In Whom We Trust

In a previous post, I said that Jim Rosenau and I underestimated the disruption to American politics from people going online. I want to expand on that, to consider whether we misunderstood the way people use the Internet, and how that affects social trust. The upshot: probably not.

Jim and I saw the Internet as a profound tool which people can use to improve their lives. Our premise was that most people would use it as such — a premise colored by our own use of the Internet as professionals committed to reason and knowledge. That commitment requires active use, however, to filter useless or bogus information.

What Jim and I did not appreciate is that not everyone uses the Internet the same way. Some — probably most — people use it much more passively, primarily for entertainment. More accurately: we knew that people use the Internet as entertainment, but failed to appreciate how passive use could matter. This may have been a mistake, at least where our research speaks to the 2016 election.

Insofar as it is an entertainment medium, the Internet might be much more like television than we anticipated. This is an important comparison: in a lecture in 1995, putnamRobert Putnam argued that the decline in social capital in the United States was due to television, at least in part. Television sucks up time that would instead be used for social and civic activity, leaving the viewer isolated and alone instead. Television also affects how people see the world around them, making them less trusting and more afraid. Putnam — in this photo from 1995, looking like he just murdered his family — expanded this argument in his famous book, Bowling Alone.
In the book, he includes the Internet with television as causing the decline in social capital.

Putnam used social trust and civic engagement as indicators of social capital. Here is a chart from the 1995 article based on his lecture, showing how social trust, newspaper reading, group membership, and voting vary based on age. The x-axis is year of birth: people born in years up to 1930 have high rates of all four, but these rates drop off sharply for people born from 1945 onward.  putnamtrust

Putnam said that this trend correlates well with the uptick in television viewing, and he found little evidence that other possible culprits explain the data. One of the consequences he pointed to was the ‘mean world effect’ — that television promotes pessimism of human nature, and a view of the world as more violent than it really is (this is also called mean world syndrome). Just offhand, I would guess this is likely a worse problem for the Internet, for those who are passive users.

For a rough look at whether the Internet is in fact making things worse, I went back into the data. I made this graph from the same General Social Survey data that Putnam used for some of his data two decades ago — it’s still being updated regularly. (I did not apply any smoothing to the graph; Putnam used 5-year moving averages). Keep in mind that this is 2014 data (voting for the 2012 election), where Putnam used 1994 data (the 1992 election). So, for example, where 70% of respondents born in 1930 voted in 1992, 100% of them voted in 2012. Granted, 2012 was a historic election, with unusual turnout across the board.  socialtrust

For someone born in 1940, social trust was around 40% in 1994, and about the same in 2014; for a person born in1950 it was 35% in 1994, but twenty years later it was just over 40%. For someone born in 1960, it was just under 30% in 1994, but over 40% in 2014. If anything, social trust has increased for a given age cohort in the last two decades. This is not definitive, but it surely does not support the idea that the Internet is making things worse. Which is a relief for Jim and me — it means we did not totally miss the forest.

Eric Uslaner, who has studied the problem far more than I have, says pretty much the same thing. From this graph, published in a Pew Research Center report, he argues that the disparity in trust between millennials and other generations is due to economic inequality. Overall, it’s hard to say from these graphs that social trust is significantly decreased from its 1987 levels (although, it turns out, there is research that shows it is in decline).


More significant than any year-to-year trend is the generational effect. It is clear that millennials trust less than Gen X, who trust less than Boomers & the Silent Generation. Uslaner specifically discounts the suggestion that Internet usage makes a difference between millennials and any other generation.

It is still possible that the difference between active and passive use of the Internet affects social trust; however, there is not enough difference from generation to generation in who uses the Internet actively and who uses it passively to show in the data. We cannot blame that difference for generational differences in social trust.

I would also point out that the study Uslander cites does not adequate address the consequence of coming of age in a post 9/11 America. Considering that millennials have been fed a steady diet of fear and xenophobia, it would be hard to imagine them coming to adulthood with a great deal of social trust.

How do we improve social trust? Obviously, fix the economy — specifically with respect to inequality. Second, civic engagement of the sort Putnam proposes. It is heartening to see so many people across generations engaged in activism, although those people were likely high social-trust voters in the first place.

Third, I do think there is a role for the Internet in increasing social trust, especially when it comes to combating harassment and bullying online, and helping cement social norms against that sort of behavior. I also think we are just beginning to renegotiate the sources of authority online: in lay terms, we’re just starting to deal with the fake news problem.

Low social trust is a crucial problem for our democracy. While it is low, it is not the lowest it has ever been. The Internet is — sigh of relief — not making things worse, as far as we can tell.

Leave a Comment

Your email address will not be published. Required fields are marked *