Privacy Policy

Research Shield™ ("we", "us", "our") respects your privacy. This Privacy Policy explains what data we collect, how we use it, and what rights you have.

1. Who We Are

Research Shield™ is a data quality assurance system for market research. We help research companies verify the authenticity of survey responses.

Data Controller:
TGM Research Pte. Ltd.
6001 Beach Road, #22-01 Golden Mile Tower
Singapore 199589
privacy@researchshield.com

Data Protection Officer:
Marcin Kaleta
privacy@tgmpanel.com
Postal correspondence: use the address above with notation "Data Protection Officer"

EU Representative (Art. 27 GDPR):
TGM Research Europe
[Address in EU Member State]
eu-representative@researchshield.com

UK Representative (UK GDPR Art. 27):
TGM Research UK
[Address in United Kingdom]
uk-representative@researchshield.com

2. What Data We Collect

2.1 Technical Data

• IP address and approximate geographic location
• Browser and device type
• Operating system
• Screen resolution
• Time zone
• Device identifier (fingerprint)

2.2 Behavioral Data

For fraud detection purposes, we analyze:

• Typing patterns and keystroke timing — the rhythm, speed, and pauses in your typing
• Mouse movements and click patterns — cursor trajectory, click locations, scrolling behavior
• Response timing — time spent on each question and navigation patterns
• Browser window activity — focus events, tab switching, window visibility

2.3 Survey Data

• Responses provided in the survey
• Session identifier

We do not collect: names, email addresses, or other directly identifying information unless they are part of the survey commissioned by the client.

3. Why We Process Data

Purpose	Legal Basis (GDPR)
Fraud and bot detection	Legitimate interest (Art. 6.1.f)
Ensuring research data quality	Legitimate interest (Art. 6.1.f)
Delivering the survey for the client	Contract performance (Art. 6.1.b)
Service analysis and improvement	Legitimate interest (Art. 6.1.f)

Legitimate Interest

Our legitimate interest is protecting the integrity of market research and our clients' interests against fraud. Processing behavioral data is necessary to achieve this purpose and does not unduly infringe on the rights of data subjects.

Clarification: Which Data Falls Under Which Legal Basis

Behavioral data (typing patterns, mouse movements, device characteristics, interaction timing) is processed exclusively under our legitimate interest in fraud prevention (Art. 6.1.f). This data is not required to deliver your survey responses to the client.

Survey response data (your answers to survey questions) is processed under contract performance (Art. 6.1.b) on behalf of the research client who commissioned the survey.

These are distinct processing operations with separate legal bases, as required by GDPR and EDPB Guidelines 2/2019.

4. Who We Share Data With

4.1 Research Clients

Survey responses and data quality information are shared with the company commissioning the research.

4.2 Sub-Processors and Their Roles

We use the following technical service providers to deliver our fraud detection service. For each, we disclose what data is shared, how they process it, and their legal role under GDPR.

Amazon Web Services (AWS)

Data shared	All data collected by Research Shield™
Purpose	Cloud hosting and data storage
Location	EU (Frankfurt, eu-central-1)
Role	Processor — processes data only on our instructions
Cross-client data pooling	No — data isolated per customer
DPA in place	Yes (AWS Data Processing Addendum)

FingerprintJS, Inc.

Data shared	Device signals: browser type, OS, screen resolution, installed fonts, WebGL renderer, audio context, canvas fingerprint, timezone
Purpose	Generate device identifier (fingerprint) for duplicate detection
Location	USA
Role	Joint Controller (Art. 26) — FingerprintJS uses aggregated, anonymized data to improve their identification algorithms
Cross-client data pooling	Yes — anonymized signals contribute to their global identification model
DPA in place	Yes + Joint Controller Arrangement (Art. 26)
Their privacy policy	fingerprint.com/privacy

Art. 26 Arrangement Summary: FingerprintJS acts as joint controller for the purpose of improving their device identification technology. We remain responsible for collecting device signals and for decisions about individual fraud assessments. FingerprintJS is responsible for the aggregated model improvement. You may exercise your rights with either party.

IPQualityScore (IPQS)

Data shared	IP address (hashed where possible), user agent string
Purpose	Query fraud reputation database, detect VPN/proxy/bot usage
Location	USA
Role	Joint Controller (Art. 26) — IPQS maintains a fraud reputation database enriched by queries from all customers
Cross-client data pooling	Yes — IP reputation scores are derived from cross-client fraud signals
DPA in place	Yes + Joint Controller Arrangement (Art. 26)
Their privacy policy	ipqualityscore.com/privacy-policy

Art. 26 Arrangement Summary: IPQS acts as joint controller for maintaining their fraud reputation database. We query their database with IP addresses; they use this data (alongside queries from other customers) to improve fraud detection accuracy across their network. We remain responsible for how we use IPQS scores in our fraud assessments. You may exercise your rights with either party.

Sentry (Functional Software, Inc.)

Data shared	Error logs, stack traces, minimal session context (PII scrubbed before transmission)
Purpose	Application error monitoring and debugging
Location	USA
Role	Processor — processes data only on our instructions for error monitoring
Cross-client data pooling	No — error data not shared across customers
DPA in place	Yes (Sentry Data Processing Addendum)
Their privacy policy	sentry.io/privacy

4.3 International Data Transfers

Some personal data is transferred to countries outside the European Economic Area (EEA), particularly the United States. We ensure appropriate safeguards for such transfers:

Recipient	Country	Primary Safeguard	Fallback Safeguard	TIA Conducted
FingerprintJS, Inc.	USA	EU-US Data Privacy Framework	SCCs (2021/914)	Yes
IPQS (IPQualityScore)	USA	SCCs (2021/914)	—	Yes
Sentry (Functional Software)	USA	EU-US Data Privacy Framework	SCCs (2021/914)	Yes

Transfer Impact Assessments (TIA)

In accordance with the Schrems II ruling (C-311/18), we have conducted Transfer Impact Assessments for each US recipient. These assessments evaluate:

• The legal framework in the recipient country
• The nature of data transferred (pseudonymized technical/behavioral data)
• Likelihood of access by public authorities
• Effectiveness of supplementary measures

TIA Conclusion: Given the technical nature of data (device fingerprints, behavioral patterns), pseudonymization, encryption, and the limited personal identifiability, the risk of harm from potential government access is assessed as low. Supplementary measures provide effective protection.

Supplementary Measures

Standard Contractual Clauses (SCCs) are legal contracts approved by the European Commission (Decision 2021/914) that bind data recipients to protect personal data to EU standards. We implement the following supplementary measures for all US transfers:

• Technical: Encryption of data in transit (TLS 1.3) and at rest (AES-256)
• Technical: Pseudonymization of personal identifiers before transfer
• Organizational: Access controls and audit logging
• Organizational: Data Processing Agreements (DPAs) with all sub-processors
• Contractual: Notification obligations if recipient receives government access requests

Adequacy Decision Contingency

If the EU-US Data Privacy Framework adequacy decision is invalidated (as occurred with Safe Harbor and Privacy Shield), we will rely on Standard Contractual Clauses as the transfer mechanism. SCCs are already in place with all US recipients as a fallback safeguard. We monitor regulatory developments and will implement additional measures as required.

You may request a copy of the relevant SCCs or TIA summaries by contacting privacy@researchshield.com.

4.4 No Advertising or Marketing Use

We do not use your data for advertising, marketing profiling, or any purpose unrelated to fraud detection. Specifically:

• We do not sell your personal data
• We do not share data with advertising networks
• We do not create marketing profiles
• We do not use data for targeted advertising
• We do not share data with social media platforms for advertising

4.5 Other Disclosures

We may disclose data when required by law, court order, or law enforcement authorities.

5. How Long We Keep Data

We apply tiered retention periods based on the data minimization principle. Each category is retained only as long as necessary for its specific purpose:

Data Category	Retention Period	Justification
Raw behavioral data (typing patterns, mouse movements)	90 days	Session fraud scoring and ML model training; derived scores retained instead
IP addresses	6 months	Short-term duplicate detection; longer retention unnecessary
Device fingerprints	24 months	Required for longitudinal device reuse detection across survey waves
Fraud scores and flags	24 months	Required for recurring fraud pattern analysis across clients
Session metadata (timing, navigation)	12 months	Quality assurance and system performance analysis
Aggregated statistics	Indefinitely	Anonymized; no personal data

After each retention period, data is automatically deleted or irreversibly anonymized. Raw behavioral data is converted to derived fraud indicators before deletion.

6. Your Rights

Under GDPR, you have the following rights:

• Access - you can obtain information about processed data
• Rectification - you can correct inaccurate data
• Erasure - you can request deletion of data ("right to be forgotten")
• Restriction - you can request limitation of data processing
• Portability - you can receive data in a machine-readable format
• Objection - you can object to processing based on legitimate interest

Granular Objection Rights

You may object to specific aspects of our processing. After survey completion, you can object to:

• Continued retention of your data — we will delete your data upon valid objection
• Cross-session use of device fingerprints — we will remove your device from our duplicate detection database
• ML training use — we will exclude your data from machine learning model improvement

Such objections will be assessed individually and can be honored without affecting your past survey participation. See Section 7.4 for the full objection process.

How to Exercise Your Rights

You can exercise your rights in two ways:

• Online: Visit our Data Rights page for detailed instructions and options
• Email: Send a request to privacy@researchshield.com

We will respond within 30 days.

Complaint to Supervisory Authority

You have the right to lodge a complaint with your local data protection authority.

7. Fraud Detection and Automated Decision-Making

7.1 How Fraud Detection Works

We use automated analysis systems to detect inauthentic responses (bots, scripts, fraud). The system analyzes behavioral patterns and technical indicators to generate a fraud risk score.

7.2 Automated Processing and Decision-Making (GDPR Art. 22)

How our system works:

• Research Shield™ generates fraud risk scores based on behavioral and technical data
• The system flags responses as potentially fraudulent
• These scores and flags are recommendations only

Human-in-the-loop architecture:

Research Shield™ does NOT make disqualification or incentive decisions. Our automated system produces fraud indicators only. Final eligibility decisions, including survey disqualification and incentive denial, are made by:

• The research client's own review process, or
• The panel provider's quality control team, or
• Human reviewers at the client organization

Why Art. 22(1) does not apply: Because Research Shield™ provides fraud scores and flags as decision-support tools only, and does not itself produce decisions with legal or similarly significant effects, the prohibition in Art. 22(1) is not triggered by our processing. The data subject's relationship regarding eligibility and incentives is with the research client or panel provider, not with Research Shield™.

Legal basis for processing: Legitimate interest in fraud prevention (GDPR Art. 6.1.f, Recital 47).

7.3 Safeguards (Binding Obligations)

Regardless of Art. 22 applicability, we implement the following safeguards as binding obligations:

• Human review available: You may request human review of any fraud flag by contacting the research client or privacy@researchshield.com. We will facilitate this request.
• Right to contest: You may contest any fraud determination and provide additional information to the research client
• Transparency: We disclose the existence and logic of automated processing in this policy
• No sensitive data: We do not use special category data (Art. 9) in fraud analysis
• Accuracy obligation: Our system is regularly tested and calibrated to minimize false positives. We maintain documentation of accuracy rates.
• Client obligations: Our contracts with research clients require them to implement appropriate human review processes for disputed cases

7.4 Right to Object (Art. 21)

Under GDPR Art. 21, you have the right to object to processing based on legitimate interest. We assess each objection on a case-by-case basis, considering your specific situation and grounds for objection.

Types of Objection We Distinguish:

Objection Type	Can It Be Honored?	Explanation
Real-time fraud analysis (during survey participation)	Limited	If you object to real-time behavioral analysis, you may still participate, but we cannot guarantee acceptance of your response. As a practical consequence, objection to real-time analysis may mean choosing not to participate.
Post-session data retention	Yes	You can object to continued retention of your data after your session has been scored. We will delete your data upon valid objection, subject to any legal retention obligations.
Cross-session profiling (device fingerprint retention for duplicate detection)	Yes	You can object to retention of device fingerprints used to detect repeat participation. We will delete this data upon valid objection.
Use for ML model training	Yes	You can object to your behavioral data being used for machine learning model improvement. We will exclude your data from training datasets.

How to Object

Send your objection to privacy@researchshield.com with:

• A description of which processing you object to
• Your grounds for objection (your particular situation)
• Any identifiers that help us locate your data (session ID, date of participation)

Our Assessment Process

We will:

1. Acknowledge your objection within 7 days
2. Assess your specific grounds against our legitimate interests
3. Provide a reasoned response within 30 days
4. If we reject your objection, explain our compelling grounds
5. Inform you of your right to complain to a supervisory authority

7.5 Relationship Between Participation and Objection

We do not operate on a "take-it-or-leave-it" basis. However, we are transparent about practical consequences:

• Before participation: If you object to all fraud detection processing, participation may not be possible as fraud detection is integral to the service
• During participation: Real-time analysis occurs automatically; objection at this stage is impractical but you retain post-session rights
• After participation: Objection to retention, profiling, and ML training can be fully honored without affecting your past participation

Your right to object is genuine and subject to individual assessment, not a pre-determined outcome.

8. Cookies, Browser APIs, and Similar Technologies

8.1 Cookies

We use essential cookies:

Cookie	Purpose	Lifetime
session_id	Survey session identification	Session
respondent_id	Response linking	24h

8.2 Browser APIs for Device Fingerprinting

Our fraud detection system (via FingerprintJS) accesses the following browser APIs to generate device identifiers:

API	Purpose
Canvas API	Generates unique image rendering signature
WebGL API	Identifies graphics hardware characteristics
Audio Context API	Generates audio processing signature
Navigator API	Collects browser and system information

These accesses are strictly necessary for fraud detection under ePrivacy Directive Art. 5(3). They do not involve marketing or advertising cookies.

8.3 Local Storage

We may use localStorage for temporary session data. This data is cleared when you close your browser or after 24 hours.

We do not use marketing cookies or tracking cookies for advertising purposes.

9. Children

Our services are not intended for individuals under 16 years of age (or under 13 in the United States). We do not knowingly collect data from children.

10. Information for United States Residents

10.1 Categories of Personal Information Collected

Under US state privacy laws (including CCPA/CPRA), we collect the following categories:

Category	Examples	Collected
Identifiers	IP address, device ID, session ID	Yes
Internet activity	Browsing behavior, interaction with survey	Yes
Geolocation	Approximate location from IP	Yes
Inferences	Fraud risk scores, quality assessments	Yes
Sensitive personal information	N/A	No

10.2 How We Use and Disclose Information

• Business purpose: Fraud detection, data quality assurance
• Sale of data: We do not sell your personal information
• Sharing for advertising: We do not share your data for cross-context behavioral advertising

10.3 Your Rights (California, Virginia, Colorado, Connecticut, Utah)

Depending on your state, you may have the right to:

• Know what personal information we collect
• Access your personal information
• Delete your personal information
• Correct inaccurate information
• Opt-out of sale/sharing (not applicable - we don't sell data)
• Non-discrimination for exercising your rights

10.4 How to Exercise Your Rights

Submit a request to: privacy@researchshield.com

We will verify your identity and respond within 45 days (or as required by applicable law).

10.5 Authorized Agents

You may designate an authorized agent to submit requests on your behalf with written permission.

10.6 Do Not Track

We do not respond to "Do Not Track" browser signals as there is no industry standard for this feature. However, our tracking is limited to fraud detection purposes only.

11. Policy Changes

We will notify of significant changes by updating the date at the beginning of this document. We encourage periodic review of this Policy.

12. Contact

For privacy-related questions, please contact:

Email: privacy@researchshield.com
Address: TGM Research Pte. Ltd., 6001 Beach Road, #22-01 Golden Mile Tower, Singapore 199589

Last updated: January 1, 2026