Talking about helping others makes a person seem warm and leads to social
approval. This work examines the real world consequences of this basic, social-cognitive phenomenon by examining whether record-low levels of public
approval of the US Congress may, in part, be a product of declining use of prosocial language during Congressional debates. A text analysis of all 124 million words spoken in the House of Representatives between 1996 and 2014 found that declining levels of prosocial language strongly predicted public dis
approval of Congress 6 mo later. Warm, prosocial language still predicted public
approval when removing the effects of societal and global factors (e.g., the September 11 attacks) and Congressional efficacy (e.g., passing bills), suggesting that prosocial language has an independent, direct effect on social
approval.As recently as 2002, public
approval of Congress was reliably over 50% and as high as 84%. In late 2013, though, public
approval reached an all-time low, with less than 10% of Americans expressing support (
1). What caused this dramatic decline in public
approval in just over a decade? One explanation is that the public held Congress responsible for societal and global problems (e.g., a weak economy) (
2,
3). A second explanation is that the public disapproves of ineffective governance. For example, public
approval of Congress tends to drop when Republicans and Democrats are polarized against one another and when Congress conflicts with the President (
4,
5). We test a third explanation that has less to do with action and more to do with talk (
6). We suggest that recent public dis
approval partly resulted from the disappearance of warm, prosocial language in Congressional discourse.Previous experimental research has shown that presenting a warm and prosocial demeanor increases social
approval (
7). People reveal a wealth of information about their feelings and intentions through verbal communication (
8–
10). The speaker’s underlying motives notwithstanding, talking about helping others makes positive impressions upon an audience (
11). We investigated whether this well-documented finding can explain public perceptions of Congress. Specifically, we asked whether the recent rise of public dis
approval of Congress is predicted by declining prosocial language of elected representatives.To measure prosocial language, we computer analyzed all 123,927,807 words spoken in session of the US House of Representatives between 1996 and 2014. Our approach was to look for linguistic markers of prosocial language; we used content analysis software (
12) to calculate the proportion of words in the target text that matched entries in a validated dictionary of prosocial words (
13). We then compared levels of prosocial language within each month of Congress with their
approval ratings by the American public (
14) and found a striking match. shows that levels of prosocial language and the public’s
approval followed the same trajectory between 1996 and 2014,
r(204) = 0.55,
P < 0.001. Notably, the language of both Democrats,
r(204) = 0.53,
P < 0.001, and Republicans,
r(204) = 0.54,
P < 0.001, predicted the public’s
approval of Congress.
Open in a separate windowProsocial language within the US Congress predicts the public’s
approval of Congress. (
A) Prosocial language represents the density of prosocial words in the in-session speeches of members of the US House of Representatives. (
B) Public
approval is Gallup survey data. Individual data points represent monthly scores. Solid lines connect 2-y session means.Public
approval peaked in the aftermath of the September 11 attacks, declined over the next 7 y, rose slightly in the wake of the 2008 financial crisis, and then declined again. Prosocial language followed a nearly identical trajectory. In the years spanning 2002 and 2014, a small (19%) decrease in prosocial language ushered in a large (75%) decrease in public
approval. The individual words whose use most strongly predicted public
approval were as follows: gentle, involve, educate, contribute, concerned, give, tolerate, trust, and cooperate.The correlation between prosocial language and public
approval does not necessarily imply that representatives’ language caused the public to approve of them. The reverse could be true: The public’s
approval could cause changes in the topics that representatives raised. If one variable were causing the other, the causal factor would have changed first and the effected factor second. We tested which variable changed first in time by assessing the association between prosocial language in the present and public
approval with time lags of up to 50 mo in the past and future (). The distribution of the associations across the time lags nearly perfectly fitted a normal distribution curve,
r(98) = 0.97,
P < 0.001. The maximum association between prosocial language and public
approval was at +6.7 mo, meaning that what Congress says today best predicts their public
approval ratings 29 wk into the future.
Open in a separate windowTime-lagged associations between prosocial language and public
approval of the US Congress. How representatives speak today best predicts their public
approval ratings 6.7 mo into the future.Another concern with the present data is the possibility that some exogenous factor (e.g., the September 11 attacks) caused changes in both language and public
approval. We reasoned that, if operative, societal and global factors would also have influenced the US President’s language and/or the economy. Alternatively, dysfunctional governance may cause both politicians’ rhetoric to be less civil and the public to disapprove (
SI Text and
Table S1). To test whether prosocial language has an independent, direct effect on the public sentiment, we ran a regression analysis in which we controlled for the effects of both societal/global factors (in the form of the President’s prosocial language, US unemployment rate, and US consumer expectations about the economy) and competent governance (in the form of partisan conflict, the number of bills that Congress passed, and Presidential vetoes). Even with these conservative controls, prosocial language within Congress still predicted the public’s
approval (