Quantitative Methods for LEL

Welcome to the main site of the course Quantitative Methods for Linguistics and English Language (Semester 1).

This website is your go-to place throughout the semester for any info related to the course.

1 Course description

This course is an introduction to quantitative data analysis (including data wrangling, visualisation and modelling) as commonly employed in linguistics, using the R software.

The following is a list of topics you will learn:

  • The basics of quantitative data analysis.
  • Data preparation.
  • Data summaries.
  • Principles of data visualisation.
  • Statistical modelling with Bayesian regression models.
  • Open Research practices and the dangers of statistics.

At completion of the course you will have gained the following skills:

  • Import common data formats, tidy and transform data.
  • Choosing and reporting appropriate summary measures.
  • Using compelling visualisations to communicate a specific message about patterns in the data.
  • Master regression models for different types of data (continuous measures and binary outcomes).
  • Using Bayesian inference to answer research questions and avoid common interpretation pitfalls of frequentist techniques.

Examples from different branches of linguistics will be used to provide you with hands-on experience in quantitative data analysis and Open Research practices.

2 Course rationale

This course is designed to help you develop the necessary skills for conducting and interpreting analyses of data as commonly employed in linguistics.

The content and objectives of the course are in response to recent advances in our understanding of the theory behind research methods.

Recent meta-scientific research has identified four important aspects of research: reproducibility, replicability, robustness and generalisability.

  • A result is reproducible when the same analysis steps performed on the same dataset consistently produces the same answer.
  • A result is replicable when the same analysis performed on different datasets produces qualitatively similar answers.
  • A result is robust when a different analysis workflow performed on the same dataset produces qualitatively similar answers.
  • A result is generalisable when a different analysis workflow performed on different data sets produces qualitatively similar answers.

See Definitions on “The Turing Way” website for a more detailed explanation.

However, based on surveys from different disciplines, we are currently facing three research crises (reproducibility, replicability and generalisability crises) by which most results are neither reproducible, nor replicable, nor generalisable (Munafò et al. 2017; Simmons, Nelson, and Simonsohn 2011; Ioannidis 2005; Yarkoni 2022).

The Open Research movement (also known as Open Science or Open Scholarship, Crüwell et al. 2019) was developed with the aim of improving our understanding of these crises and with the objective of providing researchers with guidelines and tools to produce reproducible, replicable and generalisable research.

The statistical philosophy adopted in the course is that of the New Statistics (Cumming 2013; Kruschke and Liddell 2018). The main goal of the New Statistics is to shift the attention from statistical significance to estimation with quantification of uncertainty.

3 Support

Ask for help

If at any point during the course you don’t feel comfortable with any aspect of the course, you are unsure about anything that has been covered in class or in your own time, you are struggling to keep up with the course workload, you are experiencing mental of physical distress due to a pre-existing or new illness, medical condition or disability, or you find yourself unable to access basic needs like food or housing, please do get in touch with me and/or the PPLS support (go to the PPLS UG or MSc Hub on SharePoint > Support for students > Health & Wellbeing).

We are humans first and the rapidly-changing new world we are living in now can put us under pressure. What is most important to me is that you are first and foremost healthy and able to participate to the course, and that you succeed and get the most out of the course.

It is OK not to be OK, and remember that you are not alone. Other people, teachers and students, might be struggling right now or have struggled before and might have gone through what you are going through now. Remember that support exists for you, so please do reach out. If you see somebody close to you struggling, please let them know about the available support network and encourage them to reach out.

4 Contacts

You can reach me (Stefano) at or on Teams.

If you want to book office hours with me, you can do so here: https://bit.ly/33BH84L.

Here’s the list of course tutors:

  • TBA

References

Crüwell, Sophia, Johnny van Doorn, Alexander Etz, Matthew C. Makel, Hannah Moshontz, Jesse Niebaum, Amy Orben, Sam Parsons, and Michael Schulte-Mecklenbeck. 2019. “Seven Easy Steps to Open Science: An Annotated Reading List.” Zeitschrift für Psychologie 227 (4): 237–48. https://doi.org/10.1027/2151-2604/a000387.
Cumming, Geoff. 2013. “The New Statistics: Why and How.” Psychological Science 25 (1): 729. https://doi.org/10.1177/0956797613504966.
Ioannidis, John P. A. 2005. “Why Most Published Research Findings Are False.” PLoS Medicine 2 (8): e124. https://doi.org/10.1080/09332480.2019.1579573.
Kruschke, John K., and Torrin M. Liddell. 2018. “The Bayesian New Statistics: Hypothesis Testing, Estimation, Meta-Analysis, and Power Analysis from a Bayesian Perspective.” Psychonomic Bulletin & Review 25 (1): 178206. https://doi.org/10.3758/s13423-016-1221-4.
Munafò, Marcus R., Brian A. Nosek, Dorothy V. M. Bishop, Katherine S. Button, Christopher D. Chambers, Nathalie Percie Du Sert, Uri Simonsohn, Eric-Jan Wagenmakers, Jennifer J. Ware, and John P. A. Ioannidis. 2017. “A Manifesto for Reproducible Science.” Nature Human Behaviour 1 (1): 21. https://doi.org/10.1038/s41562-016-0021.
Simmons, Joseph P, Leif D Nelson, and Uri Simonsohn. 2011. “False-Positive Psychology: Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant.” Psychological Science 22 (11): 13591366.
Yarkoni, Tal. 2022. “The Generalizability Crisis.” Behavioral and Brain Sciences 45. https://doi.org/10.1017/s0140525x20001685.