Factors Affecting Response Quantity, Quality, and Speed for Questions Asked Via Social Network Status Messages

Jaime Teevan, Meredith Ringel Morris, Katrina Panovich

Abstract

Social networking tools enable people to easily ask questions of large groups of personal acquaintances, but the effectiveness of doing so depends in part on how the question is phrased. In this paper we present a study in which 282 participants posted variants of the same question as their status message on Facebook. We analyze the quantity, quality, and speed of the responses each variant received. We find that by ending an information need with a question mark, explicitly scoping the audience, and being succinct, a person can increase the likelihood of quickly receiving many high-quality answers.

Introduction

While social networks have become a primary hub for "Social Q&A," the effectiveness of these requests is often inconsistent. Previous research had observed correlations between phrasing and response rates, but these studies were largely observational—looking at existing data where the topic, the asker’s reputation, and the phrasing were all confounded.

Researchers at Microsoft and MIT sought to isolate the impact of phrasing itself. They wanted to know if the linguistic structure of a request—independent of the person asking or the topic being discussed—could be optimized to elicit better help from one's social circle. By conducting a controlled experiment with 282 Facebook users, the authors demonstrate that seemingly minor tweaks to a status update can significantly boost the speed, quantity, and quality of the answers received.

Methods

The researchers designed a controlled field study involving 282 participants (primarily Microsoft employees). Each participant was assigned one of 12 specific variants of a question about the movie E.T.. The study varied the questions along three axes:

  1. Punctuation: Explicit question (ending in "?") vs. implicit statement (ending in ".").
  2. Length: One sentence vs. two sentences (adding the filler "Taking it easy.").
  3. Scope: No scoping vs. generic scoping ("anyone") vs. targeted scoping ("movie buff friends").
The 12 question variants used to test the impact of punctuation, length, and scoping.

Participants posted their assigned text verbatim to Facebook and provided screenshots of the responses three days later. The researchers then manually coded these responses for "directness" and "usefulness" (providing extra value like trivia or alternative suggestions).

An example of a participant's Facebook status update and the resulting comment thread.

Key Insights

  • The "Question Mark" Signal: Phrasing a request as an explicit question (with a "?") yielded an 88.1% response rate compared to 76.3% for statements. A question mark acts as a powerful social trigger that distinguishes a request for help from a mere status broadcast.
  • The Succinctness Premium: Adding conversational filler like "Taking it easy" actually hurt performance. Single-sentence questions were answered faster (68 seconds vs. 115 seconds on average) and were much more likely to receive a direct "yes/no" answer. Extra context seems to bury the "ask."
  • The "Anyone" Paradox: Even though "anyone" is a content-less scope, including it improved response metrics compared to no scoping at all. It likely triggers a psychological "call to action" that reduces the bystander effect.
  • Scoping Trade-offs: Targeting "movie buff friends" led to the highest quality and longest responses, but it also prompted some friends to post "disclaimers" (e.g., "I'm not a movie buff, but..."), potentially creating a social barrier for non-experts.

Key Results

The study found statistically significant advantages for specific phrasings:

  • Quantity: Questions received more responses (3.41) than statements (2.96). Single-sentence posts received significantly more replies (3.68) than two-sentence posts (2.54).
  • Quality: 80.4% of responses to explicit questions were deemed "useful," compared to only 63.3% for statements.
  • Speed: The time to the first response was nearly twice as fast for single-sentence questions (1:08) compared to two-sentence questions (1:55).
  • Scoping: Addressing "movie buff friends" resulted in the longest average response length (63 characters) and the highest percentage of "useful" information (78.6%).

Limitations and Future Work

The authors acknowledge several limitations. The participants were Microsoft employees, which may not represent the broader global population. Additionally, the study only used one movie (E.T.); results might differ for high-stakes or highly technical topics. Finally, the study was conducted in 2011 on Facebook; modern algorithmic feeds might react differently to these phrasing variants today. Future work could explore how these factors interact with the asker's reputation or how they translate to newer platforms and specialized Q&A communities.

Summary updated

Abstract

Social networking tools enable people to easily ask questions of large groups of personal acquaintances, but the effectiveness of doing so depends in part on how the question is phrased. In this paper we present a study in which 282 participants posted variants of the same question as their status message on Facebook. We analyze the quantity, quality, and speed of the responses each variant received. We find that by ending an information need with a question mark, explicitly scoping the audience, and being succinct, a person can increase the likelihood of quickly receiving many high-quality answers.
BibTeX
@inproceedings{teevan2021factors,
  author = {Jaime Teevan and Meredith Ringel Morris and Katrina Panovich},
  title = {{Factors Affecting Response Quantity, Quality, and Speed for Questions Asked Via Social Network Status Messages}},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year = {2021},
  doi = {10.1609/icwsm.v5i1.14159},
}
Figure