You know that feeling you get, as a fundraiser, when someone who isn’t a fundraiser starts talking about fundraising? We’ve all endured the lecture on how our totally unqualified friend/stranger would do our job. Annoying right?

Well, as a behavioural scientist, that’s how I feel whenever I hear someone, who isn’t a behavioural scientist, talk about applying that science to fundraising. It hurts! Not because I don’t welcome the interest in the field, but because what I hear is almost invariably wrong.

The Notre Dame fire is a classic example. So many have cited the huge public response as evidence of the ‘identifiable victim’ effect (that we respond better to one person than many).

Let’s put aside the fact that Notre Dame is a world-famous place of worship and tourist attraction, not a person. And we’ll also put aside the fact that when a world-famous place of worship and tourist attraction catches fire, it gets enormous free media.

The danger isn’t just that those selling you the ‘identifiable victim’ can’t even define it properly. It’s that they’re not even aware that the effect isn’t anywhere near as powerful as they make it out to be.

Even if Notre Dame was an example of the identifiable victim effect – which is not, not even a little bit – it still wouldn’t explain the huge response.

A meta-analysis of various studies on the identifiable victim effect examined its effectiveness and boundary conditions. The outcome isn’t anything like what you’re being mis-sold. The truth is the overall effect is small, at best moderate, but definitely not strong.

What’s more, the effect is not  even always present. It only seems to be significant when certain conditions apply:

• the victim is a child, not an adult, not a church
• there’s a single identified victim, not a group of identified victims (singularity effect)
• there’s a picture of the victim along with name & age, not when there’s no picture
• the victim’s source of need is poverty
• the victim is not seen as responsible for their plight
• monetary contributions were included (phew!)

So, what does this mean for you?

The effect is still valid and it’s not unimportant for practical application. You could benefit from using an identified victim with personal information. But don’t expect a miracle. You’ll get, on average, .10 Standard Deviation more support than when using anonymous or statistical victims. That’s it! You will see a trickle;  don’t expect a flood.

Expecting this and any other effect to skyrocket your revenue takes away the importance of its actual impact. Instead of acknowledging its moderate contribution, we might even conclude it doesn’t really work.

What about charities that don’t fit the above criteria like animal or conservation charities? Based on the study I mentioned above, talking about a single dog, or a single tree, or a single building, won’t do anything. Does this mean you should go all in with statistics? Not necessarily. Maybe the identifiable victim effect isn’t present, but what about psychic numbing (in short, that numbers numb any emotions of sympathy or empathy and lead to inaction)? This could still be at play and result in reduced giving.

If you want a scientist’s advice, then I say test it for yourselves. But please test it properly.

If you are interested in testing it, but are unsure about the best design or execution, don’t hesitate to get in touch. Testing is important but testing the right way is crucial. If you want to trust your findings, that is.

Kiki

P.S. If you’d like to learn more about this and many other behavioural science techniques, Lauren Merrill of Catholic Relief Services and I are doing a webinar together called Beyond the nudge: putting deep behavioral insights to work for you on June 11 at noon Eastern/9 AM Pacific/5 PM UK/6 PM Central European Time.