Slides used in the Workshop “How Bayesian Statistics Tell Us What We Want to Know”, by J.P. de Ruiter at the 2019 Conference of the Society for Text & Discourse, in New York City.

To say that our dominant statistical paradigm, “Null Hypothesis
Significance Testing” (NHST), is confusing is an understatement. It has
been shown that it often befuddles even experts. In this workshop, I
will explain the underlying logic of NHST, and why it is so confusing. I
will then introduce an alternative approach, the Bayesian framework,
which is more consistent, easier to interpret, and above all, answers
the questions that empirical scientists (even those who use NHST) really
*want* to ask.

Find slides here

Zoltan Dienes

The Bayes Factor

Zoltan Dienes

00:00 / 47:11

In this episode JP and Alex interview Zoltan Dienes. They discuss Zoltan's passion for the martial arts, 
why Bayesian inference could be more Popperian than you might think, 
and the easiest way to start using Bayesian statistics in practice.

- Follow us on Twitter: https://twitter.com/TheBayesFactor
- Follow us on Facebook: https://www.facebook.com/TheBayesFactor/
- Subscribe & leave us a review on iTunes: https://itunes.apple.com/us/podcast/the-bayes-factor/id1308207723

Notes and links:

- Zoltan’s book: https://www.amazon.com/Understanding-Psychology-Science-Introduction-Statistical/dp/023054231X
- Zoltan’s Bayes references and Bayes factor calculator: http://www.lifesci.sussex.ac.uk/home/Zoltan_Dienes/inference/Bayes.htm
- Four reasons to prefer Bayesian over significance testing: https://link.springer.com/article/10.3758/s13423-017-1266-z

Human Interaction Lab Data Session: April 25th, 2019, 2pm-4pm

The Human Interaction Laboratory Data session will meet on Thursday, March 28th 2019 in unit 2614, 200 Boston Ave. Medford, MA 02155.

If you don’t know what a data session is, you can read Saul’s description of learning about Conversation Analytic data sessions, or have a look at this resource from Arizona State University that discusses language and social interaction data analysis sessions from a variety of methodological/disciplinary perspectives: https://www.learninghowtolookandlisten.com.

See our full list of data sessions on our events page.

RSVP: julia.mertens@tufts.edu

Data: Dr. Kristen Bottema-Beutel and Dr. Gabrielle Oliviera from Boston College will be presenting data on Question-Response-Evaluation (QRE) sequences in the home interactions of a bilingual child with Autism Spectrum Disorder. We will focus on QRE sequences within interactions between a 5-year old bilingual child with ASD and his parents. In these sequences, the adult poses known-answer questions, the child responds, and the adult evaluates the response. QRE sequences are primarily structured by the questioner (i.e., the parents in our context), so we will focus on the interactive work done by parents to initiate, maintain, and close these sequences.

Keep in touch:

If you’d like information about upcoming data sessions and announcements at the Human Interaction Lab, please subscribe to our mailing list.

Human Interaction Lab Data Session: March 28th, 2019

The Human Interaction Laboratory Data session will meet on Thursday, March 28th 2019 in unit 2580, 200 Boston Ave. Medford, MA 02155.

If you don’t know what a data session is, you can read Saul’s description of learning about Conversation Analytic data sessions, or have a look at this new resource from Arizona State University that discusses language and social interaction data analysis sessions from a variety of methodological/disciplinary perspectives: https://www.learninghowtolookandlisten.com.

See our full list of data sessions on our events page.

RSVP: saul.albert@tufts.edu / 857-222-5992

Data: Our next data session will be lead by Dr. Chris Pudlinski. The data is from a collection of peer support warm lines calls. The current research interest is moments of extended silence within these calls. There are four different environments in which multiple longer silences occur: difficult/emotional topics; caller reluctance to share “bad” news; call taker as “active” listener; and call taker as “inattentive” listener. The audio data is of “inattentive” listening. This is the most problematic of these environments, one in which extended conspicuous lapses occur where call taker response is clearly expected.

Keep in touch:

If you’d like information about upcoming data sessions and announcements at the Human Interaction Lab, please subscribe to our mailing list.

Redefine or Justify? Comments on the Alpha debate

Fill in this form to receive a download link:

Reference: Redefine or Justify? Comments on the Alpha debate.

Abstract:

Benjamin et al. (Nature Human Behaviour 2, 6-10, 2017) proposed improving the reproducibility of findings in psychological research by lowering the alpha level of our conventional null hypothesis significance tests from .05 to .005, because findings with p-values close to .05 represent insufficient empirical evidence. They argued that findings with a p-value between 0.005 and 0.05 should still be published, but not called Bsignificant^ anymore. This proposal was criticized and rejected in a response by Lakens et al. (Nature Human Behavior 2, 168-171, 2018), who argued that instead of lowering the traditional alpha threshold to .005, we should stop using the term Bstatistically significant,^ and require researchers to determine and justify their alpha levels before they collect data. In this contribution, I argue that the arguments presented by Lakens et al. against the proposal by Benjamin et al. are not convincing. Thus, given that it is highly unlikely that our field will abandon the NHST paradigm any time soon, lowering our alpha level to .005 is at this moment the best way to combat the replication crisis in psychology.

Acknowledgments:

The author wishes to thank Alexander Etz, Jason Noble, and Eric-Jan Wagenmakers for their helpful comments on earlier versions of this paper.

Human Interaction Lab Data Session: Thursday 31st January 2019

The Human Interaction Laboratory Data Session will meet on Thursday, January 31st 2019 from 3 – 5 PM in unit 2580, 200 Boston Ave. Medford, MA 02155.

See our full list of data sessions on our events page.

RSVP: julia.mertens@tufts.edu

Data: In this session, Bryanna Hebenstreit (PhD Candidate at the University at Albany, SUNY) will present data set at a Habitat for Humanity build in Upstate NY. We will be analyzing a short clip showing how a volunteer gets the floor and demonstrates a technique that is presented as new for the onlookers. His demonstration is specifically targeted for the supervisor but ends up becoming a resource for all onlookers. We’ll explore how interactants ‘take over’ the main interaction and build their own talk, gestures, and embodied comportments to showcase an action to be seen by a particular interactant. Data in American English.

Keep in touch:

If you’d like information about upcoming data sessions and announcements at the Human Interaction Lab, please subscribe to our mailing list.

JP and Alex 2.0

The Bayes Factor

JP and Alex 2.0

00:00
/

55:20

In this episode JP and Alex catch up with each other and talk about how they like podcasting so far, Alex’s experience doing a joint MS/Phd in stats and cognitive science, JP’s new paper responding to the p<.005 debate, and their future plans for the podcast.

Follow us on Twitter: https://twitter.com/TheBayesFactor
Follow us on Facebook: https://www.facebook.com/TheBayesFactor/
Subscribe & leave us a review on iTunes: https://itunes.apple.com/us/podcast/the-bayes-factor/id1308207723

Notes and links:

JP’s entry into the .005 debate: https://link.springer.com/article/10.3758/s13423-018-1523-9 (preprint: https://psyarxiv.com/rbm8y)
Notes on Kullback-Leibler divergence: https://psyarxiv.com/5vhzu/

Halloween Interaction Lab Data Session: Wednesday 31st October 2018

The Human Interaction Laboratory Datasession will meet on Wednesday 31stOctober 2018 in unit 2580, 200 Boston Ave. Medford, MA 02155.

Data: This special Halloween data session will involve data from Julia Mertens’ collection of Trump/Stern interviews on the Howard Stern radio show between the 90s and the late 2010s. We’re also open to any Trump Whitehouse data you might have transcribed so feel free to bring your own. There might be candy.

See our full list of data sessions on our events page.

RSVP: saul.albert@tufts.edu / 857-222-5992

Keep in touch:

If you’d like information about upcoming data sessions and announcements at the Human Interaction Lab, please subscribe to our mailing list.

Human Interaction Lab Data Session: Wednesday 26th September 2018

The Human Interaction Laboratory Datasession will meet on Wednesday 26th September 2018 in unit 2580, 200 Boston Ave. Medford, MA 02155.

See our full list of data sessions on our events page.

RSVP: saul.albert@tufts.edu / 857-222-5992

Data: Bryanna Hebenstreit (PhD Candidate at the University at Albany, SUNY) will provide video data of an instance where a participant becomes familiar with a floorboard cutting tool on a home refurbishment site. This data provides a look at how participants arrange themselves in relationship to the tool, how participants project their usage of the tool, how other participants are brought into the interaction, and how ‘experienced’ onlookers may anticipate by the embodied conduct of a participant potential consequences of that arrangement and exclaim or otherwise attempt to avert undesirable consequences. Data is in American English.

Keep in touch:

If you’d like information about upcoming data sessions and announcements at the Human Interaction Lab, please subscribe to our mailing list.

The robot bartender will take your order now

Here’s a fun two minute interview on @sciam with Sebastian Loth, who used our Ghost In the Machine methodology to study how humans use predictive and incremental language processing to anticipate customer requests.

https://www.scientificamerican.com/podcast/episode/robot-bartender-will-take-your-order/

The report is based on this paper:

Loth S, Jettka K, Giuliani M, Kopp S, de Ruiter JP (2018) Confidence in uncertainty: Error cost and commitment in early speech hypotheses. PLOS ONE 13(8): e0201516.https://doi.org/10.1371/journal.pone.0201516