Tests and quizzes

Description

Many courses in the biological sciences include short tests and quizzes as part of their overall assessment strategy and curriculum design. Such tests are generally described as being shorter than examinations*, take place during the course of study, and if formally assessed the grades generally make a smaller contribution to the overall grade. Tests vary in terms of their principal purpose, question design and format. A unifying characteristic, however, is that tests are opportunities for students to assess their knowledge, and they provide timely and diagnostic feedback on students' learning.

Written or online tests associated with practical classes are included in this section. A distinction is made, however, with the demonstration and evaluation of practical skills. Practical assessment of this type is quite distinctive and is described elsewhere [see also: Practical assessment].

* Tests, quizzes and examinations share many common features. For the purposes of this site, we have distinguished 'examinations' on the basis that they tend to be end-of-unit tests which contribute significantly to the overall grade [see also: Examinations]. We make no distinction between a test and a quiz, as these terms are used variously and interchangeably in the biological sciences.

Approaches

Purposes

As the examples described to us illustrate, tests are diverse in design and in the role that they play in learning and teaching in the biological sciences. Fundamentally, however, they are used to provide feedback and incentives.

Tests provide students with feedback on their knowledge:
While it is true that all forms of assessment have the potential to provide students with feedback on their learning, this is particularly true of tests. A well-designed test helps students to check their knowledge and identify gaps and areas of misunderstanding. This is true of tests designed as self-assessment tasks [Examples: Peat 3; Peat 5; Plastow; Rogers 2; Shapter], as well more formal and mandatory tests [Example: Mrongovius 2].

As many tests are 'progressive', they enable students to act on the feedback in time to improve their performance in subsequent assessment tasks (including, often, the final examination). It is common for first year courses to include early, low-stakes tests as a means of providing timely feedback on progress as they adjust to the expectations of university studies.
Tests provide staff with feedback on student progress, and on the learning and teaching:
Tests and quizzes conducted during a unit of study enable staff to identify, and respond to, issues with the learning and teaching. This feedback is important as two levels:

Results for individual students: Test results provide valuable information about individual students' learning and progress. For example, poor test results contribute to identifying students at risk, particularly important in early-year levels [Example: Shapter].

Classwide results: Areas of general misunderstanding can be identified and addressed. In-class quizzes, including the use of Audience Response Systems*, can be used to good effect [Example: Wood 1], as can more formal tests such as practical tests [Example: Dow; Mrongovius 2] and in-semester tests [Example: Peat 3; Plastow].

... If it's a test I always looking at the different groups and their average mark and standard deviation to make sure that there's equity there. We also put up the marking schemes or the test answers or what we were looking for in the dissection so a student can always come back and ask for an explanation if they're not happy. ... we tend to modify marking schemes from year to year depending on our experience from a previous year. We try to improve... you know, if there's been something that the students have raised with us on a regular basis then we'll take that into account and modify from year to year. [academic]

[*see also: Engaging large classes through assessment; Online assessment]

In-semester tests are also used as incentive and reward for student diligence:

Tests that contribute toward the grade for a course are often designed to encourage students to revise and study throughout the course - not simply in the lead up to a final examination. The use of tests for this purpose varies from regular small tests such as weekly practical tests each contributing just a few percentage points, to single, 'mid-semester' tests which are much 'higher-stakes'.

Question types

Tests tend to involve questions which are answered quickly - extended answers and essays questions, common in examinations, feature less often in tests and quizzes. The typical question types are described below.

multiple-choice questions (MCQs). MCQs are extensively used in tests and quizzes. They may be combined with images and diagrams, and are the question-type most commonly used in online quizzes.
As diagnostic tools, well-designed MCQs are an effective way for students to identify misconceptions and gaps in their knowledge. In particular, where explanation of both 'correct' and 'incorrect' options is provided, MCQs are a powerful device for students' self-assessment [Example: Plastow]. Similarly, classwide response patterns to MCQs provide staff with important feedback on teaching and student learning.

As noted under Examinations, however, the challenge lies in designing questions which are appropriate to the desired learning outcomes. The simplest MCQs to prepare are those which test students ability to recall specific information - for example: the correct name of an organelle; the most numerous molecule in plant cell walls. If, however, the intent is to encourage critical analysis, synthesis or problem-solving, effective MCQs are more difficult to construct.
true/false questions. True/false (T/F) questions can effectively diagnose students' level of understanding of complex concepts, particularly when they are used to trigger discussion. For this reason, T/F questions are often used where students are working collaboratively, or as a self-assessment tool.
When used for more formal testing - such as mid-semester tests - there are issues related to reliability and ambiguity. As discussed under Examinations, with just two options, guesswork and chance are even more an issue than is the case with MCQs. In addition, the risk of ambiguity in the typically-complex statements used in T/F questions is high. We were told of two different approaches to overcoming these limitations. The first approach was to require a written justification for the choice of true or false. In this case, the questions were not dissimilar to short-answer questions. A second approach described to us involved penalty for incorrect answers. In this case, there was no deduction if fewer than 30 per cent of the responses to a particular question were correct - the reasoning being that this would help control for questions which proved confusing or ambiguous.
short-answer questions. Extensively used in examinations, questions requiring a brief written response are also widely used in tests - particularly formal tests which contribute to the grade for a course, such as practical class tests or mid-semester tests. Such questions may require an explanation or description, and often allow for the use of diagrams in addition to text. The emphasis is upon demonstrating understanding, rather than writing skills - 'note-style' writing is often accepted, and errors in grammar and spelling tolerated.
images and diagrams. Images and diagrams are commonly used in tests in the biological sciences. Such devices are used to assess knowledge of structure and function, and students' ability to analyse the data representations that form the 'language' of the particular discipline. For example, the simplest are questions requiring students to name or describe the indicated parts of a photograph, micrograph or diagram. Other tasks require more interpretation, such as explanation of data presented in a graph or the interpretation of a three dimensional structure from a series of sections. Depending upon the conditions under which the test is taken, video can also be used to good effect - practical tests and online tests are a good example.

Test format

The most obvious distinction in format is between paper-based and computer-based tests. The use of online tests and quizzes is increasingly rapidly, in part due to the uptake of learning management systems in higher education, and the fact that such systems usually have inbuilt tools for test creation and management.

[Examples: Plastow; Peat 3]

Another variation on format, and one which is growing in use, is the incorporation of informal quizzes into lecture classes. While this has always been an option, the advent of Audience Response Systems* has provided impetus to this form of assessment [Example: Wood 1]. Typically there is no grade involved, and the test is used for the multiple purposes of engaging students and providing feedback to both students and staff.

[See also: Online assessment]
[*Audience Response Systems are described further under Engaging large classes through assessment]

Conditions and requirements

The conditions under which tests and quizzes are completed relates directly to their principal purpose.

Compulsory, and under standard 'test conditions'
Particularly where the assessment contributes to the overall grade for a course, tests are typically taken sight-unseen and in a timed and invigilated environment. Depending upon the approach, such tests can provide both students and staff with feedback, and also serve as motivation to study.
In practical tests of this type, it is common for students to have access to additional resources such as laboratory notes.
Compulsory, but in students' own time and with access to resources
Tests which are not invigilated may be set as a hurdle requirement. For example, completion of an online quiz relating to library resources may be a requirement for course completion, or a test relating to practical class procedures may be a prerequisite for admission to the class itself. It is common for such tests to contribute only a very small percentage, or nothing at all, to the overall grade for a course. Such tests tend to used to encourage student diligence. The degree of feedback students receive depends very much upon the test design.
Optional
Increasingly, tests are used as a means of providing students with opportunities for self-assessment. We were told of numerous examples of tests purpose-designed for providing students with specific and diagnostic feedback on their learning. Such tests are purely formative, in that they do not contribute to grading.

Feedback

As described above, many tests are designed with feedback as the principal purpose. Therefore, the type of feedback provided by tests is an important consideration. The type of feedback provided is often directly influenced by the question type. Feedback on a multiple-choice quiz is typically quite different to that provided on a test involving short-answer questions and diagrams. In addition, providing the best possible feedback given limited time and resources inevitably results in the need to make strategic decisions.

The types of feedback provided by test results may be categorised as follows:

Grade alone - (any question type)
It is not common for a test score to be the only feedback provided to students. There is general agreement that a score alone has limited value, beyond a very general indication of relative progress.
Due to the way in which information about test results is often provided to students, however, the score may be the only feedback many students 'register'. For example, if scores are made available in advance and/or independent of other feedback, many students pay attention to the score alone. Therefore, consideration of how feedback is released can be as important as the actual type of feedback provided [Example: Mulder 4].
Identification of correct and incorrect options - (MCQs, T/F, some diagrams etc)
A common approach to feedback on MCQs is simply to make the questions available after the test, with the correct option highlighted. This is simple and is readily automated for online tests.
As discussed above, test feedback will be most effective if it is available at a time when students are engaged in thinking about the questions. Approaches include feedback for each question as students progress through an online quiz or immediately after completing such a quiz. This can also be achieved in class, if students correct their own tests as the questions are discussed in turn.

[see also: Online assessment]
Correct responses provided - (short answer questions, some diagrams etc)
For tests involving students' written responses, automated assessment is far less common. Assessment is usually done manually. A common example in the biological sciences is the assessment of practical class tests. Manually grading large numbers of test papers, and in time to return the results quickly, often involves simply 'marking' the papers (with ticks and crosses, or numerical scores for each question), and making available model answers for students to reference.
A variation on this approach is self-assessment - students checking their own responses against the model answers provided. This is an effective approach for optional, purely 'formative' tests [Example: Rogers 1].
Tailored feedback on students' written answers - (short answer questions, extended answer questions)
A far more resource and time intensive approach than providing model answers is to annotate the papers of individual students, identifying and explaining errors. Students respond well to individualised feedback of this kind. The issue of effective use of limited resources remains, however, an important consideration. It is often decided to reserve such individualised feedback for written reports, rather than tests.
Explanation of options - (MCQs, T/F)
- Correct options
  A valuable extension of feedback on MCQs is to provide a further explanation of each correct response. This is particularly valuable when the reasoning is described - the logic and the evidence that a student would draw upon to identify this as the correct option.
- Incorrect options
  A further extension to the above is provision of explanations for incorrect responses. In a well-designed question, the incorrect options will correspond to recognised misconceptions. By providing an explanation of these options, the feedback directly addresses the misconception itself [Example: Plastow].
The preparation of effective explanations of this kind requires that the author have both disciplinary expertise and insight into student learning. It is both a powerful and potentially efficient approach - just as with model answers, such explanations need only be prepared once for a particular question and can provide a valuable study resource for students.
Peer support - (any question type)
An extension of in class discussion of test questions is that which involves students from later year levels. Several universities have set up programs such as Peer Assisted Study Sessions (PASS) which enable students to discuss their test results and questions with students who have completed the course in previous years [Examples: Miller 1; Miller 2].

Issues and strategies

The need for fidelity of test results

When tests contribute in even a small way to the overall grade for a course, the issue arises of ensuring that each students' work is their own. Staff differ in their philosophies in this regard and test conditions vary accordingly - from timed, invigilated, and with questions unavailable beforehand, to 'take-home' or online tests which are untimed and rely on students' honesty in terms of authentication. Typically, however, where there is a significant contribution to course grades, tests are supervised and structured to minimise the risk of dishonest behaviour.

A simple and long-standing approach is to manage students seating position within a shared test venue. This is no longer as simple as it once was, however, with increasing class sizes. Also, if computers are used, students' work may be in clear view of one another. A strategy commonly used to manage this is to vary the order of questions and/or the order of options in MCQs. Another approach, and one that is commonly used if tests are taken by groups of students at different times, is to have versions of the test with different but 'equivalent' questions. Moderation becomes important in such situations to ensure that the versions are indeed equivalent.

Tests and quizzes are also used for purely formative purposes. In such situations, collaboration between students is typically encouraged - test questions can provide an effective trigger for productive discussion and debate.

Ensuring that feedback is both meaningful and timely

It is a reality of university teaching that resources are finite and, therefore, that the strategic use of resources is a critical aspect of assessment design. Feedback for students is most effective when it is both timely and meaningful [see also: Providing feedback]. Identifying ways in which meaningful feedback can be provided quickly and efficiently is therefore critical in the design and implementation of tests and quizzes.

Considerations regarding feedback are described in detail in the section above entitled 'Feedback'.

Maximising the investment in the preparation of test questions

The design of effective test questions requires disciplinary-expertise, insight into student learning ... and time [see also 'Question types', above]. This is particularly true for MCQs, which are challenging to prepare and are typically needed in large numbers. For this reason, the questions used in graded tests are often 'protected' and reused in later years. This poses a challenge with regard to providing feedback. Displaying the questions, with feedback, in a secure location for a fixed period of time is one solution. While perhaps not ideal, this approach can be effective - particularly if students have both opportunity and incentive to consider the feedback. Discussion of the feedback, in class or online, is one way to maximise the feedback value from 'protected' test questions.

Generic skills

The focus of tests is typically upon the development and assessment of disciplinary knowledge. Well-designed tests can, however, also be used to foster the development of generic skills and, to some extent, assess these skills. The most obvious are critical thinking, analysis and problem-solving

Authentic assessment

A number of the biologists we interviewed said thay developed tests and quizzes in realistic contexts. For instance, one academic uses a quiz to ensure students navigate through the National Aquatic Animal Disease Awareness web site [Example: Benkendorff 1]. Another creates problem sheets based on authentic scenarios [Example: Sanderson].

To reference material from this site, please use:

Harris, K-L., Krause, K., Gleeson, D., Peat, M., Taylor, C. & Garnett, R. (2007). Enhancing Assessment in the Biological Sciences: Ideas and resources for university educators. Available at: www.bioassess.edu.au

Enhancing Assessment in the Biological Sciences

Relevant Examples

Attachments