Big Question 4
Developing a Language Test Battery for Young Adults
(last update 2019-07-03)
This position fully focused on the development of the Language in Interaction test battery (LiB)
The role of processing speed in picture naming and lexical decision
Team members: Hintz, Jongman (MPI), Dijkhuis, Van ‘t Hoff, Meyer, and McQueen
Picture naming and lexical decision are frequently used tasks in psycholinguistic research. Both are considered good tools for studying lexical access, and both are speeded tasks. Given these commonalities, one would expect a high correlation between participants’ average response times (RTs) in both task. This prediction was tested in a large group of young adults with diverse educational backgrounds. Participants also completed tests assessing non-verbal processing speed, vocabulary, and non-verbal intelligence.
The results revealed a moderate correlation of the average RT in the two linguistic tasks (10% of shared variance) and moderate correlations of each of the linguistic tasks with processing speed (10% of shared variance with picture naming, 24% with lexical decision). Thus, the processes underlying picture naming and lexical decision are less closely related than one might have thought. Moreover, as RT in picture naming and lexical decision depended as least as much on general processing speed as on domain-specific linguistic processes, these tasks are not well suited, when used in isolation, to obtain a pure measure of a person’s ability to retrieve words for speaking or listening.
Correlations between performance on the picture naming and lexical decision tasks, as well as their performance on the measures reflecting general processing speed, vocabulary size and non-verbal intelligence for the two participant groups: University students and vocational college students.
- Completion of Language in Interaction test battery:
The final version of the battery will contain 29 different tests. The selection of these tests and their specific protocol has been done in interaction with other BQ4 members.
Five tests are devoted to assessing ‘Linguistic Knowledge/Representations’ (Stairs4Words, Peabody Picture Vocabulary Test, Spelling Test, Author Recognition Test, Klepel & One-Minute Reading Test). Eight are devoted to assessing ‘General Cognitive Skills’ including processing speed (auditory/visual simple/choice test, letter comparison), short-term memory (forward & backward digit span & Corsi block tapping), inhibition (antisaccade task, Eriksen flanker task), and non-verbal intelligence (Raven’s Advanced Progressive matrices). The remaining 16 tests capture ‘Linguistic Processing’ skills. Specifically, processing skills are measured at the word level and at the sentence level; for production and comprehension, respectively:
1) Word production: rapid automatized naming, verbal fluency, maximal speech rate, antonym production, picture naming;
2) Word comprehension: (non-)word monitoring in (non-)word lists; rhyme judgment, lexical decision, semantic categorization;
3) Sentence production: phrase and sentence generation, spontaneous speech;
4) Sentence comprehension: word monitoring in predictable and non-predictable sentence contexts, prediction in the visual world: morpho-syntactic features, prediction in the visual world: verb-specific selectional restrictions, prescriptive grammar task, SynTest.
- Developing tests assessing linguistic knowledge and linguistic processing skills
While the tests selected for the ‘general cognitive skills’ section already existed, many of the tests assessing linguistic knowledge and processing skills had to be developed from scratch. This required extensive piloting in the population intended to be tested. To that end, large-scale pilots were conducted at the MPI and at vocational colleges across the Netherlands. Based on the pilot data, the tests were revised and, where necessary, piloted again. Data acquisition with the full battery will begin in spring 2019.
- Technical implementation of tests
A full time programmer, based at the MPI, implemented 14 of the 29 tasks in a programming environment (called Frinex) suited to carry out experiments in a web browser via the internet. The remaining 15 are reaction time critical tasks and have been programmed in Presentation. They still need to be converted into an internet-ready format.