Pew asked people who used online dating about four unwanted behaviors in 2022. 48% said they faced at least one. Among women under 50 who used online dating, the share was 66%.
On this page
The 48% is a count of people, not incidents
Pew Research Center asked U.S. adults who had ever used a dating site or app whether four specific unwanted behaviors had happened to them there. In July 2022, 48% said at least one had. The figure counts each respondent once, even if several behaviors happened or one behavior happened repeatedly. It is therefore a share of past users who reported at least one of the measured experiences. It is not the share of dates, messages, profiles, or people responsible for the behavior. The question covers experiences at any time during someone's online-dating use, so it does not supply a current-year incident rate.
The four questions were about sexually explicit messages or images the person had not asked for, contact that continued after they said they were not interested, being called an offensive name, and a threat of physical harm. The wording matters. It names conduct rather than asking whether the respondent considers themself a harassment victim. Someone's label for an event may differ from the survey's list, while their answer to the specific event question still counts. Pew did not measure every kind of abuse that can happen on or after an app encounter. The combined 48% has the limits of this four-item list.
Why the four percentages cannot be added
Among past users in Pew's 2022 survey, 38% reported unwanted sexually explicit messages or images; 30% reported continued contact after saying they were not interested. Twenty-four percent said someone called them an offensive name, and 6% reported a threat of physical harm. Each percentage uses the past-user group as its denominator, but the groups of affected people overlap. A respondent who received unwanted content and later faced repeated contact belongs in two behavior rows while contributing only once to the 48% combined row. Adding 38, 30, 24, and 6 would count that person more than once.
The 48% combined share is lower than the sum of the four rows because of that overlap. The published table does not show exactly how many people experienced every pair of behaviors, so we cannot turn the gap into a precise overlap count. The 6% threat figure is also not a footnote to the more common categories. Threats are a distinct and serious event, and their lower share should not make them seem minor. Read each behavior row for its own question, then use the combined row only when the claim is about experiencing any one of the four.
Compare the four behaviors
The table may be wider than the screen. Focus it and use the arrow keys to reach every column.
| Behavior | All past users | Women past users under 50 |
|---|---|---|
| Unwanted sexual messages or images | 38% | 56% |
| Contact after saying no | 30% | 43% |
| Called an offensive name | 24% | 37% |
| Threatened with physical harm | 6% | 11% |
| At least one of the four | 48% | 66% |
Who got counted: Each column uses people in that group who had tried online dating. The groups overlap; women under 50 are part of all past users.
How to read it: One respondent could report several behaviors. Read the final row as people affected by at least one, not as a sum of the rows above.
Download the table (CSV) ↓Women under 50 reported a different pattern
Pew found that 66% of women under 50 who had ever used online dating reported at least one of the four behaviors. Within that same group, 56% reported unwanted sexually explicit content, 43% reported continued contact after saying no, 37% reported offensive names, and 11% reported threats of physical harm. These shares are within women under 50 with past online-dating experience. They are not percentages of all women under 50 or of all dating-app users. The subset and the all-user group overlap, so comparing their combined 66% and 48% describes a group difference within the same survey.
A group pattern does not describe every person in the group. It also does not tell us which services were used, how often a behavior occurred, or whether it happened recently. Those missing details matter if a report claims a particular platform solved or caused the difference. Pew did not assign respondents to platforms and did not test an intervention. The finding still makes the uneven burden visible: the respondents in this group were more likely to report the listed conduct than past users overall. State both the group and the time frame when quoting it.
Other group results need the same care
The 2022 Pew report also found that 64% of lesbian, gay, or bisexual past users reported at least one of the four behaviors, compared with 45% of straight past users. Again the denominator is past users in each group, and the survey groups can overlap with age and gender groups. This comparison does not identify which identity caused an experience or which app was involved. It reports how people in the named groups answered the same set of questions. Pew combined lesbian, gay, and bisexual respondents for this analysis because smaller categories could not be reported with useful precision here.
Some individual behaviors also differed by group. Pew reported unwanted sexual messages or images among 56% of lesbian, gay, or bisexual past users and 34% of straight past users. That specific contrast should not be substituted for the combined 64% and 45%, which cover any of four events. The survey has a national design and weighted responses, but breaking its users into smaller groups widens uncertainty. A careful quote keeps the survey year, the past-user condition, and the behavior named. Leaving any one out can make a true statistic sound broader than it is.
Pew reported further differences for individual behaviors in 2022: 47% of Black past users and 35% of White past users said they had received unwanted sexual messages or images. For threats of physical harm, the shares were 11% among Hispanic past users and 4% among White past users. These are comparisons within named past-user groups, not counts of every Black, Hispanic, or White adult. The report did not find racial or ethnic differences for every listed behavior. Small subgroup estimates require care, and the figures do not identify why respondents' experiences differed. They do show why an all-user average can conceal uneven exposure to a specific behavior.
Behavior, perception, and crime reports are different data
In another Pew question, 48% of all U.S. adults called online dating safe in 2022. The identical number does not mean the same people answered the two questions or that half of unsafe encounters were harmless. One 48% uses all adults and measures an opinion; the other uses past users and counts at least one listed experience. A person can report an unwanted message while still seeing online dating as generally safe. The published aggregates do not link those answers person by person on this page. Keep the harassment statistic with its past-user denominator.
The four survey items are also not police-report counts. Many experiences are never reported to police, and Pew asked about conduct rather than criminal classifications. The survey does not measure the number of perpetrators or establish how many users faced a prosecutable threat. It also cannot tell whether blocking, reporting, or platform moderation resolved an incident. Those are separate questions that require different evidence. Treat the table as a precise summary of what respondents said they experienced, within the four behaviors Pew asked about.
What someone can do with the finding
RAINN's safer-dating guidance advises ending contact, blocking, and reporting troubling behavior through the service. For an in-person meeting, it also suggests a public place and independent transport. Those are options a person can use; they do not make the harmed person responsible for someone else's actions. If a threat is immediate, seek help from local emergency services. The survey cannot tell an individual whether a particular person will behave safely, but it supports taking reports of unwanted conduct seriously and making reporting tools easy to find.
The evidence table below keeps all four behaviors and the combined result visible for the full past-user group and for women under 50. Choose the row that matches the claim you need to check. If you are looking for broader safety attitudes, the linked safety article covers Pew's separate opinion questions. If you are looking for a response to conduct you are facing, use the service's current block and report controls and trusted support. A national survey can describe the scale of reported experiences without deciding what happened in one person's case.
