Sharing software engineering research on LinkedIn

LinkedIn is the world’s largest professional network. This makes the platform the ideal place to share and discuss career updates, job vacancies, and career tips.

But LinkedIn can also be used to share software engineering research! Ideally, this leads to a win-win situation: engineers gain useful knowledge that they can use to improve the quality of their work, while researchers reach the audience they intend to support with their research and gain valuable feedback.

How often does this happen and what do such interactions look like?

In this study, the researchers used Google Search to find posts on LinkedIn that mention an ICSE or ESEC/FSE paper or an artifact repository.

Their search yielded a sample of 98 LinkedIn posts by 86 posters, of which roughly 60% had an industry affiliation and 40% had an academic affiliation. Interestingly, most posts are in English and only 61% of those were created by a paper author.

The average post is about one to two paragraphs long. Most posts (83%) focus on a scientific paper, while others focus on semi-related subjects (e.g. winning an award or attending a conference). Posts can be categorised according to their intention (side note: This categorisation is hierarchical, i.e. each intention also includes all intentions that precede it.):

1. self-promotion: the post primarily promotes the author’s own achievements or an organisation

2. paper awareness: the post raises awareness about the existence of a paper and its topic

3. results awareness: the post raises awareness about the results of the paper, e.g. important findings, recommendations or developed tools

4. results discussion: the post presents detailed results, ideally in an easily digestible manner, to discuss their implications or interpretations

Most posts (55%) mainly try to draw attention to a paper. Results are described occasionally (22%), but rarely discussed (11%).

About a third of posts include at least one image. This is often a figure from the paper (N=13), a photo of the presentation (N=11), or a screenshot of the first page (N=7). Rarely do posters go through the effort to find or create images that do a really good job at drawing attention to the post.

Each post in the sample received at least one reaction (side note: Like, Celebrate, Support, Love, Insightful, or Funny). On average, posts received almost 50 reactions, but one extremely popular outlier received 1,185 reactions. Conversely, reposts are considerably less common: over half of posts (54%) did not receive a single repost.

As one would expect, posts from academia attracted more reactions than posts from industry. Similarly, posts from paper authors attract more reactions than posts from non-authors.

The 98 posts received a total of 388 comments, of which a third were from authors of the post or the referenced paper. The study’s analysis only focusses on comments from others, as those are the only actual responses to the post.

Nearly half of the posts have no such comments at all, while almost a third have only received one or two comments. Posts from academia attract slightly more comments than posts from industry. However, most comments (70%) are from industry. This is largely due to the fact that while engineers and other types of practitioners will reply to posts from academics, the opposite rarely happens.

Not all comments are equally useful: more than half are simply congratulations on getting a paper accepted or winning an award. The more valuable types of comments (voicing interest, tagging other users, sharing own experiences related to the paper’s topic, criticism, etc.) mostly appear on posts from industry.

Good posts tend to share several characteristics. Most discuss the results of a paper in a way that’s easy to read (side note: Paragraphs and bullet points!) and understand. Virtually all posts include at least one image. It also helps if the poster has a large network and a paper topic that many engineers find interesting (which may or may not be hard to achieve).

What can the SciCom community learn from this paper?

While there is nothing inherently wrong with self-promotion, paper authors should try to “climb up” the post intention hierarchy, ideally by discussing the results of the paper: what do the results mean or what are their implications for industry? Similarly, commenters from academia would do well to not only congratulate the paper author (side note: Which is mostly certainly a good thing to do!), but actually engage with the topic by voicing support or criticism, asking questions, and discussing its implications.

Normally I’d summarise the paper in a handful of few bullet points here, but instead I’m just going to share this graphic from the paper that does a very good job of describing the results in an attractive way: