https://lingoboost.csl.sri.com/lingoturk/turk/
This site has instructions for the LingoTurk experiment we are running. This is also where users can check their current participation status. Status is a bit messy right now as we work on cleaning the system logs (e.g., we need to remove incomplete logs).
https://lingoboost.csl.sri.com/lingoturk/admin/
This site requires a login and is where the researchers login to setup their experiments and user parameters.
We have two distribution models.
Preferred Method
App Updating: The app will update automatically through the Google Play Store.
Email Requirements: This requires a gmail account the user has setup on the Play Store (the account they use when they download apps). This is akin to your Apple ID if we were using iPhone.
App Updating: The app will need to be updated manually when new versions are released and emailed to the user.
Email Requirements: No email required. This only requires a link (https://lingoboost.csl.sri.com/lingoboost-beta.apk), but then the user has to install the app outside of the official Google/Android market.
When users first start the lingoBoost app, they should enter the username that is assigned by the researcher.
If users need to adjust their username, they can do so later from the Settings menu.
lingoBoost was developed to integrate with foreign language learning applications, so testing vocabulary performance on a regular basis is a must! Test Session will retrieve all of the words the user has trained on, to date, and test them with a basic recall test (stimulus = foreign word, response = user's English translation). Users must first set a preferred volume following the on-screen instructions before starting the test.
Sleep session is the core of the app, providing a "TMR Lite" experience. This portion of the app should be run when users are ready for bed. Users will be prompted to select the volume for the white noise and spoken words/sounds to ensure that words/sounds are audible through the white noise. After pressing READY FOR SLEEP, a popup (Android Toast) will briefly let the user know that the words (or sounds) have been updated from the server.
These are parameters that are passed from the server to the client, using the langlearn-api (separate, private repo). The keys and default values (global) are listed below. These values are managed from the admin page through the API so that they can be overridden at the user level, so lingoBoost may see values that differ from those below.
| Key | Default Value | Description |
|---|---|---|
| word_delay | 10 | Time, in seconds, that should elapse between each word stimulation |
| random | True | Boolean; If True, the order of words will be randomized |
| feedback | True | Boolean; If True, correct translations will be shown during test, after each response |
| sham | False | Boolean; If True, no word stimulation will take place |
| min_words | 10 | The number of words that must have been learned/practiced since the time of the client request in order for the server to return anything, otherwise the server returns "Not enough words practiced." The client may request 1) all words, 2) words since a given date-time, or 3) words this calendar day (server-based calendar). |
| volume_dampening | 30 | Percent (integer) of the user-set volume we want white noise to decrease to, during word presentation; dampened_volume = (volume_dampening/100)*user_WhiteNoise_Volume |
| stimulation_stop | 18000 | Time, in seconds, when stimulation should stop after the Sleep Session began |
| session_start_delay | 1800 | Time, in seconds, that should elapse after beginning a sleep session before stimulation starts |
| pause_delay | 300 | Time, in seconds, that should elapse after resuming a sleep session after a pause |
| all_words | True | Boolean; If True, Duolingo users will have all the words they have ever trained sent during the Sleep session (recent*3 + old) |
Format: comma-delimited with word wav file, matched wav file (semantic), and word. No header! If no "matched" sound is needed, it can be ignored. See the following example;
ld1.wav,,lucid1
ld2.wav,,lucid2
ld3.wav,,lucid3
ld4.wav,,lucid4
We have glorious data output links on the admin page that will output csv's for both the test logs and sleep logs (Test Summary, Sleep Summary).
The examples and lists below represent the current setup. **Two new variables will be added in subsequent test logs. These variables represent test length, so that we can easily determine if a test log is complete, and whether a test session is actually the pre (first test, prior to training) or post (final test, after all test and training) tests.
All timestamps have been converted to match the user's local time. All the durations are represented in seconds.
nickname,log_type,session_hash,session_start,session_end,session_duration,log_event_time,trained_word,test_response,closest_translation,string_score,duolingo_first_practiced
| column name | type | description |
|---|---|---|
| nickname | string | unique participant ID |
| log_type | string | 'test' or 'sleep' to differentiate the logs if/when joined |
| session_hash | string | unique ID to label the test/sleep sessions |
| session_start | time stamp | the first timestamp in the log |
| session_end | time stamp | the final timestamp in the log |
| session_duration | integer:seconds | difference between session_end and session_start |
| log_event_time | time stamp | this is the timestamp of the actual responses/events within the log |
| trained_word | string | the unique word being trained in the foreign vocabulary |
| test_response | string | the participants response to the foreign word |
| closest_translation | string | the closest translation (from translations provided by Duolingo) match to the response |
| string_score | float | string comparison between the response and closest translation based on change needed to match the items; 1 = perfect match and no string distance |
| duolingo_first_practiced | time stamp | the first time that the word was practiced |
corticalre,test,8a46affe-c6e9-48bb-b4ac-dde3a6475a12,2017-10-28 20:43:04-04:00,2017-10-28 20:57:20-04:00,856.0,2017-10-28 20:43:22-04:00,nötkött,beef,beef,1.0,2017-10-27 21:32:03-04:00
nickname,log_type,session_hash,session_start,session_end,session_duration,log_event_time,trained_word,system_volume,white_noise_volume,word_volume
| column name | type | description |
|---|---|---|
| nickname | string | unique participant ID |
| log_type | string | 'test' or 'sleep' to differentiate the logs if/when joined |
| session_hash | string | unique ID to label the test/sleep sessions |
| session_start | time stamp | the first timestamp in the log |
| session_end | time stamp | the final timestamp in the log |
| session_duration | integer:seconds | difference between session_end and session_start |
| log_event_time | time stamp | this is the timestamp of the actual responses/events within the log |
| trained_word | string | the unique word being trained in the foreign vocabulary |
| system_volume | integer | user-set volume level for Android Media, adjustable throughout sleep |
| white_noise_volume | integer | user-set volume for the white noise, adjustable before sleep session start |
| word_volume | integer | user-set volume for the spoken foreign words, adjustable before sleep session start |
corticalre,sleep,1b19955a-1cf2-4ced-a173-78450153e600,2017-10-20 14:01:19-07:00,2017-10-20 14:02:31-07:00,72.0,2017-10-20 14:01:30-07:00,saltet,80,18,45
- Android Studio 3 ** Kotlin plugin
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent