Garett is wrong on *every* citation in his 🧵. I replied to each one, but now I’ll do so in a 🧵 so my responses aren’t lost. My criticism of the trust-growth literature focuses on “generalized social trust” – the trust measure from the WVS and GSS that is most prominently and
Garett wrote, “Here he cites 8 articles as evidence that survey measures of trust don't predict experimental trust. Four articles, highlighted in yellow, actually report that trust predicts trust! +” If it's not clear, the "He" is me. Garett is wrong about *every* citation. Now
Ermisch, Gambetta, Laurie, Siedler, & Uhrig (2009) wrote: "The study also asks the typical survey question aims to measure trust, showing that it does not predict 'trust' as measured in the experiment." By "typical survey question," they mean the question on generalized social
Garett relies heavily on Algan and Cachuc (2013) in his book. A&C (2013) cites Karlan (2005), an important paper testing generalized social trust in an experimental setting, among other forms of trust. See excerpts from the A&C (2013) paper: https://x.com/GarettJones/stat...
What does Karlan (2005) write about generalized social trust? He writes: "[T]he GSS questions predict default, or trustworthy action, but fail to predict savings, or trusting actions."
Garett’s quote “Our new scale,” from the Näf & Schupp (2009) paper here should have been a giveaway that the authors are not talking about generalized social trust as measured in the WVS and GSS. Here is what Näf & Schupp (2009) wrote about THE trust question that is the focus
Here is what Gächter, Herrmann, & Thöni (2004) wrote about the GSS and WVS measure of generalized social trust: "Trust as measured by the popular GSS trust question is not significantly correlated with cooperative behavior." https://x.com/GarettJones/stat...
Garett is really scrounging the bottom of the barrel here. Significant at the 10% level is all they've got. That’s a weak reed indeed, especially given the caveats in the actual paper. Consistent with what I wrote, Sapienza, Toldra-Simats, & Zingales (2013) wrote: “In summary,
If Garett were as rigorous as he claimed, he would have spotted this paper. @ho_ben let me know about it. It’s an interesting paper that relies on the GPS, an experimentally validated survey data that is different from the GSS and WVS survey. The WVS data https://academic.oup.com/qje/a...
Garett did not conduct a rigorous, thorough, or precise reading of the trust literature for his book. He presents himself as a very careful scholar, but his misreading of this literature shows otherwise.
The trust-growth literature is detailed and nuanced. If he writes about it trust-growth again, I’d happily volunteer several days of my time to thoroughly reviewing his manuscript prior to publication to correct mistakes and keep him on topic.