Do evangelical churches preach about the dangers of empathy? (Not so much)
A new PRRI survey finds most white evangelicals agree empathy can become dangerous. But across 45,000 sermons we find that evangelical preachers barely mention empathy, and almost none criticize it.
The survey: an empathy gap
PRRI’s 2026 American Values Survey asked Americans to choose between two statements: empathy is a foundational moral
value that helps build a healthy society, or empathy can become dangerous when it leads people away from building a
society guided by God’s truth. Most Americans (68%) chose the first. White evangelical Protestants were the only
religious group where a majority (54%) chose the second, and 39% agreed strongly, more than twice the national rate (17%).
Only 30% of white evangelicals strongly agreed empathy is foundational, against 52% of all Americans. PRRI’s founder,
Robert P. Jones, calls this an “empathy gap”.
That may describe people who call themselves evangelicals. It does not describe what is preached in evangelical churches.
We find no direct evidence that evangelical sermons spread the idea that empathy is dangerous. Evangelical churches mostly
don’t talk about empathy at all, while progressive and mainline churches do.
12 / 1,000evangelical sermons use the word “empathy”: 13% of the rate they cite core texts (Baptist: 6%)
48 / 1,000progressive sermons use it, 589% of their control rate; mainline 24 / 1,000 (74%)
≈ 7 of 13,458evangelical sermons frame empathy negatively (“toxic empathy”, “compassion, not empathy”)
67 / 1,000evangelical sermons speak of compassion, the same rate as mainline sermons
Silence, not opposition. Evangelical preachers rarely say “empathy”. They also rarely talk about the Sermon on the Mount or “love your enemies”. When they reach for the idea, they
say compassion, at the same rate as mainline preachers. Negative framing is close to zero in every tradition: between
0.5 and 2 sermons per 1,000. Where the debate over “toxic empathy” does come up, it is mostly progressive pulpits
reporting and answering it.
So where does the belief come from? Probably not from the pulpit. It more likely comes from outside the churches,
in circles that include many evangelicals: most plausibly political environments, such as commentary, media and
campaign messaging, where warnings about “toxic empathy” circulate. A survey measures what people who
identify as evangelical believe; sermons measure what their churches teach. Here the two diverge.
Limits: these are sermons that churches publish online and that Divining transcribes, cut to the sermon itself;
phrases are found by pattern matching, so a preacher could argue the idea without the words we search for. Sermon silence also
can’t rule out teaching in small groups, classes or conversation.
Bottom line: 54% of white evangelicals tell PRRI empathy can be dangerous, yet evangelical sermons almost
never say so. The pulpit is quiet on empathy; the warning against it is coming from somewhere else.
How we measured
Sermons only: scripture readings, prayers and liturgy are cut out using each transcript’s detected sermon span.
Every measure is compared with texts every tradition preaches (John 3:16, the Great Commission, Romans 8:28, Psalm 23): 100%
means as many sermons use it as cite those texts. Higher is more often; the scale is logarithmic; themes run from the biggest difference
between traditions (left) to the smallest. Critical = the preacher’s own negative framing (“toxic empathy”) or a swap
(“compassion, not empathy”); debate = reports the argument without endorsing it. Hover a dot or bar for the counts.
By tradition
progressivemainlineblack churchevangelical
Show the numbers
Tradition
Sermons
Control /1k
The word “empathy”
Reports the empathy debate
Compassion
Sympathy
Feeling-with phrases
Critical of empathy (negative or swap)
progressive
21,738
8
47.9 · 4.8 589%
3.3 · 0.0 41%
230.2 · 43.2 2828%
29.0 · 1.5 356%
11.8 · 0.2 145%
0.8 · 0.0 10%
mainline
8,828
33
24.2 · 2.7 74%
0.8 · 0.0 2%
65.7 · 9.1 199%
18.6 · 0.8 56%
5.3 · 0.0 16%
0.7 · 0.0 2%
black church
1,402
93
14.3 · 2.1 15%
1.4 · 0.0 2%
94.9 · 19.3 102%
18.5 · 0.7 20%
12.8 · 0.7 14%
2.1 · 0.7 2%
evangelical
13,458
93
12.2 · 1.5 13%
0.6 · 0.0 1%
66.7 · 12.9 72%
15.1 · 1.0 16%
8.7 · 0.3 9%
0.5 · 0.0 1%
Per cell: sermons mentioning it per 1,000 · sermons about it (3+ mentions) per 1,000; below, mentions as a share of mentions of the control texts.
By denomination
Denominations with enough sermons citing the control texts to compare (Baptist in gold). Dashed line = as often as the control texts.
Show the numbers
Denomination
Sermons
Control /1k
The word “empathy”
Reports the empathy debate
Compassion
Sympathy
Feeling-with phrases
Critical of empathy (negative or swap)
uu
20,900
8
49.4 · 5.0 608%
3.4 · 0.0 42%
234.9 · 44.3 2888%
30.0 · 1.6 369%
12.2 · 0.2 149%
0.9 · 0.0 11%
christian
6,997
57
15.7 · 1.6 28%
1.0 · 0.0 2%
66.7 · 11.4 117%
14.6 · 0.7 26%
7.3 · 0.0 13%
0.4 · 0.0 1%
baptist
6,279
133
8.3 · 1.4 6%
0.2 · 0.0 0%
66.6 · 15.0 50%
15.8 · 1.3 12%
10.5 · 0.6 8%
0.6 · 0.0 0%
umc
3,226
40
31.0 · 2.8 78%
1.2 · 0.0 3%
68.2 · 10.8 171%
17.0 · 0.6 43%
5.0 · 0.0 12%
1.5 · 0.0 4%
ucc
2,547
20
31.4 · 3.9 160%
0.8 · 0.0 4%
60.9 · 7.9 310%
25.5 · 1.2 130%
5.9 · 0.0 30%
0.4 · 0.0 2%
episcopal
1,996
31
1.5 · 0.5 5%
0.0 · 0.0 0%
49.6 · 2.0 162%
3.0 · 0.0 10%
2.0 · 0.0 7%
0.0 · 0.0 0%
ame
1,402
93
14.3 · 2.1 15%
1.4 · 0.0 2%
94.9 · 19.3 102%
18.5 · 0.7 20%
12.8 · 0.7 14%
2.1 · 0.7 2%
lutheran
595
62
10.1 · 1.7 16%
1.7 · 0.0 3%
67.2 · 16.8 108%
31.9 · 1.7 51%
11.8 · 0.0 19%
0.0 · 0.0 0%
Per cell: sermons mentioning it per 1,000 · sermons about it (3+ mentions) per 1,000; below, mentions as a share of mentions of the control texts.
Left out for too few control-text sermons to compare: unity, catholic, orthodox, jewish, presbyterian, mennonite.
Over time
The chart above, year by year: each point pools 3 years centred on that year, widened where a
tradition has fewer than 30 sermons citing the control texts, so thin years don’t swing the line.