A place to run translation studies from the first sign-up to the final dataset
The Stone Guest ToolKit is a web system for researchers who study how people translate, alone or in teams. It signs participants up, puts them in teams, gives them a place to talk and to write, records how they type, and turns all of it into data you can analyse.
Who it is for
It is for researchers in translation and writing studies who want to observe the translation process: not only the finished text, but how it came about. Pauses, corrections, look-ups in a dictionary, messages to teammates, changes to a shared draft.
You do not need to program. Everything a researcher does happens in a web panel: setting up a study, following it while it runs, and getting the data out at the end.
Why it exists
A typical translation process study needs many separate tools: a sign-up form, consent forms, a way to split people into teams, a chat, a shared document, a keystroke logger, questionnaires, and a lot of spreadsheet work to bring the pieces together. Every piece has its own accounts, its own files and its own idea of who the participant is.
The ToolKit puts these pieces in one system. Each participant has one code (a PID) from sign-up to the dataset, so a keystroke file, a chat message and a questionnaire answer from the same person can be joined without anyone handling names or email addresses.
What participants see
- A sign-up page with the study's information and a consent form, in Turkish and English.
- A personal portal: their tasks, their team, files to upload, questionnaires to fill in, and what they have handed in.
- Stone Guest, a small program for their computer that records their typing during a translation session.
- A team chat (Zulip) and a shared document (Etherpad), when the study works in teams.
What you see as a researcher
The panel, on its own web address. From there you build the study, watch who has signed up and who has uploaded, send messages, note problems, and run the analysis. Several researchers can work on one study, each with a role that decides what they may see and change.
The parts, in one sentence each
| Part | What it does for you |
|---|---|
| Sign-up and consent | Collects participants, checks eligibility, records consent item by item, and gives everyone a code. |
| Portal | One page per participant with everything they have to do, so you do not have to email instructions. |
| Teams | Forms balanced teams (for example by language or experience) and can reshuffle them between phases. |
| Zulip | Creates a team chat for every team, gives participants accounts with nicknames, and keeps a copy of the messages for analysis. |
| Etherpad | A shared document per team (or per person) where the translation is written; every change is kept. |
| Stone Guest | Records key presses and which kind of window is in front during a session, on the participant's own computer. |
| Questionnaires and tests | Background questionnaires, the LexTALE vocabulary test, NASA-TLX workload and your own questions. |
| Automation | Sends reminders and messages on a schedule or when something happens, and records every message it sends. |
| Analysis | Computes process indicators (the Task Segment Framework), pauses and bursts, and conversation networks, and exports tidy tables. |
| Replay and coding | Plays back a session keystroke by keystroke and lets coders label what they see. |
| Deliverables | Collects drafts and final translations as numbered versions, with deadlines. |
| Releases and reports | Packs a pseudonymised dataset for sharing and drafts the methods section of your paper. |
A study switches on only the parts it needs. A study with individual translators and no chat simply leaves Zulip and teams off.
New to this kind of research? Start with How a study runs, which follows one study from the first idea to the published data. Words you have not met before are explained in the glossary.