A recent Pew Research Center report found continued evidence for a sharp downturn in how foreigners view the United States. According to Pew, 70 percent of people outside the United States have no confidence in President Trump to do the right thing in world affairs. This negative perception has had repercussions for America’s international image, with views toward the United States at historic lows in many countries — well below the overwhelmingly positive views during the Obama presidency.

But exactly how has foreign public opinion toward the United States shifted in the first two years of the Trump administration? Addressing this question sheds light on implications for U.S. soft power — the ability to influence other countries without power or coercion, which can come from how foreigners view the United States.

Our research, detailed in a new article in Political Behavior, found that Trump does not unconditionally shape foreign opinion of the United States. Instead, citizens abroad react more to the content of a U.S. policy message — whether it is cooperative or uncooperative in nature.

How we did our research

We chose one country to serve as an instructive case study: Japan. Japan has been a target of Trump’s animosity, but its citizens had overwhelmingly positive images of the United States before Trump came into office. This makes Japanese public opinion useful for assessing whether Trump’s rise to power may have damaged foreign perceptions of the United States.

We fielded a survey experiment in Japan from April 26 to May 2, 2017, and collected more than 3,000 responses from Japanese citizens. We asked respondents several general demographic and attitudinal questions, then randomly assigned respondents to read policy message statements from the United States, each roughly 40 words long.

The messages randomly varied along three factors: 1) They were either attributed to Donald Trump or “a U.S. congressman,” 2) they contained either cooperative or uncooperative policy content, and 3) they focused on either a high-salience issue (security matters) or a low-salience one (educational/cultural exchange programs). The “source” of the message — the U.S. president or a generic member of Congress — was how we measured the “Trump effect” in this study.

Here’s an example (translated into Japanese in the actual survey) of what one of our subjects might have read: “U.S. President Donald Trump stated that Japan should pay entirely for its own protection and denounced past U.S. defense spending for the protection of Japan. He also said that the U.S. should not get involved in Japan’s defense policy.”

A cooperative policy message, in contrast, would emphasize helping Japan in the given policy area, strengthening ties and applauding the country’s past cooperative efforts. After exposing respondents to one of the eight resulting policy message combinations, we asked people how favorable they felt toward the United States.

Here’s what we found

There were a number of interesting results from our study, but one stands out: It’s the policy message that most strongly influenced Japanese perceptions of the United States, not whether the statement mentioned Trump.

On average, hearing a message coming from Trump is enough to move Japanese opinion of the United States in a more negative direction. However, the actual substance of the policy message — whether it was cooperative or uncooperative — had a much larger effect on Japanese perceptions. In fact, the effect was about four times the effect of a message coming from Trump.

We looked more closely at the results and found that when the policy message was an uncooperative one, having that message come from Trump considerably worsens Japanese opinion of the United States. But if the message was cooperative in nature, the effect of Trump being the messenger is insignificant. These effects and the difference in their magnitude do not vary much across high or low salience of the issue in question or across various demographics of respondents.

Moreover, these results are consistent with Japanese public opinion after the period of our investigation, which was only months after Trump’s inauguration. After a presidential election campaign filled with hostility toward foreign countries, including Japan, Trump took office in 2017, and Japanese favorability of the United States dropped to 57 percent from 72 percent.

One year later, however, it improved by 10 percentage points. What explains this? We believe this reflects the fact that Trump then became friendlier toward Japan on some occasions, at one point pledging close security cooperation with Japan. According to our study, a shift to a cooperative message is what will shape Japanese opinion toward the United States most strongly — even if the messenger is still Trump, as it was in this case.

So what does this mean?

We interpret these results as evidence of foreign citizens’ inclination to incorporate the clear positive or negative policy provisions and consequences into forming their opinion of the United States. For the U.S. image abroad, the message — not the messenger — is what’s critical.

We think our results have broader implications. Even in the context of a global leader like Trump, someone many analysts consider a polarizing, unusually hostile figure, we find that foreign opinion of the United States does not hinge unconditionally on its leader.

As the large, significant effects of policy content — regardless of the source — indicate, a more favorable U.S. policy approach toward other countries can still improve public opinion of the United States. These results thus suggest that negative statements from Trump have not irreparably damaged global perceptions of the United States.

Alexander Agadjanian is a research associate at the MIT Election and Data Science Lab. Follow him on Twitter @A_agadjanian.

Yusaku Horiuchi is a professor of government at Dartmouth College. Follow him on Twitter @YusakuHoriuchi.