Image Credit: Blend Images / Getty Images
The advent of online and blended programs has brought about changes for teachers and learners, and it has required adjustments in curricula. One such adjustment involves the design and development of testing frameworks that can fulfill new program needs. These frameworks must include measurements that provide reliable information for assessment and evaluation purposes.
Bachman and Palmer (2010) contend that “In the real world of language assessment use, it is becoming increasingly important, and in many cases mandatory, for test developers and users to be accountable to stakeholders.” Being accountable means being able to demonstrate that the intended use of a test is justified, which means developing tests that reflect what students have learned and been exposed to. This can be done by comparing the test specifications to the course syllabus. It is also important to explicitly state that the intended use of the test will be the one described in the specifications and to share this information with students.
The update of programs must be accompanied by an update in testing methods. Coombe and Hubley (2013) state that “The term validity refers to the extent to which a test measures what it says it measures. In other words, test what you teach, how you teach it!” If testing doesn’t match teaching, then the test isn’t valid—and the scores derived can’t provide reliable information about meeting program outcomes. If new programs have online teaching components, then the logical consequence is that testing frameworks should include online components, and these new tests must be developed with a careful focus on reliability and validity.
We started out with the design of a framework to be used in our blended programs. Because the components of our program were 50% in-person and 50% online, the design included 50% in-person testing and 50% online testing. We later applied the same framework to our fully online testing program.
Test Development Procedure
Designing an Online Test
Our programs include courses at all levels, from beginning to advanced, and therefore, our test design needed to comprise all levels as well. We followed these steps:
Step 1. Define the Test Construct
State the knowledge and abilities that the test is to measure.
Step 2. Revise the Inventory
Revise the inventory of course contents and materials that will guide the language to be used in each of the levels.
Step 3. Write Test Specifications
These provide the general instructions and details for creating the test blueprint.
Step 4. Define the Skills and Number the Tasks
As you create the test blueprint, define the skills to be tested and suggest the number of tasks to include to assess each skill.
Step 5. Weight the Tasks
Assign the relative weight that each skill or area is going to have in the total test score.
Step 6. Design the Question Types
Bear in mind the cornerstones of testing; the question types need to be the ones that students are familiar with, disregarding the inclusion of questions to which students have not been exposed.
Guidelines for Creating Test Items
At the same time, we developed the following guidelines for test developers to assist the process of creating objective test items, for example, multiple-choice or true or false tasks:
Item formats are correctly matched to the purpose and content of each item. When writing test items, we need to bear in mind the objective of each item and decide if the ideas included accurately match the format in question.
The items are written at the intended level of students’ proficiency. It is imperative to revise the level of the test input to assure that it is right and fair for the intended audience.
All parts of an item are visible at the same time. The task to be completed must be displayed on a single page or screen so that the test taker can focus on content and not on scrolling up and down or changing windows.
There is only one correct answer for each problem. Having more than one correct answer may cause ambiguity and thus affect test fairness.
Negatives and double negatives have been avoided. It may be misleading and confusing to write a negative answer as the correct one because that may be a source of confusion for the test taker.
Rubrics provide clear guidance for test correction. They show the relative weight of each test task and each item within the task. Rubrics are shared with students so they have a clear idea of how the test is to be graded.
Race, gender, and nationality bias have been avoided. It is of the utmost importance to revise items to detect possible sources of bias that would influence test impartiality.
At least one other colleague has proofread the items. It is essential to have reviewers that check all test items to ensure test validity and fairness.
There are many tools that can be used to develop online tests, and it should be up to the institution or the teacher to decide which is best for their contexts. Language management systems could be considered one of the most powerful and complete tools.
In the case of our adult courses, we opted for Moodle because it provided us with an array of possibilities that proved highly effective for our context. Our students are accustomed to using Moodle as a learning environment, and therefore validity is considered by using the same means for teaching and testing purposes. Additionally, Moodle presents several advantages over other language management systems, given that it is open source and offers a comprehensive, responsive interface that can adapt to different devices.
For our teenage courses, we went in a different direction. Our teenagers were familiar with publishers’ platforms for online work, but they did not use Moodle. Because these platforms did not offer us a testing solution, we resorted to Google Forms to provide students with a tool that was user friendly and easily accessible to all of them. The Google Forms feature of converting forms into quizzes offers valuable data analysis for teachers. Our students had already used a few Google Forms and we made sure to train them in the program before they faced their first online test in this modality. Google Forms offered us an efficient, easy-to-implement solution to test our teenage students.
In all cases, it is advisable to provide students with a mock test to run a system requirements check. Even if students are familiar with the testing tool to be used, this check provides an opportunity to familiarize them with the test mechanics and navigation, thus ensuring content validity.
It is undeniable that online teaching and testing have come to stay and will be used in education for the coming years. It is also true that online testing presents advantages for teachers and learners. We must capitalize on these advantages as they can provide an excellent way to help make a smooth transition from paper-and-pencil to online testing. Consequently, we need to develop frameworks for online testing so that teachers can have appropriate evaluation tools, ensuring that learners are fairly tested.
Bachman, L., & Palmer, A. (2010). Language assessment in practice. Oxford University Press.
Coombe, C., & Hubley, N. (2013). Fundamentals of language assessment. Cultural Affairs Office, U.S. Embassy.