Failure to Replicate the "Identified Victim Effect"

We may need to rethink and reframe charitable-giving biases.

Reviewed by Michelle Quirk

Key points

• The identified victim effect is considered a seminal robust finding.
• We failed to replicate several seminal articles on the identified victim effect.
• We need independent, preregistered, well-powered replications and comprehensive meta-analyses.

This post was written by Rajarshi Majumder, a Ph.D. student from Grenoble Ecole de Management who led the replication of the classic Kogut and Ritov (2005) study discussed in this post. Gilad Feldman was the project coordinator, is the corresponding author on the joint article, and edited this post for Psychology Today.

The Identified Victim Effect

Imagine coming across a heart-wrenching story about a young girl named Yelena, who is battling a rare, life-threatening disease. However, without prompt financial support to raise an astronomical amount of $500,000 for medication, she will face fatal consequences. Her image captures your attention and evokes empathy for Yelena, as well as a strong desire to respond to her plight by contributing to the charity that is raising funds on her behalf.

Now, consider a different scenario: a story detailing the plight of a child affected by the same rare, life-threatening disease and a plea for the same urgent financial assistance—however, without their name and picture. Previous research suggested that the second scenario does not elicit the same emotional response as Yelena’s. This is captured by a classic phenomenon called the “identified victim effect” (Small & Loewenstein, 2003), which commonly refers to the heightened willingness to contribute to victims whose identities are clearer, presented with details such as their first name or a picture, compared to unidentified victims.

Revisiting the Impact of Singularity on the Identified Victim Effect

In a seminal follow-up article, Kogut and Ritov (2005) investigated a boundary condition and the mechanisms underlying the identified victim effect and found that effect the was stronger when the target of contribution was a single victim than when the target was a group of victims, both for willingness-to-contribute and for empathic emotions. Kogut and Ritov (2005) argued that the reason is that people become psychologically numb, their attention loses focus, and their empathy loses intensity as the number of victims increases in attempting to avoid experiencing emotional distress.

Lee and Feeley (2016) conducted a meta-analysis of the literature on the identified victim effect and concluded some evidence in support of the effect (r = .05, which is commonly considered weak). Maier et al. (2023) reanalyzed that meta-analysis with updated methods examining the possibility of publication bias and concluded that there was “moderate evidence of publication bias and strong evidence for the absence of an identified victim effect.” Maier et al. (2023) also reported a further unsuccessful replication failing to find support for the seminal findings by Small and Loewenstein (2003), who were among the first to demonstrate the identified victim effect. In our search of recent literature and our interaction with scholars in this domain, we found several other failed replications of the effect.

This highlights the importance of conducting preregistered, well-powered, independent replications of seminal findings in psychology, which motivated our current replication.

Our Key Findings: Failed Replication of Kogut and Ritov (2005)

In Majumder et al. (2024), recently accepted for publication in the journal Judgment and Decision Making, we conducted a close replication and further extended the replication by adding a group-belonging manipulation as a conceptual replication of a different article by Kogut and Ritov from 2007. Our results were largely inconsistent with the original results. We found no support for effects or interactions between singularity and identifiability on willingness to contribute, distress, or empathy as single and groups of victims evoked similar distress, empathy, and willingness to contribute. This means that, identified or unidentified, single or group, people indicated donating about the same. We did, however, find support for the effect of group-belonging with lower willingness to contribute, distress, and empathic concern for identified outgroup victims compared to unidentified, identified, and identified ingroup victims.

I believe that our findings are indicative of puzzling decision-making. With a larger number of victims, we would expect that solicitations would elicit more intense emotions and more willingness to contribute since there are more individuals impacted by the same tragedy. Decisions regarding the allocation of resources for providing assistance to humanitarian crises are partly influenced by individuals' perceptions of the needs of the victims. The absence of differences in allocation of resources between single and groups of victims is perplexing and offers a nuanced perspective of decision-making for altruism and charitability. It is not that participants contributed the maximum possible amount, even if it is a hypothetical donation, but rather that they do not seem to make adjustments to their allocated donation amount depending on identifiability or singularity. If true, and robust, this holds practical implications for many aspects of public policy and advertising campaigns run by charities, nonprofits, and governments hoping to elicit the public’s help.

I joined the replication project in 2022. I was aware of the prior replication projects conducted by Gilad Feldman and my supervisor, Ignazio Ziano, and I personally believed that these mass replication projects are important, as they involved not only revisiting and expanding upon a seminal work in the decision-making science by confirming (or challenging) the previous findings but also broadening the comprehension of these psychological phenomena. I think it is important to acknowledge that our unsuccessful attempt to replicate one of the classic psychological phenomena highlights the necessity for independent, well-powered, preregistered, close replications.