Defining Interpersonal Trust

Alex Nowrasteh versus two leading experts

In a recent post replying to my speech at University of Chicago’s Freedom of Intellectual Navigation Conference, Alex Nowrasteh of the Cato Institute says that “Garett is confused” on the definitions of interpersonal and generalized trust. He defines the two terms this way:

Interpersonal trust refers to the trust you have in people personally known to you. Generalized social trust is a broader measure from the WVS [World Values Survey], GSS [General Social Survey], and other surveys mentioned below that asks about people in general with the question: “Generally speaking, would you say that most people can be trusted or that you need to be very careful in dealing with people?”

Indeed, that precise survey question has been central to scholarly research on trust; other trust questions are used from time to time but that’s the main one, and it’s routinely treated as a measure of what’s frequently called generalized trust. But as we’ll see, that same question is also routinely treated as a measure of interpersonal trust by leading researchers in the field—entirely contradicting Nowrasteh’s claim that interpersonal trust refers to trust in people you know personally.

First, let’s start with Eric Uslaner, a political scientist at the University of Maryland, and a godfather of modern survey research into trust. In his most-cited paper (with a book of the same name), “The Moral Foundations of Trust,” with over 6600 citations, Uslaner writes:

The interpersonal trust question that has been so important in much research on social capital does reflect generalized trust.

I could just stop right there but let’s continue with the quote, which studies responses to the famous World Values Survey generalized trust question and other trust questions (italics and bold here and throughout added by me):

The Pew Center… asked people whether they trusted eight groups of people–and whether they trusted “most people.” I performed a factor analysis on these trust questions and found distinct dimensions for trust in strangers (people you meet on the street and people who work where you shop) and for friends and family (your family, your boss, and people at your workplace, your church, and your club). The standard interpersonal trust question loaded strongly on the trust in strangers dimension, but not at all with friends and family.

Uslaner not only treats the famous “most people can be trusted” question as a standard measure of interpersonal trust (contradicting Nowrasteh). Uslaner then rules out any link “at all” between that interpersonal trust question and trust in friends and family. A clear contradiction of Nowrasteh’s definition of interpersonal trust.

But, you might wonder, perhaps Alex Nowrasteh of the Cato Institute wasn’t thinking about political scientists when thinking about the standard usage of the term “interpersonal trust.” Perhaps Nowrasteh was only thinking about how economists use the term. So let’s look at some papers co-authored by the economist best known for using the World Values Survey “most people can be trusted” question to study economic growth, World Bank economist Stephen Knack.

In Knack and Zak, “Building Trust: Public Policy, Interpersonal Trust and Economic Development,” (495 cites), interpersonal trust is right in the title. Do they use that term the way Nowrasteh does, to refer to “people personally known to you?” No, of course not. Here they mention statistical results involving interpersonal trust:

Our empirics show that education affects trust in three ways: by raising institutional quality (Table 1), by reducing inequality (Table 3), and directly raising interpersonal trust (Table 4).

And earlier in the paper, they note how Table 4 measures interpersonal trust:

In Table 4, the dependent variable is the percentage of a country’s respondents in the World Value Surveys (sic) who agree that “most people can be trusted.”

So Knack treats the famous World Values Survey trust question as a great way to measure interpersonal trust. Quite the opposite of what Alex Nowrasteh of the Cato Institute claims.

In a far more famous paper, “Trust and Growth” (over 5000 cites), Knack and Zak run statistical tests where they “control for levels of interpersonal trust.” Do they control for a measure of trust in “people personally known to you,” as Nowrasteh would expect? Obviously not: they use the same World Values Survey measure.

The same thing happens in Knack and Keefer’s best-known paper in the trust and growth regression literature, a paper with over 20,000 citations, “Does social capital have an economic payoff? A cross-country investigation.” I won’t provide the relevant quotes here; you’ll just have to trust me on this one. These two trust and economic growth papers by Knack are, incidentally, cited in Nowrasteh’s trust paper published in Kyklos.

In these papers by two leaders in the field of trust research, economist Stephen Knack and political scientist Eric Uslaner, the “standard” trust question is used to measure interpersonal trust. At this point, there’s a clear conclusion: Whether or not Alex Nowrasteh of the Cato Institute is personally known to you, you shouldn’t trust his definition of interpersonal trust.