ReCo – Automatic Text Response Coder
ReCo offers apps and software packages to automatically evaluate text responses using machine learning, natural language processing and large language models. Researchers can use these artificial intelligence methods to code text data, and teachers can give all learners a voice in the classroom.
About
ReCo stands for Automatic Text Coder. A number of apps and software packages that are openly and freely available for this purpose are all about this. If, for example, a student answers a test task with "The author wants to draw attention to poverty with this story!", this response can be automatically assessed with the help of ReCo.
Automatic Assessment of Short Text Responses: Shiny-App shinyReCoR
A graphical user interface called shinyReCoR, which can be installed using the R software, is currently available online. It is primarily, but not exclusively, aimed at researchers who want to evaluate or sift through their open text data from a study. The central feature is the grouping of responses with similar content (i.e. semantics). Thanks to the graphical user interface, no scripting skills are required. Instead, it offers a variety of interactive visualisations of the texts and their semantics as well as diagnostics of the trained classification models.
Semi-Automatic Coding: eco
The eco app, which is also aimed at researchers, is designed to reduce manual coding of text responses through automation. It uses the same techniques as the Shiny-app, but only if the quality of the automatic classification is satisfactory. This can save manual effort and possibly improve data quality. However, if no reliable classification is achieved for certain responses in the course of manual coding, these will continue to be coded by humans.
New R package ReCo
"ReCo was shiny, now it makes you shine."
A new R-package is also planned that will make the coding of the text responses even more innovative, flexible, and easier to automate for reserachers with scripting skills. In addition to the visualisation features known from the Shiny-app, it offers many new features, such as the use of large language models from huggingface and the combination with external machine learning packages in order to be able to apply learning alogorithms beyond clustering.
Re-Co Live Enables All Students and the Entire Audience to Participate
Teachers regularly ask students questions in class. Often, a few learners come forward and only one of them formulates a response to the question. However, if the aim is to mobilise all learners, this format defeats its purpose. Instead, the ReCo-Live app makes it possible to collect the responses of all learners life in class by having them type in their responses on mobile devices. In this way, all learners partcipate and the classroom discourse becomes more inclusive. All responses are then displayed on the teacher's dashboard and, if desired, grouped into response types based on their sematic similarity. Thus, the teacher can see typical responses as well as outstanding misconceptions or excellent responses, which ca be used to advance the classroom discourse. Similary, the app also enables speakers to create interactive moments in their talks.
Team
Profile
| Target Audience: | Educational Research, Educational Practice |
|---|---|
| Type of Content: | Digital Learning & Educational Technologies, Teaching & Learning |
| Online since: |
02/2021
|
| Department: | Teacher and Teaching Quality |
| Contact: | |
| Web Address |