QML - Week 1
Research methods overview
Welcome to QML!
Practicalities
Who am I?
Overview and week-by-week contents: https://uoelel.github.io/qml/content.html
Assessment: https://uoelel.github.io/qml/assessments.html
Weekly self-reflections in the Reflective Journal (on Learn): DUE EVERY WEEK by the following Wednesday.
Information: https://uoelel.github.io/qml/assessments.html#reflective-journal
This is a very intense course!
- We focus on quantitative methods for knowledge-oriented research. No machine learning, LLMs, language technology.
Finding help:
Office Hours.
Piazza: Q&A forum.
Research methods
Research methods: components
Group activity
Form groups of 5/6 students. I will give you a card with a research topic.
Think about the following research components in relation to the research topic and come up with a “research plan”:
Research process: What are the steps?
Project management: How would you manage roles, timeline, data storage?
Digital skills: Which tools will you need?
Philosophy: What count as evidence? What are your assumptions?
Ethics: How would you ensure personal rights are respected?
Group activity
Pick a group next to you and tell each other about your research plans.
What are the similarities and why do you think they are what they are?
How do your plans differ and how do you think individual differences across and within groups contributed to the differences?
Group activity
With your group, come up with a research question.
It doesn’t have to be fancy, just something you are interested in.
Write your research question on a post-it.
Research axes
Exploratory (hypothesis generating)
What is the role of empathy in the L2 perceptual development of intonation?
➡️ Higher empathic individuals, in comparison with lower empathic individuals, appear to be more sensitive to intonation cues in the process of forming sound-meaning associations (Casillas et al. 2023).
What is the relationship between spaces of language use, attitudes and self-reported proficiency in minorities languages?
➡️ With more spaces available to speakers, the relationship between attitudes and proficiency is weaker (Hampton and Coretta 2024).
Corroboratory (hypothesis-testing)
Raviv, De Heer Kloots, and Meyer (2021)
More structured languages (with higher systematicity) are easier to learn.
Languages created by larger groups are easier to learn.
➡️ High systematicity was advantageous for learning, but learners did not benefit from partly or semi-structured languages.
➡️ Languages that evolved in big and small groups were equally learnable.
Descriptive
The phonetic inventory of Northern Tosk Albanian (Coretta et al. 2022).
An accurate description of the distribution of reflexives and their antecedents in Yoruba, with an examination of the verbs which they co-occur with in terms of the valency classes of the verbs as well the semantic field into which the verbs can be categorized (Atoyebi and Anyanwu 2020).
Explanatory
Regularization is a phenomenon in which languages gradually reduce variability along certain dimensions. We use an artificial language learning experiment to test whether constraints on working memory during learning or production contribute to regularization (Keogh, Kirby, and Culbertson 2024).
We examine a model of abstraction from exemplars in which (a) all heard exemplars are stored (though influenced by factors like attention, decay, and interference), and (b) for language use, these exemplars are restructured into multiple layers of abstraction (Ambridge 2020).
Group activity
Think about your chosen research question. Which quadrant in the research axes coordinates do you think it belongs to?
Come to the board and place the post-it with your research question in the chosen quadrant.