Using Social Media to Recruit Seldom-Heard Groups: Reaching Women and Girls with Experience of Violence in Iran

Abstract

Social media recruitment and online surveys are valuable tools in social science research, but their effectiveness in reaching seldom-heard victims of gender violence in low-middle income (LMI) countries is under-explored. This empirical study aims to: (1) describe violence and abuse experiences and (2) assess the benefits and limitations of using social media to document violence against women and girls (VAWGs) in a LMI country to render visible the experiences of potentially isolated victims. A total of 453 Iranian women (aged 14–59, mean = 28.8, SD = 8.04) responded to an Instagram invitation for a study on women’s health and violence exposure from February 2020 to January 2022. The questionnaire covered general gendered abuse, domestic violence (DV), and forced unemployment. The analysis was performed using Stata 17. Nearly all participants reported abuse, including sexual (85.0%), psychological (83.4%), and technology-facilitated (57.4%) abuse, with 77.4% experiencing multiple forms. The street (62%) and home (52.8%) were common abuse locations. The perpetrators included known individuals (75.9%) and strangers (80.8%), with 56.7% reporting abuse by both. DV was reported by 72.6%, mainly involving psychological (73.1%), physical (53.4%), and/or sexual (17.2%) violence, with fathers (47.8%), husbands (42.7%), and brothers (40.2%) as frequent perpetrators. A quarter reported forced unemployment. Those experiencing DV and/or forced unemployment showed higher depression levels, suicidal ideation, and lower marital satisfaction. The study suggests using social media recruitment for VAWG research but cautions against overgeneralising from these data.

1. Introduction

Violence against women and girls (VAWGs), as defined by the United Nations, encompasses any act of gender-based violence (GBV) that leads to, or is likely to lead to, physical, sexual, or mental harm or suffering to women and girls, whether occurring in public or in private life (United Nations 1993). VAWGs is an issue of critical national and international significance, cutting across class, race, age, religious, and national boundaries. Due to its wide-ranging negative repercussions, including physical, psychological, economic, and social consequences for women, children, and wider communities, including the possibility of death, VAWGs is identified as one of the biggest barriers to women’s social, economic, and political participation at all levels (WHO 2021).

Surveys on VAWGs play a crucial role in understanding the nature and impact of such violence. This information is vital for supporting survivors and guiding policy and practice to effectively allocate resources in addressing inequalities (Walby 2005, 2016; WHO 2021). Worldwide surveys indicate that VAWGs is prevalent, affecting at least one in three women and girls globally (WHO 2021). While most of our knowledge about VAWGs comes from research and meta-analyses conducted in Western countries (Cafferky et al. 2018), the scale of the problem is significantly larger in low- and middle-income countries, especially in societies where gender discrimination is pervasive due to socio-cultural factors, patriarchal structures, and a lack of adequate legal support for victims (WHO 2021). In particular, women in Middle Eastern and North African countries experience a high prevalence of GBV (The World Bank 2021). However, data in these regions, where violence rates are notably high, are often seen as under-estimated, emphasising the need for cautious interpretation (WHO 2021). The normalisation of violence (Aghtaie 2016) and the stigma associated with violence in traditional or patriarchal societies may contribute to this under-estimation. In these societies, victims/survivors are often blamed for the abuse they experience, leading them to refrain from disclosing incidents to mitigate potential consequences, including further abuse (WHO 2021).

VAWGs is a pressing issue in Iran, a Middle Eastern country marked by its patriarchal structure and systematic and pervasive gender discrimination. The patriarchal and legal structure of the country perpetuates gender inequalities and reinforces societal norms that tolerate or even condone violence against women. The legal shortcomings and societal norms may discourage survivors from reporting incidents of violence or seeking legal recourse due to fear of stigma, retaliation, or being held accountable (Aghtaie 2017). Sometimes, a concern about preserving family “honour” prevents women from reporting. The notions of shame and honour are not exclusive to Muslim cultures and are common elements experienced by victims of violence within other communities (particularly sexual violence and rape) (Chantler and Gangoli 2011). However, the intersections of structural inequalities embedded within the legal system, both civil and criminal codes, the absence of a comprehensive definition of VAWGs, and traditional cultural norms within private spheres exacerbate the challenges faced by survivors seeking justice and support (Aghtaie 2017; Kar 2001). In Iran, it is estimated that every four days, a woman is killed at the hands of male relatives, often partners, ex-partners, or fathers (Lotfi 2023). Nevertheless, understanding the full extent of VAWGs in Iran is impeded by the sparsity of robust data (Nouri et al. 2012).

Iran lacks state-funded research programs specifically aimed at addressing VAWGs and DV as part of a comprehensive strategy. The only large-scale national survey on VAWGs was conducted in 2004, covering 28 provinces (Ghazi Tabatabai et al. 2005). The main findings have been published in 32 volumes; however, these publications, research data, and findings are not accessible to the general public and independent researchers. This contradicts Article 11 of the Istanbul Convention (Walby 2016), which highlights the importance of sharing such data to inform public debates and raise awareness about the prevalence and profound impact of VAWGs on their well-being.

Most of the existing research has mainly focused on intimate partner violence (IPV), often overlooking the various forms of abuse women may experience from other family members or in settings outside of their homes. Moreover, the scale of studies has been limited by the unavailability of sufficient funding, leading to small sample sizes. Larger studies often report on service users (such as those recruited through health service providers (Nouri et al. 2012; Saberian et al. 2004) or forensic medicine centres) (Aghakhani et al. 2003), potentially missing unreported cases to official bodies (Hajnasiri et al. 2016).

Furthermore, face-to-face interviews, the preferred data collection method for research on VAWGs (Walby 2005), face challenges in countries like Iran, mainly due to their high cost, dependence on government funding, and safety concerns. Ensuring confidentiality and safety, as emphasised in WHO guidelines (WHO 2001), is of vital importance in VAWGs research. However, maintaining these standards can be challenging, especially when dealing with sensitive information, such as the disclosure of violence, which requires utmost care and ethical considerations. In culturally bonded and stigmatised contexts such as Iran, revealing sensitive and personal information to an outsider (interviewer) may trigger further violence, particularly for women who live with abusive partners (Hajnasiri et al. 2016; Jafari and Parvin 2022; Nikparvar et al. 2021). While this is not limited to the Iranian context, the inadequate or absence of legal support could exasperate the situations for the victims as well as the researchers in Iran. This raises ethical and practical issues, as conducting a study aimed at accurately documenting the nature of VAWGs may put both respondents and researchers at risk of further violence. This could further discourage the use of face-to-face interviews as the primary method of data collection in countries like Iran.

These limitations have led to the realisation that even well-intentioned studies using conventional data collection methods may result in the under-reporting of violence. To mitigate funding constraints and safety concerns, paper-based questionnaires have gained traction as the primary data collection method in Iranian studies. However, the established measurement scales, rooted in global north framings, may be unsuitable for the Iranian context. Differences in concepts, expressions of experiences, social structures, and gender norms between the global north and Iran can lead to scales inaccurately reflecting social realities and relationship dynamics.

Due to the aforementioned limitations, women victims of GBV in Iran are “seldom heard” in research, leading to the inadequate acknowledgement of their voices and experiences in broader societal conversations, policy-making, and public discourse. Legal barriers, cultural norms, safety concerns, and ethical considerations contribute to their under-representation. Limited access to resources, support networks, and services further isolates these victims, hindering their engagement in research and advocacy (Faugier and Sargeant 1997; Guillory et al. 2018). Consequently, studies on this demographic are scarce, relying on qualitative methods with small sample sizes, limiting a comprehensive understanding of the diversity and breadth of experiences within this population.

Researchers facing challenges in participant recruitment, especially from “hard-to-reach” or “seldom-heard” populations, are increasingly turning to innovative approaches, such as utilising social media for recruitment and online data collection (Sandhu et al. 2023). This method has proven effective in reaching diverse and under-represented groups in research (Guillory et al. 2018; Sandhu et al. 2023), particularly amidst the COVID-19 pandemic (Littler and Joy 2021).

More specifically, ethical and secure social media recruitment has successfully collected data on hard-to-reach victims of violence, including victims of sexual assault, and on marginalised groups such as South Asian women in the UK experiencing gender-related violence (Canan et al. 2022; Sandhu et al. 2023). These successes highlight the potential of alternative approaches for researching manifestations of VAWGs in seldom-heard populations.

Despite the increasing use of social media for participant recruitment and online platforms for data collection, the majority of research in this area has been conducted in high-income countries such as the United States and the UK (Canan et al. 2022; Chenane and Hammond 2022; Sandhu et al. 2023). There is a notable scarcity of literature addressing the application of these methods to recruit and collect data from seldom-heard participants in low- and middle-income countries (LMICs), such as Iran.

Most social media platforms are accessed through mobile phones. Women’s mobile ownership rates vary across LMIC regions. In the Middle East and North Africa regions, 82% of women owned a mobile phone during the period of 2017–2019, which is lower than the LMICs in East Asia and the Pacific regions (92%) but higher than those in the Sub-Saharan Africa (74%) and South Asia (65%) regions (Rowntree 2020). As of 2019, there remains an 8% gender gap in mobile ownership across LMICs, with women being 8% less likely to own a mobile phone. However, this gap is gradually narrowing. Similarly, mobile internet usage is rapidly expanding as smartphone ownership increases, leading to a closing gender gap in internet access (Rowntree 2020).

While still sub-optimal, the growing accessibility of mobile devices and mobile internet in LMICs presents an opportunity for research on seldom-heard groups who otherwise remain unheard.

This research aimed to bridge the gap by investigating the use of social media to document the experiences of VAWGs in Iran. The goal was to amplify the voices and visibility of potentially isolated victims and understand the nature of their experiences. The study critically examines the value, benefits, limitations, and ethical implications of conducting online research and recruiting participants through social media in the context of VAWGs.

2. Materials and Methods

2.1. Participants and Data

The data were collected from a sample of 453 Iranian women aged 14–59 (mean = 28.8, SD = 8.04) recruited via an online survey created on a free online survey builder platform called Porsline. A link to a web survey (in Farsi language) framed as “a survey on women’s health and exposure to violence” was distributed via Instagram, which is the most widely used social media platform in Iran, with over 24 million users (according to state officials, who constitute around 43% of social media users in Iran (Turani 2023)).

The data were collected between February 2020 and January 2022. All the responders provided e-informed consent. It took, on average, 5–7 min to fill out the questionnaire, and participants were not offered any financial incentive. No personal information of the respondents, such as their name, address, or contact details, was included as part of the survey. Participation in the survey was voluntary. Participants had to provide a response to all items in order to submit the questionnaire. An “exit quickly” button on each survey page was provided in case the individuals were taking the survey in the same space as their partner or other family member, or if they entered the room.

The survey (hereafter called the 2022 Iran online VAWGs survey) was a cross-sectional survey aimed to characterise the nature of both domestic (including intimate partner violence and other family members) and non-domestic (e.g., stranger) violence experienced by women in Iran. Specifically crafted survey questions for this study were informed by the researcher’s (FB) extensive experience in working with women facing domestic violence in Iran and included both closed- and open-ended questions to collect quantitative and qualitative data. The composition of the survey was also informed by a rapid literature review of existing tools used in Iranian literature. Attempts were made to include relevant questions to measure different forms of GBV commonly measured in established surveys, such as physical violence, sexual violence, psychological abuse, and economic abuse. Three counsellors working with women victims of violence were asked to assess the clarity, comprehensibility, and cultural appropriateness of the questions. Their feedback was utilised to refine the survey instrument as needed. For example, based on their feedback, questions were worded to measure the experiences of violence in both public and private spheres, as well as to capture violence types more pertinent to settings such as Iran, including forced marriage, reproductive coercion, and forced unemployment. Three sections of the survey, addressing experiences of 1. general gendered abuse, 2. domestic violence, and 3. forced unemployment, are presented in this paper. The survey also included questions on mental health status (experience of depression and suicidal ideation) and marital satisfaction. The objectives of the 2022 Iran online VAWGs survey were to:

-

Collect information on the experiences of various types of violence and abuse in Iran, including the location of violence/abuse (domestic or non-domestic) and the relationships between the victims and perpetrators;

-

Ascertain the association between the experiences of different types of violence and abuse and a range of mental health outcomes.

It should be noted that the survey included questions on forced sex, forced marriage, reproductive coercion, attitude towards partner sexual abuse, and attitude towards the most effective measures to combat VAWGs, but these aspects are not presented in this paper.

The secondary analyses were approved by the committee at City, University of London, which considers medium-risk applications (Violence, Health and Society VISION consortium, ETH21220–299).

2.2. Measures

After reading information about the purpose of the study, potential female respondents were invited to participate in the research. Respondents were asked to report their socio-demographic characteristics, including age, residence (rural or urban), marital status (single, married, divorced, or divorcing), educational attainment, and employment status (free text). For the analysis and due to the small sample sizes, the divorced (n = 24) and divorcing (n = 32) categories were combined. Similarly, age was initially categorised into five categories (14–18, 19–30, 31–40, 41–50, and51–59), but due to the insufficient sample size for the 51–59 age bracket (only 6 respondents), this group was combined with the 41–50 age bracket. Free-text responses to the open question on the respondent’s highest educational attainment were manually coded into four categories (postgraduate degrees, bachelor’s or associate’s degrees, diploma, or middle or high school). Similarly, free-text responses to the open question on employment status were manually coded into three broad categories (unemployed, employed, or student).

2.2.1. Main Exposures

The survey addressed “general gendered abuse”, domestic violence, and forced unemployment in three distinct sections. See Supplementary Table S1 for the exact wordings and response options for the questions on all the variables.

General Gendered Abuse

Respondents were asked to respond to a general question on their experience of abuse. The subsequent questions collected details about the forms that the abuse took and the location where the abuse happened, available to all respondents regardless of their responses to the first general question. Sexual abuse was identified if the respondents reported inappropriate touching and/or catcalling. Psychological abuse was identified if the respondents reported insults (verbal harassment) and/or threats. Technology-facilitated abuse (tech abuse) was identified if the respondents reported at least one of the following acts of abuse: sexting (receiving inappropriate pictures or videos of a sexual nature), texting (receiving unwanted and inappropriate text messages), and receiving harassing/unwanted phone calls. A count of the number of types reported (sexual, psychological, and tech-facilitated) was derived (1, 2, and 3). The respondents were given six potential locations of abuse to choose from; however, a filter was applied, and the selection of location was restricted to a maximum of two. A binary variable was created to record the number of locations (1 and 2), where 2 could be interpreted to mean “2 or more”. The respondents were asked whether the perpetrators were a known person, a stranger, or both.

Domestic Violence (DV)

The respondents were then asked a general question about their experience with DV. The subsequent questions collected details about the forms that the DV took and the victim–perpetrator relationship, available to all respondents regardless of their responses to the general question on their experience of DV. Physical violence was identified if the respondents reported being beaten up. Sexual abuse was identified if the respondents reported that they had been sexually abused. Psychological abuse was identified if the respondent reported at least one of the following forms: insults, threats, or humiliation. A categorical variable was created to record the number of abuse forms (0, 1, 2, and 3). The respondents were then asked about their relationship with the perpetrator. The response options included the three main male family members commonly known as perpetrators of DV in Iran, including father, husband, and brother. The respondents could also choose “all three” and “none”. However, no free-text box was provided so the respondents could write down other types of perpetrators not listed as a response option (e.g., female family members or male family members other than father, husband, or brother, such as a cousin, uncle, grandfather, or son).

Economic Abuse (Forced Unemployment)

As an indicator of economic abuse, the respondents were asked whether they had experienced involuntary (forced) unemployment and, if so, the reason behind it.

2.2.2. Main Outcomes

Respondents rated their feelings of depression, suicidal ideation, and marital satisfaction using five-point scales.

2.3. Statistical Analyses

The analyses were performed in the Stata statistical package version 17 (StataCorp 2017). Frequencies and percentages were calculated for each derived variable. Chi-square test statistics were used to compare the prevalence rates between the socio-demographic groups. A series of two-sample t-tests were performed to compare the patterns of depression, suicidal ideation, and marital satisfaction in those who reported DV (at least one form) and those who reported forced unemployment with those who did not report such experiences.

3. Results

3.1. Sample Characteristics

Of the sample, 40.6% were married, 47.0% were single, and 12.4% were divorced/divorcing.

Table 1. The 2022 Iran online VAWGs survey sample socio-demographic characteristics compared with the general Iranian female population.

Socio-Demographic Characteristics	Survey Respondents	General Iranian Female Population
	n (%)
Age	Mean = 28.8, SD = 8.0	Mean = 33.8
14–18	41 (9.0)
19–30	230 (50.8)
31–40	149 (32.9)
41–59	33 (7.3)
Residence
Rural	16 (3.5)
Urban	437 (96.5)	75.81% in 2022 (Ali Asghar 2021; Macrotrends 2023)
Marital status
Divorcing or divorced	56 (12.4)	31.6% in 2019 (National Organization for Civil Registration 2020)
Married	184 (40.6)
Single	213 (47.0)
Educational attainment
Postgraduate degree	124 (27.4)	Tertiary 57% in 2020
Bachelor’s or associate’s degree	235 (52.0)
High school diploma	60 (13.3)
Middle or high school	33 (7.3)
Employment status
Employed	176 (38.8)	15.7% in 2022 (The World Bank 2022)
Unemployed	126 (27.8)
Student	151 (33.3)

compares the demographic characteristics of the respondents with the female Iranian general population obtained from online sources (e.g., World Bank and Iran Open Data).

The survey sample was younger than the general female population (the mean age for the survey sample was 28.8 years versus 33.8 years for the general population). The share of tertiary education (associate’s, bachelor’s, or postgraduate degrees) was higher among the sample (79.4%) than the general population (57.0%). A higher proportion of the study sample was in a paid job (employed) than the general female population (27.4% for the study sample vs. 15.7% for the general population). Residing in urban areas was reported by a higher proportion of the sample (96.5%) than the general population (75.8%).

3.2. Reported Abuse: Nature, Location, and Relationship to the Perpetrator

Experiences of abuse were reported by all but two respondents (99.6%). The proportion of respondents experiencing at least one act of sexual abuse was 85; individual acts of inappropriate touching and catcalling were reported by 70.6% and 71.5%, respectively. The proportion of respondents experiencing at least one act of psychological abuse was reported by 83.4%, including insults (75.1%) and intimidation/threats (50.8%). Technology-facilitated abuse was reported by more than half (57.4%), including at least one occasion of receiving unwanted abusive text messages (41.3%), phone calls (37.1%), or sexting (33.5%). Experiences of multiple forms of abuse were widespread, reported by 77.4% of the sample, while the experience of only one form of abuse was reported by 22.5% (

Table 2. Nature and location of abuse reported by the respondents of the 2022 Iran online VAWGs survey (n = 453).

Questions/Variables	n	%
Have you ever been abused?
Yes	451	99.6
No	2	0.4
What form did the abuse take?
Sexual abuse	385	85.0
Sub-categories of sexual abuse
Inappropriate touching	320	70.6
Catcalling	324	71.5
Psychological abuse	378	83.4
Sub-categories of psychological abuse
Insults (verbal harassment)	340	75.1
Threats	230	50.8
Technology-facilitated abuse	260	57.4
Sub-categories of tech abuse
Sexting (sending inappropriate pictures or videos of a sexual nature)	152	33.5
Texting	187	41.3
Harassing/unwanted phone calls	168	37.1
Number of abuse forms experienced
One	102	22.5
Two	132	29.1
All three (sexual, psychological, and tech-facilitated)	219	48.3
Where did the abuse happen?
Street	281	62.0
Home	239	52.8
Social media	113	24.9
Workplace	48	10.6
School/university	35	7.7
Other locations	52	11.5
Number of locations
1	118	26.0
2 or more	335	74.0
Who did this to you? (perpetrator)
Known person	344	75.9
Stranger	366	80.8
Both	257	56.7

).

The most common reported location of abuse was the street (62.0%), followed by the home (52.8%), social media (24.9%), and the workplace (10.6%), while school/university (7.7%) was the least common reported location of abuse. Three-quarters of the sample (74%) reported experiences of abuse in at least two locations. Regarding victim–perpetrator relationships, abuse by a known perpetrator was reported by 75.9%, and a similar proportion (80.8%) reported abuse by a stranger. Over half (56.7%) reported abuse by both a known person and a stranger (Table 2).

3.3. Experience of Domestic Violence, Its Forms, and Perpetrators

An experience of domestic violence (DV) was reported by 72.6% of the sample. Psychological violence was the most common reported form of DV reported by two-thirds of the sample (73.1%), including at least one form of humiliation (64%), insult (verbal abuse) (60.7%), and threat (47.9%). Physical violence in the form of beating up was the second most common reported form of abuse, reported by half of the sample (53.4%), followed by sexual violence (17.2%). Of the sample, 54.1% reported an experience of multiple forms of violence, with 41.3% reporting two forms and 12.8% reporting all three forms. For those who reported an experience of at least one form of DV, fathers (47.8) and husbands (42.7%) were the most commonly reported perpetrators, followed by brothers (23.1%), and 11.8% reported all three (father/husband/ brother) (

Table 3. Experience of domestic violence, its forms, and perpetrators in the 2022 Iran online VAWGs survey (n = 453).

	n	%
Have you ever been subjected to domestic violence?
Yes	353	77.9
No	100	22.1
What was the form of DV?
Physical violence (beating up)	242	53.4
Sexual abuse	78	17.2
Psychological violence (at least one form)	331	73.1
Sub-categories of psychological violence
Insults (verbal harassment/abuse)	275	60.7
Threats	270	47.9
Humiliation	290	64.0
Number of DV forms experienced
Zero	105	23.2
One	103	22.7
Two	187	41.3
Three	58	12.8
Who did this to you? (DV perpetrator) *
Father	167	47.3
Husband	150	42.5
Brother	81	22.9
All three	41	11.6
None	37	10.5

* The denominator consisted of those who reported at least one form of DV, totalling 353 respondents. * A “no response” option was available for a combination of two perpetrators (e.g., father and brother, father and husband, and husband and brother).

).

3.4. Experience of Forced Unemployment and Its Causes

Of the sample, a quarter (122 out of 453, 26.9%) reported that they had been forced to give up on employment due to violence or abuse, either at work or in the family. Of those who reported involuntary unemployment, 85.2% reported that either a family member (46.7%) or their husband (38.5%) had forced them to give up on a paid job or employment. Involuntary unemployment due to abuse at the workplace was reported by 14.7%. It should be noted that the question on “who forced you” or the “reason for forced unemployment” had a single answer format, which allowed the respondents to select one, and only one, answer choice (

Table 4. Forced unemployment and its causes.

Question/Variable	n	%
Have you been forced to give up on employment?
Yes	182	40.2
No	271	59.8
Who forced you? Reason for forced unemployment? *
Family forced me	57	46.7
Husband forced me	47	38.5
Due to abuse at the workplace	18	14.7
Experience of abuse **
Yes, one form	26	21.3
Yes, two or three forms	96	78.7
Pearson chi2 (p value) for the association between general abuse and forced unemployment	0.14 (0.7)
Experience of DV **
Yes	112	91.8
No	10	8.2
Pearson chi2 (p value) for the association between DV and forced unemployment	18.7 (0.001)

* The denominator consisted of those who reported an experience of forced unemployment due to violence or abuse (n = 122); ** the percentages present row%.

).

3.5. General Abuse, DV, and Forced Unemployment by Socio-Demographic Characteristics

Few significant associations were found between the socio-demographic characteristics and experience of abuse, DV, and forced unemployment, indicating that these experiences were similarly widespread across all sectors of the sample (

Table 5. Distribution of abuse, DV, and forced unemployment across socio-demographic sub-groups.

	Abuse			DV				Forced Unemployment
	Sexual	Psychological	Technological	Physical	Sexual	Psychological	Any DV
Age
14–18	87.8	87.8	56.1	41.5	9.8	73.2	75.6	29.3
19–30	87.4	83.0	62.2	50.9	14.8	72.2	76.5	27.8
31–40	82.5	83.2	50.3	58.4	20.1	75.2	81.21	24.8
41–59	75.8	81.8	57.6	63.4	30.3	69.7	75.8	27.3
Chi2 (p-value)	4.2 (0.24)	0.66 (0.88)	5.2 (0.16)	5.8 (0.12)	7.4 (0.06)	0.62 (0.892)	1.4 (0.70)	0.54 (0.91)
Residence
Rural	75	87.5	37.5	50.0	12.5	75	78.0	56.2
Urban	85.3	83.3	58.1	53.5	17.4	73	75.0	25.9
Chi2 (p-value)	1.3 (0.25)	0.20 (0.66)	2.68 (0.10)	0.08 (0.78)	0.26 (0.61)	0.03 (0.859)	0.08 (0.77)	7.2 (0.007)
Marital Status
Single	90.6	88.3	63.8	54.0	12.7	72.8	77.5	29.6
Divorcing/divorced	83.9	89.3	69.6	71.4	33.9	80.4	89.3	35.7
Married	78.8	76.1	46.2	47.3	17.4	71.2	75.0	21.2
Chi2 (p-value)	10.84 (0.004)	12.2 (0.002)	16.5 (0.001)	10.1 (0.006)	14.0 (0.001)	1.85 (0.397)	5.14 (0.08)	6.03 (0.049)
Educational Attainment
Middle/high School	84.8	84.8	63.6	48.5	15.1	72.7	72.7	30.3
Diploma	85.0	85.0	58.3	58.3	25.0	81.7	85.0	43.3
Associate/bachelor	84.7	83.0	57.0	55.7	17.4	74.0	79.1	27.2
Postgraduate	85.5	83.1	56.4	47.6	13.7	66.9	73.4	17.7
Chi2 (p-value)	0.04 (1.0)	0.20 (0.98)	0.60 (0.90)	3.11 (0.37)	3.7 (0.293)	4.7 (0.192)	3.94 (0.27)	13.7 (0.003)
Employment Status
Unemployed	82.5	80.9	54.0	57.9	23.8	80.9	83.3	32.5
Student	88.7	91.4	62.2	48.3	11.3	67.5	73.5	32.4
Employed	83.5	78.4	55.7	54.5	17.6	72.2	77.8	18.2
Chi2 (p-value)	2.56(0.28)	10.70 (0.005)	2.27 (0.32)	2.68 (0.26)	7.6 (0.022)	6.4 (0.041)	3.85 (0.15)	11.20 (0.004)

).

However, there were a few exceptions. For example, the respondents who were divorced or divorcing were more likely than single and married respondents to report having experienced almost all forms of abuse, DV, and forced unemployed. Additionally, those who were single were more likely than married respondents to report having experienced abuse, DV, and forced unemployment.

Another exception was that the respondents with lower educational attainment (high school diploma or associate’s degree), those who were unemployed at the time of the survey, and those who were living in rural areas were more likely than those with higher educational qualifications, those who were employed, and those living in urban areas to report having experienced forced unemployment (Table 5).

3.6. Association with the Outcome Variables

Those who experienced DV reported significantly higher levels of depression (t = 5.28) and suicidal ideation (t = 5.35) and lower levels of marital satisfaction (t = −5.64) than those who reported forms of abuse other than DV (p < 0.0001). A similar pattern was observed for those who reported an experience of forced unemployment, where those who experienced this form of economic abuse reported significantly higher levels of depression (t = 5.60) and suicidal ideation (t = 5.49) and lower levels of marital satisfaction (t = −5.22) than those who did not report such abuse (p < 0.0001) (

Table 6. Associations between exposure to DV, forced unemployment, mental health, and satisfaction outcomes.

	Sample M (SD)	Domestic Violence				Economic Abuse (Forced Unemployment)
		Yes M (SD)	No M (SD)	t (541)	p	Yes M (SD)	No M (SD)	t (541)	p
Depression	3.42 (1.17)	3.57 (1.11)	2.89 (1.20)	5.28	0.0001	3.91 (1.04)	3.24 (1.16)	5.60	0.0001
Suicidal ideation	2.86 (1.68)	3.08 (1.68)	2.09 (1.41)	5.35	0.0001	3.55 (1.56)	2.60 (1.64)	5.49	0.0001
Marital satisfaction	2.75 (1.45)	2.55 (1.42)	3.45 (1.34)	−5.64	0.0001	2.18 (1.32)	2.96 (1.44)	−5.22	0.0001

).

4. Discussion

This study reflects on addressing the benefits and challenges of conducting surveys on sensitive topics such as GBV via social media in LMI countries. It specifically captured the experiences of a seldom-heard group—women in Iran who have experienced abuse and violence. It provides unique insights into the various forms of GBV that victims have experienced, the locations of these incidents, and the identities of the perpetrators.

In the following, we discuss both the strengths and limitations/challenges of using social media platforms to recruit victims of GBV and to document the nature of their experiences, focusing specifically on the Iranian women sample when applicable.

4.1. Strengths

This study makes a methodological contribution to a subject that has been largely under-explored in the field of VAWG research (i.e., social media recruitment). It provides empirical validation for the potential/efficacy of social media platforms in enhancing conventional survey data collection methods within this field. This is important as social media platforms offer broader reach and a safer environment for data collection than traditional methods such as face-to-face interviews and paper-based questionnaires, especially in situations where these options may pose risks of violence or abuse to respondents and/or researchers or may result in under-estimation, which can further undermine the importance of the issue of VAWGs.

Furthermore, in addition to their extensive reach, the depth and breadth of the information gathered through social media platforms are unparalleled and come at little to no cost, distinguishing it from traditional research methods, which typically rely on substantial state funding.

Within the context of the Iranian survey, the survey’s innovative use of social media for recruitment facilitated access to a diverse group of female victims who had experienced various forms of abuse and violence who might have otherwise remained unheard.

The survey results on the intersecting and overlapping experiences of abuse, DV, and economic abuse in the form of forced unemployment among this sample of victims contribute considerably to our understanding of VAWGs in Iran. The findings from the survey also include data concerning VAWGs within the public sphere and outside of the non-domestic sphere. The victims in our findings disclosed experiences of sexual and psychological abuse, emphasising the need for targeted interventions to address these often overlooked and less recognised forms of abuse. These findings also support the importance of broadening interventions to include more than just those aimed at physical violence.

Additionally, the victims were equally likely to report abuse occurring in public spaces such as the streets and online as well as within private spheres such as homes. These findings are unique and of particular importance as there seems to be a lack of research focusing on VAWGs outside of domestic settings in Iranian literature. Our research not only sheds light on an under-researched aspect of this issue but also emphasises the imperative to investigate and address violence in various contexts beyond the confines of the home. Additionally, our findings indicate that experiences of abuse were widespread across various socio-demographic characteristics, highlighting that this issue affects different segments of society and not just specific groups/sub-populations.

These findings underscore the importance to inquire about and recognise both domestic and non-domestic forms of abuse for service providers, ensuring comprehensive support and customised interventions for individuals experiencing abuse from different sources. These findings also highlight the need for systemic changes, increased awareness, and enhanced safety measures in both domestic and public spheres.

In this sample of victims, the experience of domestic violence was also prevalent, with three out of four respondents reporting experience of at least one form of physical, psychological, or sexual DV. Psychological violence was the most frequently reported form of DV, followed by physical and sexual violence. Multi-victimisation was common, with two-thirds of the respondents reporting experiences of multiple forms of DV. Fathers and husbands were the most commonly reported perpetrators, followed by brothers. Victimisation by multiple perpetrators was also reported.

These findings contribute to the limited information available on DV perpetrated by family members other than the husband within Iranian literature. Most Iranian studies on DV primarily report on violence perpetrated solely by the husband or fiancé (Ahmadi Gohari et al. 2023), drawing on samples that mostly consist of married or ever-partnered women. This approach excludes single, co-habiting, or never-partnered women from the research.

Furthermore, our findings indicate that economic abuse in the form of forced unemployment due to violence or abuse, either at work or in the family, was reported by a quarter of the sample. Economic abuse has traditionally been conceptualised as a form of psychological abuse (Royal 2022) and has received relatively less attention compared to other forms of abuse. However, research has recently begun to recognise economic abuse as a distinct form of VAWGs and document the magnitude of the issue and adverse consequences of exposure to this form of VAWGs (Mellar et al. 2023; Royal 2022). In line with international research indicating that economic abuse often co-occurs with other forms of VAWGs (Royal 2022; Salimi et al. 2023), our findings indicate that the overwhelming majority of women who experienced forced unemployment (~92%) also faced at least one type of DV (physical, sexual, or psychological DV). These findings collectively underscore the intersectionality and complexity stemming from multiple sources of economic coercion, including abuse from family members, intimate partners, and workplace settings. Despite not receiving much direct attention in Iranian literature to date, our findings confirm that economic abuse is an important area for future attention.

Findings on the effects of VAWGs demonstrated that DV and abuse that limits engagement in public life (forced unemployment) have particularly strong negative impacts on women’s mental health, increasing their risk of depression and suicidal ideation. Additionally, these experiences were associated with reduced levels of marital satisfaction.

Iran is one of only a few states not to have signed the Convention on the Elimination of All Forms of Discrimination Against Women. Domestic violence is not a crime under Iranian law, and the government generally views and treats it as a “private family matter”. The lack of legal, financial, and social support for women in Iran who are in violent relationships is a significant burden for these women. Notably, the absence of legal support for women coupled with structural violence can foster an environment where violence may be perceived as an accepted part of life and, consequently, normalised for certain individuals, particularly the perpetrators.

Our findings on the GBV experienced by female victims in Iran endorse the necessity to critically challenge harmful social and gender norms by using approaches that are sensitive to women’s positions and unequal gender relations in the family and society. This requires a comprehensive approach, involving legal reforms, awareness campaigns, the economic empowerment of women, and bolstering support services for survivors. Engaging with the community, raising awareness, challenging the norms perpetuating violence, and holding perpetrators accountable for their actions are crucial steps in combatting this issue. Work is especially important in meeting the needs of women and girls who experience intersecting forms of discrimination and who are at the most risk of violence.

The findings of the Iranian survey on GBV have the potential to guide effective strategies to reduce the acceptance of violence and to identify high-risk groups requiring support through the provision of tailored intervention efforts if followed with appropriate investment and action. In the absence of such initiatives, findings have the potential to inform public debate about VAWGs, their cause, and harm to women, which ideally can lead to increased advocacy for policy changes and legal reforms to better address and prevent VAWGs. Additionally, informed public debate can ideally challenge and change harmful social norms that perpetuate VAWGs, promoting a culture that prioritises respect, consent, and gender equality.

4.2. Limitations/Challenges

While providing valuable insights into the manifestation and impacts of different forms of VAWGs among seldom-heard victims, conducting surveys on VAWGs through social media and recruiting participants via these platforms presents several limitations and challenges. Notably, digital literacy or accessibility issues can create a digital divide, excluding those with low literacy proficiency in social media use, those lacking access to the necessary technology, rural communities, and people with disabilities. The higher proportion of the respondents with tertiary education in our sample (80%) and those living in urban areas (96.5%) compared with the general population are examples of such sampling biases.

Furthermore, those most impacted by violence may be hesitant to participate in any research due to reasons such as fear, stigma, or other barriers, including mental and physical complexities resulting from abuse. While this limitation is not exclusive to online surveys and social media recruitment, such hesitancy can lead to under-reporting and further restrict the generalisability of the research findings.

Moreover, the findings of studies conducted through social media lack generalisability due to the sampling biases inherent in social media recruitment. In the context of the Iranian survey, the sample of victims who use social media may differ in their characteristics, experiences, and needs from those who do not use social media. Additionally, the experiences of those who come forward voluntarily may not accurately represent the experiences of the broader population of victims/survivors.

The study’s outreach was confined to those who followed a particular Instagram account owned by the founder and manager of a DV service. This approach was effective in recruiting victims/survivors; however, it may have excluded the experiences and perspectives of a broader and more diverse group of victims/survivors who do not engage in these particular online communities.

To expand the survey’s reach and enhance the representation of participants from diverse backgrounds and online engagement preferences, future research could seek to broaden the scope by recruiting participants from various social media platforms, such as Twitter and Facebook, and distributing the survey link through multiple channels and accounts.

Online surveys have the advantage of providing participants with a platform to respond to sensitive questions at their convenience, potentially facilitating a more honest response to the questions. This might be because the participants might feel more anonymous in an online setting, and hence, this could result in more candid responses. However, it is important to note that within the online survey, still, it is important to stress to the respondents that their answers are confidential and untraceable (Bjørn 2017). Additionally, online surveys reduce “social desirability bias” (Tourangeau 2014). However, one of the main concerns regarding online surveys is the challenge of authenticating the identity of the respondents.

Moreover, carrying out any surveys on sensitive and potentially distressing topics such as GBV requires careful consideration of anonymity and confidentiality issues (WHO 2001), but also a lack of face-to-face human support and other safety aspects are the ethical concerns to ensure the well-being of participants. In the context of the Iranian study, the participants were provided with reassurance regarding anonymity and confidentiality prior to responding to the survey questions on the participant information sheet. The participants were assured that the survey was not collecting their personal information, including names, addresses, and contact details and that their answers were untraceable. The social media account that was used for data collection included ample and detailed information on support services and helplines to ensure participants had access to relevant resources if they found the survey content distressing.

To safeguard participants’ safety and comfort and allow them to be in control, an exit option on each survey page allowed them to discontinue their participation at any point if they felt uncomfortable or no longer wished to participate. To provide participants with the option of returning to the survey (with no obligation) at a later time, their responses were securely stored.

Online surveys, in general, are concise enough to boost engagement and to prevent participants from feeling overwhelmed and, in this case, to prioritise participants’ safety and minimise respondent burden. However, the constraints inherent in designing surveys for brevity may result in overlooking important nuances and variations in the experiences of violence. Within the context of this study, the survey did not capture some of the violence forms/types that could have been pertinent to a country like Iran. For example, honour-based abuse, child marriage, female genital mutilation, and various forms of coercive control were not included in the survey questions. Similarly, questions pertaining to physical abuse were excluded from the section addressing general abuse (due to overlap with questions on sexual abuse in public spaces).

The survey also did not include questions on the experience of abuse within public spaces beyond the streets, such as parks and recreation areas, shopping centres, and public transportation. Additionally, while the single-item measure on forced unemployment identified the most prevalent forms of economic abuse, it failed to recognise other ways that economic abuse can occur.

Furthermore, the survey did not measure the severity or frequency of exposure to violence. Although these omissions possibly reduced the respondent burden, retained engagement, and boosted completion, the survey did not fully capture the gravity and implications of violence for victims, and therefore, it did not provide us with a complete understanding of VAWGs and its multi-faceted nuances.

To enhance future VAWG survey design, it is vital to balance brevity with capturing some of the key aspects. Pilot testing would allow for identifying, prioritising, and contextualising the key forms of violence relevant at both the country and local levels. Moreover, future research could enhance the level of understanding by integrating opportunities for qualitative follow-ups or open-ended questions.

Finally, in this study, some questions had limitations on the number of responses to two options for where the abuse occurred and one option for reasons for forced unemployment. While capping the responses aimed to focus the analysis on the primary settings and identify the main sources of economic coercion, it may have unintentionally led to an incomplete picture, the under-reporting of certain locations, and an under-disclosure of multiple sources of forced unemployment.

The final limitation worth noting pertains to the relatively low completion rate. Despite being accessible on the Instagram page for nearly two years, only 453 respondents completed the survey. This disparity may be attributed to the irregular posts that remind followers of the Instagram page throughout the life of the survey. Future research endeavours could benefit from the implementation of systematic regular reminders and the adoption of robust and diversified promotional strategies to increase the awareness of the surveys among the intended audience and stimulate their active participation. These strategies may entail sharing the survey post when the audiences are most likely to be active, using eye-catching images, the utilisation of scheduled posts, stories, or direct messages on various social media platforms, and using relevant hashtags to remind both followers and non-followers about the survey and motivate their involvement.

5. Conclusions

While population surveys drawing on probability sampling designs are considered the gold standard for providing robust data on VAWGs, social media recruitment can provide a viable alternative in settings where such surveys are not feasible. This method offers a promising avenue for conducting sensitive research without in-person interactions, particularly in contexts where traditional survey approaches face challenges due to funding constraints, political reasons, or safety concerns for both the participants and researchers.

The findings from the Iranian survey on VAWGs, conducted through an online survey with a sample recruited via a social media platform, were presented as an example. The findings provide valuable insights into the manifestation and consequences of different forms of VAWGs among Iranian women. The use of social media recruitment proves to be an effective approach, reaching a diverse and seldom-heard sample of victims. Understanding the dynamics of violence, the diverse nature and types of violence, potential overlaps, the identities of the perpetrators, and the settings where violence occurs offer crucial insights for targeted prevention initiatives and support services.

While the initial results are promising, we argue that caution is needed in interpreting the results of social media data or generalising from these data to the population of victims at large. Moving forward, the further exploration and refinement of social media recruitment strategies are warranted to maximise their effectiveness and validity in capturing diverse experiences of VAWGs.