The overall aim of the research presented in this thesis was to examine the notification decision, the decision made by child protection workers that reported concerns relating to a child warranted a child protection investigation. This decision is critical in the process of child protection as it determines the entry of a child into the child protection process. The current state of knowledge concerning what constitutes child maltreatment, what causes people to maltreat their children and what the consequences of this maltreatment are to the children provides no definitive answers for child protection workers concerning what information is important when making decisions. Child protection workers are asked to make decisions within this uncertain state of affairs. These issues are discussed in Chapter 2. In Chapter 3 the decision making literature is reviewed in an attempt to find some guidance about decision making for the child protection workers from the current understanding of decision making. Two fundamental theoretical approaches were identified in this literature. The first approach was predominantly concerned with mathematically modelling the relationships between the information used in making decisions and the outcome decisions. The second theoretical approach to decision making concentrated on the process of decision making, the cognitive processes that occur between the presentation of the information and the decision outcome. The research presented in this thesis approached the aims from both these theoretical approaches.
The study presented in Chapter 4 addresses the first of four principal research questions: Is there information available at the point of notification of a child protection case that can be used to predict the eventual outcome of the investigation? The information used for this study was sampled from the information held on the Central Registry data base. The statistical relationships between information collected by the intake officer when making the notification decision and the outcome of the investigation were investigated. The sample contained 1000 investigations. Fifteen variables were analysed including information about the child, the notifier, the type of maltreatment notified and the caregiver. The individual relationships between each variable and the outcome were examined first and then the multivariate relationships among all variables and outcome were examined. The results indicated that the information held on the Central Registry and used in tins study could be used to statistically predict outcome but not with enough reliability to be of any practical use. Complex interactions among the variables were found. This finding indicated that the use of simple check lists for assessing risk is inappropriate. The implications of this study for the development of statistically descriptive studies is examined.
The study presented in Chapter 5 addresses the second principal research question: What information do child protection workers employ when making the decision that a case warrants a statutory response? Child care officers responsible for the intake of cases by the Department of Family Services and Aboriginal and Islander Affairs (DFS) were interviewed about cases that had recently come to their attention. From the interviews, a content analysis identified 32 categories of information the child care officers considered important when making the decision that a case warranted a child protection investigation. When the use of these categories of information by the intake officers was examined across the cases it was apparent that the use of this information was extremely inconsistent. To examine factors that impact on these inconsistencies a final study was performed and is presented in Chapter 6.
Two research questions What factors impact on the consistency of outcome decisions made by child protection workers? and Does experience with the decision, level of difficulty of the case, and time pressure impact on the collection and integration of information when making the decision that a case warrants a statutory response? are addressed in Chapter 6. A process tracing methodology employing a computerised information board format was developed for this study. Thirty-one research participants were presented with case studies developed from the actual cases identified in the previous study and coded into 40 information cues also developed from the information categories identified in the previous two studies. Based on these information cues the research participants were asked to make the decision whether the case study warranted a child protection investigation (CP1) or not (an Intake). By using already identified case studies and information cues this study overcame one of the major criticisms levelled at previous process tracing studies, namely, that the decision alternatives and attributes are determined by the researchers and therefore have no ecological validity.
The decision processes that the research participants utilised in making these decisions were analysed for evidence on compensatory and non-compensatory strategy use. These decision making processes included the depth of search (how many cues the research participants selected), the cue latency (how long research participants spent examining the cues), the certainty rating (if the subject became less certain about their decision with the addition of new information), the importance of the individual information cues (frequency of selection of the cues) and sequence of selection of the cues (order of the selection of cues).
To examine the contingent nature of the strategy selection in decision making three factors were investigated in the experimental procedure. The three factors investigated were (a) time pressure, (b) prior experience of the decision maker and (c) difficulty of the decision task.
The data obtained were first examined with respect to the research question What factors impact on the consistency of outcome decisions made by child protection workers? For some case studies the decisions made by the research participants were highly consistent and for some the decisions were inconsistent. It was shown that the level of experience of the subject, the time condition in which the subject made a decision and the type of decision made (CP1 or Intake) had no impact on the level of consistency of the decision making. Consistency of decision making appeared to be related to the level of difficulty of the case study.
The data were examined to address the research question Does prior experience with the decision, level of difficulty of the case and time pressure impact on the collection and integration of information when making the decision that a case warrants a statutory response? A core set of five cues was identified. Generally, these five cues were the first cues selected and were selected by all research participants regardless of the time conditions or the level of difficulty of the case. The selection of additional cues appeared contingent on the three factors investigated, prior experience with the decision, level of difficulty, and time available to make the decision. In the unrestricted time condition the more experienced research participants provided strong evidence of contingent strategy selection. In the high difficulty case studies these research participants searched for more information and spend less time examining this information than in the low difficulty case studies.
In Chapter 7 an overview of the results are provided. The implications of the research findings to current decision making research and child protection practice are discussed.