Planning Tests and Assessments
Four Steps in Planning an Assessment
- Deciding its purpose
- Developing test specifications
- Selecting best item types
- Preparing items
Step 1: Decide the Purpose
What location in instruction? Or, the role of time in assessment!
- pre-testing
- readiness
- limited in scope
- low difficulty level
- serve as basis of remedial work, adapting instruction
- pretest (placement)
- items similar to outcome measure
- but not the same (like an alternative form)
- during instruction
- formative
- monitor learning progress
- detect learning errors
- feedback for teacher and students
- limited sample of learning outcomes
- must assure that mix and difficulty of items sufficient
- try to use to make correction prescriptions (e.g., review for whole group, practice exercises for a few)
- diagnostic
- enough items needed in each specific area
- items in one area should have slight variations
- end of instruction
- mostly summative –broad coverage of objectives
- can be formative too
Step 2: Develop Test Specifications
- Why? Need good sample!
- How? Table of specifications (2-way chart, "blueprint")
- Prepare list of learning objectives
- outline instructional content
- prepare 2-way chart
- or, use alternative to 2-way chart when more appropriate
- doublecheck sampling
Sample of a Content Domain (such as this course)
- trends/controversies in assessment
- interdependence of teaching, learning, and assessment
- purposes and forms of classroom assessment
- planning a classroom assessment (item types, table of specs)
- item types (advantages and limitations)
- strategies for writing good items
- compiling and administering classroom assessments
- evaluating and improving classroom assessments
- grading and reporting systems
- uses of standardized tests
- interpreting standardized test scores
Sample Table of Specifications (For content from this course)
Sample SLOs (you would typically have more) / Bloom LevelsRemember / Understand / Apply / Analyze / Evaluate / Create
Identifies definition of key terms (e.g., validity) / X
Identifies examples of threats to test reliability and validity / X
Selects best item type for given objectives / X
Compares the pros and cons of different kinds of tests for given purposes / X
Evaluates particular educational reforms (e.g., whether they will hurt or help instruction) / X
Create a unit test / X
Total number of items
Spot the Poor Specific Learning Outcomes (use previous table of specifications)
Which entries are better or worse than others? Why? Improve the poor ones.
- Knowledge
- Knows correct definitions
- Able to list major limitations of different types of items
- Comprehension
- Selects correct item type for learning outcome
- Understands limitations of true-false items
- Distinguishes poor true-false items from good ones
- Application
- Applies construction guidelines to a new content area
- Creates a table of specifications
- Analysis
- Identifies flaws in poor items
- Lists general and specific learning outcomes
- Synthesis
- Lists general and specific content areas
- Provides weights for areas in table of specifications
- Evaluation
- Judges quality of procedure/product
- Justifies product
- Improves a product
Why are These Better Specific Learning Outcomes?
- Knowledge
- Selects correct definitions
- Lists major limitations of different item types
- Comprehension
- Selects proper procedures for assessment purpose
- Distinguishes poor procedures from good ones
- Distinguishes poor decisions/products from good ones
- Application
- Applies construction guidelines to a new content area
- Analysis
- Identifies flaws in procedure/product
- Lists major and specific content areas
- Lists general and specific learning outcomes
- Synthesis
- Creates a component of the test
- Provides weights for cells in table of specifications
- Evaluation
- Judges quality of procedure/product
- Justifies product
- Improves a product
Step 3: Select the Best Types of Items/Tasks
What types to choose from? Many!
- objective--supply-type
- short answer
- completion
- objective--selection-type
- true-false
- matching
- multiple choice
- essays
- extended response
- restricted response
- performance-based
- extended response
- restricted response
Which type to use? The one that fits best!
- most directly measures learning outcome
- where not clear, use selection-type (more objective)
- multiple choice best (less guessing, fewer clues)
- matching only if items homogeneous
- true-false only if only two possibilities
Strengths and Limitations of Objective vs. Essay/Performance
Objective Items
- Strengths
- Can have many items
- Highly structured
- Scoring quick, easy, accurate
- Limitations
- Cannot assess higher level skills (problem formulation, organization, creativity)
Essay/Performance Tasks
- Strengths
- Can assess higher level skills
- More realistic
- Limitations
- Inefficient for measuring knowledge
- Few items (poorer sampling)
- Time consuming
- Scoring difficult, unreliable
Step 4: Prepare Items/Tasks
Strategies to Measure the Domain Well—Reliably and Validly
- specifying more precise learning outcomes leads to better-fitting items
- use 2-way table to assure good sampling of complex skills
- use enough items for reliable measurement of each objective
- number depends on purpose, task type, age
- if performance-based tasks, use fewer but test more often
- keep in mind how good assessment can improve (not just measure) learning
- signals learning priorities to students
- clarifies teaching goals for teacher
- if perceived as fair and useful
Strategies to Avoid Contamination
- eliminate barriers that lead good students to get the item wrong
- don’t provide clues that help poor students get the item correct
General Suggestions for Item Writing
- use table of specifications as guide
- write more items than needed
- write well in advance of testing date
- task to be performed is clear, unambiguous, unbiased, and calls forth the intended outcome
- use appropriate reading level (don’t be testing for ancillary skills)
- write so that items provide no clues (minimize value of "test-taking skills")
- a/an
- avoid specific determiners (always, never, etc.)
- don’t use more detailed, longer, or textbook language for correct answers
- don’t have answers in an identifiable pattern
- write so that item provides no clues to other items
- seeming clues should lead away from the correct answer
- experts would agree on the answer
- if item revised, recheck its relevance
Planning assessments 1