Workshop "Statistical standards for scientific discovery in linguistics: a practical introduction"

October 4-6, 2017, University of Zurich

This three-day workshop will focus on the discussion of the three main causes of the “replicability crisis” that social sciences are currently undergoing: 1) observations and experiments are carried out on samples that are too small to properly observe the investigated effects, 2) the current standard for claiming significant discoveries is based on p-values, which pose several problems (e.g. they do not indicate the size of the observed effects, they can suffer important alterations depending on small modifications of the data, eventually leading to “p-hacking”, and they are poorly understood by many researchers), and 3) strong pressure for publishing only significant results has lead to a publication bias caused by the reluctance to publish negative results, which precludes a great amount of relevant data highly needed to reject type I errors (i.e. false positives) from getting published.

Confirmed invited speakers

Harald Baayen (University of Tübingen)

Regina Nuzzo (Stanford University)

Joaquín Ordieres (Polytechnical University of Madrid)

David  Colquhoun (University College London)

Round table

Balthasar Bickel (University of Zurich)

Thoralf Mildenberger (Zurich University of Applied Sciences)

Maarloes Maathuis (ETH)

Tanja Samardžić (University of Zurich)

 

The workshop is especially aimed at doctoral and postdoctoral researchers. It has a strong practical focus, combining plenary conferences with short presentations by participants as well as practical sessions accompanied by the experts.

Funding by the UZH Graduate Campus via a GRC Grant is gratefully acknowledged.

Programme

 

Wednesday October 4th

(KAB-E-05)

Thursday October 5th

(KAB-E-05)

Friday October 6th

(RAA-E-27/RAA-E-08)

9:30-10:15

Introduction

Carlota de Benito Moreno

Danae Pérez

Albert Wall

Plenary lecture

Does data sampling matter at Big data scale?

Joaquín Ordieres Meré

(Polytechnical University of Madrid)

Plenary lecture

P-values provide poor evidence: some proposals for improving reproducibility

David Colquhoun

(University College London)

10:15-11:00

Spatial statistics

Curdin Derungs

(University of Zurich)

11:00-11:30

Coffee break

Coffee break

Coffee break

11:30-13:00

Plenary lecture

How not to fool yourself with statistics: Understanding and communicating p-values

Regina Nuzzo

(Stanford University)

Presentations by participants

Axel Bohmann

Haim Dubossarsky

Katharine Dziuk

Thoralf Mildenberger

 

Presentations by participants

Ana Estrada

Jerzy Gaszewski

Dominique Hess & Tobias Leonhardt

Sandra Schwab

Hilary S.Z. Wynne & Swetlana Schuster

13:00-14:30

Lunch

Lunch

Lunch

14:30-16:00

Plenary lecture

Regression modeling strategies for confirmatory and exploratory data analysis in the language sciences

Harald Baayen

(University of Tübingen)

Practical session 1

Practical session 3

16:00-16:30

Coffee break

Coffee break

Coffee break

16:30-18:00

Presentations by participants

Abdulhameed Aldurayheem

Nathalie Dherbey Chapuis

Andreia Karnopp

Hanna Ruch

Kyoko Sugisaki

Practical session 2

Round table

Balthasar Bickel (University of Zurich)

Thoralf Mildenberger (Zurich University of Applied Sciences)

Maarloes Maathuis (ETH)

Tanja Samardžić (University of Zurich)

Venue

The workshop takes place in two different UZH buildings: KAB (Kantonsschulstrasse 3) and RAA (Rämistrasse 59).
Rooms:
KAB-E-05 (4-5 October) Check it on the map!
RAA-E-08 (6 October until 13:00) Check it on the map!

RAA-E-27 (6 October from 14:30) Check it on the map!

R Crash course

On Tuesday 3rd October we'll be offering a Crash Course on R, so that those less familiarised with this programmation language feel more at ease when using it during the workshop.

Please install R and RStudio before attending the course and let us know if you have any trouble with the installation so that it can be fixed before the course starts.

During the course, we will focus on how to use R for data science. This means that we will not devote much time to explain the basis of R as a programming language. If you'd like to learn a bit more about this (this is not required for the course, but recommended either for before or afterwards!), we recommend you the package swirl. For doing so, please follow the instructions on the swirl website. In step 5, choose number 1 (“R Programming: The basics of programming in R”) and install it. You will get a good idea of how R works by completing lessons 1 to 7 (1: Basic Building Blocks; 2: Workspace and Files; 3: Sequences of Numbers; 4: Vectors; 5: Missing Values; 6: Subsetting Vectors; 7: Matrices and Data Frames). Have fun and contact us with any question!

Date and time: October 3rd 2017, 16:00 - 18:30.

Venue: KO2-F-173 (Karl-Schmid-Strasse 4)