Discrimination against lesbian, gay, bisexual, queer, and/or questioning (LGBTQ)* youth remains a contributing factor to the disproportionately high rates of socio-emotional distress and mental illness among LGBTQ youth, when compared to straight and cisgender youth.

This exploratory analysis compares proposed anti-LGBTQ legislation to the volume of Crisis Text Line (CTL) conversations from LGBTQ texters under age 18 and suggests that proposed anti-LGBTQ legislation may contribute to LGBTQ youth’s experiences of emotional distress. Our policy analysis shows that state legislatures proposed over 200 pieces of anti-LGBTQ legislation from 2015 to 2019. At the time of this publication, 2021 has already seen more than 200 pieces of legislation of this type in 35 states. This stark increase highlights an urgent need to understand how such legislation—whether enacted or not—impacts LGBTQ youth well-being.


Key Findings

  • In 42 states, state legislators subjected LGBTQ youth to at least one piece—and sometimes as many as 12 pieces—of proposed anti-LGBTQ legislation in a single year from 2015 to 2019. In total during this time, state legislatures introduced 215 pieces of anti-LGBTQ legislation, for an average of 43 proposed laws each year.
  • Laws creating restrictions on single-sex facilities were among the most frequent types of anti-LGBTQ legislation proposed.
  • The number of texts to Crisis Text Line in a state where anti-LGBTQ legislation was proposed increased in the four weeks after such legislation was proposed; this difference was small but statistically significant.

Background

Researchers use the concept of minority stress—a culmination of social discrimination, stigma, and anticipated rejection from friends, family, and community—to explain increased rates of mental illness among lesbian, gay, and bisexual (LGB) youth. State policy has played a critical role in the ever-evolving landscape of civil rights and discriminatory practice in the United States, and state legislative proposals to limit access to services for LGBTQ people represent one source of minority stress. Policies such as banning transgender athletes from sports, prohibiting health care for transgender youth, or allowing religious exemptions in foster care create conditions for increased discrimination and reduced access to services for individuals with marginalized identities. Research suggests that the stress associated with policies that restrict rights or remove protections for LGBTQ people has a negative impact on their mental health. Organizations that run LGBTQ-focused suicide prevention hotlines have long pointed to sharp increases in calls during times in which pieces of anti-LGBTQ state legislation are introduced. This suggests that LGBTQ-related legislation may have immediate implications for the health and welfare of young people, separate and apart from the specific provisions of the policy itself.

We sought to test whether there is any relationship between proposed anti-LGBTQ legislation and instances of LGBTQ youth in crisis. Although this analysis is exploratory, it provides important new findings that indicate an association between the introduction of anti-LGBTQ state legislation and the volume of LGBTQ youth seeking crisis support. The legislative data set that we generated is available for other researchers and stakeholders to access. This data set allows users to explore the types, frequency, and regional distribution of anti-LGBTQ legislation from 2015 to 2019.


Findings

In 42 states, state legislators subjected LGBTQ youth to at least one piece—and sometimes as many as 12 pieces—of proposed anti-LGBTQ legislation in a single year from 2015 to 2019, for a total of 215 pieces of legislation within the categories we analyzed. The largest number of state bills were introduced in 2016 (with 59), and the fewest in 2018 (with 28). Texas and Oklahoma proposed the most anti-LGBTQ legislation during the time period, with 23 and 16 bills respectively. Among the most common types of proposed legislation were prohibitions on health care for transgender youth and the provision of religious exemptions from marriage ceremonies.


 


Select topics to include:


Select all
Unselect all


Excluding trans youth from athletics
Prohibiting health care for trans youth
Religious exemptions from marriage ceremonies
Religious exemptions in foster care and adoption
Religious exemptions in health care
Religious Freedom Restoration Act (RFRAs) or First Amendment Defense Acts (FADAs)
Single-sex facility restrictions



States in the Northeast introduced fewer pieces of anti-LGBTQ legislation, while states in the South (especially Oklahoma, Texas, and Missouri) introduced the most. Additionally, over time, there has been a decrease in the number of proposed pieces of legislation, although the decrease is not linear (42 in 2015, 59 in 2016, 45 in 2017, 28 in 2018, and 41 in 2019). Although outside the scope of this study, the significant increase in proposed anti-LGBTQ legislation in 2021 thus far underscores the urgent need to understand how such legislation contributes to the minority stress of LGBTQ youth.

Proposed anti-LGBTQ legislation is associated with a small but statistically significant increase in texts to Crisis Text Line from LGBTQ youth. Using a multi-level model to compare the number of daily conversations with LGBTQ youth on days with and without anti-LGBTQ legislation, we found that the daily text volume was higher on days for which an anti-LGBTQ bill had been proposed in the past four weeks (fixed-effect of .036, p<.0206). This effect is small, corresponding to approximately 1.1 additional conversations per state per month, but is statistically significant. For more details about the approach, see our Technical Appendix.

This suggests a possible association between the proposal of anti-LGBTQ legislation and an increase in youth mental health help-seeking behavior. Future research should further investigate this relationship and explore alternative data sources.


Implications

The results of this exploratory analysis suggest that proposing anti-LGBTQ legislation may increase the number of LGBTQ youth who experience a mental health crisis. These preliminary findings suggest that the negative impacts of anti-LGBTQ legislation on LGBTQ youth may be broader-reaching than previously understood and may go beyond the specific provisions of such laws once enacted. As policymakers consider introducing legislation that limits access to or participation in services and opportunities for LGBTQ people, they must also consider the potential detrimental impacts of such policy debates on the well-being of LGBTQ youth.


Future Research

Our preliminary findings suggest that—with a larger sample of mental health crisis call or text conversation frequency data from LGBTQ youth—a future study might reveal a stronger correlation between proposed anti-LGBTQ legislation and help-seeking behavior by LGBTQ adolescents. A future study to quantitatively examine the link between proposed legislation using more sophisticated quasi-experimental methods (such as difference in difference, interrupted time series, or an event study model) would provide more conclusive evidence of the impact of legislation on LGBTQ youth’s mental health. Such a model might require additional years of data to gain sufficient statistical power. Future research should also explore whether there is a correlation between proposed supportive LGBTQ legislation and LGBTQ youth mental health.


Our Approach

Policy Scan

Our source for policy information was Politico Pro. We created search terms to identify relevant anti-LGBTQ legislation in seven topic areas: excluding trans youth from athletics, prohibiting health care for trans youth, religious exemptions from marriage ceremonies, religious exemptions in foster care and adoption, “Religious Freedom Restoration” acts/“First Amendment Defense” acts (RFR and FADA, respectively), religious exemptions in health care, and single-sex facility restrictions. These topic areas were selected based on the American Civil Liberties Union’s (ACLU) annual compilation of legislation that affects LGBT rights. While the ACLU compiles an annual list of all proposed legislation that can potentially impact LGBT individuals, the seven topic areas we selected for analysis are the most commonly proposed “types” of anti-LGBTQ policies at the state level. Overlapping legislative categories can also be found at the National Conference of State Legislatures, which tracks “bathroom bills” and Religious Freedom and Restoration acts (RFRA) specifically. Note that, while other categories of laws could be classified as anti-LGBTQ, these are not included in the analysis. We identified 215 laws across 42 states that were proposed from January 1, 2015 to December 31, 2019. Our database includes all proposed legislation, not just those enacted into law. This data source is made available for future research (readers can download this data source here).

Crisis Text Line

To measure the level of youth seeking help through CTL, we used data on all conversations with CTL from January 1, 2017 to December 31, 2019, excluding those conversations in which the texter requested their data be scrubbed. The data from Crisis Text Line include one record for each conversation** initiated with CTL during this period. These data also included area code (used to identify participants’ state*). Approximately 20 percent of respondents completed a demographic survey, which allowed us to identify respondents under age 18 who self-identified as LGBTQ. For this analysis, we included only cell phone text conversations (not including web browser-based SMS) from the 50 states and Washington, DC. The total number of conversations with LGBTQ people under age 18 was 100,594. CTL does not track users over time, so the same individual might have multiple conversations with CTL across a span of time.

Data Limitations

One key caveat to our use of these data is that a relatively small number of conversations came from youth who identified as LGBTQ (n=100,594). We also did not analyze data separately by participant race/ethnicity. Without disaggregated data, we are unable to determine 1) how the racial diversity of LGBTQ CTL users may compare to the LGBTQ population nationally, and 2) whether LGBTQ people of different races respond to or experience anti-LGBTQ legislation differently. Additionally, our analysis did not track and account for the extent to which each piece of legislation was publicized. Theoretically, we would expect that highly publicized proposed laws (for example, laws that attract substantial media or social media attention) would impact LGBTQ youth mental health much more than less-publicized laws.

* Using area code to identify participants’ state is not a perfect measure, as studies suggest that about 10 percent of cell phone users may reside in a state that differs from their area code.

** A “conversation” is defined as a series of text messages back and forth between a crisis counselor and a texter who is seeking help. 

Technical Appendix

Policy Proposal Coding

Our source for policy information was Politico Pro. We created search terms to identify relevant anti-LGBTQ legislation in seven topic areas: excluding transgender youth from athletics, prohibiting health care for transgender youth, religious exemptions from marriage ceremonies, religious exemptions in foster care and adoption, “Religious Freedom Restoration” acts/“First Amendment Defense” acts, religious exemptions in health care, and single-sex facility restrictions. These topic areas were selected based on the American Civil Liberties Union’s (ACLU) annual compilation of legislation that affects LGBTQ rights. While the ACLU compiles an annual list of all proposed legislation that can potentially impact LGBTQ individuals, the seven topic areas we selected for analysis are the most frequently proposed anti-LGBTQ provisions at the state level. Note that, while other categories of laws could be classified as anti-LGBTQ, we did not include these in the analysis. Search terms were developed based on common verbiage used within each type of anti-LGBTQ legislation. For example, to identify legislation that falls under the “excluding trans youth from athletics” topic area, the following search terms were used either individually or in combination to generate results in Politico Pro: “athletic,” “interscholastic athletic,” “sport,” “biological sex,” “biological male,” “biological female,” “born as male,” “born as female,” “sex assigned at birth,” “testosterone,” “steroid,” and “hormone.” Our searches identified all proposed state legislation within each topic area, not just those enacted into law.

We reviewed all legislation (i.e., proposed policies) identified through our searches of Politico Pro and coded whether the proposed policy included provisions that would limit access to services or programs for LGBTQ youth. In the four topic areas related to “Religious Freedom Restoration” or other religious exemptions—from marriage ceremonies, in foster care and adoption, and in health care—any policy proposal that permitted an individual, organization, or agency to deny services or care to LGBTQ people was considered explicitly anti-LGBTQ. In the “excluding transgender youth from athletics” and “single-sex facility restrictions” topic areas, any policy proposal that prohibited youth from participating in school athletic teams or using single-sex public facilities (e.g., bathrooms, locker rooms), respectively—based on their gender identity rather than their biological sex—was considered explicitly anti-LGBTQ. In the “prohibiting health care for transgender youth” topic area, a policy proposal was considered explicitly anti-LGBTQ if it prohibited gender confirmation or related procedures or treatments for youth under age 18, or if it prohibited health insurance policies from providing coverage for those procedures. Proposed policies in this topic area were not considered anti-LGBTQ if they only allowed gender reassignment or related surgeries when procedures were deemed “medically necessary.” While the “medically necessary” provision may pose barriers to transitioning for transgender youth, it also presents a clear path forward for gender confirmation or related treatments for youth under age 18.

Crisis Text Line Data

To measure the degree to which youth sought help through Crisis Text Line (CTL), we used data on all conversations with CTL from January 1, 2017 to December 31, 2019, excluding conversations in which the texter requested that their data be scrubbed. The data from CTL include one record for each conversation initiated with CTL during this period. These data also included area code, which was used to identify the participant’s state. While using area code to identify the participant’s state is not a perfect measure, studies suggest that around 90 percent of cell phone users reside in the state associated with their area code.* Approximately 20 percent of participants completed a demographic survey, which allowed us to identify their age and whether they identify as LGBTQ.

Additionally, for this analysis, we included only cell phone text conversations (excluding web browser-based SMS) from the 50 states and Washington, DC. For this analysis, we included only conversations from those who answered the demographic survey and who reported being LGBTQ and under age 18. The total number of records was 100,594. CTL does not track users over time, so the same individual might have multiple conversations with CTL across a span of time. Additionally, the contents of the conversations were not available to researchers for privacy reasons.

Analysis

To explore the difference that proposed legislation might make to LGBTQ youth’s well-being, we chose to compare, within each state, text volume in the 28 days after a piece of legislation was proposed to text volume during days without recent proposed legislation in that state. We chose the period of 28 days because we expected laws to receive more publicity in the first few weeks after they were proposed.

The table below shows an example of how the data are structured.

There were 55,845 day-state pairs (365 days, over 3 years in 51 states) in this data set; for 2,054 of these records, a law was proposed in the same state within the past 28 days. We performed an exploratory analysis of these data using a mixed effects model, specified as below:

Where:

And:

In statistical software notation, this relationship might be considered as such:

Count of Texts LGBT Youth = Intercept + Indicator of law in past 28 days + (1 + year| state)

The specified model is a mixed effects model, with an intercept and a fixed effect for the indicator variable “Law proposed in past 28 days” (this is the variable of interest), a random intercept for state, and a random slope for state-level year indicator variables. Following the often-cited logic of Searle, Casella, and McCulloch,** we specified the variable of interest as a fixed effect. Because the dependent variable is a count, we used a poisson link function to fit the model. We chose the random slope and intercept model to account for substantial difference in text volume across states, and the increase in overall text volume to the CTL over time (which might vary between states). We found a small, positive relationship between recent proposal of anti-LGBTQ legislation and daily text volume; the fixed effect here was .037 (p=.02). This effect is small, corresponding to approximately 1.1 additional conversations per state per month, but is statistically significant. The table below shows the results from this model.

We recognize that this is not the only way this relationship could have been modeled: An event study, interrupted time series, or growth curve model might also be used to further investigate. There are also more control variables that could be explored, and many alternative specifications of the multilevel model could be used.

Limitations and Future Directions

Data Limitations

One key caveat to our use of these data is that a relatively small number of conversations came from respondents who identified as LGBTQ and were under age 18. Because only 20 percent of users filled out the demographic survey, and only a minority of these users were LGBTQ youth, our sample size is relatively small. Future research could benefit from increased sample size (either by running the study over a longer period of time or augmenting the data with other sources of information on youth mental health).

Future Directions

Overall, this exploratory analysis could serve as the foundation for future research that could make more quantifiable claims about the magnitude of this relationship, or about its causal nature. Future research might:

  • Explore whether the type of legislation For example, future research could answer whether laws with more direct relevance to youth (such as transgender sports bills) have a greater impact on youth mental health.
  • Investigate alternative time periods during which a proposed law might have influenced help-seeking behavior by LGBTQ youth. For example, would we observe a stronger effect by examining only the week in which a piece of anti-LGBTQ legislation was proposed? Or would the relationship be stronger if we looked only at the time period after the law is proposed? Additionally, does the status of the law (when the bill is voted on or passed through the legislative houses) relate to youth’s help-seeking behavior?
  • Integrate multiple sources of data on youth help-seeking behavior. Future research might understand the relationship explored here in more depth by including data from other crisis services, school climate surveys, and social media, among other sources that give a broader picture of youth help-seeking behavior. Additionally, alternative sources of data might help us better understand differences due to race/ethnicity.
  • Explore alternative statistical techniques (such as difference-in-difference models, interrupted time series, or alternative specification of mixed-effects models) that might identify whether the observed difference in text volumes is attributable to proposed legislation.
  • Identify a control group. The minority stress model leads us to believe that LGBTQ youth would be much more influenced by anti-LGBTQ legislation than their straight and cisgender peers. If future research used a data source with demographic information on all respondents, we could better compare how help-seeking behavior differed between LGBTQ and non-LGBTQ youth.
  • Investigate whether a dose-response relationship exists. Does the number of laws proposed in a given time frame increase the effect on help-seeking behavior?
  • Include information on how much media attention or publicity a law received and model whether this exacerbates the effect of anti-LGBTQ legislation.
  • Explore whether the presence of past LGBTQ-affirming legislation or other protective factors dampen the potential effect of anti-LGBTQ legislative proposals.

* Jiang, N., Jin, Y., Skudlark, A., & Zhang, Z. L. (2013). Greystar: Fast and Accurate Detection of {SMS} Spam Numbers in Large Cellular Networks Using Gray Phone Space. In 22nd {USENIX} Security Symposium ({USENIX} Security 13) (pp. 1-16).

** Searle, S.R., Casella, G. and McCulloch, C.E. (1992) Variance Components. Wiley, New York.
http://dx.doi.org/10.1002/9780470316856


* This brief uses the acronym LGBTQ, which stands for lesbian, gay, bisexual, transgender, queer, and/or questioning. At times we employ variations of this acronym (such as LGB, or LGBQ), which are used intentionally to reflect the populations represented in specific data collections.