Marital Fidelity and Digital Infidelity in a UAE University-Affiliated Sample: An Exploratory Mixed-Methods Study

Abstract

This exploratory mixed-methods study examines self-reported perceptions of marital fidelity, marital infidelity, and digital boundary violations among adults recruited through one private university in the United Arab Emirates. Nationality or citizenship and direct socioeconomic indicators were not collected; participants are therefore described as a UAE university-affiliated sample rather than as Emirati or nationally representative. A quantitatively led design combined 156 complete questionnaires with 11 semi-structured interviews. Marital fidelity received very high endorsement (M = 4.50, SD = 0.47), and its observed association with perceived seriousness of infidelity was positive (r = 0.685, p < 0.001). Primary interpretation is limited to the strongly reliable marital-fidelity indicator and the comparatively better-supported seriousness indicator. Findings involving the three-item digital-infidelity index (alpha = 0.481), the two-item forgiveness/restoration index (alpha = 0.534), and the single boundary-clarity item are reported only as secondary exploratory analyses and are not used to support the principal conclusions. Women showed sample-specific differences on fidelity and seriousness, but no result is interpreted as behavioral or moral superiority. The observed correlation between seriousness and digital-infidelity judgments (r = 0.864) indicates substantial overlap; no correction for attenuation is presented because the dimensionality and error structure of the short indices were not established. Interview findings are reported as counts out of 11 rather than percentages. The findings are sample-bound, attitudinal, and hypothesis-generating.

1. Introduction

Marriage is a social institution through which intimacy, care, responsibility, kinship, and intergenerational continuity are organized. Within marriage, fidelity is commonly associated with trust, commitment, honesty, and respect for agreed relational boundaries. It is therefore more than a private moral preference: perceived violations can affect emotional security, family interaction, reputation, and decisions about whether a relationship can continue [1,2]. The topic is sociologically important and timely because digitally mediated interaction has expanded the settings in which couples negotiate exclusivity, privacy, and acceptable contact with others.

In this article, digital behavior refers to communication or monitoring conducted through social-networking sites, messaging applications, email, shared devices, and other online platforms. Behaviors relevant to fidelity may include concealed or deleted conversations, repeated intimate messaging, emotional or sexual exchanges, sharing private images, maintaining a hidden online relationship, or monitoring a partner without agreement. The meaning of any one behavior is not fixed: a private message may be regarded as ordinary by one person and as a violation by another, depending on secrecy, intent, intimacy, prior agreements, and the meaning assigned to the act within the relationship [3,4].

The United Arab Emirates provides a relevant setting because strong family and religious norms coexist with extensive digital connectivity and cultural mobility. Administrative data also indicate that marital dissolution is a significant social issue. The Federal Competitiveness and Statistics Centre recorded 4213 divorce cases in 2020, 5450 in 2021, and 5298 in 2022 [5]. These counts provide context for studying marital relationships, but they neither demonstrate a continuously rising trend nor identify infidelity as a cause of divorce. The present study therefore avoids attributing national divorce patterns to marital infidelity without direct causal evidence.

Research increasingly treats infidelity as a multidimensional and negotiated construct rather than as sexual intercourse alone. Reviews distinguish sexual, emotional, online, and secrecy-related dimensions and emphasize that couples may differ in the behaviors they count as a breach of exclusivity [6,7,8]. Digital environments intensify this ambiguity because they permit persistent, private, and easily concealed contact, while also supporting ordinary communication and connection. Prior studies associate infidelity-related social-media behaviors with jealousy, surveillance, relational dissatisfaction, and conflict, although cross-sectional findings cannot establish causality [9,10,11].

Evidence from Arab settings remains comparatively limited and is often based on psychometric work, small qualitative studies, or research conducted outside the Gulf. Existing studies nevertheless suggest that trust, religious and familial expectations, gendered socialization, and digital privacy shape how infidelity is defined and evaluated [12,13,14]. The present study contributes an exploratory case situated in the UAE, but nationality or citizenship was not collected and cannot be inferred from university enrollment or residence. Participants are therefore described as adults recruited through one private university in the UAE, not as a nationally representative Emirati sample. No claims are made about national prevalence or population-wide attitudes.

The study addresses four bounded questions in paragraph form rather than as a proposal-style list. First, what levels of marital fidelity, perceived seriousness of infidelity, and the secondary exploratory indicators are reported in this university-affiliated sample? Second, what observed association appears between fidelity and perceived seriousness, and what additional associations appear only in secondary exploratory analyses? Third, what sample-specific sex and other demographic differences appear in the better-supported indicators, and what additional patterns are confined to exploratory analyses? Fourth, how do interview accounts elaborate the survey patterns concerning trust, digital behavior, marital boundaries, protective factors, and the difficulty of restoring trust? The article first develops a concise conceptual framework, then presents the mixed-methods design and the explicit links between instruments, reports primary and secondary exploratory findings, integrates both strands in the discussion, and concludes with limitations and implications.

2. Conceptual and Analytical Framework

2.1. Defining Fidelity, Infidelity, and Digital Boundaries

Marital fidelity is defined here as adherence to the explicit and implicit commitments through which spouses regulate sexual, emotional, digital, and informational exclusivity. This definition is intentionally broad but not unlimited. It is grounded in research showing that infidelity may involve sexual behavior, emotional investment, secrecy, deception, or online intimacy and that perceptions vary across individuals and relationships [6,7,8,15]. Fidelity therefore includes not only abstention from extramarital sexual contact, but also the maintenance of agreed boundaries around emotional closeness, disclosure, and digital interaction.

Infidelity is correspondingly treated as a perceived breach of relational expectations rather than a universally identical act. Physical infidelity refers to sexual or intimate bodily contact outside the marriage. Emotional infidelity refers to directing romantic intimacy, attachment, or privileged emotional investment toward another person in a manner that displaces or conceals commitment to the spouse. Digital infidelity refers to online behavior that combines secrecy, intimacy, sexual or emotional content, or intentional boundary crossing. A digital act becomes sociologically meaningful through the couple’s expectations and the context in which it occurs, rather than through the technology alone [3,4].

2.2. Sociological Interpretation

Symbolic interactionism directs attention to how spouses define situations and negotiate the meaning of messages, privacy, secrecy, and trust. The same online act can be interpreted differently because its meaning emerges through interaction and shared or disputed expectations [16]. Bourdieu’s concepts of habitus and doxa provide a complementary account of how family upbringing, religion, gender norms, and reputation make certain understandings of fidelity appear natural or self-evident [17,18]. These dispositions do not mechanically determine behavior, but they shape what participants recognize as honorable, suspicious, permissible, or disloyal.

Bauman’s account of liquid modernity adds a macro-level lens on the instability and expanded choice associated with contemporary relationships [19,20]. Digital media can multiply opportunities for contact and comparison while blurring the distinction between public and private interaction. Used together, the three perspectives connect micro-level interpretation, culturally learned dispositions, and the wider conditions of digitally mediated social life. They are used here as sensitizing frameworks for interpreting findings, not as hypotheses that the cross-sectional data can definitively test.

2.3. Empirical Expectations and the Arab Context

Previous research consistently identifies trust, commitment, relationship satisfaction, and communication as central to the experience and consequences of infidelity [1,21,22]. Online environments add questions about concealment, surveillance, and the boundary between ordinary contact and intimate involvement [10,23,24]. Gender differences have also been reported, particularly in the salience assigned to emotional and sexual dimensions, but such differences are shaped by measurement, social expectations, and sample composition and should not be interpreted as essential traits of women or men [15,25].

In Arab and Gulf contexts, marital commitments are additionally embedded in religious, familial, and reputational fields. The Jordanian validation study by Aldahadha and Al-Momani [12] supports a multidimensional understanding of infidelity, while research on digital interaction among Arab couples highlights emerging tensions surrounding privacy and trust [13]. The present study examines whether comparable meanings appear in a UAE university-affiliated sample and uses interviews to interpret, rather than merely repeat, the survey findings.

2.4. Negotiated Device Access and Intimate Surveillance

Digital boundary-setting cannot be reduced to a simple privacy-versus-transparency binary. Research in human–computer interaction and cybersecurity shows that partners often share devices, unlock codes, and account credentials for convenience, care, household coordination, or as a signifier of trust, while the permitted uses of that access may remain explicit, implicit, or contested. Across this literature, a central distinction is that access should be mutual and consensual: permission to unlock a device for one purpose does not necessarily authorize unrestricted inspection of private messages, social-media accounts, location history, cloud storage, or communications with third parties [26,27,28].

Password sharing may also create a perceived entitlement to monitor a partner’s accounts, and research distinguishes voluntary reciprocal access from requested, coerced, covert, or persistent monitoring [29]. Covert account access, spyware, location tracking, credential misuse, and demands for unrestricted access are analytically different from negotiated sharing and may constitute intimate surveillance or technology-facilitated abuse [26,30]. Digital boundary clarity should therefore be evaluated through mutuality, informed consent, specificity of purpose, revocability, power, and relational context rather than equated with total transparency. This distinction also exposes the limits of measuring boundary clarity with a single general item.

3. Materials and Methods

3.1. Design and Mixed-Methods Integration

The study used a quantitatively led mixed-methods design with an explanatory qualitative component. The questionnaire identified patterns in attitudes and perceived boundaries, while semi-structured interviews elaborated the meanings participants attached to trust, physical and emotional infidelity, digital behavior, protection of the relationship, and restoration of trust. Mixed-methods quality depends not only on collecting two types of data but also on specifying where they are connected and how they are interpreted together [31,32]. Integration in this study occurred at the construct level when the interview guide expanded questionnaire domains and at the interpretation level through a joint display that compared convergent, complementary, and cautionary findings [33].

The two components were used to answer different but connected questions: the questionnaire estimated patterns within the recruited sample, whereas the interviews explored why certain behaviors were interpreted as breaches of trust and how participants described relational protection. The qualitative component was not used to generate population percentages or to claim thematic saturation. Instead, its role was explanatory and contextual, appropriate to the modest interview sample and the sensitivity of the topic.

The questionnaire and interview components used independent participant samples. None of the 11 interview participants had completed the questionnaire, and no individual-level matching or linkage was undertaken between the two datasets. No identifying codes were used to connect questionnaire responses with interview data.

Integration between the quantitative and qualitative components therefore occurred at the level of the research questions, conceptual domains, and interpretation of findings rather than at the level of individual participants. The questionnaire measured awareness of marital fidelity, perceived seriousness of infidelity, digital infidelity, forgiveness or restoration of trust, and clarity of marital boundaries. The interviews explored these domains in greater depth by examining how participants defined fidelity and infidelity, interpreted breaches of trust and concealed or intimate online interaction, understood emotional and digital boundary violations, and identified relational, communicative, religious, and social factors that may protect the marital relationship. The qualitative findings were used to elaborate, contextualize, and explain the quantitative patterns, rather than to trace or validate the responses of particular questionnaire participants. The use of independent samples broadened the range of perspectives included in the study, although it did not permit person-level triangulation between questionnaire scores and interview accounts.

3.2. Setting, Participants, and Scope of Inference

The study population comprised adults enrolled at one private university in the United Arab Emirates during the 2024/2025 academic year. Some participants were also employed, and the recruited group included different age categories and marital statuses. A total of 158 questionnaire submissions were received; two incomplete responses were excluded, leaving 156 complete cases. Recruitment used convenience sampling through the university setting, and the questionnaire was distributed mainly through an electronic link, with a small number completed on paper. Nationality and citizenship were not included in the demographic questionnaire. Consequently, university enrollment, residence in the UAE, or recruitment location cannot be used to classify participants as Emirati nationals.

This recruitment strategy provides access to an exploratory UAE-based university sample but does not produce a probability sample of either Emirati citizens or the wider UAE population. University affiliation may be associated with education, employment, digital access, age, and social networks. The sample was also concentrated among married participants (73.7%) and those aged 35–50 years (51.3%), so the aggregate estimates substantially reflect these groups. In addition, the study did not collect direct indicators of socioeconomic position such as household income, occupation, housing, or household wealth. The quantitative results are therefore reported only as characteristics of the recruited sample. Balance on sex or place of residence does not establish socioeconomic, citizenship, or cultural representativeness [34].

The qualitative component comprised 11 participants drawn independently from the same university setting and included variation in marital status (married, unmarried, divorced, or separated). Interviews were conducted face-to-face in a private university room or by video call according to participant preference and lasted approximately 30–40 min. Nationality was not collected for this component either. The small interview set was used to illuminate meanings and mechanisms; it is not sufficient for estimating theme prevalence, claiming saturation, or representing the range of perspectives in the UAE population.

3.3. Questionnaire and Measures

The questionnaire was a researcher-assembled instrument rather than a verbatim administration of a previously validated scale. It contained demographic questions and five substantive indicators rated from 1 (strongly disagree) to 5 (strongly agree). Composite scores were calculated as the arithmetic mean of the constituent items, with higher values indicating stronger endorsement of the named construct. The indicators comprised marital fidelity (13 items), perceived seriousness of infidelity (7 items), digital-infidelity judgments (3 items), forgiveness or restoration of trust (2 items), and perceived clarity of marital boundaries (1 item). Development was informed by the multidimensional infidelity literature and Arabic psychometric work, particularly Aldahadha and Al-Momani [12], followed by specialist review for clarity, content coverage, and cultural suitability. Appendix A discloses the operational structure and psychometric status of each indicator, while Appendix C provides the complete questionnaire codebook, exact item wording, variable and value labels, item-to-indicator allocation, scoring rules, and missing-data conventions.

Face and content review supported the relevance and clarity of the items but did not establish construct validity. No exploratory or confirmatory factor analysis, measurement-invariance testing, convergent validity assessment, or criterion validation was conducted. Internal consistency was examined using Cronbach’s alpha. The 13-item fidelity indicator showed strong reliability, whereas the 3-item digital-infidelity index and the 2-item forgiveness/restoration index showed weak internal consistency. The two-item result is interpreted cautiously because reliability estimation is especially limited for very short measures [35]. Boundary clarity was assessed with one item, for which internal consistency cannot be estimated. These three variables are retained transparently as exploratory indicators and are not presented as fully validated standalone scales or precise latent constructs.

Because fidelity is morally and culturally sensitive, the survey was vulnerable to social-desirability bias. No social-desirability scale, randomized-response method, indirect-questioning technique, or behavioral validation criterion was included. Procedural safeguards consisted of self-administration for most questionnaires, assurances of confidentiality and anonymity, neutral study instructions, voluntary participation, and the right to omit a response or discontinue participation. These procedures may reduce evaluation pressure but cannot eliminate normatively desirable responding; high scores are therefore interpreted as self-reported endorsement rather than evidence of actual marital behavior [36,37].

3.4. Semi-Structured Interview Guide and Link to the Questionnaire

The interview guide was organized around the same conceptual domains as the questionnaire. Participants were asked how they defined marital fidelity and infidelity; which physical, emotional, or online behaviors they counted as violations; how secrecy, intent, and privacy affected their judgments; what made marital boundaries clear or ambiguous; what factors protected a relationship; and whether and under what conditions trust could be restored. Open follow-up prompts allowed participants to explain examples and meanings that fixed-response items could not capture. Appendix B reports the core interview domains. The original guide did not include a dedicated, systematic module distinguishing consensual device access from coercive account monitoring, revocability of access, or technology-facilitated abuse; this omission is addressed as a conceptual limitation and a priority for future instrument development. The explicit links between the questionnaire and interview components are summarized in

Table 1. Explicit links between questionnaire domains and interview elaboration.

Questionnaire Domain	Interview Elaboration	Purpose of Integration
Marital fidelity	Meanings of loyalty, honesty, trust, and commitment	Interpret why fidelity received high endorsement and identify the social meaning of the construct
Perceived seriousness of infidelity	Consequences for trust, dignity, family stability, and continuation of marriage	Explain what participants understood as the harm represented by high seriousness scores
Digital-infidelity index (3 items)	Examples of hidden messaging, intimate exchanges, deleted conversations, and online emotional involvement	Contextualize which digital behaviors were judged as violations; do not validate a distinct scale
Boundary-clarity item (1 item)	Acceptable and unacceptable contact, privacy, disclosure, and mutual agreements	Illustrate negotiated meanings while acknowledging that one item cannot capture consent, mutuality, revocability, or coercion
Forgiveness or restoration of trust	Perceived depth of betrayal and conditions for rebuilding trust	Contextualize the low-to-moderate survey score without treating it as a fixed cultural rule
Protective factors	Trust, communication, respect, religious commitment, attention, and mutual understanding	Identify mechanisms participants believed could prevent or reduce relational vulnerability

.

3.5. Ethical Compliance, Consent, and Data Management

The questionnaire and interview data were collected during the 2024/2025 academic year as an internal course-based research project undertaken for teaching and assessment purposes and not initially intended for external publication. Formal prospective committee review was not initiated at that stage because the research team treated the project as a course assignment rather than as a study intended for external dissemination. The authors acknowledge that this was a procedural oversight rather than an ethics exemption, particularly because the project involved human participants and a sensitive topic. When the work was subsequently considered for further analysis and external publication, the authors disclosed to the Research Ethics Committee that the questionnaire and interview data had already been collected. The Committee reviewed the research procedures, participant-information and consent process, instruments, confidentiality safeguards, data-storage arrangements, proposed analyses, and publication plan. It subsequently approved the ethical use, analysis, storage, and publication of the existing data on 27 July 2026. The approval was signed by Prof. Haitham Bani Salameh, Chair of the Research Ethics Committee and Dean of Scientific Research at the University. This retrospective review does not substitute for prospective authorization, and the timing of review remains a procedural limitation.

Participation was voluntary and based on informed consent. Before taking part, all questionnaire respondents and interview participants were informed of the study purpose and procedures, the sensitive nature of the topic, the intended scientific use of the data, the measures adopted to protect confidentiality, their right to decline to answer any question, and their right to discontinue participation without penalty. Informed consent was obtained from all participants before their participation. Survey findings are reported in aggregate form, and qualitative findings are presented through de-identified themes so that individual participants are not named or readily identifiable.

Electronic questionnaire files, de-identified interview transcripts, coding files, and statistical outputs are stored in password-protected and encrypted university-approved electronic storage. Direct identifiers, contact information, and consent records are stored separately from the research datasets, and no names or direct personal identifiers are included in the analytical files. Each questionnaire record and interview transcript is assigned a unique study code. Access is restricted to the principal investigator and authorized members of the research team whose access is required for data cleaning, coding, analysis, or verification. Any file linking study codes to identifying information is stored separately and is accessible only to the principal investigator.

Where interviews were audio-recorded, recordings are retained only until transcript accuracy has been verified and direct or potentially identifying information has been removed, after which the recordings are permanently deleted. De-identified quantitative data, anonymized interview transcripts, the study codebook, and analytical files will be retained for five years after publication, or for any longer period required by Al Ain University policy. At the end of the retention period, electronic files will be securely and permanently erased, and any paper consent forms or identifying documents will be destroyed through confidential shredding.

3.6. Data Analysis

Quantitative data were coded and analyzed in SPSS 25. Analyses were organized according to the measurement strength of the available indicators. Primary interpretation emphasized descriptive results and observed associations involving the strongly reliable marital-fidelity indicator and the comparatively better-supported perceived-seriousness indicator. Tests involving the three-item digital-infidelity index, the two-item forgiveness/restoration index, and the single boundary-clarity item were classified as secondary exploratory analyses because their measurement limitations reduce the precision and likely replicability of inferential estimates. Because the archived analytical materials did not permit a defensible additional sensitivity model, the weak and single-item indicators were removed from the study’s primary inferential interpretation. The existing hierarchical regression is retained only as a secondary exploratory analysis for transparency and is not used to support the principal conclusions. Digital-infidelity judgments were not entered with seriousness because r = 0.864 corresponds to 74.6% shared variance, pairwise tolerance of 0.254, and a pairwise VIF of approximately 3.94. These are pairwise diagnostics rather than full-model VIF statistics.

Interview material was transcribed and analyzed using a descriptive thematic-analysis procedure informed by Braun and Clarke [38,39]. The analysis involved familiarization, initial coding, grouping codes into candidate themes, reviewing themes against the transcripts, and defining the final categories. The analysis did not include independent double-coding; consequently, no inter-coder reliability coefficient is reported. The core qualitative coding domains are documented in Appendix B, but the absence of an independent coding audit limits analytic dependability. Qualitative tables report only the number of interviewees, out of 11, who mentioned a category. Percentages have been removed because they can imply unwarranted precision in such a small interview set. Counts are descriptive orientation only: they are not prevalence estimates, evidence of saturation, or measures of thematic importance. Quantitative and qualitative findings were then placed in a joint display to identify convergence, complementarity, and limitations.

4. Results

All quantitative estimates in this section describe the 156 complete responses from the recruited UAE university-affiliated sample. Because nationality was not collected, they should not be read as estimates for Emirati citizens or for the UAE population. They are also self-reported attitudes and perceptions rather than verified marital behavior.

4.1. Sample Characteristics

Women comprised 53.2% of the sample and men 46.8%. The sample was concentrated among participants aged 35–50 years (51.3%), married participants (73.7%), and those with bachelor’s degrees (46.2%). Although sex and place of residence were relatively balanced, the distributions by age, marital status, and education were not. Nationality/citizenship and direct socioeconomic indicators were not measured. The descriptive results therefore predominantly reflect a middle-aged, married, university-affiliated group and should remain sample-bound. The sample characteristics are summarized in

Table 2. Demographic characteristics of questionnaire participants (n = 156).

Variable	Category	n	%
Sex	Male	73	46.8
	Female	83	53.2
Age	Under 20	6	3.8
	20–35	56	35.9
	35–50	80	51.3
	50–65	14	9.0
Educational level	Secondary	40	25.6
	Diploma	15	9.6
	Bachelor’s	72	46.2
	Postgraduate	29	18.6
Place of residence	Inside Al Ain city	79	50.6
	Outside Al Ain city	77	49.4
Marital status	Unmarried	32	20.5
	Married	115	73.7
	Divorced	7	4.5
	Separated	1	0.6
	Widowed	1	0.6

.

Note on sample identity: nationality or citizenship was not included in the questionnaire and cannot be inferred from residence, university enrollment, or recruitment location.

4.2. Scale Properties and Descriptive Results

Marital fidelity had a very high mean (M = 4.50, SD = 0.47) and strong internal consistency (alpha = 0.898). Perceived seriousness of infidelity was high (M = 4.07, SD = 0.56; alpha = 0.692). The three-item digital-infidelity index also had a high mean (M = 4.16, SD = 0.65), but its internal consistency was weak (alpha = 0.481). The two-item forgiveness/restoration index was lower (M = 2.49, SD = 0.87) and likewise weakly reliable (alpha = 0.534). The single boundary-clarity item was high (M = 4.04, SD = 1.11), but a one-item measure cannot provide an internal-consistency estimate or adequately represent mutuality, consent, revocability, coercion, and context. Accordingly, only the fidelity indicator has strong internal-consistency support; conclusions involving the other short indicators are exploratory. High means should be read as endorsement of the presented statements and may partly reflect normative or socially desirable responding, not observed behavior. These descriptive and reliability results are summarized in

Table 3. Descriptive statistics and internal consistency of study indicators.

Indicator	Items	M	SD	Cronbach’s Alpha	Interpretation
Marital fidelity	13	4.50	0.47	0.898	Very high; reliable
Perceived seriousness of infidelity	7	4.07	0.56	0.692	High; borderline acceptable
Digital-infidelity index	3	4.16	0.65	0.481	Exploratory; weak reliability; not a validated standalone scale
Forgiveness/restoration index	2	2.49	0.87	0.534	Exploratory; weak reliability; very short measure
Boundary-clarity item	1	4.04	1.11	Not applicable	Single item; internal consistency and multidimensional coverage cannot be assessed

.

4.3. Primary and Secondary Exploratory Sex Differences

Primary interpretation is limited to the better-supported marital-fidelity and perceived-seriousness indicators. Women reported higher mean fidelity endorsement than men (4.58 versus 4.40, p = 0.013, d = 0.41) and higher perceived seriousness (4.21 versus 3.92, p = 0.001, d = 0.53). These are sample-specific differences in self-reported endorsement and do not establish actual behavior, moral superiority, or population-wide gender patterns. Results for the weak digital-infidelity index, weak forgiveness/restoration index, and single boundary-clarity item are retained in

Table 4. Primary and secondary exploratory Welch t-tests comparing men and women.

Indicator	Men M +/− SD	Women M +/− SD	t	df	p	Cohen’s d
Marital fidelity	4.40 +/− 0.49	4.58 +/− 0.44	−2.51	145.30	0.013	0.41
Perceived seriousness	3.92 +/− 0.59	4.21 +/− 0.50	−3.30	142.86	0.001	0.53
Digital infidelity	4.04 +/− 0.65	4.27 +/− 0.63	−2.26	150.18	0.026	0.36
Forgiveness/restoration	2.59 +/− 0.91	2.41 +/− 0.82	1.28	145.80	0.201	−0.21
Boundary clarity	3.63 +/− 1.23	4.41 +/− 0.84	−4.56	124.87	<0.001	0.75

Table note: Fidelity and perceived seriousness provide the better-supported group comparisons. Digital infidelity and forgiveness/restoration are weak short indices, and boundary clarity is one item. Their effect sizes are secondary exploratory estimates only.

only as secondary exploratory estimates. The boundary result (d = 0.75) quantifies a difference in responses to one general statement; it cannot establish a sex difference in the multidimensional construct of negotiated digital boundaries, including consent, mutuality, revocability, privacy, coercion, or surveillance.

4.4. Secondary Exploratory Demographic Comparisons

The demographic comparisons in this subsection are retained as secondary exploratory analyses because they combine indicators of unequal measurement strength and were not adjusted for multiple testing. Most omnibus tests were not statistically significant. Age was associated with the weak forgiveness/restoration index at p = 0.046, but this isolated result is fragile because the age groups were highly unequal, the under−20 group contained only six participants, and no reported post-hoc comparison or multiplicity adjustment was available. It is not used to support the principal conclusions. The corresponding p-values are reported in

Table 5. Secondary exploratory p-values for demographic comparisons.

Demographic Variable	Marital Fidelity	Perceived Seriousness	Digital Infidelity	Forgiveness/Restoration
Age	0.877	0.262	0.326	0.046
Educational level	0.642	0.963	0.705	0.728
Marital status	0.159	0.254	0.240	0.142
Place of residence	0.290	0.221	0.286	0.313

Table note: These demographic comparisons are secondary exploratory analyses and were not adjusted for multiple testing. Results involving the weak forgiveness/restoration index should not be treated as stable subgroup effects.

.

4.5. Primary and Secondary Exploratory Observed-Score Correlations

The primary observed-score association was between marital-fidelity endorsement and perceived seriousness of infidelity (r = 0.685, p < 0.001). Correlations involving the weak digital-infidelity and forgiveness/restoration indices or the single boundary item are reported as secondary exploratory associations. Random measurement error may attenuate correlations involving weak indices, so their observed magnitudes may underestimate associations among the intended constructs. However, corrected correlations are not reported because the unidimensionality and error structure of the short indices were not established. A mechanical attenuation correction for the seriousness-digital correlation would yield an inadmissible value above 1.0, showing that the required assumptions are untenable and reinforcing the interpretation of severe observed-score overlap rather than a precisely estimated relationship between distinct latent constructs. The observed seriousness-digital correlation was r = 0.864 (r² = 0.746; pairwise tolerance = 0.254; pairwise VIF approximately 3.94). The observed-score correlations are summarized in

Table 6. Primary and secondary exploratory observed-score correlations.

Variables	r	p
Marital fidelity × perceived seriousness	0.685	<0.001
Marital fidelity × digital infidelity	0.621	<0.001
Marital fidelity × forgiveness/restoration	−0.243	0.002
Marital fidelity × boundary clarity	0.358	<0.001
Perceived seriousness × digital infidelity	0.864	<0.001
Perceived seriousness × boundary clarity	0.380	<0.001
Digital infidelity × boundary clarity	0.256	0.001

Table note: Correlations involving the digital-infidelity and forgiveness/restoration indices are secondary exploratory observed-score associations. Correlations involving boundary clarity refer to one statement. Measurement unreliability may attenuate some observed coefficients, but corrected correlations are not reported because the required psychometric assumptions were not established.

.

4.6. Secondary Exploratory Regression Analysis

Because the archived analytical materials did not permit a defensible additional sensitivity model, the existing hierarchical regression is retained only as a secondary exploratory analysis for transparency. The expanded model contains the weakly reliable forgiveness/restoration index and the single boundary-clarity item; its coefficients are therefore measurement-dependent, sample-specific estimates and are not used as the evidential basis for the Abstract, Discussion, Conclusions, or practical recommendations. Within this exploratory model, perceived seriousness had the largest observed standardized coefficient (beta = 0.632, p < 0.001), but it should not be interpreted as a definitive predictive effect because it was estimated alongside psychometrically weak covariates. Digital infidelity remained excluded because of its redundancy with seriousness. The model does not establish causality, mediation, or a hierarchy of validated predictors. The coefficients for this secondary exploratory model are reported in

Table 7. Secondary exploratory hierarchical regression of marital-fidelity scores.

Variable	B	SE	Beta	p
Model 1: demographic indicators
Female	0.190	0.077	0.202	0.015
Age	0.032	0.067	0.047	0.634
Educational level	−0.029	0.035	−0.066	0.413
Outside Al Ain city	0.099	0.076	0.105	0.197
Married	−0.107	0.103	−0.100	0.300
Model 1 fit: R2 = 0.062; adjusted R2 = 0.031
Model 2: demographic + attitudinal indicators
Female	0.001	0.060	0.001	0.982
Age	0.089	0.049	0.132	0.071
Educational level	−0.042	0.026	−0.095	0.108
Outside Al Ain city	0.020	0.056	0.022	0.719
Married	−0.085	0.076	−0.079	0.267
Perceived seriousness	0.530	0.054	0.632	<0.001
Forgiveness/restoration	−0.073	0.033	−0.134	0.027
Boundary clarity	0.043	0.028	0.102	0.122
Model 2 fit: R2 = 0.515; adjusted R2 = 0.488

Table note: This regression is reported only as a secondary exploratory analysis. It includes a weakly reliable two-item index and a single-item indicator, and its coefficients are not used to support the principal conclusions. Digital-infidelity judgments were excluded because of severe overlap with perceived seriousness.

.

4.7. Qualitative Meanings of Infidelity

The interview analysis elaborated how participants understood the survey domains. Ten of 11 participants described infidelity as a breach of trust, nine referred to a physical relationship outside marriage, six identified emotional involvement with another person, and five referred to a violation of the couple’s marital commitment. The latter means crossing mutually understood commitments and relational boundaries concerning exclusivity, honesty, and loyalty. Four participants also framed the violation in explicitly religious terms. Categories overlap because one person could mention several meanings. Counts describe this interview set only and should not be ranked as population prevalence or as proof of thematic saturation. The qualitative categories and participant counts are presented in

Table 8. Meanings of marital infidelity mentioned in interviews (n = 11).

Meaning Mentioned	Participants (n)
Breach of trust	10
Physical relationship outside marriage	9
Emotional relationship with another person	6
Violation of mutually understood marital commitment and relational boundaries	5
Violation framed as a religious or faith-based commitment	4
Deterioration of the family relationship	4
Absence of attention	3
Desire for and persistent thinking about another person	2
Lack of honesty	2
Absence of loyalty between spouses	2
Hidden relationship	2
Emotional shock or personal insult	1

Table note: Counts indicate the number of participants, out of 11, who mentioned each category. Categories are non-exclusive; counts are not percentages, prevalence estimates, or measures of thematic importance.

.

Interviewees also described technology as a double-edged context. Digital platforms can support routine contact between spouses, but they can also facilitate concealed, repeated, or emotionally intimate communication outside the relationship. The relevant mechanism in the accounts was not Facebook or any single platform; it was the combination of access, secrecy, persistence, and ambiguous boundaries across digital media. This formulation removes the earlier repetitive emphasis on Facebook-related jealousy.

4.8. Protective Factors

Mutual trust and honest communication were each mentioned by 7 of 11 participants. Respect and religious commitment were each mentioned by five, followed by frank communication, sustained attention, and mutual understanding, each mentioned by four. These counts indicate which categories appeared in this interview set; they do not estimate their prevalence in the UAE population or establish that a more frequently mentioned theme is analytically more important. Participants generally described prevention as relational and value-based rather than as surveillance or unilateral restriction. The reported protective factors are summarized in

Table 9. Protective factors mentioned in interviews (n = 11).

Protective Factor	Participants (n)
Mutual trust	7
Honest communication	7
Respect between spouses	5
Religious commitment/deterrence	5
Frank communication	4
Sustained attention	4
Mutual understanding	4
Sincere love	3
Renewing affection	2
Awareness of both partners	2
Contentment	2
Consideration of feelings	2
Responsibility and coping with life pressures	1
Psychological and emotional support	1
Commitment to moral values	1
Emotional communication	1
Spending time together	1
Avoiding behavior that creates suspicion	1
Upbringing, customs, and traditions	1
Meeting partners’ needs	1
Lowering the gaze	1
Sacrifice for relationship continuity	1

Table note: Counts indicate the number of participants, out of 11, who mentioned each factor. Multiple factors could be coded for one participant; the table is descriptive rather than inferential.

.

4.9. Joint Display of Quantitative and Qualitative Findings

Table 10. Joint display integrating questionnaire patterns and interview elaboration.

Quantitative Pattern	Qualitative Elaboration	Integrated Interpretation
Very high fidelity endorsement and positive observed association with seriousness	Infidelity was most often described as a breach of trust and a violation of marital commitment	Primary quantitative interpretation is limited to the better-supported fidelity and seriousness indicators; qualitative accounts contextualize trust and commitment.
High digital-index mean and severe observed-score overlap with seriousness (r = 0.864; r2 = 0.746)	Participants emphasized hidden, intimate, or repeated digital contact and ambiguity over acceptable behavior	Secondary exploratory finding: weak reliability and overlap prevent treatment of digital infidelity as a distinct validated construct; attenuation correction is not defensible.
Primary sex differences on fidelity and seriousness; secondary exploratory difference on one boundary statement	Interview themes stressed secrecy, trust, and clear limits, but n = 11 was too small for a gendered thematic comparison	Only fidelity and seriousness inform the principal interpretation. The single-item boundary result is hypothesis-generating and cannot represent multidimensional digital boundaries.
Low-moderate score on the weak forgiveness/restoration index	Definitions emphasized deep damage to trust, dignity, and commitment	Secondary exploratory only; qualitative accounts contextualize perceived damage to trust but do not validate the short index.
Observed association with a single boundary-clarity statement	Participants described acceptable conduct through mutual understanding and disclosure; the interviews did not systematically probe consensual device access versus coercive monitoring	Secondary exploratory only; future measures must distinguish mutual, specific, revocable sharing from surveillance, entitlement, and coercive access.
Trust and communication most frequently cited as protective	Participants described relational maintenance through openness, respect, attention, and mutual understanding	Protective mechanisms were relational and communicative, with religious and moral commitments providing an additional framework

makes the link between the two strands explicit. Because the questionnaire respondents and interview participants were independently recruited, the qualitative findings cannot be interpreted as direct explanations of particular survey scores or as person-level validation of the questionnaire. Instead, they provide contextual interpretations of the broader quantitative patterns at the level of shared constructs and research questions.

5. Integrated Discussion

5.1. Fidelity as Trust and Relational Commitment

The central integrated finding is that fidelity was understood as a structure of trust and commitment rather than as the absence of sexual contact alone. The primary quantitative result was a positive observed-score association between fidelity endorsement and perceived seriousness. Qualitatively, breach of trust was the most common definition, followed by physical and emotional relationships and violations of marital commitment. This interpretation aligns with multidimensional accounts in which secrecy, emotional displacement, and the violation of expectations can be as consequential as a particular physical act [1,6,7].

Symbolic interactionism helps explain why the meaning of infidelity depends on how spouses define the act and its implications. A message, concealed conversation, or emotional attachment is not inherently assigned one meaning by all couples; it becomes a violation when it contradicts shared or expected definitions of commitment. Bourdieu’s framework adds that these definitions are shaped by culturally learned dispositions concerning family, religion, dignity, and reputation. In the interview accounts, the language of trust and commitment connected private interaction to broader moral and familial meanings.

5.2. Digital Infidelity, Negotiated Access, and Intimate Surveillance

The digital-infidelity index had a high mean and a large observed correlation with perceived seriousness, but these results are secondary exploratory findings. The weak reliability of the three-item index increases measurement error, and the very high overlap with seriousness prevents a claim that digital infidelity was measured as a distinct validated construct. Although unreliability can attenuate observed correlations, a defensible disattenuated estimate cannot be calculated because the short indices have not demonstrated unidimensionality or a stable error structure. Interview accounts emphasized secrecy, intimacy, repeated contact, and concealment, which is consistent with work showing that online boundary violations are evaluated through their relational meaning rather than a strict online/offline distinction [3,4].

The HCI and security literature further refines this interpretation by distinguishing consensual access from surveillance. Shared passwords or device access may represent convenience, care, or trust when it is mutual, specific, and revocable; the same technical access can become a boundary violation when it is covert, coerced, repurposed, or used for persistent monitoring [26,27,28,29]. The study’s single boundary-clarity item cannot distinguish mutuality, informed consent, scope of permission, revocability, power imbalance, location tracking, or account monitoring. Future measurement should therefore separate at least five dimensions: consensual device sharing, credential sharing, expectations governing use, non-consensual monitoring, and technology-facilitated coercion. This approach is more precise than treating transparency as inherently protective or privacy as inherently suspicious.

5.3. Interpreting Sex Differences Without Essentialism

Women reported higher mean scores on the better-supported fidelity and seriousness indicators, with small-to-moderate sample-specific standardized differences. Results for the weak digital-infidelity index and the single boundary statement are retained only as secondary exploratory findings. In particular, the numerically largest difference concerned one general boundary-clarity statement and cannot be generalized to consent, mutuality, revocability, privacy expectations, coercion, surveillance, or digital-boundary competence. None of the observed score differences establishes that women behave more faithfully or that the pattern applies to the wider UAE population. Differential social-desirability pressures also cannot be ruled out.

Prior research has reported gender differences in judgments of emotional and sexual infidelity, but findings vary with definitions and social context [15,25]. Bourdieu’s perspective suggests that gendered socialization and unequal reputational consequences may influence sensitivity to secrecy and boundary violations. However, the present interview sample was too small for a defensible gender comparison, and the secondary exploratory regression is not used to explain the sex differences. The mechanism remains a question for future research rather than a settled conclusion.

5.4. Trust Restoration and Protective Factors

The observed negative association between fidelity and forgiveness/restoration is a secondary exploratory result and should not be interpreted as evidence that reconciliation is impossible. The two-item indicator was weakly reliable, so measurement error may reduce precision and replicability. Participants’ qualitative definitions emphasized the depth of the perceived violation. Research on recovery after infidelity shows that rebuilding trust is a process involving acknowledgment, meaning-making, accountability, and renegotiation of expectations rather than an immediate decision to forgive [40,41]. The present findings do not measure actual post-infidelity outcomes.

Trust and honest communication were the most frequently cited protective factors, followed by respect, religious commitment, attention, and mutual understanding. These themes support an emphasis on relationship maintenance and explicit boundary-setting. For family counseling and premarital education, the practical implication is not to impose a universal list of prohibited digital acts, but to help couples discuss privacy, contact with former partners, intimate messaging, deleted conversations, device monitoring, and expectations of disclosure before conflict emerges.

5.5. Contribution and Practical Implications

The study contributes a bounded sociological account of how adults recruited through a private university in the UAE linked fidelity to trust, digital conduct, and relational boundaries. Its principal quantitative contribution is the descriptive pattern of high fidelity endorsement and its positive observed association with perceived seriousness, interpreted alongside qualitative accounts of trust, secrecy, commitment, and relationship protection. Findings involving weak or single-item indicators, including the hierarchical regression, are retained only as secondary exploratory evidence. The HCI/security literature demonstrates why negotiated access and intimate surveillance must be separated conceptually.

Universities, family centers, and counseling services could use these findings to develop culturally responsive discussions of digital relationship literacy. Programs should distinguish privacy from concealment, encourage explicit agreements about online interaction, and address the emotional and social meanings of trust. Such initiatives should avoid moralizing assumptions and should be evaluated empirically with diverse samples, validated Arabic measures, and outcome data.

6. Limitations and Scope of Inference

Several limitations define what the study can and cannot conclude. First, the questionnaire used convenience recruitment from one private university. Nationality/citizenship was not collected, so participants cannot be characterized as Emirati and claims are limited to a UAE-based university-affiliated sample. The sample was concentrated among married participants (73.7%) and those aged 35–50 years (51.3%), and direct socioeconomic indicators were not collected. The study therefore cannot assess representation across nationality, income, occupation, housing, or household wealth. Second, 156 questionnaires provide exploratory evidence but are insufficient for national or fine-grained subgroup inference, especially where categories were very small.

Third, the qualitative component contained only 11 interviews. It provides contextual interpretation but does not establish saturation, prevalence, or the range of meanings across national, regional, or socioeconomic groups. Coding was not independently duplicated, so no inter-coder reliability coefficient is available. Fourth, the topic is morally sensitive and no formal social-desirability measure or indirect-response design was included. Fifth, the questionnaire was researcher-assembled and did not undergo factor or criterion validation. The digital-infidelity and forgiveness/restoration indices had weak internal consistency, and boundary clarity was measured with one item. Low reliability reduces precision and may attenuate observed correlations, but corrected correlations are not defensible because dimensionality and error assumptions were not established. The archived analytical materials did not permit reconstruction of a defensible reduced sensitivity model; the existing regression is therefore classified as secondary exploratory and does not support the principal conclusions. Finally, the cross-sectional design identifies associations but cannot determine causality, mediation, behavioral fidelity, or change over time.

A further procedural limitation concerns the timing of institutional ethical review. The project began as an internal course-based assignment not initially intended for external publication, and formal prospective committee review was not initiated. The authors acknowledge this as a procedural oversight rather than an ethics exemption. When publication was later considered, the authors disclosed that data collection had already occurred, and the Research Ethics Committee reviewed the consent process, instruments, confidentiality protections, data management, analyses, and publication plan before approving the ethical use, analysis, storage, and publication of the existing data on 27 July 2026. This retrospective approval does not replace prospective authorization.

7. Conclusions

Within this UAE university-affiliated sample, marital fidelity received very high self-reported endorsement and showed a positive observed association with perceived seriousness of infidelity. Interview participants most often described infidelity as a breach of trust and also recognized physical, emotional, digital, relational, and religious dimensions. Women reported higher endorsement on fidelity and seriousness, but these differences are sample-specific, may reflect differential normative responding, and should not be interpreted as behavioral or moral superiority. Results involving the weak digital-infidelity and forgiveness/restoration indices, the single boundary item, and the hierarchical regression are secondary exploratory findings and do not support the principal conclusions. Trust and communication were the most frequently mentioned protective factors in the 11 interviews.

The findings suggest that contemporary discussions of fidelity should address physical exclusivity, secrecy, emotional intimacy, privacy, disclosure, negotiated device access, and non-consensual monitoring. They do not support generalization to Emirati citizens or to the UAE population. Future research should use multi-site or probability-based sampling, record nationality and socioeconomic position, include validated social-desirability controls, recruit larger and more diverse qualitative samples, use independent coding or transparent peer audit, and validate multidimensional Arabic measures. Digital-boundary instruments should distinguish consensual and revocable sharing from covert surveillance and coercive access. These improvements are necessary before stronger conclusions can be drawn about family values and digital practices in the UAE.