Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

How to Engage Your Readers? Generating Guiding Questions to Promote Active Reading

Peng Cui, Vilém Zouhar, Xiaoyu Zhang, Mrinmaya Sachan

TL;DR

The paper addresses how in-text guiding questions can promote active reading by examining their use, functions, and impact. It introduces GuidingQ, a 10,577-question dataset from textbooks and research articles, and develops a taxonomy of five question roles to analyze discourse and interaction with readers. It then presents three generation paradigms (Pipeline, Multitask, Joint) with a QA-focused objective, showing that joint generation with role prediction best captures human questions and improves readers' memorization and comprehension in a human study. The findings highlight the importance of inter-question relationships and position-aware generation, offering practical insights for crafting engaging, reader-centered texts while noting potential cognitive load and domain-generalization considerations.

Abstract

Using questions in written text is an effective strategy to enhance readability. However, what makes an active reading question good, what the linguistic role of these questions is, and what is their impact on human reading remains understudied. We introduce GuidingQ, a dataset of 10K in-text questions from textbooks and scientific articles. By analyzing the dataset, we present a comprehensive understanding of the use, distribution, and linguistic characteristics of these questions. Then, we explore various approaches to generate such questions using language models. Our results highlight the importance of capturing inter-question relationships and the challenge of question position identification in generating these questions. Finally, we conduct a human study to understand the implication of such questions on reading comprehension. We find that the generated questions are of high quality and are almost as effective as human-written questions in terms of improving readers' memorization and comprehension.

How to Engage Your Readers? Generating Guiding Questions to Promote Active Reading

TL;DR

The paper addresses how in-text guiding questions can promote active reading by examining their use, functions, and impact. It introduces GuidingQ, a 10,577-question dataset from textbooks and research articles, and develops a taxonomy of five question roles to analyze discourse and interaction with readers. It then presents three generation paradigms (Pipeline, Multitask, Joint) with a QA-focused objective, showing that joint generation with role prediction best captures human questions and improves readers' memorization and comprehension in a human study. The findings highlight the importance of inter-question relationships and position-aware generation, offering practical insights for crafting engaging, reader-centered texts while noting potential cognitive load and domain-generalization considerations.

Abstract

Using questions in written text is an effective strategy to enhance readability. However, what makes an active reading question good, what the linguistic role of these questions is, and what is their impact on human reading remains understudied. We introduce GuidingQ, a dataset of 10K in-text questions from textbooks and scientific articles. By analyzing the dataset, we present a comprehensive understanding of the use, distribution, and linguistic characteristics of these questions. Then, we explore various approaches to generate such questions using language models. Our results highlight the importance of capturing inter-question relationships and the challenge of question position identification in generating these questions. Finally, we conduct a human study to understand the implication of such questions on reading comprehension. We find that the generated questions are of high quality and are almost as effective as human-written questions in terms of improving readers' memorization and comprehension.
Paper Structure (39 sections, 6 equations, 8 figures, 16 tables)

This paper contains 39 sections, 6 equations, 8 figures, 16 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Taxonomy of Guiding Questions
  4. GuidingQ Dataset
  5. Source Texts
  6. Construction Pipeline
  7. GuidingQ Analysis
  8. What is the distribution of question roles?$^{\spadesuit}$
  9. Where are guiding questions answered?$^{\spadesuit}$
  10. How are questions related to other text units?$^{\spadesuit}$
  11. Guiding Question Generation
  12. Task Formulation
  13. Data Preparation
  14. Modeling
  15. Pipeline.
  16. ...and 24 more sections

Figures (8)

  • Figure 1: We generate interconnected questions with diverse rhetorical functions during reading to engage readers and improve comprehension.
  • Figure 2: The construction pipeline of GuidingQ. Important outputs are highlighted
  • Figure 3: Distribution of different question roles.
  • Figure 4: Position distribution of different questions. We omit Arouse Interest questions as they are in titles by definition.
  • Figure 5: The average distance (in terms of sentence numbers) between a question and its farthest evidence.
  • ...and 3 more figures