I voted "no" for industrial manslaughter. To the extent such policies have an effect, they divert attention away from safety and towards protection against prosecution.

Simon,
A good post. The paper you're working from is really just a review / opinion piece, but the guy who wrote it, Frank Guldenmund, knows what he is talking about.

The real questions about "Safety Culture", which we should ask about any safety idea:

1. Is there anything meaningful and useful that organisations can do _with_ the concept, that they couldn't (or wouldn't) do _without_ the concept?

2. What's the evidence that these meaningful and useful things are actually good for safety?

Despite its popularity as an idea, safety culture fails these tests. It provides some new measures for safety, but no one really knows what to do with the results of safety culture surveys except pat themselves on the back, or gasp in surprise that their front-line staff give them a different answer to their management staff.

Safety culture provides a few levers to pull for safety, but they are levers that we had even without the concept - leadership, attitude, and behaviour. You don't need the concept of safety culture to tell you that those things matter.

Personally, I find it hilarious that anti-intellectual, "safety is just common sense" practitioners want to dismiss the researchers who are providing evidence and argument to criticise safety culture. Do they think the concept "safety culture" just grew on a tree somewhere? It's a 1970s organisational science theory that was adapted for safety by an Israeli academic in 1980, that became popular after it was referenced in a propaganda fight between the USSR and USA after Chernobyl. But let's not listen to people criticising it,because they're just "academics".

I have a family acquaintance (let's call them "Charlie") who likes to do things for other people. The things Charlie doees aren't what other people would actually want (and are actually sometimes against their express wishes) and usually impose extra work on other people to take part. Charlie then gets annoyed if everyone else isn't grateful for all the effort Charlie put in.

Even with the day described from your own point of view, you're taking away a day of operations to supply people with a set of information that you have decided that they need to have. It doesn't sound like you're being at all discriminating, trying to "cover off" everything, so for any given participant there's a good chance that a lot of the information will be irrelevant to their role, or a repeat of things they've heard before. You're "bringing in" people to do the training, so the quality, consistency, and cohesion of the presentations could be all over the place.

Now, I'm not saying that your day will actually be like this, but that's a typical experience people have of training days. That's what some of them will be primed to expect. The question you should ask is not "Why are they negative?" but "What am I doing to subvert their expectations?" They are entirely right to be skeptical, it's up to you to prove that they are wrong, so that they come up to you afterwards and say "Thanks, that's not at all what I expected. Good job."

A starting point might be to ask them about their experience of previous events, so you can make sure you don't repeat the things that have created their current attitude.

(I'm often one of the people "brought in" to speak at company safety days. The experience is highly variable. Some are planned months in advance by professional event planners who vet every presentation and ensure they blend into each other with a consistent yet interesting message. Others have mix of internal "death by powerpoint" and external "here's my generic talk I give at 200 workplaces a year" with bad AV and poor time co-ordination.)

I think it's worth pointing out that EVEN IF you think that reported injury counts are a good measure of safety, dividing them by another uncertain number (the hours worked) can only ever make them a worse indicator.

Any organisation that genuinely cares about the number of injuries as a statistic should report the raw number, and then, if they think any change is due to a change in exposure, discuss that. As Denise points out, hours worked is a terrible measure of exposure anyway. Who is spending the hours? Where are they spending the hours? What type of work are they doing? Whether you divide by 100,000, 200,000 or 1m, you're claiming that all of those hours are fundamentally the same. It's a factually incorrect claim, and we have a professional obligation not to communicate misleading information about risk.