Objective Analysis - Principles and General Procedures


It may be worth starting the objective analysis be first classifying objectives according to their various different characteristics. Objectives may be different if they are quantitative, binary-event, or qualitative. A further classification could be to measure them according to the attainment of the objectives.

A quantitative objective may be measured either by deterministic or probabilistic methods. A deterministic measurement is made when a definite attainment of an objective is determined from numerical data (i.e. to build 150 low income houses, to increase the schools’s teacher-student ratio to 40 per 1000).

A probabilistic measurement is made when collected data are insufficient to determine with certainty that an objective has been attained. This is the case when data is collected on only a sample of the target population (i.e. to produce the product with no more than 1% defects, to persuade at least 90% of drivers to wear seatbelts).

A binary-event objective clearly occurs or does not occur. Logical measurement is used as a basis for determining whether a binary-event objective has occurred (i.e. to acquire a new subsidiary, to complete the construction of the hospital).

Qualitative objectives are those judged subjectively to decide if they have been attained. Axiological measurement or measurement that is judgemental yet more of less evident maybe accomplished through interviews (i.e. to improve the appearance of a product, to improve the health of the target groups).


The major assumption underlying the objective tree is the hierarchical relation between objectives. The objectives for a project may be uncertain or ambiguous because they have not been articulated by the interested parties and because goals are not constant over time. The objective tree technique assumes that the objectives higher in the tree are less variable over time and that they are shared by a larger number of interest groups. The assumption that qualitative objectives can be sub-divided into quantifiable sub-objectives is implicit in the technique. Its validity does not affect the use of the objective tree to explicitly reveal goals and ends, whether they are reasonable or not.

During the formulation stage of an objective tree it is important that the participants do not strive for perfection. Since the work is usually undertaken as a group exercise initially a lot of effort is needed merely to get the process started. The interaction between the objectives often only becomes apparent after the initial framework of the tree has been developed. As the tree begins to evolve during the planning process, the participants begin to consider more carefully and they also refine the tree step-by-step. The following step provide a very generalised approach for developing an objective tree:


Step 1: Generate an initial list of objectives:

Define the problem area (use of brainstorming technique).
Identify the people who will be involved in designing or directing the project.
Elicit their project related objectives.
Identify as many project objectives as possible, without attempting to structure the objectives.

Step 2: Identify an overall objective

Identify an overall objective for the project, to which all other objectives will relate. This objective will reflect a value judgement and it will need some form of measurement (e.g. axiological). Examples of such objectives include: “to meet the needs of the community” or “to achieve equality in housing benefits”.
This objective is positioned are the first level of the tree or even in the centre of the tree. All other objectives will be positioned below this one (means for achieving the objective) or above it (ends result of the objective).

Step 3: Extend the tree one level down

Select objectives for the next level down fro the list generated in step 1 or generate additional objectives by asking “what are the sub-objectives necessary to accomplish these objectives”. This is the branching rule for this type of tree diagram.Draw lines on the tree to connect these lower level objectives to the objective they help to achieve.

Step 4: Extend the tree to the next lowest level

Choose one of the objectives listed at the current lowest level of the tree. Identify sub-objectives that help to achieve it.Repeat steps 3 for all other objectives at the lower level most recently constructed. Another level of objective results when all the objectives in this level are dealt with.

Step 5: Review the tree or hierarchy:

Review the tree constructed in order to determine whether objectives are missing; an intermediate objective level may need to be added; whether the tree needs to be further extended upwards (additional ends relationships); an objective at a lower level is seen to achieve more than one objective at some higher level. In this case, redefinition of the objective is necessary.If the tree appears to be complete it is possible to move to step 6, otherwise it is necessary to return to step 4.


Step 6: Check the measurability of the lowest level objectives

Take an objective at the lowest level objective tree. Ask the question: Is this objective measurable? This is the stopping rule for the tree diagram. The measurability of an objective depends on two processes: 

  • The selection of a measure or units by which the attainment of objectives will be assessed (this should be objectively verifiable).
  • The design of a measurement scale and data collection process, to aid in determining the degree to which an objective may be reached.

Generally the objective at the lowest level will be quantitative or binary-event. Quantitative objectives generally have a numerical threshold to indicate what performance is acceptable.
If the lowest level objective is not measurable, then extend the objective down one more level, i.e. return to step 4.
Repeat step 5 for each of the lowest level objectives.


Something missing, unclear, misleading or a typo? Help us to make this page better!
Upon approval, the MethodFinder team will publish your comment here (* mandatory fields):