Loading a phrase...

Entry Speed 0.00 Average 0.00
Error Rate 0.00% Average 0.00%
Hide Introduction

WebTEM: A Web Application to Record Text Entry Metrics

Ahmed Sabbir Arif

Last update: Let me know if you are having issues with this version and want to use older one

WebTEM is a Web application to record/gather text entry metrics. It works with almost all operating systems, browsers, and text entry techniques. Yet to verify compatibility, press the Check button and type a chunk of text to confirm that it detects all insertions and deletions. Press the Back button to return to the settings. To use the application, first enter a valid email address, study-specific information, and keyboard type. The application emails all logs to the email address provided and then deletes all data from the server. The study-specific settings are used for recordkeeping. Then select a phrase set, number of phrases in the session, performance metrics, and optional features. The most popular performance metrics and features are preselected for your convenience. Arif & Stuerzlinger, 2009 reviews and analyzes the most popular text entry metrics. International Metrics for non-Latin scripts are displayed when a non-Latin corpus is selected and calculated using the output stream convention proposed by Sarcar et al., 2015. The application displays the phrases in a randomized order. Press the Start button when ready to start the study. Make sure that the device has access to the Internet throughout the study.

The application generates one information (.INFO) and two tab-delimited files (.TSV) to record all settings, timestamped events, and performance metrics, respectively. Each row of the metrics log represents a phrase and each column a metric. The last row holds average values. Events are recorded as [time,text,event,duration,insertion,deletion], where time is the number of milliseconds since January 1, 1970, text is the current state of the transcribed text, event is a user or system action (tap, autocorrection, etc.), duration is the time for the action in milliseconds, and insertion and deletion are character(s) entered and deleted by the action, respectively. Read this following paper or watch this video to know more about WebTEM.

Ahmed Sabbir Arif and Ali Mazalek. 2016. WebTEM: A Web application to record text entry metrics. In Proceedings of the 2016 ACM International Conference on Interactive Surfaces and Spaces (ISS '16). ACM, New York, NY, USA, 415-420. DOI: https://doi.org/10.1145/2992154.2996791

Feedback & Consultation/Collaboration

Contact me at to report bugs or request new features, metrics, and/or phrase sets for your user studies; if you are seeking consultation or want to collaborate on a research project evaluating a text entry technique(s).

Log Files
Study Data

Keyboard Type Virtual Keyboard
Physical Keyboard
Phrase Set The quick brown fox jumps over the lazy dog
500 Short English Phrases MacKenzie & Soukoreff, 2003
200 Memorable English Phrases Vertanen & Kristensson, 2011
500 Children's English Phrases Kano et al., 2006
Show International Sets
111 Bengali Phrases Without Numbers & Punctuations Arif & Fardeen, 2016
Metrics Presented Text (PT)
Transcribed Text (TT)
Presented Text Length (|PT|)
Transcribed Text Length (|TT|)
Input-Stream Length (|IS|)
Words per Minute (WPM)
Characters per Second (CPS)
Error Rate (ER)
Minimum String Distance Error Rate (MSD ER)
Keystrokes per Character (KSPC)
Corrected Error Rate (CER)
Total Error Rate (TER)
Input Time (IT) signifies the time (in seconds) users took to enter a phrase
Visual Scan Time (VST) signifies the time (in milliseconds) users took to visually scan a recently completed phrase, mainly to proofread, before submitting it
Cursor Control Count (CCC) is the total number of times users repositioned the cursor using the arrow keys, the mouse, direct touch, or a digital pen to correct errors or to edit text in the process of inputting a phrase
Backspace Count (BC) is the total number of backspaces per phrase
Corrective Action Count (CAT) is the total number of corrective actions, particularly backspaces and cursor control, per phrase
Corrective Action Rate (CAR) is the ratio (%) of the total number of corrective actions to the total number of actions per phrase
Total Corrective Action Time (T-CAT) is the total time (in seconds) the user took to perform the corrective actions per phrase
Corrective Action Time Ratio (CAT-R) is the ratio (%) of the total time the user took to perform the corrective actions to the total time s/he took to enter a phrase
Prediction Rate (PR) is the ratio (%) of the total number of characters automatically entered by the predictive system and the total number of characters entered in a text entry episode. It does not account for incorrect predictions, since they are difficult to identify
Presented Text Output Stream Length (|OSP|)
Transcribed Text Output Stream Length (|OST|)
International Metrics Characters per Second (CPSint)
Error Rate (ERint)
Minimum String Distance Error Rate (MSD ERint)
Keystrokes per Character (KSPCint)
Corrected Error Rate (CERint)
Total Error Rate (TERint)
Additional Options Disable predictive features
Present all phrases in lowercase
Ignore letter case and extra spaces in metrics calculation
Present all phrases without special characters
Display the number of phrases entered
Display performance summary
Hide presented text when users start typing
Force error-free submissions (includes auditory feedback)


The application is distributed in the hope that it will be useful, "AS IS" without warranty of any kind, either expressed or implied, including, but not limited to, the implied warranties of merchantability and fitness for a particular purpose. The entire risk as to the quality and performance of the program is with you. Under no circumstances shall the authors have any liability hereunder for any loss or damage caused directly or indirectly from the use of this application.