A successful assessment answers the referral question. This kind of assessment typically involves a case conceptualization (i.e. clear description of problems). The goal of assessment is not necessarily diagnosis or reducing descriptions of a child’s functioning to a test score.

The model of evidence-based assessment is guided by three principles:

1. Every decision made during assessment should be guided by the most current and best available research.
2. Results from tests should be used only for making interpretations for which they have been validated.
3. The assessment process should be guided by a hypothesis-testing approach (i.e. address the referral question by developing hypotheses based on research).

The use of the evidence-based approach to assessment involves the need to include an assessment of a child’s psychological context. In the meta-systems approach, an understanding of the various systems involved with the child or available to children and families are considered in a case conceptualization and ultimate intervention plans.

The child’s context is important for understanding the child’s adjustment (1) and for understanding the child’s assessment information (2). Testing should be construct-centred, rather than diagnostic centred or test-centred, in evidence-based assessment. Construct-focused testing means that there is a focus on the primary and secondary difficulties and their apparent underpinnings.

Regular assessment of change increases treatment fidelity and improves treatment outcomes. In child mental health settings, regular assessment should be the norm. The criteria by which treatment progress is evaluated should be measurable (1), only measures that are sensitive to change should be used for treatment monitoring (2), the criteria for evaluating treatment outcome should be meaningful (i.e. defined in child’s relative functioning) (3) and the criteria for evaluating treatment outcomes must be feasible (4).

Meaningful outcomes can be established using baseline data on the referral issues.

There are several ethical guidelines for psychological assessment with children:

• The psychologist should have appropriate training for the methods to be used.
• The psychologist should consider the client’s background in interpreting assessment results.
• The psychologist should receive informed consent before initiating the assessment.
• The psychologist should consider to whom assessment feedback should be provided.
• The psychologist should take appropriate steps to maintain the client’s confidentiality.
• The psychologist should obtain releases to provide information from the assessment to outside parties.

Psychological assessment of children necessitate the use of multiple methods of gathering information on the constructs of interest. There is empirical evidence for the initial selection of instruments but not for the integration of the data from multiple instruments. The developmental context of the child should be taken into account when selecting methods. An instrument needs to be able to distinguish between typical and atypical development. Basic psychometric properties also need to be taken into account. The reliability of instruments should hold for the population of interest.

Clinical utility refers to the extent to which a measure will make a meaningful difference in relation to diagnostic accuracy, case formulation considerations and treatment outcomes. Incremental validity refers to the extent to which the addition of a measure provides unique additional information that aids in the assessment process. Assessment techniques vary in their utility for certain interpretations.

Unstructured clinical interviews are unreliable but can still provide the clinician with invaluable information about the client but do not allow for conclusions about the extent to which the child’s difficulties are significant relative to same-aged peers. The interview should be guided by the most recent research. The structured clinical interview is more reliable and often contains content validity. The disadvantages of the structured interview are the administration time (1), the reliance on the informant’s subjective report (2), the lack of direct translation between meeting diagnostic criteria and the most appropriate intervention (3) and the fact that norm-referenced interpretations are not possible (4).

Behavioural observation gives information on possible antecedents of (problematic) behaviour. Behavioural observation can be conducted informally through the observer’s descriptions of the child’s behaviour or formally through structured observational systems. Reactivity is an issue with observation. Older children are more likely to recognize an observer and adjust their behaviour accordingly, making observation better for younger children. Behavioural observations provide direct collection of relevant evidence on the child’s behaviour but miss relevant internal states that play a role in the child’s functioning.

Intelligence tests can provide unique and important information in terms of the level at which the child’s verbal and non-verbal reasoning abilities have developed. However, these tests require specialized training to administer and score and take more time than other assessment techniques. Intelligence tests can be used to assess a variety of disorders.

Behaviour rating scales allow for assessing a large number of constructs and age-based comparisons on constructs of interest. There may be reporting biases in behaviour rating scales. This could be countered by employing validity scales.

Broadband rating scales are rating scales that have a number of subscales assessing different domains of functioning. These rating scales are typically focused on evaluating the presence of problems in adjustment. Child self-report rating scales are closely aligned with parent and teacher counterparts and are derived from adult inventories. Single-domain scales are not as widely used because they are geared toward the evaluation of a specific problem.

Laboratory tasks are designed to elicit performance that will help confirm or disconfirm the presence of a specific problem. Laboratory tasks should not replace other elements of assessment but should be compensatory because the validity can be questioned.

It is essential that an assessment battery includes procedures that provide data from multiple informants who interact with the child in different settings and who may have different perceptions of the child’s adjustment. It is also important to use different methods so that the strengths of one method can compensate for the limitations in another.

A parent is believed to be the most crucial informant for children before adolescence. Parent reports become less useful when the child gets older but the parent can still provide a developmental history. Parent ratings may be influenced by psychopathology (1) and by the view that the parent takes regarding the cause of the child’s difficulties (2).

Many emotional and behavioural problems are most evident and cause the greatest level of impairment at school. This means that obtaining information from teachers is vital in assessment. However, there are limitations to teacher reports. Teachers are less likely to observe anti-social behaviour (1), the age of the child influences the usefulness of teacher reports (2) and older children have multiple teachers (3).

Child self-report can be used as long as social desirability is taken into account. The child’s perception of the construct being evaluated may reveal relevant information about the case. Peer report is rarely used due to limited access, time and ethical constraints. Peers provide a unique perspective on the child’s social functioning and may reveal interpersonal issues that inform intervention efforts.

Institutional records (i.e. school, detention centre) can provide a clear indicator of impairment in the setting from which records are obtained.

The informant’s situation, attributions, type of construct, the child’s age and the child’s gender should be considered when integrating information from different sources. There are several steps for integration of data:

1. The clinician should document all clinically significant findings across constructs and informants.
2. Any areas in which convergence is evident across sources are noted and likely to point to an area of concern.
3. The clinician should try to determine the reasons behind any discrepancies.
4. The clinician should develop a hierarchy of problems from primary to secondary.
5. The clinician should determine the relevant information that should be in the assessment report.