The active control arm is an email-based intervention, EMAILS, in which existing public resources for supporting, affirming, and protecting LGBTQ+ students are emailed to participants.

The investigators will divide the number of participants who adhere to the PLANTS intervention by the total number of school staff who were enrolled at schools randomly assigned the PLANTS intervention and multiply by 100 to obtain a percentage. Adherence is a composite variable ranging from 0-100% comprised of: 55% for online module completion (based on the number of completed items divided by the total number of items offered); 45% for Live Zoom Event attendance (where each event is 15%).

Participants will complete the eleven-item Active-Empathic Listening Scale at baseline and follow-up. Each item uses a seven-point Likert scale for responses, coded from 1 ("never or almost never true") to 7 ("Always or almost always true"). Mean scores will be calculated for all eleven items at each time point. The minimum possible mean value is 1 and the maximum mean value is 7. Higher values represent greater active-empathic listening. The mean scores from baseline will be subtracted from the mean scores from follow-up to produce the change in mean scores. Higher values represent greater increases in active-empathic listening from baseline to follow-up.

The investigators have developed items pertaining directly to PLANTS' self-efficacy change objectives. Participants will complete each item using a five-point Likert scale, where 1 represents "not at all certain" and 5 represents "extremely certain." Mean scores will be calculated for all items at each time point. The minimum possible mean value is 1 and the maximum mean value is 5. Higher values represent greater self-efficacy. The mean scores from baseline will be subtracted from the mean scores from follow-up to produce the change in mean scores. Higher values represent greater increases in self-efficacy from baseline to follow-up.

Participants will complete the 16-item Teacher Bystander Intervention Model in Traditional Bullying scale, which has five subscales. Each response will be coded on a scale of 1 to 5, with 1 representing "strongly disagree" and 5 representing "strongly agree." Mean scores will be calculated for all items at each time point. The minimum possible mean value is 1 and the maximum mean value is 5. Higher values represent greater skills related to bystander intervention in bullying. The mean scores from baseline will be subtracted from the mean scores from follow-up to produce the change in mean scores. Higher values represent greater increases in bystander intervention skills from baseline to follow-up.

Participants will complete the 16-item Teacher Bystander Intervention Model in Cyberbullying scale, which has five subscales. Each response will be coded on a scale of 1 to 5, with 1 representing "strongly disagree" and 5 representing "strongly agree." Mean scores will be calculated for all items at each time point. The minimum possible mean value is 1 and the maximum mean value is 5. Higher values represent greater skills related to bystander intervention in cyberbulling. The mean scores from baseline will be subtracted from the mean scores from follow-up to produce the change in mean scores. Higher values represent greater increases in bystander intervention skills from baseline to follow-up.

The investigators have modified the language of the Gay Affirmative Practice Scale to incorporate school-oriented words instead of therapy-oriented words to measure self-efficacy for school staff working with SGMY. Participants will complete the nine-item scale at baseline and follow-up, with each item using a five-point Likert scale ranging from "strongly disagree" (1) to "strongly agree" (5). Mean scores will be calculated for all items at each time point. The minimum possible mean value is 1 and the maximum mean value is 5. Higher values represent greater self-efficacy. The mean scores from baseline will be subtracted from the mean scores from follow-up to produce the change in mean scores. Higher values represent greater increases in self-efficacy from baseline to follow-up.

Investigators assess myriad safety outcomes including parent backlash, social media backlash, school board backlash, suspension or removal from employment, censorship of LGBTQ+ literature/history/stories or removal of books with LGBTQ+ representation from school libraries, and emotional discomfort with intervention and control conditions at follow-up. Response options include frequency of each event occurrence (none, once, twice, 3-9 times, and 10 or more times). Investigators will report the overall frequency of events, frequency of each type of event, and percentage of school staff reporting any event.

Comparison schools will receive emails with publicly available resources for supporting LGBTQ+ students as a control intervention.

EMAILS: The active control arm is an email-based intervention, EMAILS, in which existing public resources for supporting, affirming, and protecting LGBTQ+ students are emailed to participants.

Total of all reporting groups

Duke University Religious Index (DUREL)

Analyzed as the mean response to the item "How often do you attend religious meetings (e.g. church)?" where 1 means "Never" and 6 means "More than once/week".

Duke University Religious Index (DUREL)

Analyzed as the mean response to the item "How often do you spend time in private religious activities, such as prayer, meditation or Bible study?" where 1 means "Rarely or never" and 6 means "More than once a day".

Duke University Religion Index (DUREL)

Calculated as the mean response across 3 items (ex. "In my life, I experience the presence of the Divine (i.e., God)"), each assessed from 1-5, where 1 means "Definitely not true" and 5 means "Definitely true". Higher values indicate more intrinsic religiosity.

.

Marlowe-Crowne Social Desirability Scale (13-item)

Each participant receives a score from 0-13, where 1 point is added for a "True" or a "False", depending on the item. An example item is "No matter who I'm talking to, I'm always a good listener." A higher score indicates a tendency towards socially desirable responses.

At follow-up, participants will complete the ten-question System Usability Scale (SUS) about PLANTS. Each question uses a five-point Likert scale for responses, coded with a value from 1 (Strongly agree) to 5 (Strongly disagree). The odd-numbered question responses will be reverse coded. All items will be summed and that sum will be multiplied by 2.5 to produce the final SUS value (ranging from 0-100). The higher the computed value, the more useable the intervention.

Intervention participants who completed the follow-up survey

At follow-up, participants will complete the four-question Intervention Appropriateness Measure (IAM) about PLANTS. Each question uses a five-point Likert scale for responses, coded with a value from 1 (Completely disagree) to 5 (Completely agree) and averaged to produce the mean appropriateness value. The minimum possible mean value is 1 and the maximum mean value is 5. The higher the computed mean value, the more appropriate the intervention.

Intervention participants who completed the follow-up survey

At follow-up, participants will complete the four-question Feasibility of Intervention Measure (FIM) about PLANTS. Each question uses a five-point Likert scale for responses, coded with a value from 1 (Completely disagree) to 5 (Completely agree) and averaged to produce the mean feasibility value. The minimum possible mean value is 1 and the maximum mean value is 5. The higher the computed mean value, the more feasible the intervention.

Intervention participants who completed the follow-up survey

The investigators will divide the number of school staff who consent to participate by the total number of school staff who are invited, then multiply by 100 to get a percentage.

The investigators will divide the number of school staff who complete the follow-up survey by the total number of school staff who were enrolled, then multiply by 100 to get a percentage.

The investigators will divide the number of participants who adhere to the PLANTS intervention by the total number of school staff who were enrolled at schools randomly assigned the PLANTS intervention and multiply by 100 to obtain a percentage. Adherence is a composite variable ranging from 0-100% comprised of: 55% for online module completion (based on the number of completed items divided by the total number of items offered); 45% for Live Zoom Event attendance (where each event is 15%).

Intent-to-treat intervention population

Participants will complete the eleven-item Active-Empathic Listening Scale at baseline and follow-up. Each item uses a seven-point Likert scale for responses, coded from 1 ("never or almost never true") to 7 ("Always or almost always true"). Mean scores will be calculated for all eleven items at each time point. The minimum possible mean value is 1 and the maximum mean value is 7. Higher values represent greater active-empathic listening. The mean scores from baseline will be subtracted from the mean scores from follow-up to produce the change in mean scores. Higher values represent greater increases in active-empathic listening from baseline to follow-up.

Participants who completed the follow-up survey

The investigators have developed items pertaining directly to PLANTS' self-efficacy change objectives. Participants will complete each item using a five-point Likert scale, where 1 represents "not at all certain" and 5 represents "extremely certain." Mean scores will be calculated for all items at each time point. The minimum possible mean value is 1 and the maximum mean value is 5. Higher values represent greater self-efficacy. The mean scores from baseline will be subtracted from the mean scores from follow-up to produce the change in mean scores. Higher values represent greater increases in self-efficacy from baseline to follow-up.

Participants who completed the follow-up survey

Participants will complete the 16-item Teacher Bystander Intervention Model in Traditional Bullying scale, which has five subscales. Each response will be coded on a scale of 1 to 5, with 1 representing "strongly disagree" and 5 representing "strongly agree." Mean scores will be calculated for all items at each time point. The minimum possible mean value is 1 and the maximum mean value is 5. Higher values represent greater skills related to bystander intervention in bullying. The mean scores from baseline will be subtracted from the mean scores from follow-up to produce the change in mean scores. Higher values represent greater increases in bystander intervention skills from baseline to follow-up.

Participants who completed the follow-up survey

Participants will complete the 16-item Teacher Bystander Intervention Model in Cyberbullying scale, which has five subscales. Each response will be coded on a scale of 1 to 5, with 1 representing "strongly disagree" and 5 representing "strongly agree." Mean scores will be calculated for all items at each time point. The minimum possible mean value is 1 and the maximum mean value is 5. Higher values represent greater skills related to bystander intervention in cyberbulling. The mean scores from baseline will be subtracted from the mean scores from follow-up to produce the change in mean scores. Higher values represent greater increases in bystander intervention skills from baseline to follow-up.

Participants who completed the follow-up survey

The investigators have modified the language of the Gay Affirmative Practice Scale to incorporate school-oriented words instead of therapy-oriented words to measure self-efficacy for school staff working with SGMY. Participants will complete the nine-item scale at baseline and follow-up, with each item using a five-point Likert scale ranging from "strongly disagree" (1) to "strongly agree" (5). Mean scores will be calculated for all items at each time point. The minimum possible mean value is 1 and the maximum mean value is 5. Higher values represent greater self-efficacy. The mean scores from baseline will be subtracted from the mean scores from follow-up to produce the change in mean scores. Higher values represent greater increases in self-efficacy from baseline to follow-up.

Participants who completed the follow-up survey

Investigators assess myriad safety outcomes including parent backlash, social media backlash, school board backlash, suspension or removal from employment, censorship of LGBTQ+ literature/history/stories or removal of books with LGBTQ+ representation from school libraries, and emotional discomfort with intervention and control conditions at follow-up. Response options include frequency of each event occurrence (none, once, twice, 3-9 times, and 10 or more times). Investigators will report the overall frequency of events, frequency of each type of event, and percentage of school staff reporting any event.

Comparison schools will receive emails with publicly available resources for supporting LGBTQ+ students as a control intervention.

EMAILS: The active control arm is an email-based intervention, EMAILS, in which existing public resources for supporting, affirming, and protecting LGBTQ+ students are emailed to participants.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Self-Efficacy in Identity-Affirming Practices, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Self-Efficacy in Identity-Affirming Practices, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Self-Efficacy in Inclusivity in Restrooms and Changing Options, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Self-Efficacy in Inclusivity in Restrooms and Changing Options, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Interpret the Event as an Emergency, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Interpret the Event as an Emergency, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Accept Responsibility for Intervening, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Accept Responsibility for Intervening, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Know How to Intervene, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Know How to Intervene, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Implement Intervention Decisions, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Implement Intervention Decisions, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Interpret the Event as an Emergency, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Interpret the Event as an Emergency, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Accept Responsibility for Intervening, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Accept Responsibility for Intervening, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Know How to Intervene, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Know How to Intervene, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Implement Intervention Decisions, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.

Threshold for statistical significance: <0.05

To test for within-arm statistical significance of the change in Implement Intervention Decisions, we used linear mixed models accounting for within-person clustering using random effects. We reported the beta coefficients and 95% confidence intervals.