The Influence of Text Variation on User Engagement in Cross-Platform Content Sharing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hu, Yibo, Jin, Yiqiao, Ye, Meng, Divakaran, Ajay, Kumar, Srijan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915275674222592
author Hu, Yibo
Jin, Yiqiao
Ye, Meng
Divakaran, Ajay
Kumar, Srijan
author_facet Hu, Yibo
Jin, Yiqiao
Ye, Meng
Divakaran, Ajay
Kumar, Srijan
contents In today's cross-platform social media landscape, understanding factors that drive engagement for multimodal content, especially text paired with visuals, remains complex. This study investigates how rewriting Reddit post titles adapted from YouTube video titles affects user engagement. First, we build and analyze a large dataset of Reddit posts sharing YouTube videos, revealing that 21% of post titles are minimally modified. Statistical analysis demonstrates that title rewrites measurably improve engagement. Second, we design a controlled, multi-phase experiment to rigorously isolate the effects of textual variations by neutralizing confounding factors like video popularity, timing, and community norms. Comprehensive statistical tests reveal that effective title rewrites tend to feature emotional resonance, lexical richness, and alignment with community-specific norms. Lastly, pairwise ranking prediction experiments using a fine-tuned BERT classifier achieves 74% accuracy, significantly outperforming near-random baselines, including GPT-4o. These results validate that our controlled dataset effectively minimizes confounding effects, allowing advanced models to both learn and demonstrate the impact of textual features on engagement. By bridging quantitative rigor with qualitative insights, this study uncovers engagement dynamics and offers a robust framework for future cross-platform, multimodal content strategies.
format Preprint
id arxiv_https___arxiv_org_abs_2505_03769
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle The Influence of Text Variation on User Engagement in Cross-Platform Content Sharing
Hu, Yibo
Jin, Yiqiao
Ye, Meng
Divakaran, Ajay
Kumar, Srijan
Social and Information Networks
Artificial Intelligence
Information Retrieval
In today's cross-platform social media landscape, understanding factors that drive engagement for multimodal content, especially text paired with visuals, remains complex. This study investigates how rewriting Reddit post titles adapted from YouTube video titles affects user engagement. First, we build and analyze a large dataset of Reddit posts sharing YouTube videos, revealing that 21% of post titles are minimally modified. Statistical analysis demonstrates that title rewrites measurably improve engagement. Second, we design a controlled, multi-phase experiment to rigorously isolate the effects of textual variations by neutralizing confounding factors like video popularity, timing, and community norms. Comprehensive statistical tests reveal that effective title rewrites tend to feature emotional resonance, lexical richness, and alignment with community-specific norms. Lastly, pairwise ranking prediction experiments using a fine-tuned BERT classifier achieves 74% accuracy, significantly outperforming near-random baselines, including GPT-4o. These results validate that our controlled dataset effectively minimizes confounding effects, allowing advanced models to both learn and demonstrate the impact of textual features on engagement. By bridging quantitative rigor with qualitative insights, this study uncovers engagement dynamics and offers a robust framework for future cross-platform, multimodal content strategies.
title The Influence of Text Variation on User Engagement in Cross-Platform Content Sharing
topic Social and Information Networks
Artificial Intelligence
Information Retrieval
url https://arxiv.org/abs/2505.03769