Book/Planning A Test

A test should provide a good assessment of the results of the learning and, as such, is often a part of the learning process. For the preparation of learning and its subsequent evaluation, it is therefore necessary to be able to define as precisely as possible what the graduate should be able to do, i.e. what the learning objectives are. In practice, however, actual learning only approaches these objectives. In some areas, students even entirely fail to achieve the learning objectives, but at the same time they often acquire other, unplanned, knowledge and skills. The set of competencies the graduate actually acquires are called learning outcomes.

In terms of content and scope, a test we use to verify whether the graduate has acquired the desired knowledge and skills, should correspond as best as possible to the learning objectives, and at the same time, the outputs of the learning should correspond as much as possible to the test. In practice, this will never be a perfect match. To make it as good as possible, it is necessary to properly plan both the learning and the test. If students are aware of the objectives they need to achieve and how their performance will be assessed, they are more motivated to study. If, on the other hand, the content of the test or exam differs from the actual content (i.e., the outputs) of the learning, students feel cheated and label the testing as unfair, often replacing the motivation to learn with the motivation to simply get "the answers required on the test".

Another reason to carefully plan tests is the fact that we often cannot test all students at once and need to create multiple parallel forms (versions) of the test. If we follow the same, sufficiently detailed plan it becomes relatively easy to achieve equivalence of all the versions.

Blueprint

One of the best proven methods of preparing a test plan is the construction of a so-called blueprint . The first step is to create a spreadsheet with rows that match the content objectives. This step is relatively simple – it is mostly based on the course syllabus, the order of chapters in the primary textbook, etc. Each line corresponds to one topic. It is advisable that the breakdown of topics is sufficiently detailed – one lecture, lesson or chapter in the textbook usually corresponds to several lines with partial subtopics.

The columns of the blueprint (specification table) correspond to the aspects (more precisely, the learning domains) from which we can look at the topics. Finding views that can describe most of the content objectives (i.e., table rows) at once is key, and often the most difficult step in creating a blueprint. The most general advice is to base it on the objectives of Bloom's Taxonomy, e.g.:

• Knowledge – for example, knowledge of terminology, definitions, naming of a certain phenomenon;
• Understanding – for example, comparing, interpreting graphs and data;
• Application – for example, solving a new problem using an analogy;
• Analysis – for example, identification of causes and consequences, ability to explain a certain phenomenon;
• Synthesis – for example, predicting the outcome of an event, estimating the consequences.

Since the domains created according to Bloom's taxonomy are relatively detailed, and because there are just too many of them for the creation of shorter tests, sometimes simpler diagrams are used, e.g. acquiring knowledge – application – problem solving. In some fields, common aspects naturally follow from the material taught and are relatively easy to find, e.g. in clinical fields of medicine, the columns can often be labeled etiology – symptoms – diagnosis – treatment – prognosis, etc.

In any case, the blueprint should be constructed so that as many combinations of rows and columns as possible make sense. Individual fields then contain the planned number of items, the type of items or the method of testing. It is not necessary to fill in all the fields of the blueprint, but the test plan should be balanced – there should not be any empty or almost empty row or column, and no larger blank areas should remain.

When filling out the number of items, we already take into account the total scope of the test. It may happen that some fields of the blueprint, corresponding to less essential knowledge and skills, will be used only in some versions of the test, and other versions will contain other items instead. In any case, a test plan constructed in this way will help ensure that the number of items devoted to a certain topic reflects its importance and that all aspects of a certain problem are adequately tested.

The two-dimensional blueprint is often very detailed and extensive, so it resembles a technical drawing in its dimensions – that's why the term blueprinting was used for this method of test planning. The detailed blueprint is usually not public, it is used only by the test creators. Its publication could lead to the speculative behavior of test takers, who would concentrate their preparation more on the test itself than on acquiring the knowledge and skills needed for practical work. On the other hand, the content of the test (i.e. the lines of the blueprint) should always be published, as well as the share of each area in the total scope of the test.