Predicting Text Readability from Scrolling Interactions

Dataset containing scroll interactions of 598 partcipants reading advanced and elementary texts from the OneStopEnglish corpus. Participants answer multiple choice reading comprehension questions from OneStopQA. Study details and preliminary experiments can be found in our paper.

Interface Examples

Text Interface	Comprehension Questions

Participant Demographics

Proficiency	(%)	Education	(%)	Age range	(%)	Hours spent reading English	(%)
Native	46.69	Graduate	53.67	18 - 24	18.23	0 - 4	24.20
Near-native	14.75	Undergraduate	39.51	25-34	57.19	5 - 9	22.87
Advanced	27.78	High School	3.59	35-44	13.38	10 - 14	11.72
Intermediate	9.83	Vocational	2.65	45-54	8.02	15 - 19	7.18
Beginner	0.95	No formal	0.57	55+	3.17	20 +	33.84

Data Description

Columns (elementary / advanced)	Example	Description
elementary_text	WNL Rwanda-ele.txt	Title of the text shown to the participant as named in the OneStopEnglish corpus
elementary_read	read	Value indicating whether it is likely the article was read from the number of scrolls
elementary_reading_time_ms	92909	Total time spent reading the article in ms
elementary_reading_event_ms	['15', '5408', '6122' …]	Elapsed time for a scroll event, in the example the first scroll action occurs after 15ms and the second after 5408 ms
elementary_reading_scroll_event	['0', '200', '500' … ]	Y axis distance from the start of the article, in the example the first scroll action moves the text 200px from the top of the article
elementary_reading_paragraph_visibility	['0_100,1_47', '0_53…]	Percentage of paragraphs visible whilst reading, in the example paragraph 0 is 100% visible and paragraph 1 is 47% visible then as scrolling occurs paragraph 0 is 53% visible
elementary_reading_span	['A1_97', 'A1_100…]	Shows the percentage of the distractor / answer spans that were visible to the participant. The example shows that the first comprehension question is 97% visible and then becomes fully visible (100%) as scrolling occurs

Citation

@inproceedings{gooding-etal-2021-predicting,
    title = "Predicting Text Readability from Scrolling Interactions",
    author = "Gooding, Sian  and
      Berzak, Yevgeni  and
      Mak, Tony  and
      Sharifi, Matt",
    booktitle = "Proceedings of the 25th Conference on Computational Natural Language Learning",
    month = nov,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.conll-1.30",
    pages = "380--390"}