(CNN) – According to the Federal Bureau of Investigation, a crime becomes a hate crime when there’s an added element of bias. For example, a murder is labeled a hate crime if the victim was killed because he was gay. Or black. Or Muslim.

“Hate crimes are different from other crimes. They strike at the heart of one’s identity,” FBI Director James Comey said in a speech at the Anti-Defamation League National Leadership Summit in 2014. “They strike at our sense of self, our sense of belonging. The end result is loss: loss of trust, loss of dignity and, in the worst case, loss of life.”

They’re motivated by bias, said David Stacy, government affairs director of the Human Rights Campaign.

“In a hate crime, the victim is targeted because of his characteristics,” Stacy said. “These are bias-motivated crimes, and often they are much more violent than traditional crimes.”

Counting the hate

The Civil Rights Act of 1968, the first law that dealt with hate crimes, protected people against attacks related to their race, religion or national origin. Violence triggered by gender, sexual orientation, gender identity or disability was not charged as a hate crime until 2009. The FBI says that a hate crime can target either a person or property and can be motivated in whole or in part by biases, meaning offenders can have other motives besides hate.

The Hate Crime Statistics Act of 1990, signed by President George H.W. Bush, required the US attorney general to report hate crimes annually. The law was inspired by a violent attack on three black men in Howard Beach, a mostly white neighborhood in Queens, New York, in 1986.

Walking out of a pizzeria, the men were confronted by a gang of white teens, who called out racial slurs and chased the men. One man, Michael Griffith, 23, was killed by oncoming traffic. Another was severely beaten but survived. The third managed to run away.

To track attacks such as these, the FBI now collects voluntary reports of hate crimes from local jurisdictions across the country.

Since the data collection began, the FBI has published hate crime statistics from 1996 to, most recently, 2015. In 2015, there were 5,818 hate crime incidents reported, the majority of which were biased toward race and ethnicity. There were about 340 more hate crimes in 2015 than in 2014.

“Hate crimes occur during a period of heightened rhetoric, like a presidential election,” said the Human Rights Campaign’s Stacy. Whenever a vulnerable group is given national attention — whether the attention is positive or negative — people who are biased against the group may lash out, he explained.

The FBI reports that the numbers of hate crimes have decreased over several years, with the exception of 2015. Regardless of the year, however, the hate crime count is underreported to the FBI, explained Stacy. Some state, city and local police agencies simply don’t collect or disclose the data.

“Crimes happen, and the cop on the beat checks the motivation of the crime to state in the police report,” Stacy said. “Often, they say there’s not enough evidence to say it’s a hate crime. There’s an underreporting of hate crimes in general.”

Tracking hate crimes against the LGBT community

Stacy joined the Human Rights Campaign in 1999, an important time for the LGBT community. The previous year, Matthew Shepard, 21, a gay student at the University of Wyoming, was beaten, tortured, chained to a fence and left for dead by two men he had met at a bar.

The case caused a stir, but it wasn’t clear whether Shepard’s attack was an anti-gay hate crime. The two offenders confessed to pretending to be gay in order to lure Shepard into their truck, but they claimed their motive for the murder was robbery. However, LGBT activist groups nationwide, including Shepard’s mother, Judy, argued that the student was killed because he was gay.

Stacy took the activists’ message to Congress, where they lobbied for laws that would protect victims attacked because of their sexual identities. They succeeded 11 years later, in 2009, when President Obama signed the Matthew Shepard and James Byrd Jr. Hate Crimes Prevention Act, which expanded hate crime laws to include sexual orientation, disability and gender bias.

Sen. Bob Casey, D-Pennsylvania, recently proposed a law that would prevent hate crime offenders from buying guns. This bill comes in response to the mass shooting at Pulse nightclub, a gay club in Orlando where nearly 50 people were killed in June.

“If you have proven you will commit criminal acts based on hate, you absolutely should not have access to a gun. It’s common sense,” Casey said at a news conference after the shooting.

“No one in America should ever be afraid to walk down the street holding the hands of the person they love,” remarked President Obama when he signed the 2009 Hate Crimes Prevention Act.

The act “was a huge victory for us” Stacy said. Still, he noted, much work needs to be done to raise awareness of anti-LGBT hate crimes.

As of now, the Hate Crimes Prevention Act protects LGBT victims from violent crimes only where the federal government has jurisdiction. Since most crimes are tried at the state level, many victims aren’t protected based on their gay or transgender identities.

The Movement Advancement Project, a Denver-based think tank that advocates for LGBT equality, maintains an online map of which states have hate crime laws that protect LGBT people. According to the map, 17 states have hate crime laws that cover sexual orientation and gender identity. Thirteen states have laws that cover only sexual orientation.

The project reports that 15 states have laws that cover neither sexual orientation nor gender identity, while five states (South Carolina, Georgia, Indiana, Arkansas and Wyoming) have no hate crime laws in place. Without protections in these states, it’s hard to track how many hate crimes against LGBT people occur.

In 2013, the FBI began recording hate crimes motivated by gender and gender identification biases — for example, attacks on transgender people. Crimes motivated by gender identification rose from 31 in 2013 to 114 in 2015, according to FBI reports.

Stacy says the lack of support for LGBT victims of hate crimes is troubling. In an environment that doesn’t recognize hate crimes motivated by homophobia, he explained, victims don’t feel safe reporting attacks. And because these crimes aren’t always recorded by the jurisdictions, FBI data do not reflect these incidents, Stacy said.

The 2015 data reveal that anti-LGBT hate crimes rose by about 5% but remained below 2013 levels. Still, the Human Rights Campaign suggests that this increase is an understatement: “The number (of LGBT hate crimes) likely only represents a fraction of such cases given that thousands of law enforcement agencies throughout the country did not submit any data,” it said in a statement.

Tracking hate crimes against Muslims

Three years after Shepard’s death, in the wake of the attacks of September 11, 2001, a wave of anti-Islamic hate crimes surfaced. In 2001, 481 incidents of hate crime were recorded. Incidents, as defined by the FBI, may include multiple offenses. Comparatively, only 28 incidents against Muslims were recorded in 2000, according to the FBI.

The increased attacks against people who appeared Muslim showed most drastically in the religion-biased crime count. Before 2001, crimes motivated by ethnicity made up a little more than 10%, on average, of the total number of single-biased hate crimes. In 2001, ethnicity-biased crimes counted for more than 20% of the total, according to the FBI.

After 2001, the numbers of hate crimes against Muslims dropped, but they remained higher than before. Between 2010 and 2013, attacks against Muslims made up about 13% of religion-biased offenses. Comparatively, anti-Jewish crimes counted for about 60% of those offenses.

But in 2014, anti-Muslim offenses increased again, counting for more than 16% of the religion-biased offenses. Interestingly, this was the only religion within that category of hate crimes that increased in the 2014 FBI report, and the total hate crime count in 2014 was the lowest it had ever been.

In 2015, the hate crime total increased by 356 single bias incidents, and anti-Muslim incidents surged again. Over 20% of attacks on religious groups in 2015 targeted Muslims, according to FBI data. The FBI reported 257 hate crimes against Muslims in 2015, an increase from 154 crimes in 2014.

Why the surge? The Southern Poverty Law Center, an Alabama-based organization that monitors hate crimes across the country, points to the 2015 terrorist attacks in Paris and the “exploitation of these attacks by right-wing media and political figures” in the US as encouraging hatred toward Muslims, according to recent blog post.

The law center also noted that the anti-Muslim hate crime count in 2015 may be even higher because a handful of anti-Sikh hate crimes (a new category for 2015) were also reported last year. Because Sikhs are sometimes mistaken for Muslims, an offender may have intended to commit a hate crime against a Muslim but instead attacked a Sikh. In this case, the crime would have been recorded as an anti-Sikh crime.

The center suggests that President-elect Donald Trump may have influenced the uptick in hate crimes.

“What was likely even more important (than terrorist events in Europe) was Trump’s attacks on Muslims, including his infamous call for a ban on Muslims entering the US,” it said in a blog post.

The law center continued, “In addition, many of Trump’s aides and surrogates embraced similar rhetoric. The Trump campaign also maligned blacks, Latinos and other minorities.”

Tracking hate crimes against blacks and minorities

Historically, about half of all hate crimes have been racially charged, according to the FBI reports. Of these race-biased crimes, the majority have targeted black people. In 2015, there were 1,745 anti-black hate crimes reported, far outnumbering the 613 anti-white incidents and 299 anti-Hispanic or Latino incidents.

The recent report reveals a nearly 8% increase in hate crimes against blacks from 2014 data, which cited 1,621 anti-black hate crimes. Prior to 2015, the race-biased crime numbers had steadily decreased since 2008. In 2008, the same year Obama was elected, there were 2,876 anti-black hate crimes reported, nearly 37% of the total number of hate crimes that year.

“There was an uptick in hate crimes (against black people) after Obama’s election,” confirmed Heidi Beirich, director of the Southern Poverty Law Center’s Intelligence Project. That year, the law center saw churches with predominantly black congregations burned to the ground, among other incidents, she said. With a black president in the White House, “people were angry,” she explained.

Though the number of race-biased hate crimes has decreased over many years, tensions between races were still strained in the years following Obama’s election. Incidents of police officers killing young black men, including Trayvon Martin, Michael Brown and Eric Garner, launched the Black Lives Matter movement in 2013, which advocates for racial justice.

But the hate didn’t stop: In 2015, nine people were killed in a traditionally black church in Charleston, South Carolina, by a white supremacy supporter. The shooter, Dylann Roof, was charged with 33 federal counts, including nine counts of violating the Hate Crimes Prevention Act resulting in death and three counts of violating the act in an attempt to kill. Roof confessed to the killings and was convicted December 15 on all 33 charges.

South Carolina does not have hate crime laws, so Roof couldn’t be charged locally for hate crimes. The city of Charleston reported four hate crimes in 2015, according to the FBI. The report doesn’t specify whether Roof’s crime was among those four hate crimes.

At the Southern Poverty Law Center, Beirich tracks hate incidents from current media headlines and individual reports from victims. As the center’s data are updated daily with new hate crime incidents, she can observe how the crime counts ebb and flow depending on the political climate.

Beirich compares the spike in race-biased hate crimes after Obama was elected to the surge of crimes after the tumultuous 2016 election.

“Now, (the hate crime offenders) are happy,” Beirich said. When Trump won the US election, she explained, the reaction among a segment of his supporters was similar to reactions to Brexit in Europe: a heightened sense of nationalism and increased hate for minorities, particularly blacks and immigrants. In the week after Trump’s win, the law center counted more than 700 hate incidents, the majority of which targeted black people and immigrants, it said.

In addition to tracking hate crimes, the law center monitors active hate groups in the country, including white supremacy groups, black separatists and general hate groups. Actions by these groups — or in reference to them, such as a spray-painted swastika on a building — contribute largely to the hate crime count.

According to the law center, there were 892 active hate groups in 2015, an increase from 784 in 2014. Of the hate groups, about 20% are KKK branches, 20% black separatists, 20% general hate, 10% neo-Nazis, 10% racist skinheads and 10% white nationalists. Neo-Confederate and Christian identity groups made up the rest of the total. General hate groups include groups whose biases aren’t specified and groups with multiple biases, such as both anti-black and anti-gay.

“Reacting to demographic changes in the US, immigration, the legalization of same-sex marriage, the rise of the Black Lives Matter movement, and Islamist atrocities, these people fostered a sense of polarization and anger in this country,” the law center commented in a blog post.

Invisible hate crimes

Beirich and Stacy emphasized that FBI hate crime statistics are grossly underreported, mainly because the hate crime reports are handed in voluntarily. Several large cities, including Honolulu and Portland, Oregon, didn’t report in 2015, according to Stacy.

“There was a big drop in jurisdictions reporting,” Stacy said. In past years, participating agencies have increased: About 12,000 submitted data in 2001, about 12,400 in 2005, about 15,000 in 2010 and about 15,500 in 2014. But in 2015, fewer than 15,000 agencies reported, according to the FBI. Of those that didn’t report, 21 were cities with populations of more than 100,000 people, the Human Rights Campaign notes in a statement on their website.

However, even the data cities do report may not be reliable. Every year since the FBI began tracking hate crimes in 1996, more than 80% of jurisdictions have reported “zero” hate crimes. In 2015, only 11.6% of participating law enforcement agencies reported hate crimes. The vast majority of participating jurisdictions — 88.4% — said not a single hate crime occurred in their cities.

“It’s bogus,” said Beirich. Not acknowledging hate crimes skews the data, she explained, because some states appear to have much more hate than others. “States that are good at reporting, like California and New Jersey, look like they have a lot of hate crimes, while most states in the deep South don’t report hate crimes.”

Areas with histories of racial tensions like Mobile, Alabama, and the entire state of Mississippi reported no hate crimes for 2015, according to the FBI.

In the places where there are no local hate crime laws, hate crimes must be charged at the federal level, as was the case for Roof in South Carolina. However, charging someone with a hate crime in a jurisdiction without local hate crime laws can be much more difficult than in places with the laws, Stacy says.

Stacy adds that jurisdictions don’t admit hate crimes, which may be why the FBI numbers seem to be severely understated. Part of the problem, Stacy suggests, is that police officers don’t acknowledge that crimes are hate crimes when they happen, either because there are no local hate crime laws or the officers are biased themselves.

“Communities are much more likely to report bias-motivated crimes when they don’t trust the police,” Stacy said. In these areas, he continued, “there’s much more willingness to report hate crimes.”

Conversely, in communities that strongly support law enforcement — such as traditional, right-wing Southern towns — the public is not as likely to question an officer’s judgment, according to Stacy. This compounds the problem: First, hate crimes are less likely to be acknowledged in these districts, which creates an unsafe space for vulnerable individuals; secondly, since the crimes aren’t reported even if they occur, areas with high rates of hate crimes may be overlooked by the FBI.

Even for the states that reported hate crimes in 2015, the numbers were still exceedingly low, according to Stacy and Beirich. Wyoming filed just two incidents, while Arkansas reported five. Comparatively, Maine, whose population is half the size of Arkansas, reported 38 hate crimes, according to the FBI.

Bringing hate crimes to light

By tracking hate crime in real time, the Southern Poverty Law Center wants to expose hate incidents that may not be accounted for by the FBI.

Beirich suggests hate crime data are better represented elsewhere. “Instead of looking at the FBI data, we should look at the crime survey data” administered by the Bureau of Justice Statistics, she said. In 2012, the bureau conducted a hate crime survey that measured hate crimes perceived by victims.

The survey was administered to people age 12 or older and asked every six months, for a three-year period, whether they had experienced victimization, which could be nonfatal personal crimes (rape or sexual assault, robbery, aggravated and simple assault, and personal larceny) or household property crimes (burglary, motor vehicle theft and other theft). They were also asked what they perceived as the bias for the attack (anti-race, anti-LGBT, etc.) and whether they reported the crime to the police. Responses from the sample of people who took the survey were then weighted to represent the total population of the US.

After calculating the survey responses, the bureau estimated that 293,800 hate crimes occurred in 2012: 50 times more crimes than the FBI reported in 2012. The survey estimated that 60% of the hate crimes submitted weren’t reported to the police.

“It’s our responsibility to get out this information,” Beirich said. “We need a much more aggressive program of reporting.”

The Bureau of Justice Statistics numbers differ from the FBI statistics because they are victim-reported, rather than reported by the police. They are also estimated from a sample population, which means there is a margin of error between the estimated number of hate crimes and the real number, according to the survey. Additionally, since it was victim-reported, bias-motivated murders were not included in the Bureau of Justice Statistics report.

The Southern Poverty Law Center accumulates both victim-reported hate crimes and hate crimes reported by the media, but these data are updated daily, and the center does not keep an annual report of total hate crimes. Of the three sets of hate crime data, the FBI statistics are the only statistics that track physical hate crimes across the country year after year.

In his 2014 speech, FBI director Comey addressed the flaws with the hate crime statistics. “We need to do a better job of tracking and reporting hate crime to fully understand what is happening in our communities and how to stop it,” Comey said.

“There are jurisdictions that fail to report hate crime statistics. Other jurisdictions claim there were no hate crimes in their communit, a fact that would be welcome if true,” Comey continued. “We must continue to impress upon our state and local counterparts in every jurisdiction the need to track and report hate crime. It is not something we can ignore or sweep under the rug.”

The Human Rights Campaign proposes requiring districts to report hate crimes. “This would provide a more complete picture of hate based violence in the United States and allow for targeted efforts to address areas with high levels of crime,” the campaign said in its statement.

Stacy makes an analogy between hate crime and domestic crime, using the example of a woman who is beaten by her husband. If there’s a domestic abuse center in town, she is much more likely to report the attack. Without a support system available, however, the woman will probably suffer in silence.

“The same thing is true for hate crimes,” Stacy said. “If the services are there, victims are more likely to report.”