Video observations of sensitive caregiving “off the beaten track”: introduction to the special issue

ABSTRACT

This introduction to the special issue on video observations of sensitive caregiving in different cultural communities provides a general theoretical and methodological framework for the seven empirical studies that are at the heart of this special issue. It highlights the cross-cultural potential of the sensitivity construct, the importance of research on sensitivity “off the beaten track,” the advantages and potential challenges of the use of video in diverse cultural contexts, and the benefits of forming research teams that include local scholars. The paper concludes with an overview of the seven empirical studies of sensitivity in this special issue with video observations from Brazil, Indonesia, Iran, Kenya, Peru, South Africa, and Yemen.

KEYWORDS:

• maternal sensitivity
• cross-cultural
• video observation

One of the biggest current challenges of attachment theory is to test its universality claims outside of the (urban) Western world, not just sporadically, but consistently. The starting point of such an endeavor must be to examine whether or not the theory’s descriptors of both child and caregiver behavioral patterns can be identified across cultures, and whether or not culture-sensitive reformulations of its basic tenets are needed to capture patterns that look different but might serve the same social developmental functions (Mesman, 2018; Mesman et al., 2018). This special issue aims to carry on the seminal observation work of infant caregiving by Mary Ainsworth in Uganda (Ainsworth, 1967), incorporating ethnographic as well as more quantitative approaches to early caregiver–child interactions from an attachment perspective in seven cultural contexts, with a special focus on video observations of sensitive responsiveness.

The construct of sensitive caregiving, or sensitivity – defined as a caregiver’s ability to notice and respond to a child’s signals in a way that fits with the child’s needs (Ainsworth et al., 1974) – is a particularly important one in attachment research. First, it refers to the part of the caregiving process that precedes actual attachment development, and plays a role in predicting individual differences in attachment quality (Lucassen et al., 2011; De Wolff & Van IJzendoorn, 1997). Second, recent ethnographic work suggests that the sensitivity construct allows for culture-specific manifestations of sensitive responsiveness, facilitating its use in different cultural contexts (Mesman et al., 2018). Third, there appear to be variations between communities in overall sensitivity levels that are consistent with theories about the influence of cultural-contextual characteristics such as affluence, and the general availability of social-emotional resources (Mesman, Van IJzendoorn et al., 2016). Fourth, studies of observed sensitive caregiving in Western and non-Western communities have reported significant within-group variations in sensitivity levels, that also relate meaningfully to both caregiver and child characteristics (for an overview, see: Mesman, Van IJzendoorn et al., 2016).

In the study of caregiving practices in general and sensitivity in particular, observations play an important role (Mesman & Emmen, 2013). Indeed, standardized observations are seen as the gold standard in the assessment of parenting (Haws & Dadds, 2006) as they are able to uncover complex and unconscious behavioral patterns that cannot be measured through self-reports. Many aspects of sensitive responsiveness (such as comforting a crying child or returning a child’s smile) are likely to be part of the intuitive automatic parenting system rather than the planned system that reflects conscious choices (Mesman, 2018), and are thus not easily self-monitored. This is especially relevant for sensitive parenting given that noticing and interpretation of a child’s signals are central to its definition. Logically, a parent who shows low levels of sensitivity because of a lack of awareness or understanding of a child’s signals, is also unaware of their inabilities in these areas (i.e., one can hardly be aware of being unaware). This phenomenon is also known as the Dunning-Kruger effect in the social psychology literature, which describes that lack of competence in certain domains often goes together with a lack of metacognitive ability to realize this incompetence, and even inflated self-evaluations about performance in the domain in question (Kruger & Dunning, 1999). This effect is clearly relevant to sensitivity, emphasizing the need for independent evaluations by trained observers for the reliable assessments of sensitive caregiving.

Finally, although many valuable ethnographic studies have been carried out with in-vivo observations (including Mary Ainsworth’s Uganda study), the use of video allows for establishing intercoder reliability, multiple reviews of relevant interactions, more in-depth analyses of multiple behavioral modalities, and discussions between local experts and attachment researchers about the meaning of specific observations. Especially when observing a community that has not yet been studied before in terms of parenting, being able to review the videotaped interactions multiple times enhances the likelihood of uncovering not only the most notable but also more subtle patterns of parenting behavior that might escape the researcher in live observations. In this type of exploratory studies, it also allows the researcher to go back to previous videos when after observing the first set they notice a certain pattern that was not yet represented in the coding system. Adding relevant behavioral categories or contextual variables to the coding system to capture unexpected behavioral patterns is crucial to incorporating culturally specific behaviors in our cross-cultural understanding of sensitivity, Thus, the current special issue specifically focuses on video observations of sensitivity, and provides evaluations of the use of video in different cultural contexts.

The current body of observational research on sensitive caregiving is almost exclusively based on urban Western (middle-class) samples, with some notable exceptions of observational work in other parts of the world, often from a more anthropological angle (e.g., the seminal work “Hunter-gatherer childhoods,” edited by Lamb & Hewlett, 2005), but also from a developmental attachment perspective (e.g., Heng et al., 2018; Jin et al., 2012; Liang et al., 2015; Mesman et al., 2018; Tomlinson et al., 2005). The predominance of Western middle-class samples is endemic in most areas of research in the behavioral sciences (Henrich et al., 2010), but particularly so in subfields using video observations as their major assessment tool. This state of affairs is not surprising, given the many cultural and practical barriers that have to be overcome when conducting a study in which parents and children are videotaped in non-Western communities. Extensive preparation and cultural expertise is needed to access such communities and then establish a working relationship that will allow for data collection in the first place. Further, distrust, fear, and cultural taboos regarding being videotaped are important issues to address and overcome, as well as participants’ interpretation of the “assignment” during videotaping and the extent to which their behaviors can be considered at least somewhat naturalistic under these circumstances. No wonder video-observation studies “off the beaten track” are rare. Nevertheless, it is extremely important that we continue to attempt such studies, and find ways to employ the valuable tool of video observations of parenting in cultural contexts that are sorely underrepresented in the literature to date. This special issue aims to do just that and explicitly evaluates the use of video in a variety of cultures.

The sensitivity construct was developed within the context of attachment theory with its traditional focus on infancy, and as such has been mostly studied in early childhood. Although there is no inherent reason to study sensitivity only in relation to young children – indeed it has been fruitfully studied in samples well into adolescence (e.g., Allen et al., 2003) – there are some reasons to prefer young children for the (cross-cultural) study of sensitivity. First, there is evidence that sensitive caregiving in early childhood has ramifications for child adaptation across developmental stages well into adolescence (e.g., Haltigan et al., 2013), suggesting that for understanding developmental pathways, an early childhood focus is particularly valuable. Second, the study of early childhood has the advantage of minimizing the educational and other institutional influences on the family system. Third, culturally specific socialization goals become increasingly prominent as children get older, and parenting is likely to become more complex because of children’s more advanced skills and needs, as well as increased societal involvement. For these reasons, the studies in this special issue al focus on sensitive caregiving in early childhood.

Summarizing, this special issue has two main aims: (1) to provide insight into the feasibility of videotaping parents and children in different cultural contexts for the study of sensitive caregiving in early childhood, identifying both obstacles and potential ways to overcome these; (2) to enhance our understanding of the occurrence, nature, and role of caregiver sensitive responsiveness to young children in non-Western cultural. Both aims are represented in each of the empirical papers, but have specific features depending on the research context. Each of the seven empirical papers describes experiences with the use of video to assess parent–child interactions in that particular community, as well as substantive insights into sensitive caregiving in a certain cultural context, addressing specific aspects of the sensitivity construct and/or relations with other salient variables. The papers represent samples from impoverished urban areas in Indonesia, Yemen, South Africa, and Brazil, and from rural communities in Iran (also urban), Peru, and Kenya. The unique study characteristics and goals are listed in Table 1, and the shared methodological characteristics of the seven empirical studies are highlighted below.

Table 1. Overview of studies in this special issue

Paper	Infant age in months	Urban/Rural	Sample size	Video per infant^a	Degree of observation structure^b	Main study topic
1. Kenya (Mesman et al.)	7–23	R	7	120	-	Exploring nature of sensitivity in Gusii community
2. Peru (Fourment et al.)	4–21	R	12	180	-	Exploring nature of sensitivity in mothers in rural Peru
3. Brazil (Ribeiro-Accioly et al.)	2	U	22	15	+	Exploring correlates of maternal sensitivity in urban Brazil
4. Iran (Asanjarani et al.)	18–60	U/R	26	30	+	Exploring maternal sensitivity in urban versus rural Iran
5. South Africa (Dawson et al.)	3–15	U	50	20	++	Comparing cultural fit of Ainsworth scale vs MBQS in S-A
6. Yemen (Alsarhi et al.)	24–60	U	54	15	+	Testing correlates of sensitivity in veiled mothers in slums
7. Indonesia (Rahma et al.)	24–60	U	98	15	+	Testing predictors of maternal sensitivity in Indonesian slums

^aduration in minutes

^bcompletely unstructured (-), somewhat structured (+), very structured (++)

The most important methodological issue in a study of sensitivity across cultures is which observation instrument to use. As described by Mesman and Emmen (2013), many instruments include more specific criteria of the type of behaviors and activities that are deemed sensitive (such as smiling, teaching, talking) than the original Sensitivity versus Insensitivity scale developed by Mary Ainsworth. The original scale in fact does not include any descriptions of specific concrete behaviors, and instead speaks of patterns of behaviors that are child-centered and fit with the signals and needs of the child, leaving open the question of how a caregiver might do this. This makes the Ainsworth scale particularly useful for observing sensitivity across different cultural context as it does not presume specific behavioral manifestations of sensitive caregiving, but rather leaves room for variations in how sensitivity is expressed and through which communicative modality (Mesman et al., 2018). As such, the Ainsworth scale can capture a wide range of sensitive responses and does not penalize caregivers for not showing common Western expressions of sensitivity, such as verbal responsiveness and face-to-face positive affect, which are far less common in many non-Western communities.

In addition, the scale does not favor one particular interaction activity over another, because the nature of the scale implies that any activity could be done sensitively or insensitively (except of course extremes such as physical abuse). Whereas play is a highly common interactional feature of parent–child interactions in Western cultures (and almost always has a teaching element), this is highly uncommon in many others where play takes place only in the realm of children amongst themselves, not one that features adults (as indeed it was in most Western countries two or three generations ago as well!). Shared time between parents and children in many places revolves around household or livelihood-related activities such as washing, cooking, or tending to livestock. Such “chores” can in essence have the same relational meaning as playing a game in that they both represent what is deemed important and adaptive in a particular cultural context, and provides an opportunity for social learning. Whether such moments of interaction between a parent and a child are conducted in a way that is mindful of the child’s needs and wishes is the key question that is reflected in the sensitivity construct, especially as measured with the Ainsworth scale that does not specify activities or contexts.

Finally, the Ainsworth sensitivity scale has traditionally only been used to assess caregiver behavior towards infants in their first year of life. However, the same characteristics that make these scales so well suited to use across cultures, also make it particularly useful for application to older age groups. The basic behavioral descriptions in the scale refer to the general notion of adapting one’s behavior to a child’s signals without necessarily referring specifically to caregiving behaviors that are limited to infancy. Some descriptions in the scales are particular to infants, but they are all supplementary and function more as examples than as the main coding anchors. Simply taking these out and replacing “B” (for Baby in the original scales) with “child” immediately makes clear that the scale can easily be used for sensitivity towards older children as well.

Methodology

It is important to note that the empirical studies described in the separate papers of this special issue were designed more or less independently from each other and thus differ in their specific sample characteristics (apart from geography), observation procedures, and coding methods. The choices made for every individual study were informed by considerations such as local community practices and care systems, programmatic focus of particular research groups, and/or practical issues related to resource access. Nevertheless, the seven studies do share several methodological aspects that represent a common underlying approach to the cross-cultural study of sensitive caregiving.

Procedures

Working with local partners is crucial to research in different cultural communities for many reasons that have also been emphasized by Mary Ainsworth with regard to her work in Uganda (Ainsworth, 1967). The empirical studies in this special issue were conducted in teams with researchers from each of the seven countries – and whenever possible even the specific regions – represented in this special issue. Importantly, the local partners were more than assistants, as they were trained scientific partners who were involved in all research phases, from study design to data collection and the writing process. Some as PhD students with scholarships temporarily based in a Western country, but collecting data in their country of origin, and others as researchers living and working locally and collaborating internationally. There are several important advantages of working closely with local researchers. First, there are clear practical advantages to conducting sample recruitment and data collection with people who know the local culture and language. Second, the interpretation of video materials and results is greatly enhanced and deepened by the continuous involvement of local scholars who can provide explanations and information that would be very difficult to generate with Western scholars only. Third, actively involving local scholars as integral members of the research team contributes to local scientific knowledge and skills development and allows for active participation in the international scientific discourse, in scholarly contexts where access to expert training, state-of-the-art research methods, and international publication opportunities is often limited. Fourth, local researchers are more likely to have the necessary networks and knowledge to effectively disseminate research results to local professionals and communities who might benefit from the findings. Although not its primary purpose, this special issue does hope to illustrate the benefits of this approach.

All of the empirical studies in this special issue focus on early childhood, representing children aged 0–6 years old. Four studies focus on infancy with samples up to age 24 months (Brazil, Kenya, Peru, South Africa), and three studies include children from toddlerhood until age 6 years (Indonesia, Iran, Yemen). Sample sizes vary widely, which also reflects differences in study goals and scope that are discussed in more detail in the individual papers, but are summarized in Table 1. The studies described in the empirical papers employ different observational procedures, ranging from relatively short observations (i.e., 15 minutes per family) to quite extensive ones (up to 3 hours per family). Most studies represent relatively naturalistic observations in the sense that caregivers were free to what they did during the videotaping, although some situational constraints were in place in certain studies. This varied from no constraints at all (i.e., following the child around regardless of where they were and who was with them), to some structure (e.g., caregiver and focus child are asked to stay indoors), to high structure (caregiver was given specific tasks). Such procedural issues are discussed in more detail in each of the papers.

Observational methods

All studies in this special issue are tied together by the use of two specific observational coding systems, namely the Ainsworth Sensitivity versus Insensitivity scale (Ainsworth et al., 1974). In addition, most studies also included observations of caregiver warmth, physical contact, verbal expression, and a set of scales designed to observe participants’ camera awareness to assess the possible effects of videotaping on participant behavior. These scales are all described in detail below, and somewhat more briefly in the empirical papers, where also study-specific issues of training and reliability are discussed. If additional coding methods or other measures were used, these are described in the empirical papers.

Warmth

Caregivers’ warmth was coded as expressed physically (hugs, kisses, caresses, gentle holding), verbally (terms of endearment, praising, expressing love and affection), or with facial expressions (smiling). The smiling had to be directed at/shared with the child; general smiling or smiling at others than the focus child was not scored. It is important to note here that observing smiling of veiled women could be unseen but in some cases it was with difficulty spotted from the mothers’ eyes movement and general interaction with child. Warmth was coded on 5-point scale ranging from 0 = no warmth, to 4 = very high warmth (warmth is shown throughout the video and almost the entire interaction is characterized by this warmth).

Physical contact

Caregivers’ physical contact with the child included touching and holding the child, regardless of the quality of the physical contact. Hair brushing and washing the child were also scored as physical contact. The scale included the following scores: 0 = low physical contact (just a few times and almost all of those are brief), 1 = medium physical contact (up to half of the duration of the video), 2 = high physical contact (more than half the duration of the video).

Verbal expression

Caregivers’ verbal expression included talking and whispering that was aimed at the child. The expressions were coded on 3-point scale: 0 = low verbal expression (a few words here and there, or only whispering), 1 = medium verbal expression (regular talking but also silent episodes), 2 = high verbal expression (talking almost throughout the video, with few and generally only very short breaks).

Camera awareness

This scale consists of three subscales reflecting the extent to which caregivers were observed to be clearly aware of the camera and being filmed. These scales were newly developed to deepen our understanding of the impact of using video on the behaviors of caregivers in different cultural contexts. The sum of the three subscales described below form a total score reflecting overall camera awareness.

(1) Looking at the camera. The number of times that a caregiver was looking at the camera during the video observation was counted in the first 15 minutes of observation (which in several studies constitutes the entire observation time). Looking at the camera included behaviors such as mother looking deliberately at the camera as if she was checking something, seeking approval, or wondering what to do. Fleeting glances that were almost unavoidable simply because the camera was in the mother’s line of vision were not counted. Looking at the camera was coded according to a 3 – point scale; 0 = never or rarely (once or twice, briefly), 1 = several times, (3–5 times, mostly briefly), 2 = Many times (more than 5 times). (2) Talking about being filmed. Talking about being filmed includes caregivers mentioning to the child or others in the room that the child is being filmed, talking to the camera person about being filmed. In this scale we also scored instances where a caregiver talks about how long the filming is taking, trying to get the child to do things explicitly for the camera, or worrying about what people (who see the video) will think if the child behaves a certain way. The videos were coded according to 3-point scale (i.e., 0 = never, 1 = sometimes (once or twice), 2 = several times (more than twice). (3) Expressing insecurity. Insecurity expressed by the caregiver included explicitly saying to the researcher that she did not know what to do, asking what to do next, or asking whether a certain activity was right or not while being filmed. The expressions were coded in 3-point scale (i.e., 0 = never, 1 = sometimes (once or twice), 2 = several times (more than twice).

Outline of the special issue

This special issue is structured to reflect a progression from small-scale qualitative studies aimed at describing the nature of sensitivity in specific cultural contexts, to more quantitative studies examining patterns of sensitivity in relation to locally salient social-contextual variables (see also Table 1). The first paper is a small-scale pilot study with observations of sensitivity among seven Gusii families in Kenya (Mesman, Basweti, & Misati, this issue), who represent a particularly interesting ethnic group in this regard, as they have previously been described as showing little sensitivity in infant care. This study is explorative and descriptive. The second paper also describes a small-scale pilot study focusing on 12 families in two indigenous communities in rural Peru (Fourment, Nóblega, Conde, Nuñez del Prado & Mesman, this issue) who have to our knowledge never been observed from an attachment perspective. Salient themes discussed in relation to sensitivity are networks of caregivers, maternal multitasking, and flexible caregiving routines. This study is explorative and descriptive. The third paper explores patterns of socio-contextual factors and maternal sensitivity in a small at-risk sample of 22 mothers in urban Brazil (Ribeiro-Accioly, Seidl-de-Moura, Mendes, & Mesman, this issue). The fourth paper also describes a small-scale exploratory study among 26 families on maternal sensitivity in rural and urban Iran, focusing on behavioral correlates of sensitivity to elucidate differences in modalities of expressing sensitivity across these different regions (Asanjarani, Abadi, Ghomi, & Mesman, this issue).

The fifth paper examines the nature of maternal sensitivity in a sample of 50 mothers from a South African township, comparing Ainsworth sensitivity observations to ratings using the Maternal Behavior Q-sort, and analyzing differences in definitions between the two measures to elucidate their relative cultural-sensitivity (Dawson, Bain, & Mesman, this issue). The sixth paper describes a unique study among 62 mothers in the slums of Yemen, investigating the particularly interesting issue of using video observations to assess sensitivity in mothers who are fully veiled, addressing the validity of such observations in relation to social-contextual variables (Alsarhi, Rahma, Prevoo, Alink, & Mesman, this issue). The seventh paper investigates maternal sensitivity in a sample of 98 mothers in an urban slum in Indonesia, in relation to mothers’ own childhood experiences of maltreatment, and current sociodemographic risk factors (Rahma, Alsarhi, Prevoo, Alink, & Mesman, this issue). Finally, a closing discussion paper will reflect on the conclusions regarding the main aims of this special issue, lessons learned, and remaining challenges in video observation research on culture and sensitivity.

Disclosure statement

No potential conflict of interest was reported by the author(s).