mess, but it should be enough to get you started.

Bibliography

Qualtrics Tips – A HTML5 and JavaScript Mirror-Tracing Task


Social relationships can confer power on some to control the fates of others. There is a large and growing body of social-psychology research examining the psychological effects of power, with findings documenting profound changes in cognition, motivation, moral reasoning, and more.

One intriguing finding is that having social power can actually improve motor skills. Specifically, Bugmer & Englich (2012) have shown that manipulating power substantially improves power at mini-golf and darts (10.1177/1948550612452014).

I’ve been working over the past two years to replicate this finding. While it was easy enough to replicate mini-golf with real, live participants, I wanted to test the study online, with a much larger and more diverse pool of participants. But how to measure motor skill online? My solution was to develop an online version of the classic mirror-tracing task, using HTML 5 and Javascript.

A paper describing the results is finally in published (10.1371/journal.pone.0140806) with three student co-authors:

Cusack, M., Vezenkova, N., Gottschalk, C., & Calin-Jageman, R. J. (2015). Direct and Conceptual Replications of Burgmer & Englich (2012): Power May Have Little to No Effect on Motor Performance. PLOS ONE, 10(11), e0140806. doi:10.1371/journal.pone.0140806, http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0140806

Below you can see a sample in action. The top box is the mirror–the bottom box is the drawing pad. Move the mouse into the green target in the drawing pad–this starts the trial. From this point on, you mouse will leave a mirrored trail in the mirror-box. Try to trace to the red target while staying in the line of the figure in the mirror. When you are within the line, the trail will be red–go outside the line and the trail is blue. You’ll see your score as a % of time within the line just below the drawing pad. The trial ends automatically when you reach the green target. Note that this WordPress theme uses a bunch of white-space—you may need to increase or decrease the zoom on your browser to get both boxes to show within your screen without scrolling

I kludged this task together even though I’ve never previously written javascript before. I relied heavily on examples found via Google. The final produce works, but could really use a code review both a) to clean up all the horrible on-the-fly decisions I made in getting this to work, and b) to credit the sources I used in taping this together.  Still, there are some cool features to this task:

  • The final drawing is saved back to a server as an image, complete with score–this allows you to visually inspect performance to get a feel for how different participants are scoring.
  • The scoring is fully automated.
  • The script runs fine within a Qualtrics survey, and Qualtrics can read the scores back from the script.
  • As expected, I found a consistent negative correlation between performance and age–this indicates the validity of the task and provides a useful covariate for reducing within-subjects noise.
  • As expected, performance improves from first to second-trial.
  • With this task, one can analyze overall performance and performance change across trials (though fatigue seems to set in within a few closely-spaced trials for most participants)
  • Difficulty can be varied by changing line thickness.
  • Any arbitrary line tracing can be uploaded–I matched all mine for number of turns and total line length.
  • It’s a fun task!

If you’d like to adopt this for your own studies, please shoot me an email and I’ll be happy to send you the code and some instructions for how to use it (or you could probably just rip it off directly from this page).  When (if) I get the paper using this task published, I’ll post the reference for citation.

Mirror Tracing Demo

Bibliography

2014×3 – Transcriptional correlates of long-term habituation

Third paper of the year for the lab (gasp!) is now out in Learning and Memory (10.1101/lm.036970.114).

The focus of the project is habituation, considered the simplest and most ancient form of memory. Long-term habituation requires changes in gene expression, but to date there is almost nothing known about what specific changes are required to encode and store a long-term habituation memory.

We’re not the first to try to tackle this issue, but it turns out to be a very difficult topic for study. Habituation is typically very site specific, occurring only at the site of training. This implies a relatively discrete set of neurons encode the memory, and that presents a real problem for qPCR and microarray analysis, because the signal from memory-encoding neurons could easily be washed out from signal from non-encoding neurons, glia, etc.

Our strategy was to develop a new, automated protocol for inducing long-term habituation over the entire body of an Aplysia. With the help of a tinker-toy set, a windshield-wiper motor, a relay box, an old computer with a parallel port, and some qBASIC programming (blast from the bast), we developed a slug car wash–an apparatus we could place over the tanks to repeatedly (though gently) brush Aplysia without any need for human intervention during training. We made a video to show off the system, which you can see here.

The slug car wash turns out to work great. We tracked the development of habituation over repeated rounds of training and saw a classic pattern of behavior–robust decreases in behavior at the end of each round of training, substantial overnight recovery (forgetting), but a progressive development of a persistently decreased response within 3 days of training. Importantly, we could observe habituated responding when stimulating the animal at the head, the siphon, or the tail. Moreover, the effect sizes were huge. So it was pretty clear that the slug car wash was producing the high impact we were looking for. In addition, we found that pattern of training really does matter–when training has breaks between sessions and is spaced out over 3 days it is extremely effective; massing all the same stimulation together into a single one-day session (at a slightly higher rate to squeeze it all in) produced neither long-term nor short-term habituation. This is a useful finding because it gave us an additional no-memory control, one which could ensure any molecular correlates identified are specific to memory formation, not just to the activity induced by brushing.

So what’s changing transcriptionally? We decided to focus on the pleural ganglia containing the VC nociceptors. These are relatively high-threshold neurons, and are probably not carrying the bulk of the activity induced by the brush. Unfortunately, though, no one yet knows *where* in the Aplysia nervous system to find the cell bodies of the low-threshold neurons that mediate light touch (probably in the periphery). Not to worry, though–we did record from the VCs in reduced preps and found that they do actually get some activation from the brush: about 1/4 fired APs, and most of the rest got lots of IPSPs from off-center stimulation.

To track transcriptional changes, we used the custom-designed microarray we recently developed in the lab (25117657). Some quick words about methods: We again used a large-ish sample size (n=8/group; can you believe that n=3/group is still common in microarray!?). We also used very high statistical standards by adopting the ‘treat’ function in limma which allows you to specify a reasonable null hypothesis (e.g. at least 10% regulation in either direction, rather than the standard practice of testing against a null of no regulation). Adopting a more reasonable null enables you to test for statistical and practical significance at the same time, and we’ve found that transcripts which pass such a rigorous test generalize very well to new samples. We’ve been finding R and limma surprisingly easy to use, which is pretty fantastic for free software.

Anyways, back to the data. The microarray results were a bit of a bummer. Out of over 20,000 transcripts tested, only *one* came up as strongly regulated. Bummer. Another 20 transcripts came up as regulated if you use a standard null hypothesis, but, as expected, none of these validated.

Although the microarray results were not what we hoped, we did further explore the one regulated transcript, and it turns out to be quite interesting. From sequence alignment, it seems to be an Aplysia homolog of cornichon, an auxiliary subunit for AMPA receptors. In invertebrates, cornichon seems to limit trafficking to AMPA receptors to the membrane and therefore reduces glugatmate-induced currents(24094107). Note that this is precisely the type of effect that could produce behavioral habituation. Moreover, one of the few known molecular correlates of long-term habituation is a decrease in surface expression of glutamate receptors (14573539). Fits perfectly!

To ensure that cornichon is truly regulated in our paradigm, we did some additional follow-ups. First, we used qPCR to check cornichon levels not only in the microarray samples but in an additional, independent set of samples. Sure enough, we confirmed up-regulation of cornichon in the pleural ganglia 1 day after training. In addition, we checked levels in massed animals, who display no memory after training. In this case, cornichon was actually slightly down, and was significantly different than in the regularly trained animals. So, cornichon is quite specifically and consistently up-regulated after long-term habituation training. As far as we know, this is the first specific transcriptional correlate of long-term habituation to be identified.

Needless to say, we’re quite proud of this work. It wouldn’t have been possible without two of the most talented undergrads we’ve had in the lab: Geraldine Holmes and Samantha (Sami) Herdegen. Geraldine was the most diligent slug trainer in the history of the lab. For this paper alone she ran over 48 animals, testing each 8 times a day for 3-5 days–that’s a whole lot of behavior to monitor! Sami, of course, has been the qPCR wizard in the lab, testing lots and lots and lots and lots of transcripts for regulation. It’s no surprise that both are on to bigger and better things, Geraldine is now in a PhD program in Canada and Sami is soon to start pharmacy school. We also had contributions from John Schuon (when he could fight his way in for some qPCR; now off to medical school), Ashly Cyriac (who helped start the project before heading off to pharmacy school), Jamie Lass and Catherine Conte. Congrats!

As has now become the norm for the lab, all the raw data from this study been posted online at the Open Science Framework: https://osf.io/6ew4i/.

Bibliography

Sluglab Strikes Again – New paper tracing dynamics of learning-induced changes in transcription

A nice way to wrap up 2014–we have a new paper out (25486125) where we trace learning-induced changes in transcription over time and over different location in the CNS. We think it’s a nice follow-up to the microarray paper, because:

  • We show that some transcriptional changes are likely occuring in interneurons and motor neurons, not just in the VC nociceptive sensory neurons.
  • We found some transcripts which, like Egr, are rapidly *and* persistently up-regulated by sensitization training (GlyT2, VPS36, and an uncharacterized protein known for now as LOC101862095). We’re interested in such transcripts because they could be related to memory maintenance
  • We were able to better test the notion that CREB supports memory maintenance. So far, our evidence continues to go against this hypothesis, with no long-lasting changes detected in the VC sensory neurons nor in the pedal ganglia.
  • As a methodological point, we found that microdissecting out the VC cluster really really improves signal:noise for identifying transcriptional changes induced by learning. This is exciting–most work on the molecular mechanisms of memory uses tissue samples representing homogenous cell types. Zooming in on a single cell type of known relevance for storing the memory really enhances the power of the analysis.
  • We re-rested the four novel transcripts identified in our microarray paper from earlier this year (25117657). All four validated again! Moreover, all 4 were specifically up-regulated in the VC nociceptors (and some elsewhere as well). Another good indication that we’re on the right track with our microarray approach.
  • Another 3 student co-authors on this paper! We’re especially proud of Sami, Catherine, and Saman.
  • The paper is free on PLOSE ONE: http://dx.plos.org/10.1371/journal.pone.0114481. Also, you can download our raw data to examine for yourself at the Open Science Framework: https://osf.io/ts9ea/.

    Bibliography