Introduction

According to the World Health Organization (2014), international estimates on the occurrence and prevalence of child maltreatment in a family context vary, among other factors, according to the definitions of abuse and neglect employed, which play a central role in decision-making on referrals and the remaining assessment process (Arruabarrena and De Paúl 2012; Rodrigues et al. 2015). For this reason, in recent decades, a number of different studies have been done on the definition of maltreatment (e.g., Calheiros 2006; English et al. 2005), with its type (i.e., classification into types and subtypes) and severity being the most commonly studied aspects (Herrenkohl 2005; Litrownik et al. 2005). In general, these studies confirm the lack of social consensus over what forms of parenting are dangerous or unacceptable (Cicchetti and Manly 2001) and which inappropriate parenting behaviours should be considered maltreatment (Wolfe and McIssac 2011). Indeed, although a consensus already exists with regard to the multifaceted definition of maltreatment—physical abuse, sexual abuse, neglect, emotional/psychological abuse—the differentiation between poor parenting and maltreatment within the parental behavior continuum is still a key issue for definition, identification and assessment (Wolfe and McIssac 2011).

There are also differences in the specificity and degrees of severity given to the various subtypes across different samples of professionals and laypeople (Giovannoni and Becerra 1979; Peterson et al. 1993; Portwood 1999; Runyan et al. 2005; Korbin et al. 2000), underscoring the need for operating definitions integrating the different social conceptions of the problem (National Research Council 1993; Schmid and Benbenishty 2011). This need is particularly important, since laypeople and community professionals are among the primary agents in identifying and referring situations of risk/hazard (e.g., school; police; health or social services, etc.) (CNPCJR 2013; USDHHS 2013). However, with a few exceptions (e.g., Simarra et al. 2002), the search for integration in common-sense and technical definitions has been overlooked in empirical research.

In fact, according to the American agency Children’s Bureau, in 2012 (USDHHS 2013), more than half of the referrals were made by community professionals (58.7 %, e.g., educators; authority figures; healthcare workers) and the remainder by unclassified (23.3 %, e.g., anonymous reports) and non-professional sources (18 %, e.g., family members; neighbours), with this referral pattern remaining consistent in the prior 4 years.

In European countries (e.g., Portugal; Spain; United Kingdom), the pattern is similar (CNPCJR 2013; Gilbert et al. 2009). Furthermore, since child maltreatment is a public crime in many European countries and American states (i.e., not dependent on the submission of a complaint by the victim, and able to be submitted by anyone, with police entities and public workers obliged to report cases of which they become aware while performing their duties), the reporting systems have been streamlined (e.g., online) to facilitate and encourage community involvement in its detection.

Some authors question the feasibility and effectiveness of the legal obligation for the community to report cases of suspected child maltreatment (Melton 2005), bearing in mind, among other aspects, the negative effects of often unsubstantiated over-reporting to child protection services. Along these lines, others say that, if the community did not play a proactive role, many children would continue to suffer indefinitely without intervention (Mathews and Bross 2008), arguing that over-reporting and under-reporting are two realities that must not be disassociated. If, after investigation, many cases are proven to be unfounded, the circumstances of many children never become known to child protection services due to biased interpretations and assessments (Besharov 2005). As such, a number of studies have shown that the lack of knowledge and ability to recognize cases of maltreatment has, among other aspects, been one of the main barriers to its referral, thus pointing to the need for operating definitions of maltreatment and objective guiding criteria as one of the possible responses to this problem (Alvarez et al. 2005; Gilbert et al. 2009; King and Scott 2014; Pietrantonio et al. 2013).

Some studies show that assessing the severity of abusive practices is among the key variables in recognizing these cases (Egu and Weiss 2003) and in decision-making on the case’s eligibility for technical monitoring (Arruabarrena and De Paúl 2012; Molina 2010); as such, the lack of consensus on levels of severity has also been cited among the major problems (Gambrill 2008; Munro 2005). However, according to what we know and with few exceptions (e.g., Smith 2006), there is a lack of studies analysing the assessment of severity in abusive practices at the community level.

Finally, another underlying challenge in the process of defining maltreatment revolves around the cultural and geographic variability in parenting practices and child upbringing (e.g., Fallon et al. 2010). In fact, although the National Research Council pointed in 1993 towards the need for studies in this regard (Barnett et al. 1993; Litrownik et al. 2005), the most relevant research has been done in the United States and Canada (e.g., Herrenkohl 2005), and there are very few studies in Europe differentiating and describing levels of maltreatment severity (e.g., Arruabarrena and De Paúl 2012). In this context, the adoption of definitions from different socio-cultural contexts may result in judgments and interpretations of maltreatment cases that are out of line with their socio-cultural reality.

To minimize these problems, in the present studies, we analysed the conceptions of laypeople and community professionals to seek an operating definition of maltreatment which integrates them, and which distinguishes between various types of abusive practices. We also analysed the severity allocated to the various contents of each subtype to obtain indicators for distinguishing between different degrees of severity. Two studies were carried out for this purpose. In study 1 (qualitative), we sought to define maltreatment in terms of types by jointly analysing the conceptions of laypeople (by analysing interviews) and community professionals (by analysing statistical summary reports). In study 2, a questionnaire was used to assess the allocation of severity to the contents from Study 1, bearing in mind the various descriptors of each subtype of maltreatment, through a quantitative study with interns in the area of social sciences and health, i.e., future community professionals.

Study 1

Method

Participants

We interviewed 123 participants, mostly female (62.6 %) aged 18–68 (28.5 % 25 and under; 35.2 % aged 26–35; 17 % aged 36–45 and 19.3 % 46 and over). Less than half (32.5 %) of the participants had completed higher education (29.3 % secondary education and 38.2 % basic education). With regard to professional status, based on Portuguese Classification of Occupations (Instituto Nacional de Estatística, 2010), 25.2 % belonged to middle or higher-level staff (e.g., teachers, technicians of electronica), 22 % worked in services (e.g., administrative staff); 9.8 % were specialized workers (e.g., hairdressers, mechanics); 8.1 % were non-specialized workers (e.g., cleaning services, kitchen assistants) and 32.5 % were not actively employed (e.g., students, retired, unemployed). Thirty-nine percent had professional experience with children, but none of the participants were involved in youth and child protection services or had professional contact with child maltreatment.

Procedure

Participants were recruited through convenience and snowball sampling from workplaces and professional training services not related to children and youth protection. Although it was a convenience sample we recruited participants in places where it was possible to have the highest diversity levels regarding age, education and socio-economic status. Prior to the interview, participants were informed that the objective of the study was to collect their opinions about the meaning of parental maltreatment. It was highlighted that there were no right or wrong answers and that we were interested in the opinions of participants. In order to allow the content analysis, individual interviews, lasting an average of 10 min, were recorded in audio format and subsequently transcribed to text. Confidentiality and anonymity were guaranteed for the data gathered, and informed consent was obtained for participation and recording. Given the sensitivity of the subject and the possibility of people having experienced abuse themselves, in the case participants were distressed by the emotional or social content of the interviews there was a set of measures to respond to any disclosures of abuse. The interviews were conducted by two experienced professionals in the child protection system and family violence (i. e, one clinic psychologist and one social worker) at the participants’ workplace or professional training services, in Portugal.

With regard to gathering statistical summary reports, a collection of institutions was chosen according to whether statistical summary reports on the referral of children with signs of abuse existed within their departments. Access and authorization for consulting the reports were obtained through institutional directors, while likewise ensuring the confidentiality and anonymity of the data obtained.

Measures

With regard to the collection of information with laypeople, semi-structured interviews were conducted with a script including direct questions on socio-demographic status (e.g., age, sex, academic background and profession, contact with child maltreatment) and open-ended questions on the definition of abuse and neglect in the parent–child relationship/education (e.g., “What do you consider to be an abuse in the parent–child relationship/education?”; “What do you consider to be a neglect in the parent–child relationship/education?”).

With regard to the corpus of analysis for a technical definition, nine annual reports of first-rate community services were analysed, six from hospital institutions and three from community welfare services working with families. The statistical summary reports, describing detailed indicators of maltreatment (e.g., percentage of burns, bruising, malnutrition, abandonment, verbal violence) show the collective situations of maltreatment referred by these institutions to the competent authorities, and were drawn up by social workers (i.e., psychology, social service and sociology) and healthcare workers (i.e., medicine, nursing and speech therapy), and were based on the case records of 516 children being monitored at these institutions (two institutions monitor children aged 0–4; four institutions receive children aged 0–11; and the remaining institutions monitor children aged 0–17).

Data Analyses

To create a categorical conceptual scheme of maltreatment, the corpus of analysis, comprising material obtained from the interviews and described in the statistical summary reports, underwent a consensual qualitative research method (Hill et al. 1997). This consisted of a thematic content analysis (Braun and Clarke 2006), using a bottom-up procedure, with categories and subcategories based on the data semantic content, i.e., in reference and relevant to a single theme. With this criterion, the “keyness” of a certain category or subcategory was not dependent on its frequency, but on whether it captured something important in relation to the definition of maltreatment. Also prevalence was counted at the data level (i.e., a content can appear anywhere in each individual interview or statistical report) and not in terms of the number of different participants/reports who referred that item. Therefore, the set of record units (words or phrases) was organised by the research team into categories (types) and subcategories (subtypes) according to their semantic meaning and a coding system was developed. Through this process 1235 record units were obtained, 1065 from the interviews, and 170 from the statistical summary reports.

Next, to evaluate the categorization system’s reliability through inter-rater agreement, around one-fourth of the record units (randomly chosen) were categorized by four independent judges (psychologist, teacher, physician and social worker) with professional experience in the child protection system, using the parameters established in a dictionary created by the researchers for this purpose as a reference. The coding system had good inter-rater agreement indices (Cohen’s kappa = .81, p < .001).

Finally, given the nature of the corpus of analysis (material obtained from 123 interviews and 9 statistical summary reports) we used quotes to illustrate how each source contributed to this definition issues, and we reported the relevance of the record units within categories.

Results

Definition of types and subtypes of Abuse, Neglect and Sexual Abuse. The 1235 record units obtained were categorized into 6 types and 20 subtypes of abuse—physical abuse (14.9 %; two subtypes); psychological abuse (29.9 %; six subtypes); educational maltreatment (7.4 %; two subtypes); neglect—lack of physical provision (28.7 %; six subtypes); neglect—lack of supervision (16.1 %; four subtypes); and sexual abuse (2.9 %)—bearing in mind parental omissions and behaviours, together with the consequences for the child (see Table 1).

Table 1 Categorization system for maltreatment (N = 1235)
Full size table

Physical Abuse

This type of abuse refers to the use of violence and physical aggression, and includes two subtypes. The subtype aggressive physical interaction (78.3 %) includes violent physical acts by parents as coercive/punitive methods of upbringing (e.g., “beating the child to educate him/her”, “spanking, hitting”), as well as observable physical wounds on the child (e.g., “belt marks”, “bruises”, “fractures”). In turn, the subtype physical violence methods (21.7 %) refers to how the abuse was perpetrated (“violently shaking the child”, “slaps”, “putting in boiling water”). Note that the content of both subtypes was cited in both the interviews (i.e., laypeople) and the statistical summary reports (Table 2), although issues involving serious consequences for the child such as “burnt child,” “bruises” “trauma”, “injury”, “fractures”, “retina bleeding” and “perforation of the tympanic” were mostly cited in the statistical summary reports.

Table 2 Categorization system for maltreatment by laypeople and professionals
Full size table

Psychological Abuse

This type includes six subtypes, and revolves around parent actions/omissions that may affect the child’s emotional needs and harm his/her psychological development. The subtype conflictual family environment (8.9 %) refers to the acts of parents prohibiting the child’s relationship with other family members (e.g., “the parents do not get along with the grandparents, and do not let them see their grandchildren”) and the child’s exposure to a disorganized and violent family environment (e.g., “he/she witnesses domestic violence”). The subtype unresponsive attachment Figs. (22.5 %) relates to parents’ actions showing disinterest and a lack of attention to the child’s emotional needs (e.g., “do not stimulate”, “lack of contact”), as well as emotional rejection and unpredictability (e.g., “inconsistent and disconnected reactions”, “emotional rejection of the child”). The subtype aggressive verbal interaction (20.3 %) refers to verbal repression and aggression through insults and threats (e.g., “constant yelling without reason”, “belittling”, “they do not let them speak”). The subtype age inappropriate autonomy (20.1 %) relates to parent expectations that are out of line with the child’s responsibilities (e.g., “they do not acknowledge that they are children”), and encouraging the performance of tasks beyond their developmental phase (e.g., “forcing minors to perform tasks unsuited to their age”, “not allowing them to play”). All of the above subtypes were described in the interviews as well as in the statistical summary reports (see Table 2). The subtype coercive discipline methods (20.3 %) refers to the use of intimidating (e.g., “creating situations of fear”) and restrictive disciplinary techniques (e.g., “depriving the child of freedom by locking him/her in rooms or other locations”), and was cited by both sources, although much more in the interviews. The subtype harsh evaluation patterns (7.9 %) describes both the parents’ disinterest in the child’s performance (e.g., “they are not concerned about academic performance”), as well as strict and critical assessments in this regard (e.g., “they are never satisfied with what the child does”, “they humiliate the children”), as well as blaming the child for family problems (e.g., “they accuse the child of their divorce”) and was less cited by both sources.

Note that the content of all subtypes was similar in both the interviews (i.e., laypeople) and the statistical summary reports.

Educational Maltreatment

This type includes two subtypes, and describes parents’ actions that may affect the development of children’s citizenship and academic education. The subtype fostering child deviant behaviours (55.4 %) includes parent actions promoting children’s exposure to and involvement in illegal and inappropriate activities (e.g., “taking drugs in front of them”, “begging”, “child labour”), and exposure to and reinforcement of deviant models (e.g., “inciting them to violence”, “accompanying marginal groups”). All the contents were cited in the interviews and in statistical summary reports, although issues involving alcohol and drug consumption were cited only in the statistical summary reports (e.g., intoxication due to children’s consumption of substances was only referred to in the reports). Finally, the subtype lack of school monitoring (44.6 %) describes parent actions showing disinterest for the child’s academic involvement and direction (e.g., “they do not control schedules”, “they do not keep pace with the child’s education”), together with those promoting absence and dropping out from school (e.g., “they do not take the child to school”), and were cited by both sources.

Neglect—Lack of Physical Provision

This type of maltreatment describes shortcomings in basic care involving the child’s physical needs, together with the respective damages observed. This type of maltreatment is divided into six subtypes, according to lacking type of care: inadequate hygiene (15.5 %) (e.g., “do not bathe”, “the child has parasites”, “skin diseases caused by dirtiness”), inadequate clothing (8.5 %) (e.g., “dirty clothes”, “oversized or undersized clothing”, “clothing inappropriate for the time of year”); inadequate housing conditions (16.6 %) (e.g., “the child lacks an appropriate place to sleep”, “the living conditions are so bad that the child has frequent respiratory infections”); lack of physical health monitoring (30.1 %) (“no health surveillance”, “lack of routine doctor appointments”, “inappropriate medications”); lack of mental health monitoring (13.2 %) (e.g., “failure to help them when they have some sort of difficulty”, “do not take them to services that may help their poor learning and developmental conditions”); and inadequate feeding (16.1 %) (e.g., “incomplete meals”, “the child is hungry, and the parents do not provide food”, “poor nutrition”, “failure to provide food to the point that the child becomes sick”). Generally speaking, the content of all subtypes was cited in the interviews as well as in the statistical summary reports, although more frequently in the latter (with the exception of mental health monitoring), which mentioned a collection of specific issues with regard to children’s physical health (Table 2). The content cited exclusively in the statistical summary reports, among other things, included: skin lesions due to a lack of hygiene; lack of routine doctor appointments; growth deficiencies; food poisoning and malnutrition due to an inadequate diet.

Neglect—Lack of Supervision

This type of maltreatment includes four subtypes where parent omissions jeopardize the child’s safety, given his/her specific developmental needs. The subtype unattended developmental needs (16.1 %) refers to a lack of appropriate supervisory measures, particularly in view of the child’s development phase and behavioural profile (e.g., “they leave the children with siblings who do not know how to take care of them”). The subtype lack of supervision (37.7 %) considers a situation where children are left without reliable adult supervision (e.g., “the children don’t go to school, and stay alone at home”, “they are out in the street”). Insecurity in the environment (16.1 %) refers to a lack of safety assessment where the children spend prolonged periods of time with potential immediate physical hazards (e.g., “leaving drugs or other harmful products in sight”, “playing in a hazardous area”). Finally, the subtype inadequate supplementary supervision (30.2 %) includes situations with a lack of appropriate care for children, by alternative caregivers, while the parents are absent or physically or mentally impaired. Generally speaking, the content of all of the subtypes was cited in both the interviews and statistical summary reports, although with less relevance of lack of supervision and inadequate supplementary supervision in the latter. With regard to the subtype insecurity in the environment, the irreparable consequences of serious accidents were cited exclusively in the statistical summary reports.

Sexual Abuse

This type of abuse (2.9 %) has no subtypes, but does include any sexual attempt and/or contact with children for the purposes of sexual gratification (e.g., “they exploit the child with pleasure”) or economic advantage (e.g., “they put the child up for prostitution”, “they use the child for pornographic purposes”), with or without physical or psychological coercion (e.g., “rape”, “incest”), and exposure to pornographic material or acts (e.g., “abnormal sexual practices”), cited both in the interviews and the statistical summary reports.

Discussion

In general, the definition obtained includes the different types and subtypes of maltreatment referred to in the literature, pointing towards a multifaceted understanding of the constructs, and adapting to the structure suggested by other studies and classification systems (e.g., Barnett et al. 1993; English et al. 2005; Fallon et al. 2010). Furthermore, it includes content related to parent behaviour (i.e., acts and omissions), observed damages (defined primarily by health professionals), and potential danger to the child, similar to other studies (e.g. Barnett et al. 1993; Herrenkohl 2005).

A little bit surprising was the categorization of “fostering child deviant behaviours” and “lack of school monitoring” in the same category. However, the content analysis that made up the subcategory of “lack of school monitoring” indicated that most quotes (21/36) are parental acts related to child education and school attendance, that foster child’s deviant behaviour, such as “school dropout”, “parents’ lack of interest for what children do”, “parents do not send child to school”, “they do not put the child in school”. Another aspect that may have been important in this categorization was the fact that school dropout is an act of parental responsibility that is directly punishable by law in Portugal (unlike other neglect or mistreatment acts).

Along these lines, despite the existing consensus in defining subtypes, this study found a distinct but supplementary contribution in the nature of the content and degree of specificity of the information furnished by each of the sources (i.e., professionals and common sense). In this regard, the main differences are in educational maltreatment and neglect from the standpoint of provision and supervision, where the statistical summary reports cite more aspects related to the acts’ consequences for the child (e.g., serious accidents, namely irreparable consequences of the lack of safety) and specific issues on the child’s and family’s physical health (e.g., alcohol and drug consumption; skin lesions due to a lack of hygiene; lack of routine medical visits; and deficient growth, food poisoning and malnutrition) compared to laypeople. In relation to the above aspects, the results thus seem to show also that the content cited describes different levels of severity within each subtype.

Study 2

Method

Participants

The participants were 159 interns in the areas of Education (50.3 %), Psychology (30.2 %) and Health – medicine and nursing—(19.5 %), the majority female (80.5 %), aged 22–56 (M = 25.22; SD = 6.65). With regard to contact with situations of abuse, 30.2 % of the respondents had previous professional contact with cases in this area, 20.1 % said they had knowledge of close situations and 8.2 % cited personal experience with situations of maltreatment.

Procedure

Participants were recruited through convenience sampling from social and health care institutions related to children and youth protection. The interns were chosen because they had a recent formation in this area, they were being trained in specialized institutions and they would be the future community professionals. Data were collected at Portuguese public institutions in the areas of Medicine, Nursing, Psychology and Education. Before filling out the questionnaires, it was explained to the participants that the objective of the study was to classify different descriptors of maltreatment according to their perceived degree of severity. The questionnaires were answered in person and in group, guaranteeing the confidentiality and anonymity of the data. As in study 1, given the sensitivity of the subject and the possibility of people having experienced abuse themselves, in the case participants were distressed by the emotional or social content of the questionnaire there was a set of measures to respond to any disclosures of abuse.

Measures

To create a scale of severity for abuse based on the record units obtained in Study 1, we followed a top-down procedure, using the proposal of Barnett and collaborators (1993, Maltreatment Classification System—MCS) as a reference. In this system most items are operationally defined by five different levels of severity for each subtype of maltreatment (ranging from inadequate parental act/omission to potential damage, and “observable” consequences of abusive behaviours in children). This scale was translated and adapted based on a discussion panel comprising the principal researcher and four technicians from the Commissions for the Protection of Children and Young People (social worker, attorney, physician and teacher). Therefore, 242 units of analysis obtained in Study 1 (corresponding to around one-fourth of the record units, and distributed over the previously identified types and subtypes of abuse), were categorized by these technicians on a five-level scale (1–5) of increasing severity. The record units obtained in the material under analysis, but not appearing in the categorization system, were categorized by the judges based on their semantic meaning.

The results showed that the majority of subtypes gathered from the material in Study 1 did not present indicators corresponding to the five degrees of severity proposed by the American version (Barnett et al. 1993). In fact, in the categorization process, we were only able to identify a correspondence between the five levels proposed by Barnett and collaborators and the indicators of severity obtained in the subtypes aggressive physical interaction and inadequate feeding. Three levels of severity were identified in subtypes: physical violence methods; unresponsive attachment figures; aggressive verbal interaction; lack of school monitoring; inadequate hygiene; inadequate clothing, inadequate housing conditions; lack of physical health monitoring and lack of mental health monitoring. Four levels of severity were identified in the subtypes: age inappropriate autonomy; coercive discipline methods; harsh evaluation patterns; fostering child deviant behaviours; insecurity in the environment; sexual abuse. Finally, only two levels of severity were identified in the subtypes conflictual family environment and lack of supervision, and just one level in the subtypes unattended developmental needs and inadequate supplementary supervision. We also found that in the majority of the subtypes, the distribution of record units was concentrated in the lower levels of severity (1 and 2).

Given that the correspondence between the five levels proposed in the Maltreatment Classification System (MCS) only occurred in two of the defined subtypes, in building a scale of severity, four levels of severity were defined (i.e., simple phrases describing the characteristics of each degree of severity). As such, in the subtypes where the record units did not describe content related to four of the five levels of severity proposed by Barnett et al. (1993), MCS indicators were used; in the subtypes where four levels of severity were found, the content was maintained, and in the subtypes where the content analysis resulted in five levels, we chose to combine two of the extreme levels of the MCS.

In this manner, the scale of severity built from the material gathered in Study 1, supplemented with the descriptors of Barnett et al. (1993), differentiated four levels of severity per subtype of maltreatment (example of descriptors of the subtype aggressive physical interaction: (1) They hit the child without touching the neck or head, and without leaving marks, or only leaving small marks; (2) They leave several marks or a highly visible mark on the child’s body, without touching the neck or head; (3) They cause small burns, scratches or minor cuts to the body, or leave marks on the head, face or neck; (4) They inflict wounds causing hospital treatment or hospitalization). Similar to Barnett et al. and taking into account the nature of each maltreatment subtype, we intended to create a continuum of severity, whose main criterion was the intensity of the act/omission, which ranged from parental risky acts/omission with potential damage and the consequences for the child.

The four-levels scales, grouped according to the corresponding subtype, were presented randomly to the participants, who were asked to classify them according to their perceived degree of severity on a scale of 1–4 (1 − less serious to 4 − the most serious).

Results

We used Kendall’s coefficient of concordance to analyse the consensus between participants in assessing the four levels of severity presented per each subtype of abuse, on the whole and in paired groups (Table 3).

Table 3 Description and ranking of descriptors of severity, W values and means
Full size table

When considering the assessment of the four levels of severity as a whole, most subtypes of abuse have acceptable and good significance values (W between .33 and .92), indicating that participants ranked them in a rather consensual manner. Assessment means ranged approximately from 1 to 4 in all of the subtypes, except in the subtype age inappropriate autonomy (psychological abuse), where the mean varies between 2.38 and 2.69, with a non-significant W value (W = .01; χ2 = 5.19; df = 3; p > .05), showing a lack of consensus between participants. Note that the levels of severity assessed with a lesser degree of consensus involved the subtypes unresponsive attachment figures (psychological abuse) (W = .33), aggressive verbal interaction (psychological abuse) (W = .40), and inadequate hygiene (neglect—lack of physical provision) (W = .44), as opposed to sexual abuse (W = .92).

When considering the assessment of the different levels of severity in paired groups (levels 1 and 2; levels 2 and 3; levels 3 and 4), the analysis revealed that nine subtypes were not evaluated in a consensual manner. Between levels of severity 2 and 3, there were consensus problems in the subtypes insecurity in the environment (neglect—lack of supervision) (W = .022; χ2 = 3.45; df = 1; p > .05); inadequate hygiene (neglect—lack of physical provision) (W = .009; χ2 = 1.45; df = 1; p > .05); inadequate feeding (neglect—lack of physical provision) (W = .017; χ2 = 2.59; df = 1; p > .05); unattended developmental needs (neglect—lack of supervision) (W = .005; χ2 = .78; df = 1; p > .05) and physical violence methods (physical abuse) (W = .034; χ2 = 5.02; df = 1; p > .05). In turn, between levels of severity 1 and 2, there were problems in the subtypes lack of physical health monitoring (neglect—lack of physical provision) (W = .000; χ2 = .006; df = 1; p > .05) and aggressive verbal interaction (psychological abuse) (W = .000; χ2 = .000; df = 1; p > .05). Finally, between levels of severity 3 and 4, there were agreement problems in the subtypes unattended developmental needs (neglect—lack of supervision) (W = .007; χ2 = 1.09; df = 1; p > .05) and inadequate housing conditions (neglect—lack of physical provision) (W = .015; χ2 = 2.32; df = 1; p > .05).

Discussion

The results showed that, in the public and technical opinions, a consensual evaluation of severity in situations without signs of immediate, clear and observable damages to the child (e.g., age inappropriate autonomy) was more difficult, as well as when involving parental domains with less discussion in the public spectrum or in dimensions more recently acknowledged as abusive, either academically or socially (e.g., neglect).

Indeed, the fact that the dimensions of physical abuse and sexual abuse portray parental acts whose consequences to the child are more evident, and which enjoy greater public prevalence (i.e., frequent media dissemination of sexual abuse cases), may contribute to increased public awareness of these situations and, as a result, a greater ease in identifying, recognizing and differentiating their severity by the community. Furthermore, psychological abuse and neglect are less consensual areas, suggesting that they may be subject to less community awareness (e.g., Korbin et al. 2000). In fact, bearing out the results of other studies (e.g., Peterson et al. 1993; Portwood 1999), the perceptions of the severity of neglectful practices in supervising children gather less consensus among the participants; as such, it should be noted that identifying inadequate supervision is complex, bearing in mind the difficulty of assessing parent omissions, together with a lack of clear standards for leaving children unsupervised (Peterson et al. 1993). In general, there are no clear, agreed upon standards to differentiate between acceptable parental practices and those that cross the line into child maltreatment (Cicchetti and Manly 2001). This situation has been further complicated regarding acceptable versus maltreating parenting in cases of neglect or psychological abuse (Barnett et al. 1993).

General Discussion

The literature has underscored the need for conceptual schemas structured over the maltreatment of children that streamline the recognition and referral of these cases, since laypeople and community professionals, as those making the referrals, may have biased interpretations of these situations, leading to the under-reporting or over-reporting of cases (Mathews and Bross 2008). The decision to report a case of parental maltreatment has been characterized as complex, ambiguous and full of errors and uncertainty. That is even more the case for instances of parental neglect in which, although the long-term effects may be detrimental (DePanfilis 2006), the physical proofs are hard to obtain (Dickens 2007; Rodrigues et al. 2015). Understanding the decision of reporting neglect cases is particularly pertinent in Portugal, where the concept is absent in the law and institutionally undervalued in comparison with other forms of maltreatment like physical or sexual abuse (Torres et al. 2008).

The results obtained in these two studies highlight the importance of cultural values and social contexts (i.e., professional versus community) in understanding the phenomenon and its conceptualizations regarding child maltreatment (Barnett et al. 1993; Calheiros 2013; Knutson 1995), not only in terms of category content, but also in describing the severity of its different indicators.

The present results show that, although the subtypes are highly similar to those which had been defined in the analysis of the records of American technicians, the content of the majority of the subtypes in study 1 do not have the same degree of specificity, namely psychological abuse and lack of supervision (in which some subtypes included only two or three descriptors). In fact, except for the area of physical abuse, which is described more specifically when compared with the content proposed by Barnett and collaborators (1993)—the reason for including a new subtype in our version (subtype of physical violence methods)—the majority of the subtypes do not include its descriptive specificity. Also, it can be concluded that participants assessed the increased severity of abusive practices with little consensus in nearly half of the subtypes, with a less consensual evaluation in relation to a subtype of psychological abuse. Finally, we concluded that the main discrepancies are between middle levels of severity (i.e., 2 and 3), especially in the subtypes of maltreatment related to neglect, namely lack of physical provision and lack of supervision.

Along these lines, an understanding of community standards is essential in optimizing social intervention policies. One of the most important stages of social intervention, on a par with prevention and intervention, is avoiding the often late detection of situations of children at risk, already under circumstances of serious neglect and abuse. Therefore, clear definitions of abuse and neglect, considering the continuum of inadequate parent practices, enable decision-making on the need for intervention without having to be directly based on the extreme severity of maltreatment episodes.

The observed variability in how primary referral agents define which parent behaviours are abusive, and which constitute more serious practices, underscores the importance of undertaking strategies encouraging social awareness on the characteristics of this phenomenon with a view to avoiding biased interpretations of situations and minimizing the problems of over-reporting, under-reporting and unsubstantiation and, consequently, promoting more effective intervention for protecting children and young people.

A continuation of this work will allow a definition of referral parameters and the scheduling of preventive interventions in situations of risk in Portugal, as well as also allowing the decision-making process on the referral of maltreated children to be based on a clearer and more objective assessment than that which is currently being done.

The next phase of this research will be to make the definitions of child maltreatment obtained in the present studies applicable to the community area by laypeople and professionals. In addition, as the definition framework suggested by this research includes the perceptions of professionals and laypeople, those definitions must be validated over time, since views change and new information emerges.

Some limitations may be cited in relation to these studies. First, on studies 1 and 2 we used a convenience sample. Second, the questioning of the subjects on the ranking of severity was done in relation to the indicators of each subtype, and not in relation to the different subtypes of abuse and neglect. Finally, in both studies, children’s age as an indicator of their development has not been included. In proposals for future work, it thus seems essential to pursue research incorporating in the sample different groups of professionals and considerations on the children’s age in the definitions and allocation of severity, so as to define what constitutes maltreatment, taking developmental stages of children into account. Other limitation is the lack of information about participants’ parenting experience (Portwood 1999). Thus in future studies it should be analysed if the fact of being a parent have influence in the maltreatment types and severity definition.

In addition, although we consider the role of cultural context and community values in defining abuse and neglect especially important, we must not overlook the existence of communities that may display abusive behaviours while not constituting a problem in some specific sociocultural context. In such circumstances, the subjective views of certain groups or community standards and beliefs seem largely invalid as defining criteria. This is yet another reason, along with understanding social norms, for using scientific knowledge on which conditions or circumstances put children at risk and promoting a two-way street in a social construct for the problem: from common sense to scientific and vice versa.