Current Child Welfare Systems Are Not Designed to Offer the Best Outcomes for Children and Families

Probably everyone reading this would agree that child welfare agencies should be designed to protect children and build strong, resilient families. However, current Western child welfare systems are inherently designed to produce accountability and defensible risk assessments rather than safeguarding or permanency for children. This structural reality cannot simply be laid at the feet of workers’ capabilities or intentions. In today’s political climate, child welfare organizations tend to do what they are designed to do: provide oversight and scrutiny. Child welfare practice is under intense public scrutiny, and when outcomes are negative and children are hurt, fault-finding is the norm. There is nothing politicians fear more than news headlines about child deaths. Organizations operating under these conditions have been nudged into defensive practices rather than learning and prevention. Organizations seek to avoid blame, appear in control, and reduce uncertainty (Hood, 1991; Lewis, 2015). These features promote the appearance of responsible and businesslike organizations and are thus favoured over features that enable them to manage uncertainty, encourage professional discretion, and focus on learning and improvement (Munro, 2011). Current child welfare systems have learned how to complete paperwork more effectively than to ensure services keep children safeguarded. Today’s child welfare systems are perfectly designed to deliver the results they are designed to deliver (Deming, 2000).

Today’s Western Child Welfare Systems Are Designed for Defensibility Rather Than Learning

Compliance-dominant child welfare systems are not merely producing procedural rigidity; they operate within a design logic that prioritizes organizational survivability over adaptive performance. Argyris’s (1977) distinction between single-loop and double-loop learning is not merely descriptive here but diagnostic of system architecture. Single-loop learning prevails because the system’s governing variables, its implicit goals, are not child safety and well-being reliability but error avoidance, reputational protection, and procedural defensibility. When error avoidance is the criterion, rule adherence becomes rational, even if it undermines the promoted goals of family resilience and children’s safeguarding. Following serious incidents, the predictable organizational response is to introduce additional documentation requirements, procedural safeguards, layers of oversight, and training mandates (Munro, 2011). From a systems perspective, these actions function as risk-absorption mechanisms rather than learning mechanisms, but they do not enhance the organization’s ability to sense indeterminate risk, tolerate ambiguity, or manage complex relational interventions. The mechanism is causal: if performance is evaluated based on audit visibility, actors invest effort in creating visible compliance rather than operational reliability.

This dynamic illustrates goal displacement, where the means of demonstrating control become the operational end (Merton, 1957). Documentation ceases to be a tool for reflection and becomes evidence of organizational fit within the prevailing mindset. The system thereby shifts from learning-oriented adaptation to symbolic assurance production, consistent with institutional theory’s argument that formal structures often function to maintain legitimacy rather than improve technical performance (Meyer & Rowan, 1977). Procedural intensification following failure demonstrates sensitivity to external pressures but decreases internal capacity.

Double-loop learning would involve challenging fundamental presuppositions, such as definitions of risk, distribution of authority, decision-making under conditions of ambiguity, and definitions of success (Argyris, 1977). However, double-loop learning threatens system stability by calling into question whether existing structures may be contributing to the problem. In high-blame environments, this is politically and psychologically unsafe. Edmondson (2018) demonstrates that learning systems depend on psychological safety to surface error and uncertainty. In child welfare, the consequences of visible error, legal, professional, and reputational, create powerful disincentives to disclose system weaknesses. Thus, the system’s error management regime becomes one of error suppression rather than error learning.

The result is a set of reinforcing feedback loops where increased scrutiny leads to tighter procedures, tighter procedures increase administrative load, administrative load reduces time for reflective practice, reduced reflection diminishes early detection of drift, drift accumulates until a crisis occurs, and crisis triggers further procedural tightening. This is a classic example of a fix that fails in systems thinking (Senge, 1990). The short-term fix, more control, exacerbates the long-term vulnerability by weakening the system’s adaptive capacity. These dynamics produce what can be described as organizational self-protection, where information that challenges the adequacy of system design becomes threatening because it implies systemic rather than individual accountability. Workers soften records to protect themselves, managers align reporting with inspection expectations, and near misses go unreported. Learning signals are filtered out, creating an illusion of stability. The system becomes less sensitive to early warning signs while appearing more controlled. In the context of high-reliability theory, the organization shifts from a preoccupation with failure to one with defensibility (Weick & Sutcliffe, 2007). Institutional pressures further reinforce this design. Regulatory, political, and media environments reward visible compliance and punish visible uncertainty (DiMaggio & Powell, 1983). Organizations therefore converge on similar procedural models, not because these are most effective, but because they are most legitimate. Unfortunately, immense anxiety-driven pressures drive homogenization around defensible practice, crowding out experimentation and contextual adaptation.

The unintended consequence is that systems learn to optimize for the defence of actions taken rather than the prevention of risks. Staff learn to showcase their practice in ways that highlight alignment with statutory responsibilities and policy rather than in ways that construct conditions that make harm less likely to occur. Adaptive capacity atrophies over time. The agency can become procedurally sound while operationally fragile. What looks like bureaucratic overhead is actually an intelligent adaptation to the distortion of governing variables. You cannot have a learning system without changing what drives performance, how systems are held accountable, and the conditions of psychological safety. If systems are not redesigned, calls for creativity will be co-opted by the conformity feedback loop and disabled. There is no resistance to learning and improving; it is simply safer and easier not to.

Administrative Expansion and the Displacement of Professional Work

Administrative growth in child welfare should not be interpreted as operational inefficiency or poor managerial discipline. It is better understood as a structural adaptation to institutional risk and accountability pressures. As documentation becomes the primary evidence of work, it assumes the status of a proxy for performance. Work that is not recordable becomes organizationally invisible, while recordable activity becomes equated with professional contribution. This shifts the system’s governing variable from the impact on children’s safeguarding to the documentation of compliance activity.

Administrative intensification reflects what institutional theorists describe as the expansion of ceremonial structures, formalized procedures that demonstrate rationality and control to external audiences (Meyer & Rowan, 1977). These structures signal accountability and legitimacy but are often loosely coupled from technical performance. In child welfare, recording systems, approval processes, and procedural milestones function as legitimacy artifacts. They reassure regulators and the public that the organization is operating in a disciplined manner, regardless of whether those routines enhance the reliability of safeguarding. This produces a classic case of goal displacement (Merton, 1957). Documentation, timelines, and procedural milestones, originally intended to support practice, become the operative definition of practice. The means of evidencing work become the purpose of the work. As performance evaluation is limited to what can be audited, practitioners rationally allocate effort to system navigation rather than relational intervention. The system’s design thus reconfigures professional labour from engagement work to defensibility work.

From a systems perspective, this shift also reflects a substitution dynamic. Because relational safeguarding is uncertain, contextual, and hard to prescribe, documentation is predictable, measurable, and auditable. Under high-stakes accountability pressure, organizations replace the uncertain with the controllable. This does not happen because professionals are lazy or uncaring; it happens because they are sensible human beings responding to the incentives by which they will be evaluated (Lewis, 2015). The perverse outcome is that the hardest, riskiest, and most important work of safeguarding, building trust, mobilizing networks, negotiating risk, and facilitating behavioural change is systematically under-resourced. Research has also shown that frontline child welfare practitioners spend substantial portions of their time on administrative and recording tasks, reinforcing how organizational systems displace direct relational work (Holmes et al., 2013). Munro (2011) suggested that excessive bureaucracy corrodes professional judgment by limiting discretion and overwhelming cognitive capacity. Here, both cognitive and structural processes are at play. Cognitively, when submerged in administrative tasks, professionals have less attention to absorb complex information, less time for reflective deliberation, and begin to think like rule-appliers rather than critical thinkers. Structurally, practitioners become processors of risk, converting real family life into tick boxes compatible with system categories rather than facilitators of safeguarding.

This creates reinforcing loops whereby the demand for documentation crowds out time for relational work. Less time for relational work increases anxiety about how well families are coping and the need to know, leading to more boxes to check and closer supervision, which leaves even less time for meaningful interaction with families and children. This is the archetype of shifting the burden (Senge, 1990). Systems trade short-term controls for long-term capabilities. The more risk management paperwork that is piled on, the more administratively stable the system feels, but the less capable it becomes of preventing errors. As paperwork expands, professional identity changes. If the ability to do well is measured by how well one works within the system, competence becomes about doing paperwork instead of helping families. After a while, conversations about high-quality documentation replace learning conversations about how to keep children safeguarded. Professionals become what the incentives drive them to be.

Performance Regimes and Organizational Distortion

Child welfare performance measurement regimes should be conceptualized not only as tools of accountability but also as technologies of governing that establish certain ways of realizing, problematizing, and imagining success within organizations. Managerial performance measurement systems introduced through New Public Management reforms reconceptualized the delivery of public services in terms of outputs that could be compared and controlled (Hood, 1991). The purpose of these reforms was to help agencies become more transparent and better able to address performance, but in child welfare, such initiatives create a structural tension between the espoused goals, prevention and best outcomes for children and families, and the actual systemic conditions.

Lewis (2015) argues that performance systems do not simply reflect work; they reorganize it. Metrics function as attention-directing devices, signalling what matters and, by omission, what does not. When indicators stand in for outcomes, organizations orient behaviour toward indicator optimization. In child welfare, measures such as assessment completion times, visit frequency, plan documentation quality, and case closure rates serve as surrogate endpoints for safeguarding. Yet these are process indicators, not direct measures of children’s lived safeguarding or network reliability. This produces what sociologists describe as reactivity to measurement (Espeland & Sauder, 2007). Once measured, behaviours change to improve measured performance, regardless of substantive impact. The causal mechanism is incentive alignment. Workers and managers respond to what is visible, reviewable, and career relevant. Effort is reallocated toward metric compliance, even when this displaces more meaningful but less measurable work. Over time, indicators undergo goal substitution (Merton, 1957). The metric ceases to be a proxy and becomes the target itself. Timely completion of assessments becomes synonymous with safeguarding diligence, and completeness of documentation becomes equated with professional rigour. As Power (1997) observed in audit cultures, organizations increasingly focus on the production of process rather than the outcome of the process. The system’s operational logic shifts from prioritizing safety and well-being to prioritizing compliance.

Murphy’s (2022) assertions demonstrate how this plays out in practice. When job security depends on producing a recordable audit trail, it makes sense for practitioners to focus on record-keeping rather than on messy relational work. This is not a failure of ethics but an adaptation to the accountability environment. Performance regimes thus act as behavioural regulators, narrowing the scope of professional discretion. Inspection systems intensify these effects by reinforcing external audience orientation. Success becomes defined through audit outcomes, ratings, and regulatory approval. Institutional theory predicts that under such scrutiny, organizations will prioritize legitimacy-preserving activities (Meyer & Rowan, 1977). As scrutiny increases, ceremonial compliance structures expand. These structures are highly visible but loosely coupled to operational effectiveness. Systems thinking reveals the reinforcing feedback loop. Performance indicators emphasize procedural milestones, procedural milestones drive administrative focus, administrative focus reduces capacity for reflective and relational work, reduced capacity for reflective and relational work weakens safeguarding reliability, incidents occur, and incidents lead to tighter performance controls. This is another fix that fails dynamically (Senge, 1990). The short-term fix, more measurement, erodes long-term resilience.

Another consequence is the systemic focus on risk aversion. Professionals working in high-public-performance regimes tend to make decisions based on what is least likely to be criticized rather than on what may be best suited to particular circumstances. In other words, decisions are based on policy rather than context. This echoes findings from research showing that audit and inspection cultures promote risk-averse behaviour and punish innovation (Power, 1997). Applied to child welfare practice, this can look like conducting interventions defensively so that one can check the box but not build family capacity. This stands in contrast to high-reliability theory. In high-reliability systems, indicators are built to illuminate weak signals and variation in work, not audit theatre (Weick & Sutcliffe, 2007). Performance information is used in learning loops rather than reputation loops. But in child welfare, measurement systems operate as control architectures. Information travels up to scrutinize and blame, not across organizations for learning and improvement. When performance becomes the purpose, a distorted set of behaviours follows. If all that matters is what is measured, then safeguarding quality comes to mean something achievable within patterns of administrative consistency. The system will grow more skilled at producing safe cases for audit and less able to facilitate safeguarding for children and families. Until performance systems are redesigned to emphasize outcome reliability, attend to relational indicators, and measure network effectiveness, organizations will continue to reward activities that feel successful on paper.

Families Experience Surveillance Rather Than Safeguarding

Safeguarding practice is relational and requires collaboration, trust, transparency, and the co-production of responsibility for children’s welfare. As child welfare systems become more centrally organized around defensibility, auditability, and the minimization of organizational risk, these relational dynamics are reshaped. Contact is reframed through surveillance and documentation, checking whether people have followed plans, corroborating evidence where possible, or searching for indicators that increase risk. Families feel monitored rather than supported. Safeguarding becomes surveillance. This is not accidental. When systems are oriented toward limiting organizational risk, the best way to do so is to extract information that can be traced and audited. Asking questions about people’s histories, writing things down when disclosures are made, checking whether people have done what they say they will, and taking photographs of indicators of risk, all of these actions make sense from an accountability perspective. They do not feel like safeguarding work for families. Families who have contact with child protection experience it as intrusive, controlling, and fear-inducing (Merritt, 2021). This is not because professionals intend it, or because families are simply reacting emotionally to others’ initiatives. From a systems perspective, people are less likely to trust someone whom they feel is watching them. They may hide aspects of their lives that they are struggling with from someone they do not trust or do not care for. If they hide things from the professionals working with them, those professionals are less able to accurately assess the risks posed to children. Uncertainty in assessment leads professionals to feel they need to increase surveillance. More checking means less trust. As soon as someone behaves defensively, the system views them through a risk lens.

When compliance behaviours are used as indicators of safety and well-being, they become unreliable. Families may attend appointments, present well, or verbally agree with plans while privately disengaging. The system confuses observable compliance with underlying capacity. Procedural adherence becomes equated with safeguarding, even though safeguarding reliability depends on sustained behavioural change and network engagement beyond professional oversight. Institutional theory helps explain why systems drift toward surveillance logics. Under conditions of high scrutiny, organizations prioritize activities that demonstrate vigilance and control (Meyer & Rowan, 1977). Monitoring becomes a visible assurance mechanism. Yet this visibility comes at relational cost. Families adapt to the accountability structure by managing impressions rather than sharing vulnerabilities. The result is information distortion, in which the system’s knowledge of family functioning becomes less accurate as oversight increases.

Evidence-based literature on relational safety shows that lasting change does not occur unless family members and professionals alike experience psychological safety. Families will withhold risk-relevant information when they feel judged, fear the loss of their children, or worry about the loss of reputation. Edmondson’s (2018) work on psychological safety is relevant at the relational level. When families interacting with professionals lack conditions that foster open dialogue about mistakes and failures, learning is not possible for either the family or the system. This unintentionally causes systems to reach short-term goals of behavioural compliance at the expense of long-term safeguarding reliability. Families act differently when being watched than they do when they are not. If family members do not internalize safeguarding plans and informal networks do not reinforce them, safety and well-being will always rely on professional surveillance. As soon as that surveillance is reduced, risks will return. Therefore, when system design is oriented around monitoring families rather than partnering with them, it changes the goal of professional-family interaction from building capacity to verifying risk. Practitioners become watchers rather than cultivators of long-lasting safeguarding systems. Defensible information becomes more important to the system than relational partnership. If systems were designed to achieve outcomes, they would look very different at the interaction stage. The focus would be on collaboratively understanding risk, developing shared accountability, and connecting family members to their support networks. Monitoring would be used to learn about what works, not to control it. Without restructuring systems to do this, child welfare will continue to create relationships that do not allow for the openness and collaboration needed for true safeguarding.

Organizational Change Perspective: Why Reform Efforts Stall

Viewed through the lens of organizational change theory, recurring setbacks are not failures of implementation but failures of change design. Meaningful change interventions require alignment across strategy, paradigms, structure, systems, culture, and leadership actions (Burke & Litwin, 1992; Covey, 1989). Child welfare reform efforts tend to target the surface level of practice models, tools, and training while maintaining the underlying performance structure: accountability processes, reporting requirements, inspection guidelines, budget policies, and risk management protocols. These structures condition what is rewarded, so they hold more behavioural influence over how people work than any new practice model will (Burke, 2018). When the underlying conditions of work remain the same, the proposed change is absorbed into the existing system. The shiny new intervention lands in the organization, gets processed through existing bureaucratic logic, and is translated into something that makes sense within prevailing paradigms. The new tool or practice model becomes just another hoop to jump through. Reflective supervision becomes another piece of paperwork. A networking initiative becomes another box to check. This is not resistance because the intervention is not being pushed out. It is being assimilated. Through the lens of systems thinking, this is homeostasis. The system weakens the agitation and returns to the status quo (Senge, 1990). The status quo has been built to defend itself, not to learn.

Kotter (2012) points out that change efforts flounder when organizations focus on urgency as a means of avoiding crisis rather than on creating a sense of purpose around a powerful performance vision. In child welfare, energy for reform tends to coalesce around tragedy, inspection failure, or political scandal. The unstated objective becomes reputation management rather than operational improvement. Staff perceive that they must change to avoid becoming the next headline story. Creating urgency in this way reduces cognitive capacity and limits experimentation. When threatened, humans tend to act out of fear, make poor decisions, work in silos, standardize practice, and reduce discretion. This is the antithesis of adaptive change. Most change efforts also ignore the incentive structure. What is rewarded, measured, and sanctioned does not change if promotion, performance scores, and organizational credibility remain tied to compliance with audit criteria. Staff will understandably focus on those processes despite protestations to the contrary. Getting structure right is often more influential than leadership espousing the new change effort (Burke, 2018). If incentives do not change, change will remain superficial.

Psychological safety is another missing precondition. Edmondson (2018) demonstrates that learning cultures depend on environments where individuals can surface uncertainty, question assumptions, and report errors without fear. In child welfare’s high-blame environment, reform occurs alongside heightened scrutiny. Staff perceive reform as an additional risk rather than a learning opportunity. This produces defensive compliance, visible adherence to reform processes without substantive engagement. The result is reform layering. Each cycle of change adds procedures, tools, and reporting requirements without removing older structures. Complexity accumulates. Administrative burden grows. Cognitive overload increases. The system becomes less able, not more able, to learn. This is an example of what Senge (1990) calls the fix-that-fails archetype. In trying to fix things, organizations implement short-term solutions and new programs that compound the long-term problem by creating more structural inertia. Institutional theory allows this argument to be pushed further. Organizations under the spotlight feel pressure to appear legitimate to external stakeholders (Meyer & Rowan, 1977). Reform then takes on symbolic importance; organizations must appear responsive and current. They adopt new models because they represent change, not necessarily because they alter performance architecture. The longer this continues, the better organizations become at pretending to reform.

Leadership Misalignment

Professional practice expertise often frames collective understanding of leadership within child welfare systems. Leaders often rise through practice pathways and maintain deep passion and investment in children and families, but they are seldom trained or mentored in organizational design, performance architecture, or systems governance. The challenge is not one of values or intentions but of definitions. If leadership is viewed through the lens of practice, leading becomes the act of senior-level practice supervision rather than the design and governance of the environment in which practice takes place. Systemically speaking, this distinction matters greatly. Variables embedded in the architecture of complex organizations ultimately determine outcomes far more than individual effort. How work is designed, how information is shared, how decisions are enabled, how performance is measured, and how learning is embedded all govern outcomes far more than individual performance, focused on case quality, policy compliance, and professional supervision (Senge, 1990). To change how people think and act requires changing the system they work within. Doing things within the current system maintains the status quo and does not change the variables that govern behaviour (Argyris, 1977).

High-performing systems evaluate leadership differently by holding leaders responsible and accountable for clarifying purpose, aligning measurement with purpose, stress-testing system capabilities, and closing the loop so operational learning feeds into system redesign (Deming, 2000). Variation in performance is understood as a signal of the system’s state, not as proof that someone did something wrong. Within child welfare, however, performance problems tend to be reframed as issues of adherence, proficiency, or effort at the worker level. Leaders get pulled into thinking about performance in terms of individual compliance rather than system improvement. This pressure to perform also comes from political forces. In times of high public visibility, leaders are judged on their capacity to show that systems are under control, manage reputation, and provide assurances to regulators (Meyer & Rowan, 1977). There are minimal incentives to lead in a generative way, as less attention is paid to the hard work of developing system capacity when most available bandwidth is spent managing risk optics. Over time, this cycle reinforces itself. As visibility increases, leaders centralize decision-making authority and dictate processes to reduce variation. Standardization decreases variation but also reduces local decision-making authority and innovation. As adaptability to local complexity decreases, systems become more brittle and more likely to fail. When they fail, organizations centralize further and become more managerial in their approach, but less leadership-oriented from a high-performing systems perspective.

Heifetz’s (1994) framework on technical versus adaptive leadership helps clarify this dilemma. Technical problems are solved with current procedures and expertise. Adaptive problems require questioning basic assumptions, shifting authority and responsibility, and changing core values. Child welfare leadership is pulled toward technical solutions, new procedures, guidance, and oversight because they are safer and more easily understood by external stakeholders. Asking leaders to do adaptive work, such as redesigning how accountability is understood or what successful outcomes actually look like, is politically dangerous. Adaptive leadership is easily derailed. The unintentional result is that leadership amplifies the very system dynamics it is trying to control. When leaders focus on compliance assurance, they create a performance architecture that displaces learning even further. When they focus on individual performer levels, they make structural drivers invisible. Leadership becomes custodial rather than generative over time, settling for system stability rather than building system capacity to improve outcomes. This is not a problem of individual leaders. It is produced by how the leadership role is constructed, how leaders are selected, how they are evaluated, and how they are supported. Leaders are not rewarded for, and often are not even trained to, design learning systems, align metrics with outcomes, and scale cultural practices that promote psychological safety. Until leadership is redefined as system stewardship, reforms will continue to focus on practice while leaving untouched the upstream structures that determine it.

Synthesis: System Design, Not Practitioner Failure

The preceding analysis demonstrates that the persistent underperformance of child welfare systems is not attributable to insufficient commitment, training, or professional intent. It is the predictable outcome of a system whose governing variables prioritize defensibility, legitimacy, and reputational risk management over adaptive learning and safeguarding reliability. Each structural element examined, learning suppression, administrative expansion, performance measurement distortion, surveillance-oriented relational climates, reform absorption, leadership role misalignment, and institutional self-preservation, functions not as an isolated weakness but as part of an interlocking system.

Systems thinking clarifies that these dynamics reinforce one another. Performance regimes direct attention to measurable procedural outputs, thereby incentivizing administrative expansion. Administrative load reduces time for reflective and relational practice, weakening early detection of safeguarding drift. Reduced detection increases uncertainty, prompting tighter procedures and oversight. Increased oversight intensifies surveillance climates, eroding trust with families and further degrading information quality. Poor information reduces the system’s learning capacity, leading to reform initiatives that add procedural layers without addressing root causes. These reforms are absorbed into existing accountability structures, increasing complexity and further privileging compliance. Leadership operating under institutional scrutiny focuses on risk optics and assurance rather than structural redesign. Institutional pressures reward visible control, reinforcing legitimacy-preserving routines. The cycle stabilizes around bureaucratic survivability rather than outcome reliability.

Transforming how child welfare systems work with children and families is not a technical challenge; it is an adaptive one. Technical challenges have technical solutions. As long as systems operate from the same theories of change about control, accountability, and risk, they will continue to do more of the same, react when the system fails, add more rules and procedures, and return to stability. Child welfare systems are not failing children and families; they are working exactly how they are designed to work.

References

Argyris, C. (1977). Double-loop learning in organizations. Harvard Business Review, 55(5), 115–125.

Burke, W. W. (2018). Organization change: Theory and practice (5th ed.). Sage.

Burke, W. W., & Litwin, G. H. (1992). A causal model of organizational performance and change. Journal of Management, 18(3), 523–545.

Covey, S. R. (1989). The 7 habits of highly effective people: Powerful lessons in personal change. Simon & Schuster.

Deming, W. E. (2000). Out of the crisis. MIT Press.

DiMaggio, P. J., & Powell, W. W. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 48(2), 147–160.

Edmondson, A. C. (2018). The fearless organization: Creating psychological safety in the workplace for learning, innovation, and growth. Wiley.

Espeland, W. N., & Sauder, M. (2007). Rankings and reactivity: How public measures recreate social worlds. American Journal of Sociology, 113(1), 1–40.

Heifetz, R. A. (1994). Leadership without easy answers. Harvard University Press.

Holmes, L., McDermid, S., Jones, A., Ward, H., & Munro, E. (2013). How social workers spend their time in frontline children’s social care in England. Journal of Children’s Services, 8(2), 123–133.

Hood, C. (1991). A public management for all seasons? Public Administration, 69(1), 3–19.

Kotter, J. P. (2012). Leading change. Harvard Business Review Press.

Lewis, J. M. (2015). The politics and consequences of performance measurement. Policy and Society, 34(1), 1–12.

Merritt, D. H. (2021). How do families experience and interact with child protective services? Children and Youth Services Review, 121, 105858.

Merton, R. K. (1957). Social theory and social structure. Free Press.

Meyer, J. W., & Rowan, B. (1977). Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83(2), 340–363.

Munro, E. (2011). The Munro review of child protection: Final report—A child-centred system. UK Department for Education.

Murphy, C. (2022). If it’s not on the system, then it hasn’t been done: Ofsted anxiety and the social organization of case recording in child protection. Child & Family Social Work, 27(3), 563–572.

Power, M. (1997). The audit society: Rituals of verification. Oxford University Press.

Senge, P. M. (1990). The fifth discipline: The art and practice of the learning organization. Doubleday.

Weick, K. E., & Sutcliffe, K. M. (2007). Managing the unexpected: Resilient performance in an age of uncertainty (2nd ed.). Jossey-Bass.