A Manifesto for Reality-Based Safety Science: CHAS Safety Journal Club, 9/1/20

The Science of Safety Journal Club met on 9/1/2020 to discuss the paper A Manifesto for Reality-based Safety Science which came out in the journal Safety Science earlier in 2020.

The first two authors of the paper also conduct a podcast titled “The Safety of Work” in which they discuss papers from across the field of safety science. You can listen to their discussion of this paper (~1 hour) by clicking on “Ep. 20 What is reality-based safety science?” at

Below are notes of the opening and general discussion that took place during the Journal Club. We’d love to hear your thoughts about this conversation in the comments below! If you would like information about joining the club or presenting a paper there, sign up on our Google form here Contact Jessica Martin jessica.a.martin@uconn.edu with any questions about this group.

Topic/Paper: A Manifesto for Reality-Based Safety Science

Presenter: Jessica A. Martin

Jessica’s summary of the paper

  • All research programs have a theoretical hard core with a contestable set of auxiliary hypotheses (this statement is based on the ideas in Imre Lakatos’s method of evaluating scientific progress (Lakatos, 1978)
  • The auxiliary hypotheses link the hard core of theory (e.g. Newton’s three laws of motion) with the observed world by providing ways to measure, test and apply the hard core. The auxiliary hypotheses form a protective belt around the hard core – empirical anomalies are accommodated by adjusting the auxiliary hypotheses rather than by rejecting the hard core.
  • A research program is progressive under 2 conditions:
    1. Each new theory must have greater empirical content than its predecessors,
    2. at least some of this novel content must turn out to be true.
  • There is room for a new theory to make wrong predictions, particularly if these can be explained by adjusting the auxiliary hypotheses. However, once a program bogs down in constant adjustment of auxiliary hypotheses to explain away wrong predictions, at the expense of novel true content, the program has become degenerate.
  • To the extent that safety science makes progress, it does so by adopting and customizing progressive research programs from related fields. The problem is that once those programs become part of safety science, they usually cease making progress. 
  • In other words, in terms of novel and confirmed empirical content, “safety science” is usually where research programs come to die.
  • The Hope:
  • We have stopped growing empirically growth – the problem we are facing in safety science stems from a lack of evidence production
  • The Professional Safety field consists of a variety of stakeholders
    • Empiricists: all knowledge is based on experience derived from the senses
    • Theorists: concerned with the theoretical aspects of a subject
    • Practitioners: actively engaged in an art, discipline, or profession
  • To address this challenge, theorists, empiricists, and practitioners need to come together.
  • The Manifesto for Reality-based Safety Science: where theory is grounded in rigorous observations of existing practice, and where practice is based on established theory

Safety Science Research Manifesto

 Notes and Ideas from Breakout Room 1

  • How much safety research aligns with chemical education research is amazing! (we had two Chem Ed researchers in our group)
  • Chem Ed Research does do observations of work
  • Seems that leaving out “main fields” is an issue;
  • Accident investigation involves the “Why did the researcher make that choice”
  • “We” might not be the best safety examples
  • “I understand what I am doing…but others do not fully understand the situation I am in”
  • Observing is not grounds for just doing it yourself
  • Get comfortable with where we are at.
  • Trust, communication
  • Cameras for observation? TALK TO YOUR HR DEPT and  IRB if research is being published
  • Reviewing near miss/safety concerns a good idea for safety research Look at what is reported vs actual accidents at your institution.
  • Know what is HIPAA protected
  • At CSU…we find what is reported with near misses does not really indicate what accidents have ended up happening
  • Theory vs Practitioner (general sense of conversation)·       How to share…peer reviewed not peer reviewed
  • Make safety approachable, equitable, and useful

Points from Breakout Room 2 Discussion

  • Is Safety Science actually a science?
  • Safety practice is often driven by learning from bad experiences
  • Are other scientific approaches more appropriate for safety scientists?
  • Susan Silbey (Professor of anthropology at MIT): studying how new EHS system at MIT works using an ethnographic approach
  • When attempting to “do science” on a safety question, it feels like a “chicken or egg” problem. What are you going to measure? Who decides how the work should be done? And if the theory dictates how the work should be done, then you are using the theory to measure itself – so the logic becomes circular. 
  • You can perform a bunch of incorrect behaviors, and still not get injured.
  • Performing safety assessments in academia are the exception, not the rule.
  • There are many more practitioners than researchers in the field, and what is useful to the researchers is not necessarily useful to the practitioners.
  • Tracking “near misses” can be much more informative than tracking accidents; although even here, what constitutes a “near miss” can be extremely unclear.
  • Collecting data on edge events/accidents can be entirely too complex, making it difficult to get good data.
  • How does one share “safety experience” with the wider community without going through the peer review process that looks for statistical analysis, control groups, etc.

Return to larger group discussion

  • Field is practitioner heavy
  • Safety is driven by individual learning experiences
  • If it’s not being published…then no one is aware…how/where to publish
  • Validating your own theory
  • Is this a science in a proper sense?
    • Should safety science be separated from other sciences?
    • It had gone separate already.
    • Cross field interactions.
  • System. Person. System interacting with the Person.
  • You should not divorce from the other fields
  • Be involved with culture, make sure not just for white, cis-gender, hetero…example of reducing sexual harassment in field campaigns

Do you have an idea for future papers? Contact Jessica Martin jessica.a.martin@uconn.edu

Leave a Reply

Your email address will not be published. Required fields are marked *