Methods for assessing children’s narration
by Natalia Gagarina, Nathalie Topaj
The ability to tell stories is an indispensable part of adequate functioning in society. It has an impact on social well-being and forms the basis of human communication. Oral narratives build a bridge between oral and written language (Hadley, 1998) and are useful for the early identification of children at risk for late reading development (Reese et al., 2010; Suggate et al., 2011).
Different methods can be used to assess children’s narration depending on the age of participants, available materials, the goals of the assessment, etc. The tutorial provides a critical overview of the different ways of eliciting children’s narratives, discusses the advantages and disadvantages of various methods, and presents a picture-based instrument for assessing narrative production and comprehension. This instrument is based on the multidimensional theory of narrative organization. It enables the assessment of narrative skills in monolingual and multilingual children aged 3 to 12 years and is also used with adolescents and adults. The Multilingual Assessment Instrument for Narratives (MAIN) enables the assessment of comprehension and production in three elicitation modes: Model story, Retelling and Telling. MAIN contains four parallel stories, each with a carefully designed sequence of six pictures based on the multidimensional model of story organization. The stories are controlled for cognitive and linguistic complexity, parallelism in macro- and microstructure, as well as for cultural appropriateness and robustness.
In the tutorial, various ways of using MAIN with different populations will be presented and practiced.
Ten easy steps to conducting transparent, reproducible meta-analyses for infant researchers
by Loretta Gasparini
Meta-analyses provide researchers with an overview of the body of evidence in a topic, with quantified estimates of effect sizes and the role of moderators, and weighting studies according to their precision. This workshop provides a guide for conducting a transparent and reproducible meta-analysis in the field of developmental psychology, using freely available programs, in 10 steps:
(1) Choose a topic for your meta-analysis
(2) Formulate your research question and specify inclusion criteria
(3) Preregister and document all stages of your meta-analysis
(4) Conduct the literature search
(5) Collect and screen records
(6) Extract data from eligible studies
(7) Read the data into analysis software and compute effect sizes
(8) Visualize your data
(9) Create meta-analytic models to assess the strength of the effect and investigate possible moderators
(10) Write up and promote your meta-analysis.
Meta-analyses can inform future studies, through power calculations, by identifying robust methods and exposing research gaps. Attendees will also be introduced to MetaLab, a platform where datasets across multiple topics of developmental psychology can be synthesized, and datasets can be maintained as a living, community-augmented meta-analysis to which researchers add new data, allowing for a cumulative approach to evidence synthesis.
Exploring early language via Wordbank and the MacArthur-Bates Communicative Development Inventories: A hands-on workshop
by Michael Frank, Virginia Marchman
In this course, we will examine early language learning through the lens of the MacArthur-Bates Communicative Development Inventories (CDIs), parent report forms for studying child language acquisition. We’ll first introduce CDI forms, their variants, and their strengths and weaknesses. We’ll then introduce Wordbank, an open repository for CDI data. We’ll use the R statistical programming language to download and analyze data from Wordbank. We’ll explore how to create reproducible pipelines with Wordbank data and reproduce analyses on grammar/lexicon correspondences from Bates & Goodman (1997). We’ll end by presenting new research directions using Wordbank and the CDI as outcome measures for evaluating models of language learning. Prerequisite: some familiarity with R; prior exposure to the tidyverse family of packages is desirable.
Daylong data: Raw audio to transcript via automated & manual open-science tools
by Elika Bergelson, Marisa Casillas, John Bunce, Jessica Montag, Alex Cristia, Loann Peurey, Marvin Lavechin
Recent technological innovations have made collecting long naturalistic recordings of children’s home environment far simpler than in the past. However, the raw output of such recordings is not immediately usable for most analyses. Simultaneously, speech technology and machine learning tools have improved considerably over the past decade, making it feasible to use such tools with increasingly diverse and noise-laden data. Specifically, we now have open-source options for automated diarization of these long recordings into voice classes, with high accuracies for the key child and female adult voices, and moderate or low accuracies for male adults and other children’s voices. For English, there are also off-the-shelf low-cost systems that allow fairly accurate transcription, at least for some sections. For all other goals, human annotation is necessary. This means that, today, deciding whether speech is directed to the child or to others, and whether the child’s vocalizations are cries, simple or complex utterances, and often, the content of those utterances, still needs to be done manually. In this workshop, we present the key tools required for each one of these tasks, and provide tips and tricks for working with daylong data.
Praat for child speech researchers: acoustic analysis, stimulus preparation, and running experiments
by Václav Jonáš Podlipský
Topics covered:
1. Measuring acoustic properties of child and adult speech (fundamental frequency and its contours, formant frequencies, duration of speech segments and pauses). Annotating (labeling) recordings and collecting the measure values (by hand or using a script).
2. Preparation of experimental auditory stimuli. Recording and adjusting their acoustic properties (intensity, manipulating duration and fundamental frequency, filtering…)
3. Preparing and running experiments in Praat. Using and adjusting experimental templates that come with Praat (forced-choice, likert scale, goodness judgements). Scripting experiments from scratch using Praat Demo Window.
Writing an Acquisition Sketch: Step-by-step guidelines from the Sketch Acquisition Project.
by Shanley Allen, Evan Kidd, Lucinda Davidson, Rebecca Defina, Birgit Hellwig
Despite a proud history of crosslinguistic comparison (e.g., Slobin, 1985 – 1997), child language research is mostly skewed towards English and a handful of other Indo-European languages (Kidd & Garcia, 2022). Widening our evidential base is thus an important priority, made even more urgent by the fact that we are currently living in a period of rapid language endangerment and loss. In this workshop we will introduce the Sketch Acquisition Project, which presents a model for writing a broad-spectrum sketch of an understudied language, with the aim of increasing data coverage in the field. The workshop will go through the process of collecting a sketch corpus and writing the sketch. Attendees are encouraged to bring their already collected data to work on during the workshop.
New Developments in CHILDES and PhonBank
by Brian MacWhinney, Yvan Rose
This workshop will include 4 hours devoted to CHILDES data and methods and 2 hours devoted to PhonBank data and methods.
The CHILDES segment of the workshop will have begin with a brief review of resources at the website and the basic use of CLAN programs. Next we will provide instruction in the use of five analytic methods:
1. The Batchalign system for automatic speech recognition and word-level diarization of child language data. This system can reduce transcription time by up to 75%.
2. The use of TalkBankDB database search and retrieval system for analysis of particular aspects of lexicon and syntax.
3. The KIDEVAL system for automated profile analysis of a single child, groups of children, or longitudinal data.
4. The Collaborative Commentary system for teaching about child language development or the development of collaborative analysis in research groups.
5. The use of the Stanza program inside Batchalign for automatic morphosyntactic analysis based on the Universal Dependencies (UD) framework for over 100 languages.
The PhonBank segment will begin with a brief summary of the most significant advances in the development of the database. This will be followed by a practical demonstration of new functions and user interfaces introduced in Phon 4, most of which aim at higher compatibility between CLAN and Phon, setting the stage toward more analytic integration between the two programs.
We will conclude with a dedicated period for comments and questions from the participants.