Perception and Reality
Understanding Crime Concerns in the United States
April 2026
By Ernesto Lopez, MPA, Council on Criminal Justice, and
Amanda Graham, PhD, University of Texas, San Marcos
While partisan rhetoric has long dominated the nation’s public safety discourse, Americans have been left to wonder: Is crime actually going up or going down? If it’s dropping, why do so many people believe it isn’t, or remain afraid of becoming a victim?
This report uses data from the Gallup Social Survey to examine what factors aside from crime itself, such as the racial diversity of one’s neighborhood and feelings of financial optimism, may influence perceptions of crime and fear of crime. It also explores how those perceptions and fears have changed over time, with some data dating back to 1965, and advances prior research with new findings on how political ideology shapes crime concerns.
The report is the first in a series of research projects by the Council on Criminal Justice (CCJ) that sheds light on what drives public perceptions of crime and what might help explain discrepancies between those perceptions and official crime data. CCJ continues to delve deeper into explaining crime concerns by examining what particular crimes people are afraid of, what makes people feel safe, and how additional factors, such as disorder, general anxiety, and media consumption, are associated with crime concerns.
Key Takeaways
- Changes in national perceptions and fear of crime have not aligned with shifts in overall national crime rates in most years since the mid-1960s, but crime concerns have responded to major shifts in the homicide rate.
- Property crime rates and household victimization exert more influence on crime concerns than violent crime rates.
- Political partisanship does not significantly account for people’s fear of crime or perception of crime levels once all variables, such as county-level crime rates, respondent characteristics (e.g., age, race), and neighborhood characteristics (e.g., poverty), are included in research models. On average, however, people who are more politically conservative believe crime is higher and are more fearful of it.
- Higher presidential approval ratings correlate with the perception that crime is the same or lower than the prior year, both nationally and locally. The same is true for congressional approval ratings.
- People with a more optimistic financial outlook generally believe crime rates are stable or dropping and are less afraid to walk alone at night.
- An analysis of these political and economic factors revealed that their effects on crime perceptions and fears have generally been consistent over time.
- Neighborhood characteristics such as poverty and the share of population under 18 seem to explain the likelihood of being afraid of crime. But they do not generally appear to influence perceptions that national or local crime is higher than the prior year.
Measuring Crime Concerns
The terms perceptions (perceptions of crime) and fear (fear of crime) are used throughout this report. These concepts are distinct from one another and represent only two measures within a broader range of crime-related concerns.1 The term perception refers to a person’s belief about whether the actual level of crime has increased, stayed the same, or decreased.2 Crime perceptions in this study are captured at the national and local level. Fear, on the other hand, is a psychological response typically centered on personal vulnerability.3 When discussing both perceptions and fear, the report uses the term crime concerns or crime sentiments.
Key Crime Concerns Measures Used in this Report
- National Crime Perceptions: “Is there more crime in the U.S. than there was a year ago, or less?”
- Local Crime Perceptions: “Is there more crime in your area than there was a year ago, or less?”
- Fear of Crime: “Is there any area near where you live—that is, within a mile—where you would be afraid to walk alone at night?”
Please see the supplemental methodology report for the full list of variables, definitions, and coding used in the analysis.
Introduction
It has long been the case that citizens perceive crime as rising even during prolonged periods of decline. But recently, the mismatch between perceptions and the reality of how many crimes are reported by law enforcement has become an issue of extraordinary political significance. Leading Republicans have painted a picture of cities plagued by runaway crime that requires urgent crackdowns. Many Democrats, by contrast, argue that data showing downward crime trends make such drastic enforcement actions unnecessary. Research by CCJ has found that many types of crime have been falling in recent years. FBI data will likely confirm that homicide rates in 2025 were the lowest they’ve been since 1900.
Much has been written about perception and fear of crime, including how such concerns are influenced by personal victimization or neighborhood disorder. There is some evidence, too, that broader social anxieties, such as those surrounding the economy, may influence how people feel about crime.
But there is limited research on the role that broader national influences, such as presidential and congressional approval levels, partisanship, and ideology, play in shaping crime concerns. In past research, the relationship between political measures and crime sentiments has generally focused on punitive attitudes or policy preferences. One example: studies investigating the relationship between political party and support for the death penalty.4
Given today’s highly polarized climate, examining how broad national factors shape perception and fear of crime provides insight into how widely publicized political opinions on public safety may amplify or diminish Americans’ concerns, thereby influencing support for a wide array of criminal justice policies.5
This report has two main sections:
- The first section presents long-term crime sentiments and crime rates over time. It uses national opinion survey results dating back to 1965 to track how changes in crime concerns have moved with changes in crime.
- The second section examines the relationship between crime concerns, individual characteristics, and features of respondents’ local area (zip code and county) using data from 2009 to 2021.
Each section relies on different data and research methods. Details covering the data, methodology, and regression output can be found in the supplemental methodology report.
Long-Term Trends, 1965 - 2024
Key Takeaways
- The share of people who believed the national crime rate rose from one year to the next decreased throughout the 1990s and generally stayed steady—averaging about 69%—from 2005 through 2024.
- The share of people who believed crime was up in their area followed a similar trend. On average, however, about 20% more people believed crime was increasing nationally than believed it was rising locally.
- The share of people who stated they were afraid to walk alone at night (35%) in 2024 was the same as it was in 1968.
- Changes in people’s perceptions of crime at the national and local levels usually do not align with changes in the overall levels of reported and unreported crime. But crime sentiments do appear to track large shifts in the homicide rate.
This section uses fear of crime statistics from 1965 through 2024 and crime perception statistics from 1989 through 2024 to broadly assess how crime sentiments have changed over time, and how those changes compare to overall crime rates and violent crime rates. Although the data on crime sentiments spans several decades, the Gallup survey did not ask about the topic every year. Because of these missing years, this section presents an overview of the trends rather than a more rigorous analysis.
Figure 1 illustrates how perceptions of crime at the national and local levels, along with a common measure of fear—being afraid to walk alone at night—relate to the overall crime rate. The first year Gallup provided data on perceptions of national and local crime was 1989. From 1989 to 2024, the year 1992 had the highest share of people (89%) who said crime was higher than it had been the prior year. During the 1990s, the percentage of people who said crime was higher nationally than it had been the previous year generally tracked the movement of crime levels. As the rate of crime fell, the share of people who said crime was higher also declined to a low of 42% in 2001. But this apparent relationship between national crime perceptions and crime trends did not continue through the 2000s. Between 2005 and 2024, a period during which crime rates fell nearly every year, the share of people who said crime was higher than the previous year increased to an average of 69%.
Local crime perceptions followed trends similar to those for national crime perceptions, as the share of people who believed crime was higher in their local area declined in the 1990s and then stabilized in the 2000s, even as crime rates were falling. But there were two major differences. First, about 20% more respondents, on average, said national crime was higher than the prior year than said local crime was higher. The second difference is that the highest share of people who said local crime was higher than it was the prior year peaked at 56% in 2022; in contrast, the share that said crime was higher nationally peaked in 1992 at 89%.
Gallup data on fear of walking alone at night, which is how fear of crime is commonly measured, has been available since 1965. Figure 1 displays the percentage of people who say they are afraid to walk alone at night in relation to the violent crime rate. The violent crime rate, rather than the overall crime rate, is used for this analysis because it represents a more direct threat to individual safety than the property crime rate.
The largest share of people expressing fear of walking alone at night was recorded in 1982, with 48% of respondents indicating they were afraid. This percentage then declined, experienced a moderate increase (44%) in 1992, and then generally decreased through 2001, paralleling the pattern seen for the national and local perception measures. From 2001 to 2021, the share of people who said they are afraid to walk alone at night fluctuated between 30% and 37%, and then rose to a recent peak of 40% in 2023. Overall, the trajectory of the level of fear measurement roughly corresponds to that for the proportion of people who believe crime has increased in their area, and is somewhat similar to the trend in violent crime.
Figure 1. Crime Concerns and Crime Rates
A major limitation of official crime statistics from the Uniform Crime Reports (UCR) is that they only count police-recorded offenses. Excluding homicide, most violent crimes, such as robbery and aggravated assaults, are typically reported to police less than seven times for every 10 incidents, meaning that a consequential number of offenses are not recorded. Furthermore, the reporting rate fluctuates each year. It is possible that as people’s concerns about crime change, their likelihood of reporting an incident to police may change as well. To overcome this limitation in the UCR, this report uses the National Crime Victimization Survey (NCVS) to track violent victimization trends and crime concerns.
Figure 2 shows the violent victimization rate (aggravated assault, rape/sexual assault, robbery, and simple assault) from 1996 to 2024, using the NCVS and the Gallup crime sentiment data over this period. The general patterns align with the UCR data. The share of people who believe crime is higher across the U.S. and locally, as well as fear of crime levels, declined along with victimization levels in the 1990s. But this consistency does not continue through 2024. For instance, in 1996, when the violent victimization rate was 79.8 (per 1,000 people), 43% of people said they were afraid to walk alone at night. By 2024, the violent victimization rate had dropped to 23.3, about a third of the 1993 level, and 35% of people said they were afraid to walk alone at night.
Figure 2. Crime Perceptions, Fear of Crime, and Violent Victimization (NCVS)
Figure 3 shows the three crime concern measures (national and local perceptions, as well as fear of crime) alongside the national homicide rate. Since homicide is the most serious crime and receives the most media attention, it seems logical that crime sentiments would follow homicide trends more closely than either the overall crime rate or the violent crime rate. The statistics presented in Figure 3 support this, showing that overall, the sentiment measures align more closely with the national homicide rate than with the rates for overall crime (Figure 1) and violent victimization (Figure 2).
Specifically, the rise in homicides between 2019 and 2020 and the subsequent decrease from 2023 to 2024 closely track the shifts in crime perceptions at the national and local levels as well as the changes in fear of crime. There are similarities over time between local crime perceptions and the national homicide rate, although the share of people who believed crime was rising in their area dropped from 43% in 2019 to 38% in 2020, even as the homicide rate increased during that timeframe by 31%. As for the fear of crime measure, the proportion of people who say they are afraid to walk alone at night also generally tracks the homicide rate from the late 1960s through the rate’s decline in the 1990s. Mirroring the pattern seen with local crime perceptions, the share of people afraid to walk at night fell from 2019 to 2020, despite the increase in homicides during that period.
Figure 3. Crime Sentiments and the National Homicide Rate
When homicide and other offenses were declining in the mid-1990s, most people believed that crime was higher. The same was true from 2000 to 2014. Although homicides remained relatively stable during this period and other crimes were decreasing (except between 2000 and 2001), a majority of the public still thought crime was worse than it had been the year before. This finding was remarkably consistent over the 35-year study period. With the exception of two years during that timeframe (2000 and 2001), more than half of people surveyed annually said they believed that crime was up. In short, even when crime concerns match crime trends, the public is still often mistaken in thinking that crime is rising. The lone exception to this pattern may be during large shifts in homicide rates.
Key Factors Associated with Americans' Crime Concerns
This section uses respondent data from the Gallup Poll Social Series to examine the relationship between crime sentiments and a variety of factors, including individual-level influences and local context. Since 2001, this Gallup poll has surveyed at least 1,000 U.S. adults annually. Using these data enables analysis of how individual factors, such as political party identification, relate to perceptions and fear of crime. Because respondents’ zip codes are included in the data, zip code and county-level characteristics are combined with individual responses to explore how local contexts may relate to crime sentiments.
The findings are organized around three outcome variables:
- Belief that crime in the U.S. is higher than in the prior year;
- Belief that crime is higher in the respondent’s local area than in the prior year; and
- Fear of walking alone at night.
The relationships between respondent-level characteristics, contextual factors (the zip code and county measures), and crime sentiments are examined incrementally by adding more measures with each new step. This approach helps highlight variables that may initially seem strongly related to crime sentiments, such as violent crime, but may not actually be strongly connected once other factors are considered. Since the models include multiple factors, the relationship between two measures, such as partisanship and national crime perceptions, may differ from that found in publications that examine only these factors and do not rely on a more rigorous analysis.
The initial models included county crime rates. Subsequent models added victimization information and other respondent characteristics and beliefs, and then integrated unemployment and zip code measures. For ease of discussion, this analysis focuses on the full models, those with all the included variables; partial models are mentioned when there are noteworthy findings. The visual presentation of the results is divided into two sets. The first set discusses crime, victimization, political measures, and economic beliefs, while the second set covers other characteristics included in the models. Lastly, the term “statistically significant” refers to findings that are unlikely to be due to random chance.
A Note on the Direction of Effects
It is worth noting that in some cases, the direction of an effect may be unclear. That is, although many variables likely influence crime concerns, concerns may also influence some of those factors. For example, crime concerns do not affect a person’s race, but a person’s race may affect their crime concerns. However, political and economic measures may be more likely to be influenced by concerns. For instance, someone may approve of the president because they believe crime is down. Table A1 in the appendix shows the likely direction of influence.
Since Gallup does not ask all questions every year, not all years of the Gallup Poll Social Series are included in the analyses. The results presented in this section combine multiple years into one analysis to identify robust relationships between crime sentiments and the variables of interest. These results are based on approximately 7,500 respondents for each outcome from 2009 to 2021. Additionally, while Gallup fields surveys throughout the year on a variety of social issues, not all of those items are asked during the collection of crime perceptions. For example, expected Christmas spending and the belief that it is a good time to find a job are the only consumer and economic attitudes captured in the same survey as crime perceptions.6 Full statistical methodology and results are available in the supplemental methodology report. Since the Gallup data cover multiple years, individual yearly effects were also analyzed. These results are discussed at the end of this section.
Crime Rates, Household Victimization, Political Views, and Economic Beliefs
Key Takeaways
- Higher county property crime rates, higher household victimization, more conservatism, and expecting to spend less or the same on Christmas were associated with greater odds of saying that crime is higher in the U.S. and locally compared to the prior year.
- Approving of the president’s performance, approving of Congress, and believing it was a good time to find a job were associated with the perception that crime in the U.S. and locally was the same or lower than it was the previous year.
- Higher county property crime rates, higher household victimization, more conservative views, and expecting to spend less on Christmas were factors that were associated with people saying they feared walking around their neighborhood at night.
- Approval of Congress’s performance (but not the president’s) and the belief that it was a good time to find a job were associated with a decrease in the odds of being afraid to walk alone at night.
- County violent crime rates and political party affiliation were not significantly associated with any of the three crime sentiments once all factors were included in the models.
This section presents regression results examining the relationship between crime sentiments and county-level crime, household victimization, political views, and economic beliefs. A positive number indicates that the measure was associated with an increased likelihood of stating that crime in the U.S. is higher, while a negative number indicates a response that crime was either the same or lower than in the prior year. Interpreting the numbers is straightforward for most variables. For example, believing it is a good time to find a job was associated with 30% lower odds of saying that crime was higher this year than last year. But for some variables, such as the property crime rate, immediately interpreting the size of the effect required additional analysis. The graphs presented in this section are the results of the full models, which contain all the independent variables. Please see the methodology document to examine the results of the partial models.
Perception that Crime is Higher in the U.S.
Examining county crime levels, increases in violent crime were associated with the perception that crime had risen nationally in the partial models, but this association was not statistically significant in the full model, when all variables were included. In contrast, increases in property crime remained significantly related to a higher perception that crime had increased in the U.S., even after accounting for neighborhood characteristics.7
While the county’s violent crime rate showed no such relationship, greater household victimization—defined as the number of different types of crimes experienced by a household member in the past year—was associated with the belief that crime in the U.S. was more prevalent this year compared to last year. Each additional household victimization raised the odds of believing crime was higher in the U.S. by about 30%, even when considering county crime rates and other factors.
As for political variables, Republicans were more likely than Democrats to say that crime is higher in the U.S., but this difference was no longer significant after economic factors were included in the model. There was also no statistically significant difference between independents and Democrats in terms of their likelihood to say crime is rising. Other political factors, however, showed a significant link with perceptions of national crime levels. On a five-point scale from very liberal to very conservative, a one-step increase in conservatism (e.g., moving from very liberal to liberal or conservative to very conservative) was associated with a 16% increase in the odds of stating that crime is higher in the U.S.8
Approval ratings for the president and Congress were inversely related to perceptions of national crime. Approving of the president’s performance was associated with a 55% decrease in the odds of saying crime was higher in the U.S. That is, when people approved of the president, they were less likely to believe crime had gone up.
There were also significant associations between a person’s economic views and their perceptions of crime. Compared to people who said they would spend less on Christmas this year than last year, those who planned to spend the same or more were less likely to claim that crime is higher in the U.S., after accounting for other factors. Similarly, believing it was a good time to find a job, versus believing it was not a good time, was associated with a 30% increase in the odds of stating that crime was either the same or lower in the U.S.
Figure 4a. Relationship Between National Crime Perceptions and Crime, Victimization, Political Views, and Economic Beliefs
Perception that Crime is Higher Locally
As seen with the national crime perceptions, initial modeling showed that higher county rates of violent and property crime were associated with the belief that crime had increased locally (Figure 4b). When all variables were added to the model, however, violent crime no longer had a significant relationship with local crime perceptions. Additionally, an increase in types of household victimization was associated with a rise in perceived crime at the local level. Specifically, an increase of one type of victimization incident in a household was associated with a 56% increase in the odds of a respondent stating that crime was higher locally.
The results for the political variables were also similar to those seen in the national crime perceptions models. In the initial modeling, Republicans were more likely than Democrats to say crime was higher in their local area, but this effect was no longer significant after economic factors were added to the model. Comparing local crime perceptions of independents and Democrats, there was no significant difference between them in the full model. Other political variables, however, still showed a significant link with perceptions of local crime. For example, greater conservatism (compared to those who identified as more liberal) and approval of the president and Congress were associated with the perception that local crime was higher. A one-point increase in conservatism was associated with a 10% rise in the odds of reporting that crime was higher locally compared to last year. Conversely, approval ratings for the president and Congress were inversely related to perceptions of local crime. Approving of the president’s job was connected to a 36% decrease in the odds of saying crime was higher locally; approving of Congress’ job was associated with a 24% decrease in those odds.
There were also significant links between economic optimism and beliefs about crime. After accounting for other factors, respondents who said they would spend more or the same on Christmas were associated with a 36% and 29% decrease, respectively, in the odds of believing crime was higher in their area compared to last year. Similarly, those who thought it was a good time to find a job were significantly less likely to say crime was higher in their area compared to last year.
Figure 4b. Relationship Between Local Crime Perceptions and Crime, Victimization, Political Views, and Economic Beliefs
Fear of Walking Alone at Night
Figure 4c illustrates the regression results for the fear of crime (walking alone at night) model. A positive result means the variable is linked to a higher likelihood that respondents report being afraid to walk alone at night. Respondents could only answer “yes” or “no” to the following question: “Is there any area near where you live—that is, within a mile—where you would be afraid to walk alone at night?”
As seen with the national and local crime perception models, higher county rates of violent and property crime were associated with the fear of walking alone in initial models, but violent crime was no longer a significant factor when additional measures were included. Additionally, an increase in types of victimization experienced within a household was linked to fear of walking alone. Specifically, an increase in one type of household victimization was associated with a 51% increase in the odds of a respondent stating they were afraid to walk alone at night.
The results for the political variables are similar to those for national and local crime perceptions, except for presidential approval. In the partial models, Republicans were less likely than Democrats to report being afraid to walk alone at night, but this difference became insignificant after including economic factors. There was no notable difference in fear of crime between independents and Democrats. Other political variables, however, showed a significant association with fear of walking alone at night. More conservatism was linked to higher fear. A one-point increase in conservatism corresponded to an 11% increase in the odds of respondents fearing walking alone at night. Although congressional job approval was associated with less fear, the model showed no significant relationship between presidential approval and fear of crime once all factors were accounted for.
The results for the economic perception measures were mixed. Compared to people who said they would spend less on Christmas, those who planned to spend the same amount were less likely to say they feared walking in their neighborhood at night; there was no significant difference on the fear of crime measure between this group and people who said they would spend more on Christmas. As for the belief that it was a good time to find a job, it was associated with a 27% decrease in the odds of being afraid to walk alone at night.
Figure 4c. Relationship Between Fear of Walking Alone at Night and Crime, Victimization, Political Views, and Economic Beliefs
Other Individual and Contextual Characteristics
Key Takeaways
- Older adults, Black and Hispanic adults, those who state a religious affiliation, and people who favor the death penalty were more likely to say crime is higher in the U.S. than in the prior year. Men, people with higher levels of education, those with higher incomes, and those who lived in zip codes with a larger share of non-White (non-Hispanic) population were less likely to say that crime had increased.
- People favoring the death penalty and those with a gun at home were more likely to say that crime is higher in their local area than in the prior year. Men and people with higher levels of education were less likely to say local crime was up.
- Older adults, men, gun owners, individuals with higher incomes, people employed full-time, and those with children were less likely to be fearful of crime; non-Hispanic White adults were more likely to be afraid to walk alone at night than Hispanic and Black people and members of other racial groups. People who declared a religious affiliation were more likely to be afraid to walk alone at night.
This section presents regression results examining the relationship between crime sentiments and other variables, such as respondents’ race and neighborhood characteristics, that past research suggests are related to crime concerns. A positive number indicates that the measure is associated with an increased likelihood of believing crime is higher, while a negative number indicates a response that crime is either the same or lower. Interpreting the numbers is straightforward for most variables. For example, respondents identifying as male are linked to a 50% decrease in the odds of saying that crime was higher in the U.S. this year compared to last year. But for some variables, such as population, the impact and its magnitude cannot be interpreted without additional analysis.
Perception that Crime is Higher in the U.S.
A look at respondent characteristics such as race and age reveals that several factors were associated with perceptions of crime at the national level (Figure 5a). Older adults, those with a religious affiliation, as well as Black and Hispanic adults, were more likely to say crime was higher in the U.S. than in the prior year. Favoring the death penalty was associated with a 47% increase in the odds of believing crime in the U.S. had increased from the previous year, while men, people with higher incomes, those with higher education levels, and those living in zip codes with a higher share of non-White (non-Hispanic) population were less likely to say crime had gone up. For example, having some college education, compared to high school or less, was tied to an 18% decrease in the odds of saying crime was higher in the U.S. this year than last.
Full-time employment, marital status, and having children under 18 were not significantly related to perceptions of national crime. Likewise, the region of the country where the respondent lived, their county’s unemployment rate, and zip code-level measures of population, median income, percentage of juveniles, percentage of single-headed households, percentage in poverty, and residential mobility (living in a different home than a year ago) were not associated with national crime perceptions.
Figure 5a. Relationship Between U.S. Crime Perceptions and Other Factors
Perception that Local Crime is Higher
The research identified several other factors that are associated with perceptions of local crime levels (Figure 5b). Favoring the death penalty, for example, was associated with a 19% increase in the odds of believing that crime was higher locally, after controlling for other factors; owning a firearm was associated with a 15% increase in those odds. Age was not significantly associated with local crime perceptions.
Men, compared to women, were more likely to say that crime had dropped or was the same locally, relative to the prior year. And higher levels of education were significantly associated with a decreased likelihood of saying that crime was higher locally than it was the year before. Race and ethnicity, however, did not show a significant connection to perceptions of local crime. Similarly, income, marital status, employment, having children under 18, and religious affiliation were not related to differences in perceptions of local crime.
Other variables that did not appear to shape perceptions of local crime included region and the county unemployment rate, as well as zip code-level measures of population size, median income, percentage of juveniles, percentage of non-White (non-Hispanic), percentage of single-headed households, percentage in poverty, and residential mobility (living in a different home than a year ago).
Figure 5b. Relationship Between Local Crime Perceptions and Other Factors
Fear of Walking Alone at Night
Some respondent-level factors associated with fear of walking alone (Figure 5c) differed from those connected to perceptions of crime at the national or local level. Older respondents and men were less likely to fear walking around their area at night. Adults from Hispanic, Black, and other racial groups were less likely to fear walking alone at night compared to non-Hispanic Whites. People with higher incomes, those employed full-time, and those with children under 18 were also less likely to fear walking alone at night. Even after accounting for other factors, those who shared a religious affiliation, compared to those who did not, were 25% more likely to be afraid of walking alone at night. Unlike the findings for perceptions of crime at the national and local levels, favoring the death penalty did not appear to be related to fear of walking alone at night in any of the models.
While geographically based characteristics had no influence on national and local crime perceptions, many of them were found to be associated with fear of walking alone at night. Regionally, respondents in the Midwest were less likely to fear walking alone at night than those in the Northeast. A population featuring a higher share of juveniles was also associated with a lower likelihood of being afraid of walking alone at night. Conversely, a higher population, the percentage of non-White (non-Hispanic) residents, the percentage of single-parent households, and poverty were associated with a greater likelihood of being afraid to walk alone at night. For example, a 1% rise in the poverty rate at the zip code level was associated with a 2% increase in the odds of respondents reporting fear of walking alone at night. The percentage of people who lived in a different house one year ago showed no significant connection to fear of walking alone at night.
Figure 5c. Relationship Between Fear of Walking Alone at Night and Other Factors
Analyzing Associations Over Time
Using the Gallup Social Series data, it is possible to examine how these associations may change over time by estimating predicted probabilities. For example, it is possible to generate the likelihood, in any given year, that a respondent will say that crime is higher in the U.S. if they disapprove of the president, while accounting for the other individual and community factors presented in the results discussed above. These predicted probability estimates were generated using data through 2023, focusing on presidential approval, congressional approval, political party, ideology, and Christmas spending, for all three outcomes of interest (national crime perceptions, local crime perceptions, and fear of crime). The estimates show that none of the results varied significantly over time. For example, those who disapproved of the president were no more likely to believe that crime was higher in the U.S. or locally, or to be more afraid to walk alone at night, in 2009 than in 2023. In other words, the effects of these variables generally remained consistent over time. Additional details are available in the supplemental methodology report.
Situating the Findings in Past Research
Public concerns about crime have been the subject of extensive study, and this report contributes to the existing literature in several ways.9 This section briefly highlights several key findings and how they add to past research.
The consistent positive association between household victimization and perceptions and fears of crime is one of the most robust findings in this area of scholarship. That finding is supported by the current study, which found victimization to be associated both with the perception that crime was higher than in the prior year and with fear of crime.
When it comes to how violent and property crime rates influence public sentiments on safety, there are essentially two hypotheses. One holds that violent crime could lead to more fear and the perception that crime is higher because of the serious harm violence can inflict on people’s lives.10 The second suggests that property crime could be connected to increased fear and a belief that crime is rising, due to the high volume of property offenses.11
The results in this report support the volume perspective, showing that an increase in property crime rates was associated with an increase in both national and local crime perceptions, as well as increased fear of walking alone at night. In the study sample, county property crime rates were roughly seven times higher than violent crime rates. These rate differences are similar to the differences seen nationally.
To date, only two prior studies have examined county crime rates and crime concerns, and neither found a significant relationship.12 The county crime rates produced in this study are similar to those in previous studies using county-level data, but the sample sizes in those two studies were 1,034 and 1,572, compared to 7,509 in the final sample (encompassing the models that included all the variables) here. The larger sample size might help explain differences in findings, because smaller-sample studies require larger effects to reach statistical significance.
The connection between economic perceptions and crime perceptions documented in this report was generally anticipated, although the specific measures, like “good time to find a job” and “Christmas spending,” were understudied in previous research. A review of past studies indicates that fear of crime is strongly linked to other general anxieties. For example, earlier research shows that higher personal economic insecurity and poorer perceived national economic health are associated with increased fear of crime.13
As for political variables, much of the prior research on factors such as partisanship has been used to examine impacts on punitive attitudes, with limited attention focused on fear or perceptions of crime. Basic descriptive results show a significant difference between Republicans and Democrats.14 But the findings from this research suggest that one’s political party does not significantly influence crime sentiments after accounting for several factors, including political ideology, which is significantly associated with crime concerns. Although there is overlap between ideology and party, the two are not perfectly aligned, and the ideology measure is nuanced (e.g., conservative compared to very conservative); this distinction appears to be more important in explaining crime concerns.
Some of the other findings of this report are consistent with past research. For instance, the strong link identified between gun ownership and fear of crime has been well-documented. Findings related to the impact of other individual traits, such as race, education levels, and attitudes toward crime, also align with earlier studies. Additionally, the zip code-level findings of higher poverty rates, single-parent households, and non-White populations are consistent with past research. The scholarship consistently finds that these contextual factors commonly associated with crime (e.g., poverty rates) also tend to be associated with increased fear of crime.15
Conclusion
Partisan disputes about crime levels and effective responses are nothing new in the United States, but the intensity of the conflict has reached new heights in recent months. Amid that debate, this research uncovers how broad national factors may influence public perception and fear of crime over time, and how such concerns align with actual crime rates.
Across all three outcomes measured for this report—perceptions of national crime, perceptions of local crime, and fear of walking alone at night—several consistent patterns emerged. First, household victimization and higher local property crime rates stand out as consistent predictors of heightened crime concern and fear, even after accounting for a wide range of demographic, political, and other contextual factors. People who had experienced more victimization in their household, or who lived in areas with higher property crime rates, were more likely to believe that crime was rising and to report feeling unsafe.
Second, a more positive economic outlook is associated with individuals’ perceptions of crime trends and personal safety. Specifically, the analysis showed that people who believed it was a good time to find a job and expected to maintain or increase holiday spending were more likely to say crime was stable or declining, and were less likely to report fear of walking alone at night. Although these measures capture only a portion of economic sentiment, they suggest that broader financial optimism is linked to more positive views about public safety.
Gender differences also stood out amid the findings, for their consistency across the measured outcomes. Men were more likely than women to believe that crime was stable or declining, both locally and nationally, and were less likely to express fear about walking alone in their area at night.
Taken together, the report’s findings demonstrate that crime concerns are complex. Public perceptions of crime are shaped not only by crime itself—including personal or household victimization—but also by broader political attitudes (e.g., presidential and congressional approval levels) and economic outlook (e.g., Christmas spending). And while higher property crime rates were consistently associated with crime perceptions and fear, violent crime rates were not significant predictors once other factors were taken into account. Despite that finding, the long-term trend analyses presented in this report suggest that large shifts in homicide rates may influence public views on crime, indicating an area for future research.
Another notable finding is the stability of these relationships over time. The influence of political ideology and economic perceptions on crime concerns remained largely consistent between 2009 and 2023. In other words, despite increasing political polarization and economic uncertainty, the connections between these dynamics and Americans’ sentiments about crime have changed little, a reality that is easy to overlook.
For policymakers and practitioners, this research underscores an important point: Public concerns about crime do not always reflect changes in crime alone. Crime concerns may also be shaped by broader social anxieties and economic conditions. Future research should continue to explore these relationships and what specific offenses the public are concerned about. Understanding the interplay of these forces can help policymakers interpret public opinion more accurately and design responses that address both crime and the factors that influence how people perceive it.
Appendix
This table shows the potential direction of effects between the independent variables used in the analysis and crime concerns. An arrow pointing in one direction means that the variable (e.g., race) can influence perceptions and not the other way around. Likewise, the relationship between crime rates and household victimization is also very likely to be unidirectional. The variables with two arrows are ones where there is more likely to be a bidirectional relationship, where it is not possible to identify if the independent variable is influencing crime concerns or if crime concerns are influencing the independent variable. These directions are unrelated to the regression analysis results.
Table A1. Expected Relationship Between Independent Variables and Crime Sentiments
About the Authors
Ernesto Lopez is a senior research specialist with the Council on Criminal Justice. His work currently focuses on analyzing and interpreting crime and criminal justice data to understand patterns and shifts across the broader crime landscape.
Amanda Graham is an associate professor in the Department of Criminal Justice and Criminology at Texas State University. Her research interests focus on policing, police-community relationships, public opinion, and measurement.
Acknowledgements
Stephanie Kennedy, along with other members of the Council on Criminal Justice team, provided support on this report.
This paper was produced with support from the Harry Frank Guggenheim Foundation, Stand Together Trust, and CCJ’s general operating contributors.
Suggested Citation
Lopez, E., & Graham, A. (2026). Perception and reality: Understanding crime concerns in the United States. Council on Criminal Justice. https://counciloncj.org/perception-and-reality-understanding-crime-concerns-in-the-united-states/
Endnotes
1 For a discussion of related concepts, such as fear and anger, see: Drakulich, K., & Baranauskas, A. J. (2021). Anger versus fear about crime: how common is it, where does it come from, and why does it matter?. Crime, Law and Social Change, 76(5), 451-472. https://doi.org/10.1007/s10611-021-09973-y
2 Hartnagel, T. F. (1979). The perception and fear of crime: Implications for neighborhood cohesion, social activity, and community affect. Social forces, 58(1), 176-193. https://doi.org/10.1093/sf/58.1.176
3 There is no settled definition or measurement of fear of crime, but fear is generally accepted to be an emotional, not cognitive state, and can be formless or specific. See Doran, B. J., & Burgess, M. B. (2011). Putting fear of crime on the map: Investigating perceptions of crime using geographic information systems. Springer Science & Business Media. https://doi.org/10.1007/978-1-4419-5647-7
4 For example, see, Vollum, S., Longmire, D. R., & Buffington-Vollum, J. (2004). Confidence in the death penalty and support for its use: Exploring the value-expressive dimension of death penalty attitudes. Justice Quarterly, 21(3), 521-546. https://doi.org/10.1080/07418820400095891
5 For a discussion on crime concerns and public policy see: Pickett, J. T. (2019). Public opinion and criminal justice policy: Theory and research. Annual Review of Criminology, 2(1), 405-428. https://doi.org/10.1146/annurev-criminol-011518-024826
6 While there is not clear evidence that these specific factors are used to measure personal financial situations (Christmas spending) and broader economic perception, research on planned expenses and ability to find a job are commonly used. See the U.S. Consumer Confidence report for more information. The Conference Board. U.S. Consumer Confidence (n.d). https://www.conference-board.org/topics/consumer-confidence/
7 Additional models were tested, replacing the county’s violent crime rate with the county’s homicide rate for all three outcomes, and the results were identical to those of the models that include violent crime.
8 While there is overlap between political party and ideology, ideology can further refine an individual’s placement within a political group. For instance, a Republican can identify as very conservative, moderate, or even liberal.
9 From 1997 to 2022, 926 scholarly articles were published on the topic of fear of crime. This number does not include studies that focus solely on other crime-related sentiments, such as anger or perceptions. Hart, T. C., Chataway, M., & Mellberg, J. (2022). Measuring fear of crime during the past 25 years: A systematic quantitative literature review. Journal of Criminal Justice, 82, 101988. https://doi.org/10.1016/j.jcrimjus.2022.101988
10 Zimmering, F., & Hawkings, G. (1997). Crime is not the problem: Lethal violence in America. Oxford University Press.
11 Hipp, J. R. (2013). Assessing crime as a problem: The relationship between residents’ perception of crime and official crime rates over 25 years. Crime & Delinquency, 59(4), 616-648. https://doi.org/10.1177/0011128710382264. Original source was not available for review: Skogan, W. G., & Maxfield, M. G. (1981), Coping with crime: Individual and neighborhood reactions (Sage Library of Social Research; Vol. 124). SAGE.
12 Drakulich, K., & Baranauskas, A. J. (2021). Anger versus fear about crime: how common is it, where does it come from, and why does it matter?. Crime, Law and Social Change, 76(5), 451-472. https://doi.org/10.1007/s10611-021-09973-y.
Robinette, J. W., Piazza, J. R., & Stawski, R. S. (2025). Community crime and safety: An investigation of gender differences in the daily stress process. Journal of Community Psychology, 53(2), e23158. https://doi.org/10.1002/jcop.23158
13 Britto, S. (2013). ‘Diffuse anxiety’: The role of economic insecurity in predicting fear of crime. Journal of Crime and Justice, 36(1), 18-34. https://doi.org/10.1080/0735648X.2011.631399; Singer, A. J., Chouhy, C., Lehmann, P. S., Stevens, J. N., & Gertz, M. (2020). Economic anxieties, fear of crime, and punitive attitudes in Latin America. Punishment & Society, 22(2), 181-206. https://doi.org/10.1177/1462474519873659
14 Castellano, J. (November 19, 2025). Are Americans Worried About Crime? It depends on how they voted. The Marshall Project. https://www.themarshallproject.org/2025/11/19/gallup-crime-perceptions-political-polarization
15 See Doran, B. J., & Burgess, M. B. (2011). Putting fear of crime on the map: Investigating perceptions of crime using geographic information systems. Springer Science & Business Media. https://doi.org/10.1007/978-1-4419-5647-7
