Our analysis followed a mixed methods design to integrate quantitative and qualitative insights. This was implemented through a 2-stage process: (1) a descriptive synthesis guided by the PAGER (Patterns, Advances, Gaps, Evidence for practice, and Research Recommendations) Framework to generate higher-level insights [19]; and (2) a content analysis using both structural topic modeling (STM) [20] and reflexive thematic analysis.

The PAGER Framework provided a structured approach to synthesize and report findings [19]. We applied this framework to systematically categorize the descriptive characteristics of the included studies and their measures across 5 dimensions:

To analyze the substantive content of the measures, STM and reflexive thematic analysis were conducted on the same set of questionnaire items. This approach aligns with the methodological principle that qualitative and computational text analyses are complementary and analogous processes for identifying patterns in textual data [21]. STM provided a quantitative and reproducible overview of latent conceptual patterns across the full dataset, while reflexive thematic analysis offered qualitative depth, capturing detailed meaning, contextual nuances, and specific question types that purely frequency-based analysis might overlook.

STM is a text-mining technique that identifies latent topics within a text corpus. Each questionnaire item was treated as a separate document. Text preprocessing included lowercasing, removal of punctuation and digits, whitespace normalization, lemmatization, synonym mapping (eg, “photos or images or pics” → picture), and stop-word removal. We fitted STM models with the number of topics (K) ranging from 3 to 10, evaluating model fit using held-out likelihood, residuals, lower bound, semantic coherence, and exclusivity. After identifying K=7, 9, and 10 as offering the best balance of fit, we estimated unsupervised STMs (without covariates) and selected K=9 for final reporting based on balanced fit and conceptual interpretability. All analyses were conducted in R (R Foundation) using the stm (modeling and diagnostics), ggplot2 (visualization), and textstem (lemmatization) packages. This analysis yielded a quantitative map of the dominant conceptual patterns across all measurement items. Interpretive synthesis was performed to translate statistical topics into higher-order conceptual dimensions.

Reflexive thematic analysis was conducted following the 6-phase approach by Clarke and Braun [22]: (1) familiarization with the measurement items; (2) initial code generation; (3) theme search; (4) theme review; (5) theme definition and naming; and (6) report production. Analytical memos documented interpretive decisions and theoretical insights throughout, ensuring thematic robustness and grounding in the data. This independent analysis provided rich, detailed descriptions of question types, contextual nuances, and underlying constructs, complementing and contextualizing the broader statistical patterns identified by STM.

The findings from these 2 analyses were integrated to identify gaps and misalignment within the TFSVA measurement field. Consistent with scoping review methodology, risk-of-bias assessment was omitted [16], although only peer-reviewed articles were included to maintain quality standards.

Nonconsensual sexting, defined as sending or receiving sexually explicit messages or images without one’s permission, was the most extensively studied form of TFSVA (130/218, 59.6%). In terms of age group, 56 out of 130 (43.1%) measures targeted adolescents, 39 out of 130 (30%) focused on young adults, and 32 out of 130 (24.6%) focused on adult populations; 2 out of 130 (1.5%) measures targeted both adolescents and young adults, and the remaining 1 out of 130 (0.8%) included all age groups. Three nonconsensual sexting measures specifically targeted female participants, and none of the measures specified the sexual orientation of the target population.

Nearly all studies related to nonconsensual sexting (n=127) focused on assessing the behavioral domain, with questions primarily addressing the frequency and targets of sexting. Six studies measured attitudes toward nonconsensual sexting, exploring general perceptions and perceived risks associated with the practice. Additionally, 7 studies investigated the motivations behind such behavior, examining reasons for engagement, including sexual purposes, emotional influences, body-image reinforcement, and relationship enhancement.

In terms of psychometric properties, 25 measures were validated, of which 17 reported validity analyses, such as exploratory factor analysis or confirmatory factor analysis. All validated measures reported reliability, including internal consistency (eg, Cronbach α or McDonald ω). Among the validated measures, the Sexting Motivation Scale [24], the Sexting Behavior Scale [25], the Sexting Questionnaire [26], and the Sexting Attitude Scale [27] were the most commonly used. All of these demonstrated strong validity (ie, satisfactory exploratory factor analysis or confirmatory factor analysis) and reliability (ie, Cronbach α >0.8). However, these tools were predominantly validated among adolescents and young adults (18/25, 72%), with limited application to adults or diverse cultural groups.

Online sexual harassment refers to any unwelcome or unwanted sexual behavior, advances, or communication that occurs through digital platforms, including social media, messaging apps, email, or online gaming environments [28]. This form of harassment encompasses a broad range of behaviors, such as sexualized cyberbullying, nonconsensual sharing of intimate content, sexually explicit comments, cyberflashing, and sexual coercion facilitated by digital means. Among the included studies (n=218), 27 (12.4%) measured online sexual harassment. In terms of age group, 11 out of 27 (41%) targeted adolescents, 8 out of 27 (30%) targeted young adults, 6 out of 27 (22%) targeted the adult population, and the remaining 2 (7%) included all age groups; 6 out of 27 (22%) measures were distributed to the female population only, and 1 targeted males only. One measure specifically focused on lesbian, gay, bisexual, transgender, queer (or questioning), and others not explicitly listed (LGBTQ+) individuals and examined sexuality-based cyberbullying [29].

Almost all measures included behavioral outcomes (26/27), with 1 study also measuring the motivation outcome. One study explored the attitude outcome from a bystander perspective; 19 out of 27 (70%) measures focused on exposure to violence, 3 (11%) on perpetration, and the remaining 5 (19%) on both aspects. Most questions in these studies inquired about personal experiences of online sexual harassment, such as receiving offensive sexist messages, encountering negative comments about one’s gender or sexual orientation, or being subjected to inappropriate sexual remarks about one’s body on social media. One measure with a motivation outcome asked about reasons for online sexual harassment.

As many as 11 (41%) measures demonstrated satisfactory validity and reliability. The Online Sexual Victimization Scale was the most commonly used measure [26].

IBSA, such as the nonconsensual sharing of intimate images, was assessed in 30 out of 218 (13.8%) studies; 5 out of these 30 (17%) studies targeted adolescents, 13 (43%) targeted young adults, and 6 (20%) targeted adults; 4 (13%) studies included participants across all age groups, and the remaining 2 (7%) did not specify the target population. One study focused on the male population only, while the others did not specify gender or sexual orientation.

Most measures focused on behavioral outcomes (n=25), while 4 studies assessed attitude outcomes and 1 included motivation outcomes. Among the 30 studies, 14 (47%) focused on IBSA exposure, 3 (10%) on perpetration, 12 (40%) included both exposure to violence and perpetration, 1 (3%) focused on a bystander perspective, and 1 (3%) included all perspectives. The measurement questions explored experiences related to sending, receiving, or requesting intimate, suggestive, or provocative images or videos (eg, “I have sent/received/requested intimate/suggestive/provocative images or videos”), reasons and motivations for sharing such content (eg, “I have sent dick pics hoping to receive sexy pictures in return”), and attitudes toward IBSA behaviors (eg, “If a man sends a nude or sexual image to a partner, he can’t expect it will remain private”).

A total of 6 (20%) measures were validated, demonstrating high validity and reliability. Among these, the Sexual Image-Based Abuse Myth Acceptance Scale was the most widely adapted and utilized [30].

Online grooming refers to the process by which an individual, typically an adult, establishes an emotional connection with a child or adolescent through digital platforms, such as social media or messaging apps, with the intent of exploiting them for sexual purposes [31]. Of the 218 measures, 13 (6.0%) were designed to examine online grooming. Of these 13, 7 (54%) measures targeted adolescents, 2 (15%) targeted adults, 1 (8%) targeted young adults, and the remaining 3 (23%) focused on both adolescents and young adults. No studies specifically targeted sexual and gender minorities.

All studies focused on behavioral outcomes, with 1 also exploring motivations underlying such behavior. Of the 13 studies, 11 (85%) assessed online grooming exposure, while the remaining 2 (15%) examined both online grooming exposure and perpetration. Most measurement questions centered on the frequency and experiences of online grooming (eg, “In the past 12 months, how often has an adult asked me for pictures or videos of myself with sexual content?”, “Someone has made sexual jokes or comments about me”).

In these studies, 3 (23%) scales were validated with high reliability and validity, among which the Questionnaire for Online Sexual Solicitation and Interactions With Adults was the most commonly used [32].

Cyber sexual dating abuse, encompassing behaviors such as monitoring, control, and harassment within romantic relationships, was addressed in 8 out of 218 (3.7%) studies. Of these 8, 5 (63%) focused on adolescents and 3 (38%) on adults. One study specifically targeted sexual and gender minority girls and feminine teens.

All studies focused on behavioral outcomes, with 5 (63%) specifically addressing exposure and 3 (38%) exploring both exposure and perpetration. Some questionnaire content overlapped with items in IBSA and sexual aggression in dating or intimate relationships (eg, “Sending and/or uploading photos, images, and/or videos with intimate or sexual content without permission,” “Pressured my partner to send sexual or naked photos of him or her to me”).

Of these 8 measures, 7 (88%) were validated, demonstrating high reliability and validity. Among these, the Cyber Dating Abuse Questionnaire was the most frequently adapted scale [33].

Since 2018, the term TFSVA has gained prominence as a comprehensive descriptor for a wide range of technology-mediated sexual harms [28]. Of the 218 studies, 10 (4.6%) explored TFSVA as a broader category: 2 (20%) targeted adolescents, 7 (70%) targeted adults, and the remaining 1 (10%) had no age restrictions. No studies specifically targeted sexual and gender minorities.

Of the 10 measures, 9 (90%) focused on behavioral outcomes, and 1 (10%) on attitudes toward TFSVA. Among these studies, 6 (60%) examined TFSVA exposure, while 4 (40%) assessed both exposure and perpetration. The measures encompassed various forms of abuse, including IBSA, online sexual harassment, and gender- or sexuality-based harassment, often grouping these behaviors under the umbrella of TFSVA. Questionnaire items focused on experiences of TFSVA (eg, “Nude or seminude image taken without permission,” “Unwanted sexual experience with someone met online,” “Receiving sexuality- or sexual identity-based offensive and/or degrading messages, comments, or other content”).

Of the 10 measures, 3 (30%) were validated, with the 21-item Technology-Facilitated Sexual Violence Victimization Scale being widely adapted and utilized [34].

Topic interpretation primarily explains the topic words that describe each topic. STM identifies 9 topics based on high-frequency words. Figure 4 presents the top topics generated from unsupervised STM.

Topics were synthesized into 3 core dimensions representing the content of existing measures of TFSVA: (1) consent violation and coercive control; (2) status degradation and demeaning harassment; and (3) contextual and relational framing. Details of the measurement focus and representative topics for each theme are shown in Table 3.

The first dimension, consent violation and coercive control, encompasses measures of acts that override consent, breach authorization, or apply pressure. It is constituted by topics 3, 4, and 5. For instance, topic 5, the most prevalent topic, captures the nonconsensual distribution of intimate images, anchoring a central TFSVA construct in consent violation. Topics 3 and 4 complement this by measuring coercive tactics, with topic 3 focusing on digital threats and topic 4 capturing direct sexual pressure.

The second dimension, status degradation and demeaning harassment, captures measures of acts intended to insult, humiliate, and lower the survivor’s social standing. Represented by topics 2, 6, and 9, it includes unwanted sexual advances that may carry attributions of blame (topic 2), psychological abuse and humiliation (topic 6), and body shaming in social media spaces (topic 9).

The final dimension, contextual and relational framing, defines the specific relational contexts and vulnerable populations that characterize TFSVA. It is represented by topics 1, 7, and 8. Topics 1 and 8 frame abuse within intimate relationships, measuring pressured sexting and revenge porn, respectively. Topic 7 captures online grooming and the sexual exploitation of minors.

Topic prevalence by types of TFSVA is shown in a heatmap (Figure 5). First, topic 5 (nonconsensual image distribution) demonstrates consistently high probability across all TFSVA types, suggesting it is a core element in TFSVA measures. Second, the relatively lower overall prevalence of topic 7 (online grooming and sexual exploitation of minors) may indicate that this form of abuse is underrepresented in general TFSVA instruments or is more often specialized in measures targeting child sexual exploitation. Third, the high specificity of certain topics—such as topic 8’s strong association with IBSA and the notable absence of topics 3 (digital threats) and 6 (psychological abuse) from the same type—indicates that IBSA is operationalized primarily around the violation of consent for visual content rather than coercive or psychological tactics.

To integrate the findings from STM and thematic analysis, an integrated evidence and gap map was developed and is presented in Figure 6. This conceptual synthesis does not propose a direct correlation but serves as an evaluative framework. It positions the 3 core measurement dimensions identified through STM against the 5 themes derived from thematic analysis. For each dimension, judgment was made regarding its coverage of each thematic construct. The resulting matrix reveals a structural pattern in which robust measurement of behavior differs significantly from deficits in assessing attitudes, motivations, tactical dynamics, and harm-related components.

This review confirms the definitional fragmentation of TFSVA noted in prior literature. Previous studies have observed that TFSVA encompasses a diverse range of behaviors studied in isolation using inconsistent terminology [28]. Henry and Powell’s [28] foundational work identified a critical lack of empirical research on TFSVA against adults and noted that only a few existing studies provide reliable data on the nature, scope, and impacts of TFSVA. Nearly a decade later, our synthesis provides empirical confirmation of this fragmentation. The field has produced an abundance of measures, yet they remain focused on specific behavioral subdomains with minimal conceptual integration.

Three key findings from our analysis directly inform the need for an updated definition. First, the narrow behavioral focus of existing measures reveals that current conceptualizations fail to capture the full spectrum of TFSVA. Second, our analysis identified emerging forms of TFSVA, such as AI-generated deepfakes, that are absent from existing measures. Third, our findings show that TFSVA operates within broader structural contexts that shape both the nature of abuse and its consequences.

Grounded in this evidence, we propose an updated definition that also engages with distinctions raised in restorative and transformative justice scholarship [46]. Building on this foundation, we define TFSVA as any form of sexual violence, abuse, coercion, or exploitation that is initiated, facilitated, intensified, or sustained through digital technologies. Such acts may result in psychological, reputational, social, economic, physical, or sexual harm, though the presence or absence of harm is not definitional to the act itself.

This formulation separates the act from its potential consequences. It acknowledges that (1) individuals may experience TFSVA without immediate measurable harm yet still warrant recognition and support; (2) the same act may produce different harms across individuals and contexts; and (3) harm may manifest differently across cultural and geographical settings [46].

The definition encompasses specific acts documented in our review, including IBSA, AI-generated sexual exploitation, cyberstalking leading to physical violence, and algorithmically amplified harassment. It remains adaptable to emerging forms not yet captured in existing measures. Crucially, it recognizes that TFSVA operates within systems of gendered, racialized, and intersectional oppression. These dimensions, which often transcend digital or physical boundaries and jurisdictional borders, are currently absent from existing measures.

Additionally, the definition embeds recognition of systemic oppression and acknowledges that technology is not merely a neutral tool. For instance, algorithms can amplify harassment, platform design may facilitate sexual abuse, and networked technologies enable harm across jurisdictional boundaries [47]. These dimensions follow directly from a definition centered on technological and structural contexts.

The increase in TFSVA publications observed in our temporal analysis suggests growing academic and societal recognition. A notable trend is the decline in studies on specific forms, such as sexting and IBSA, from 2011 to 2017. The use of the holistic term “TFSVA” began to rise sharply from 2018, suggesting a conceptual shift in which previously distinct behaviors are being consolidated under a broader umbrella [3]. Additionally, the decline in specific terms may reflect a move away from fragmented terminology toward more comprehensive approaches that capture the spectrum of technology-facilitated sexual harms. Alternatively, it may indicate that earlier research on sexting, IBSA, online sexual harassment, and online grooming reached a saturation point, prompting scholars to explore broader themes or emerging forms of TFSVA. Further investigation is needed to determine whether this trend reflects evolving academic priorities or a response to new technological developments that blur the boundaries between specific subtypes of TFSVA.

A core finding of this review is the gap in demographic representation within TFSVA measures. Henry and Powell [28] observed that TFSVA may be predominantly gender-, sexuality-, and age-based, with young women being overrepresented as survivors. However, our analysis reveals that the field lacks the measures needed to adequately investigate these patterns.

The gendered nature of TFSVA is well documented, with women and sexual minorities disproportionately affected [48]. Emerging evidence also suggests that men experience TFSVA, albeit in different forms and with varying impacts [49,50]. A recent study assessing the content validity of a TFSVA measure further highlighted significant variations in how individuals of different genders and sexual orientations interpret and perceive TFSVA [51].

Yet, our synthesis reveals 2 critical gaps that prevent the field from advancing this understanding. First, existing measures may lack content sensitivity to diverse identities. Items often use generic language that may not capture experiences specific to sexual and gender minorities. Second, even if measures could capture diverse experiences, they have rarely been validated with diverse populations. Of the 319 studies, 314 (98.4%) did not specify participants’ sexual orientation, and only 5 (1.6%) targeted gender minority populations. The implications of these gaps for measure design and validation are addressed in the subsequent discussion.

Moreover, the geographic focus of TFSVA research is heavily skewed. For instance, our results indicate that 260 (81.5%) of the 319 included studies were conducted in Western, high-income countries, while studies from Asian, African, and low-income contexts remain scarce. This imbalance is critical, as cultural norms and contextual factors, such as varying digital literacy, internet access, and legal frameworks, can significantly influence how TFSVA is perceived and experienced [3]. Consequently, existing measures may lack validity in these regions, highlighting the need to develop culturally adapted measures that account for specific social, legal, and technological environments.

The thematic analysis of questionnaire content confirmed that existing measures capture behavioral data and certain attitudinal and motivational factors, but lack depth. Our integrated synthesis places these content gaps within a structural framework, revealing substantial limitations in addressing critical dimensions, notably evolving technological threats, intersectional vulnerabilities, and institutional factors.

Emerging technologies present challenges for TFSVA measures. Our analysis showed that emerging forms of TFSVA, such as AI, deepfakes, zoom-bombing or digital intrusion, and abuse via smart devices, were not covered in the measurement content. These gaps render existing questionnaires outdated and create a persistent development lag, as the evolution of abuse tactics outpaces instrument updates [51]. To remain relevant, future questionnaires should be dynamically updated to integrate assessments of emerging tactics, such as GPS stalking and AI-facilitated abuse.

Results from our analysis reveal limitations in capturing intersectional experiences at 2 levels. First, in terms of measure design, the content of existing measures lacks sensitivity to how TFSVA manifests across individuals of different genders and sexual orientations. Items often use generic language (eg, “Have you received unwanted sexual images?”) that does not account for specific experiences among sexual and gender minority populations. This limits the validity of these measures for diverse groups. Second, in terms of study application, our descriptive synthesis shows that most included studies did not specify participants’ sexual orientation. This indicates that studies rarely conduct and report the intersectional analyses needed to understand disparities.

Existing literature on other sexual violence measures offers concrete guidance for improving inclusivity. For example, the Daily Heterosexist Experiences Questionnaire includes items that specifically assess discrimination and harassment tied to sexual minority identity, such as “People assuming you are heterosexual because you have children,” moving beyond generic violence exposure questions [52]. Similarly, measures of intimate partner violence for sexual and gender minority populations adapt items from the Revised Conflict Tactics Scale to capture experiences unique to these populations, including identity-based coercion and outing as a tactic of abuse [53]. These examples demonstrate that inclusive measures require consultation with diverse communities during item development to ensure content validity [52]. A second strategy is the use of language that reflects the lived experiences of specific groups rather than assuming universal applicability [53]. A third is validating measures separately across different populations to establish measurement equivalence [53]. Incorporating these strategies into TFSVA measure development is crucial for capturing nuanced harms, such as gendered and sexuality-based abuse [5]. Future measure development should prioritize inclusive language and item wording, community engagement in the design process, and rigorous validation across diverse gender and sexual orientation groups.

A related and critical gap is the lack of cross-cultural validity. Our findings show that most validated measures were developed in Western, high-income countries. To move forward, the adaptation of TFSVA measures should follow a rigorous framework to ensure validity across different contexts [54]. This involves (1) ensuring conceptual equivalence through co-creation with local experts and cognitive testing; (2) conducting statistical measurement invariance testing (eg, multigroup confirmatory factor analysis); and (3) establishing criterion validity. This structured process moves beyond simple translation to develop measures that are valid for cross-cultural research and intervention.

Findings from the thematic analysis of measurement content indicate that current TFSVA-specific measures critically underassess structural and institutional factors, such as interactions with legal systems and digital platforms. There is a notable absence of items within TFSVA measures assessing key mediators of survivor outcomes, such as a survivor’s trust in, or experience with, platform reporting systems, legal authorities, or support services. It is important to note that this finding is limited to the content of TFSVA-specific measures. Our review did not capture whether included studies administered additional, standalone measures assessing institutional trust, platform responsiveness, or help-seeking behaviors alongside their TFSVA measures. Therefore, although these constructs are absent from TFSVA-specific measures, we cannot draw conclusions regarding their overall availability or use elsewhere in the literature. Nevertheless, the absence of these constructs from TFSVA-specific measures has important implications. A growing body of evidence demonstrates that institutional responses profoundly influence survivor outcomes [55]. Research on situational betrayal theory has shown that platform reporting processes can themselves become sources of secondary trauma when they are hostile, opaque, or ineffective [56]. Institutional responses also shape bystander behaviors and help-seeking pathways [57]. For instance, bystanders’ willingness to intervene is influenced by their perceptions of safety, responsibility, and the anticipated consequences of action, which are shaped by broader institutional and social norms [57]. Similarly, survivors’ decisions to seek support are mediated by their trust in institutional responses and fear of further harm during or following disclosure [57]. Consequently, when TFSVA measures are used in isolation, they can quantify the abusive act but cannot assess the institutional and structural context in which harm is experienced and mediated. To build an evidence base for effective interventions, future TFSVA measure development should consider integrating items that assess constructs such as institutional trust, platform responsiveness, and multilevel help-seeking behaviors.

Our analysis of TFSVA-specific measures revealed limited attention to psychological and relational impacts, such as long-term trauma, reputational harm, or intimate partner dynamics. Results from STM and thematic analysis indicate that when impacts are included, they tend to frame consequences as isolated emotional states rather than examining how they interact with, or amplify, other mental health challenges [58]. However, this finding is limited to the content of TFSVA measures, as we did not capture whether included studies administered separate validated instruments assessing psychological outcomes alongside their TFSVA-specific measures. Future research should examine how TFSVA measures are used in conjunction with established mental health measures. Additionally, TFSVA measure development should consider integrating items that capture long-term and relational impacts to provide a more comprehensive understanding of survivors’, perpetrators’, and bystanders’ outcomes.