Identifying and Classifying Harassment, Threats, and Trolling in Social Media and Community Surveys

In today's interconnected world, social media and community surveys have become essential tools for businesses and organizations to engage with their stakeholders. However, these platforms also present unique challenges, particularly for organizations serving diverse communities and addressing sensitive topics. This article explores the importance of social media presence and community surveys, while also addressing the critical issue of managing online threats and harassment.

The Power of Social Media and Community Surveys

Social media platforms and community surveys offer unprecedented opportunities for organizations to:

• Connect directly with their audience.

• Gather valuable feedback and insights.

• Build brand awareness and loyalty.

• Address concerns and issues in real-time.

• Foster a sense of community among stakeholders.

For organization serving a diverse clientele including LGBTQ+ organizations, these tools are invaluable for understanding and meeting the needs of various communities.

Case Study: LGBTQ+ Community Needs Assessment in Lancaster County

The launch of the LGBTQ+ Community Needs Assessment in Lancaster County serves as a compelling example of the power and challenges inherent in community surveys, particularly those addressing sensitive topics in the digital age.

As the Lancaster LGBTQ+ Community prepared to celebrate Pride in June 2024, FIC Human Resource Partners, in collaboration with four local LGBTQ+ organizations, embarked on an ambitious project. Their goal was to conduct a comprehensive needs assessment that would give voice to the LGBTQ+ community and inform future service, event, and program decisions.

The resilience of Lancaster's LGBTQ+ community was being tested on multiple fronts.

The Lancaster LGBTQ+ Community had recently been rocked by a rather vocal and nasty response to a planned Drag Queen Story Hour at the Lancaster Public Library.   After efforts were made to increase security and create a wall of love to protect attendees from an angry protest, the Story Hour was ultimately cancelled due to a bomb threat. The level of anger, acrimony and the very real threats had left the community shaken. But that would not deter the local LGBTQ+ Organizations or community members from planning and attending a month of LGBTQ+ celebrations and events. It was against this backdrop of celebration, challenge, and resilience that the needs assessment survey was launched.

The survey, carefully crafted to safeguard data integrity and protect respondents, went live on June 1st, coinciding with the start of Pride Month celebrations. FIC's team intentionally designed the survey to minimize the potential for trolling and harassment, a reflection of their commitment to creating a safe space for honest feedback.

Each partner organization received a unique set of survey links, QR codes, and embed codes – a strategic move that allowed FIC to monitor responses closely and quickly identify any links being circulated by potential trolls or anti-LGBTQ+ groups. As the organizations shared these links across their social media channels, the survey began to gain traction within the community.

However, the digital landscape proved to be as toxic as ever. One of the partner organizations found itself besieged by online trolls shortly after sharing the survey. Despite this digital onslaught, the survey team’s vigilant monitoring revealed a pleasant outcome – only a single trolling or harassing comment had made its way into the actual survey responses.

As the survey entered its second week, the local LGBTQ+ community was rocked by unexpected news. An unaffiliated LGBTQ+ organization announced its imminent closure, thrusting the community's needs into the media spotlight. The four partner organizations rose to the occasion, fielding interviews with television and print media. They released individual statements and a joint press release, reaffirming their commitment to the community and encouraging participation in the ongoing needs assessment survey.

Just days before the local pride festival, while largely successful, had been temporarily disrupted by a security scare involving a suspicious package.

This surge in attention brought both opportunities and challenges. Anticipating a potential increase in trolling and harassment, FIC’s survey team swiftly implemented additional safeguards to fortify the survey's integrity. Their foresight paid off – in the wake of these changes, only one more instance of trolling was identified in the survey responses.

As June progressed, the survey became more than just a data collection tool – it evolved into a symbol of the community's determination to be heard and to thrive in the face of adversity.

With the survey scheduled to run until the end of the month, FIC and the partner organizations remained committed to their goal. They continued to navigate the complex interplay of social media engagement, community outreach, and data security, all while striving to create a comprehensive picture of the LGBTQ+ community's needs in Lancaster County.

Challenges of Public Engagement

While social media and surveys offer numerous benefits, they also come with significant challenges, particularly for organizations dealing with sensitive topics or serving historically excluded, underrepresented, and marginalized communities. Some of these challenges include:

• Trolling and harassment

• Spread of misinformation

• Privacy concerns

• Managing diverse and sometimes conflicting viewpoints

• Potential security threats

Understanding Threat Actors and Their Impact

In managing social media presence and conducting surveys, particularly on sensitive topics, it's crucial to understand the various types of individuals who might pose a threat to the integrity of the process or the safety of participants. Here, we categorize these threat actors, describe their behaviors, and outline the potential harm and chaos they can cause:

Disruptive Trolls

(Low-level threat)

Behavior

Individuals whose primary intent is to disrupt the survey or conversation without necessarily being hostile.

Identification

• Responses are off-topic or irrelevant to the survey questions.

• Lack of serious or coherent answers.

• May use mild insults or mockery, but not targeted harassment.

Potential Harm

• Decreased quality of survey data.

• Wasted time and resources in managing responses.

• Minor frustration for genuine participants.

Example

"This survey is stupid. I'm just here for the lolz."

Hostile Trolls

(Low-level threat)

Behavior

Individuals who intend to harass, demean, or provoke others.

Identification

• Responses contain slurs, insults, or other hate speech.

• Language is aggressive and inflammatory, but not threatening violence.

• Targets specific individuals, groups, or the LGBTQ+ community as a whole.

Potential Harm

• Emotional distress for targets and other participants.

• Creation of a hostile online environment.

• Discouragement of genuine participation.

• Potential escalation of conflicts.

Example

"All you [slur] are going to hell. Stop shoving your agenda in our faces."

Threatening Trolls

(Medium-level threat)

Behavior

Individuals whose intent is to cause emotional harm or fear through threats.

Identification

• Responses contain threats or allusions to violence.

• Language is intended to scare, intimidate or cause emotional distress.

• Threats are vague or implied rather than specific.

• No indication that steps have been taken to carry out the threat.

Potential Harm

• Significant emotional and psychological distress.

• Chilling effect on free expression and participation.

• Potential for escalation to real-world harassment.

• Need for increased security measures.

Example

"You better watch your back. People like you get what's coming to them."

Potential Threat Actors

(Medium-level threat)

Behavior

Individuals who make specific threats indicating potential for real-world action.

Identification

• Threats are specific and targeted.

• Some details are provided about a potential attack.

• Still in the "planning" stage, no actions taken yet.

Potential Harm

• Severe emotional and psychological trauma.

• Disruption of daily activities and operations.

• Necessity for involvement of law enforcement.

• Potential for property damage or physical harm.

Example

"I'm going to come to your office and teach you a lesson you'll never forget."

Probable Threat Actors

(High-level threat)

Behavior

Individuals who make direct, specific, and plausible threats and indicate steps have been taken toward carrying them out.

Identification

• Threats are explicit, targeted, and actionable.

• Specific claims of weapons access, stalking behaviors, or attack plans.

• Language indicates attack is imminent and in progress.

Potential Harm

• Immediate danger to life and safety.

• Widespread panic and chaos.

• Long-lasting trauma for targeted individuals and communities.

• Severe disruption of operations and community activities.

• Extensive involvement of law enforcement and security personnel.

Example

"I have a gun and I've been watching your office. If you don't close down, I'll do it myself."

The presence of these threat actors in online spaces can create a cascade of negative effects. Beyond the immediate harm to targeted individuals or groups, they can:

• Erode trust within communities.

• Discourage participation in important discussions and surveys.

• Force organizations to divert significant resources to security and moderation.

• Create a climate of fear that stifles open dialogue and expression.

• Potentially escalate online conflicts into real-world violence.

Identifying and managing these threats requires a multi-faceted approach

• Vigilant monitoring of online spaces and survey responses.

• Well-trained staff capable of recognizing and appropriately responding to different threat levels.

• Robust security protocols and emergency response plans.

• Collaboration with law enforcement when necessary.

• Creation of supportive networks for affected individuals and communities.

By understanding the nature of these potential threat actors and their impacts, organizations can better prepare to maintain the integrity of their online presence and protect the well-being of their communities.

Understanding the Use of "Troll" in Threat Designations

In FIC's threat classification system, the term "troll" is used deliberately, drawing on two distinct but related definitions. This dual meaning enhances the descriptive power of the threat designations, capturing both the nature of the individuals involved and their actions. Let's explore these definitions and their relevance to online threat assessment:

1. Troll as a Noun: The Individual

In internet culture, a "troll" refers to a person who intentionally initiates conflicts, provokes emotional responses, or disrupts normal discussions in online communities. This usage originated in the 1990s and has become widely recognized in the context of online behavior.

Relevance to Threat Designation: By using "troll" in this sense, the threat designation immediately evokes the image of an individual who is not participating in good faith. It suggests someone whose primary goal is to cause disruption or distress, rather than to engage in meaningful dialogue or provide genuine feedback.

2. Troll as a Verb: The Action

"To troll" is a verb meaning to fish by trailing a baited line behind a boat. In the context of online behavior, it metaphorically refers to the act of posting inflammatory, insincere, digressive, or off-topic messages to provoke an emotional response or disrupt normal discussions.

Relevance to Threat Designation: This definition emphasizes the active nature of the behavior. It's not just about who these individuals are, but what they do. The use of "troll" in this sense captures the deliberate, ongoing nature of their disruptive actions.

3. Combining Noun and Verb: A Comprehensive Descriptor

By using "troll" in the threat designations (e.g., "Disruptive Troll," "Hostile Troll"), FIC's classification system achieves several important goals:

Conciseness: It encapsulates both the identity of the threat actor and the nature of their actions in a single, widely understood term.

Accuracy: It accurately describes both who these individuals are in the online ecosystem and what they're doing to cause disruption.

Flexibility: The term can be easily modified (e.g., "disruptive," "hostile") to denote different levels or types of trolling behavior.

Cultural Relevance: It uses terminology that is widely recognized and understood in the context of online interactions, making the threat designations immediately comprehensible to those familiar with internet culture.

Action-Oriented: By invoking both the noun and verb forms, it implies that these are not just static labels, but descriptions of ongoing, active behavior that needs to be addressed.

This dual-meaning approach in the threat designation system helps FIC and its partners quickly understand and communicate about the nature of potential threats in their online surveys and social media interactions. It provides a nuanced yet accessible framework for classifying and responding to disruptive online behaviors, from mild annoyances to serious security concerns.

FIC's Comprehensive Approach to Survey Security and Integrity

In conducting the LGBTQ+ Community Needs Assessment, FIC Human Resource Partners implemented a robust set of measures to ensure the security, integrity, and validity of the survey data. These measures not only protected the survey from potential threats but also safeguarded the privacy and safety of participants.

Technical Security Measures

Response Identification: Each survey response was assigned a unique identifier, allowing for precise tracking and analysis without compromising participant anonymity.

IP Address and Geolocation Collection: To detect potential multiple submissions or coordinated attacks, FIC collected IP addresses and geolocation data. This information was encrypted and accessible only to the data analysis team.

Multiple Links with Unique Identifiers: FIC provided each partner organization with a set of unique survey links, QR codes, and embed codes. This strategy allowed for tracking the source of responses and quickly identifying any links being circulated by potential bad actors.

Data Encryption: All collected data was encrypted both in transit and at rest, ensuring that sensitive information remained protected from unauthorized access.

Survey Logic and Piping: Advanced survey logic was implemented to adapt questions based on previous responses, making it more difficult for trolls to provide nonsensical or contradictory answers.

Policy and Procedural Measures

Clear Public-Facing Data and Privacy Policies: FIC developed and prominently displayed comprehensive data usage and privacy policies, ensuring transparency and building trust with participants.

Qualifying Questions: The survey included qualifying questions to ensure that respondents met the criteria for participation and to filter out potential bad actors.

Raw Data Access Restrictions: Access to raw survey data was strictly limited to the data analysis team, with stringent protocols in place for data handling and storage.

Daily Response Monitoring: FIC implemented a system of daily monitoring of survey responses to quickly identify and address any emerging issues or threats.

Response Evaluation Process: Suspect responses were quarantined and evaluated by trained staff. Responses identified as trolling, harassing, or threatening were disqualified from the dataset.

Threat Management

Threat Classification System: FIC developed a comprehensive system for classifying different levels of threats, from low-level trolling to credible threats of violence.

Staff Training: All team members involved in monitoring and analyzing survey responses received training on identifying and classifying potential threats.

Threat Documentation: A secure list was maintained documenting all trolling, harassing, and threatening responses, including threat designations and all relevant identifying data.

Law Enforcement Liaison Protocol: FIC established clear protocols for reporting credible threats to law enforcement, including procedures for disclosing relevant information in the event of an act of violence.

Continuous Improvement

Throughout the survey period, FIC continuously evaluated the effectiveness of these measures and made adjustments as necessary. This adaptive approach allowed them to respond quickly to emerging challenges and maintain the highest standards of data integrity and participant safety.

By implementing this comprehensive set of security measures, FIC demonstrated its commitment to conducting ethical, secure, and valuable research, even in the face of potential online threats and challenges. This approach not only protected the integrity of the LGBTQ+ Community Needs Assessment but also set a high standard for future survey projects dealing with sensitive topics or vulnerable populations.

Case Study: Evaluating and Classifying Threat Levels in Survey Responses

To illustrate how FIC's threat evaluation and classification system works in practice, let's examine actual responses received during the LGBTQ+ Community Needs Assessment survey. These examples demonstrate the nuanced approach required when assessing potential threats in survey responses.

Example 1: Hostile Troll

Response ID: 194

Session ID: 1717771491_66631ce37682c3.16490044

Response Content:

1. "mental health assistance for f****d up people”

2. “how many f****d lqtb idiots there are”

3. “stop pushing this s**t”

4. “Mental Disability"

Evaluation Process:

Content Analysis: The response contains offensive language and derogatory terms directed at the LGBTQ+ community. It expresses hostility towards the survey's purpose and the community it serves.

Intent Assessment: The language used suggests an intent to provoke and demean, rather than to contribute meaningfully to the survey.

Threat Level Determination: While highly offensive, the response does not contain specific threats of violence or indicate potential for real-world action.

Classification: Hostile Troll (Low-level Threat)

Rationale:

This response meets the criteria for a Hostile Troll as it:

• Uses inflammatory and derogatory language.

• Targets the LGBTQ+ community as a whole.

• Intends to harass and provoke.

• Does not include specific threats or indications of potential real-world action.

Action Taken:

The response was disqualified from the survey dataset and added to the list of documented hostile responses for future reference and pattern analysis.

Example 2: Disruptive Troll

Response ID: 262

Session ID: 1718792918_6672b2d6d67c56.47401066

Response Content:

1. "We are a conservative community. These organizations are sin and pride destruction.”

2. “Leave our children out of your plans. God's plan is true love. And the only way to Heaven. Your group is misleading whole generations of people and they will be lost for Eternity."

Evaluation Process:

Content Analysis: The response expresses opposition to LGBTQ+ organizations based on religious beliefs. It includes accusatory language but does not use explicit slurs or directly abusive terms.

Intent Assessment: While the intent appears to be expressing disagreement and religious viewpoints, it also serves to disrupt the survey's purpose by introducing off-topic arguments unrelated to identifying the needs and experiences of the LGBTQ+ population in Lancaster County. In fact, the presence of this response supported the level of fear and lack of safety LGBTQ+ respondents expressed in relation to religious attacks and harassment.

Threat Level Determination: The language is oppositional and mildly insulting but does not contain threats of violence or indicate potential for immediate real-world action.

Classification: Disruptive Troll (Low-level Threat)

Rationale:

This response meets the criteria for a Disruptive Troll as it:

• Introduces off-topic arguments into an LGBTQ+ community needs assessment survey.

• Uses mildly insulting language.

• Serves to disrupt the quality of the survey data by providing irrelevant information.

• Does not contribute meaningfully to the survey's intended purpose.

• While expressing opposition, it doesn't rise to the level of a Hostile Troll due to the absence of explicit slurs or directly abusive language.

Note:  Had this been a different type of community survey with a different focus this response may not have been classified as a Disruptive Troll. The context of the survey, its intended audience, and its purpose all factor into whether this particular response met the definition of Disruptive Troll.

Action Taken:

The response was disqualified from the survey dataset to maintain data quality. It was added to the list of documented disruptive responses for future reference and pattern analysis. This type of response also informed potential adjustments to survey questions or instructions to minimize off-topic submissions in future surveys.

Example 3: Disruptive Troll

Response ID: 333

Session ID: 1719070766_6676f02e281879.13536444

Response Content:

1. Respondent provided answers that were strings of letter numbers and symbols with no apparent meaning.

2. Respondent indicated participation in groups meant for both LGBTQ+ Youth and LGBTQ+ Elders.

3. Respondent provided first and last names that were both multi-word responses.

Evaluation Process:

Content Analysis: The response contains a mix of nonsensical phrases, repetitive patterns, and unconventional use of symbols and characters. While some answers are within expected parameters (e.g., "Pangender" for gender), others are clearly nonsense, off-topic, or invented terms.

Intent Assessment: The respondent appeared to be deliberately providing nonsensical or exaggerated responses rather than genuinely engaging with the survey questions.

Additional Investigation: A web search was conducted using the name provided by the respondent.

Threat Level Determination: While the responses are disruptive to the survey's purpose, they do not contain any hostile language, threats, or indications of potential harm.

Classification: Disruptive Troll (Low-level Threat)

Rationale:

This response meets the criteria for a Disruptive Troll as it:

• Provides off-topic and irrelevant answers to many questions

• Uses nonsensical phrases and symbol patterns repeatedly

• Appears to be intentionally disrupting the survey rather than providing genuine feedback

• Does not contain hostile language or threats

Action Taken:

Standard Action Taken

The response was disqualified from the survey dataset to maintain data quality. It was added to the list of documented disruptive responses for future reference and pattern analysis. This type of response also informed potential adjustments to survey questions or instructions to minimize off-topic submissions in future surveys.

Additional Action Taken

In addition to the actions that are taken for responses classified as Disruptive Troll, the survey team conducted a web search of the provided name and identified that it was used across several social platforms to engage in similarly disruptive behavior.

Implications of the Additional Action taken

Cross-platform Behavior Pattern: The discovery of similar disruptive behavior across multiple platforms suggests a consistent pattern of trolling, rather than a one-time occurrence.

Enhanced Threat Assessment: While the initial classification remains "Disruptive Troll (Low-level Threat)," the cross-platform nature of the behavior provides valuable context for future interactions or responses from this individual.

Improved Screening Measures: This information can be used to develop more robust screening measures for future surveys, potentially including checks against known troll accounts or patterns.

Data Integrity Confirmation: The external validation of the troll-like behavior confirms the decision to disqualify the response, ensuring the integrity of the survey data.

This additional step in the evaluation process demonstrates the importance of thorough investigation and the value of cross-referencing information when assessing potential threats or disruptive behavior in survey responses.

Key Takeaways from These Examples

Nuanced Evaluation and Spectrum of Responses: Surveys on sensitive topics may receive a wide range of responses, from supportive to oppositional, subtly disruptive to outright hostile. The spectrum of disruption can range from subtle opposition to nonsensical submissions, and each type requires a tailored approach. Each response, regardless of its apparent threat level, requires careful evaluation within its context.

Intent vs. Content vs. Impact: While offensive language is a clear indicator of hostility, more subtle forms of disruption, such as introducing irrelevant arguments or providing nonsensical answers, can also compromise survey integrity. The intent of respondents may vary, but the primary concern for survey administrators is the impact on data quality and the overall integrity of the research.

Balancing Inclusivity and Data Integrity: Survey designers face the challenge of being inclusive of diverse identities and experiences while maintaining the integrity and usefulness of the data collected. This balance is crucial for ensuring that the survey accurately represents the community while filtering out disruptive or misleading responses.

Data Quality and Security Considerations: Threat assessment in surveys isn't just about identifying potential security risks; it's also about maintaining the quality and relevance of the data collected. Each response must be evaluated not only for its potential threat level but also for its contribution to or detraction from the survey's objectives.

Context Matters: Understanding the broader context of the survey – its purpose, target audience, and the current social climate – is crucial in evaluating and responding to potentially disruptive submissions. This contextual awareness helps in distinguishing between genuine responses that may seem unusual and intentionally disruptive ones.

Continuous Improvement and Adaptive Strategies: These examples demonstrate the need for ongoing refinement of threat classification systems, survey methodologies, and response strategies. Each disruptive response provides an opportunity to refine survey questions, add validation checks, or improve instructions for future surveys. This adaptive approach ensures that the survey process becomes more robust and effective over time.

Consistent Classification and Evaluation System: By carefully evaluating each response and applying a consistent classification system, organizations can ensure the integrity of their survey data while also maintaining awareness of potential threats or areas of concern within the community. This systematic approach aids in both data analysis and security management.

By keeping these takeaways in mind, organizations conducting sensitive surveys can better navigate the challenges of data collection, ensure the safety of participants and staff, and maintain the integrity of their research. This approach not only improves the quality of the data collected but also helps in building trust within the community being surveyed.

Best Practices for Managing Online Presence and Surveys

To navigate these challenges effectively, organizations should consider implementing the following best practices:

• Develop a comprehensive social media policy.

• Train staff on identifying and responding to different types of online threats.

• Implement robust security measures for social media accounts and survey platforms.

• Use moderation tools and techniques to manage comments and responses.

• Establish clear guidelines for survey participants.

• Partner with local law enforcement and security experts for high-level threats.

• Prioritize the safety and well-being of staff and community members.

• Regularly review and update security protocols.

For organizations that may be targeted for trolling, harassment, and threats it's crucial to develop policies that:

• Protect employees and clients from discrimination and harassment.

• Promote a diverse and welcoming online environment.

• Address the unique challenges faced by historically excluded, underrepresented, and marginalized groups.

• Demonstrate a commitment to equality and social justice.

• Ensure the physical safety and security of employees and clients.

• Detail procedures for identifying and responding to threats and dangers.

Conclusion

Navigating the complex landscape of social media and community surveys requires a delicate balance between openness and security. By understanding the potential threats, implementing best practices, and fostering an inclusive environment, organizations can harness the power of these tools while mitigating risks. As we continue to rely on digital platforms for engagement and research, it's crucial to remain vigilant, adaptive, and committed to creating safe spaces for all community members.

If your organization is looking for a secure, thorough, well managed survey or assessment partner, our Nuance Culture Consulting team is here to help!

Contact us today to learn how we can help.