Last week, in North Tyler, a black, transgender woman named Ty Underwood was shot and killed in her car in an apparent homicide. Since the murder, Underwood’s friends have maintained that the killing must have been a hate crime, carried out because of Underwood’s gender identity. But her murder won’t be designated or investigated by the Tyler Police Department as a hate crime. (The department has yet to name or arrest a suspect in the slaying.) Unlike federal hate crime legislation, Texas’s hate crime act doesn’t incorporate crimes motivated by the victim’s gender identity or perceived gender identity.

Before 2001, when the then–newly elected governor Rick Perry signed the James Byrd Jr. Hate Crimes Act into effect, Texas classified hate crimes as offenses “motivated by bias or prejudice.” The brutal killing of Byrd in 1998, however, prompted discussion about bolstering the state’s hate crime laws with language that specifies certain groups of people. The 2001 bill expanded the definition of hate crime in Texas as any crime “motivated by a victim’s race, religion, color, sex, disability, sexual preference, age, or national origin.”

Eight years later, in 2009, U.S. Congress passed the Matthew Shepard and James Byrd Jr. Hate Crimes Prevention Act. The act added protections for crimes motivated by gender identity and required that the FBI start including crimes that were committed with a gender identity bias in the annual Uniform Crime Report. Those numbers were published by the FBI for the first time at the end of last year, when the 2013 UCR data was released.

What did these new numbers reveal? According to the report, 33 out of the total 7,242 hate crime victims in 2013 (0.5 percent), were the targets of gender identity–biased offenses. It’s a number so small it’s almost negligible—especially when compared to a reported 3,563 victims of racially-biased hate crimes.

A number that small is also implausible. News stories from 2013 give us reason to believe that there were considerably more than 33 transgender victims that year, and reports released by organizations like the New York City Anti Violence Project give a much higher number. So why aren’t these crimes reflected in the FBI’s data? A lot of that comes down to the way the data is compiled and how police departments in states like Texas—which don’t recognize gender identity–biased hate crimes—choose to label and investigate these offenses.

While the FBI’s UCR data is useful for providing a snapshot of the state of crime in the United States for any given year, the numbers are notoriously underreported. The data relies on reports from thousands of participating police departments across the country and includes only those offenses that are known to the police. This creates what criminologists refer to as the “dark” or “hidden figure of crime,” or crimes and offenses that remain unreported.

The Bureau of Justice Statistics’ National Crime Victimization Survey seeks to fill in the gaps by using census data to survey a large sample of Americans about whether or not they were victims of certain crimes. According to NCVS-provided data from 2012, “an estimated 293,800 nonfatal violent and property hate crime victimizations occurred against persons age twelve or older residing in U.S. households.” This number, which doesn’t include homicides, is much bigger than the number provided by the UCR for that same year.

While a great number of hate crimes committed in a given year are lost to that dark figure of crime, still more are lost because of nuanced definitions in hate crime laws in states like Texas, where gender identity isn’t included in the hate crime definition at all. Crime report data in Texas from 1999 through 2013 is available on the DPS website. Each year includes a section dedicated to tracking hate crimes, as per a law passed in 1993 by Senator Rodney Ellis, a Democrat from Houston, that mandates the tracking and recording of hate crime statistics in the state. What the FBI’s report from 2013 has that the Texas report of the same year lacks is the number of crimes committed on a gender-identity bias. If gender identity–biased hate crimes technically don’t exist in Texas, these numbers are not only being left out of state-level reports but they’re also being left out of the FBI’s data as well.

Those missing numbers represent people like Ty Underwood, whose murder—unless there is some kind of federal intervention—will be calculated as a part of homicide statistics but won’t ever be classified as a hate crime.

Some of Underwood’s friends have organized on Twitter under the hashtag #JusticeForNunne, a play on words inspired by one of Underwood’s nicknames. They want to see a suspect captured and convicted for the murder of their friend, and they want to see justice for the transgender community at large, in which individuals are victims to a disproportionately higher incidence of hate crimes each year.

Hate crime data is collected, published, and made publicly available because unlike other crimes, hate crimes threaten entire groups or communities of people. The numbers published in state- and national-level reports are reflective of shifts in culture and perception. For example, the number of anti-Arab and anti-Islamic hate crimes in Texas alone increased from 7 in 2000 to 93 in 2001.

Similar to the way doctors use quantitative data, like a patient’s body temperature or blood pressure, as signs of sickness or disease, policy makers use the statistics from hate crime reports as signs of systemic problems in  counties and states. If the tools used to collect the data aren’t working properly—or if certain statistics are ignored entirely, like gender identity-biased offenses— it’s impossible to get an accurate read.