PLANNING AND CONSTRUCTING.pdf

165
Chapter 6
PLANNING AND CONSTRUCTING
CLASSROOM TESTS
Learning is not attained by chance; it must be sought for with
ardor and attended to with diligence.
—Abigail Adams
Issues and Themes
Testing children in school is one aspect of good teaching. The classroom
teacher and his or her students receive the most benefit when the testing is
formative. Formative tests are ongoing and fit into the teaching–learning
cycle. They are designed to provide information about the details of learning
that underlies the progress of individual learners. This information can be
used to tailor the instruction to the individual student’s needs. These mea-
sures tend to be short quizzes, mini-tests, or even informal assessments.
Tests that follow instructional units of study are summative by design.
These summative assessments are also part of the teaching–learning cycle
and are critically necessary for reporting progress to the parents and for
recording the success of each individual.
The central task of teachers in constructing classroom tests is to assure
that the measures are valid. One important element of validity involves the
06-Wright-45489.qxd 11/20/2007 5:18 PM Page 165

content of the test and the test’s coverage of the curriculum. Schools that
perform well on state assessment tests have studied the state’s learning stan-
dards, mapped the educational goals and objectives related to those stan-
dards, and have built a local curriculum that teaches those standards for
learning. The summative tests developed by classroom teachers measure
progress toward achieving the goals on objectives outlined in the curriculum.
This is facilitated by the use of a two-dimensional test blueprint. The blue-
print documents how many questions will be developed to measure each
objective and the level of cognition the child will need to answer those ques-
tions successfully.
There are various formats that can be used in designing a test measuring
achievement. One division is between the use of test items that ask the
student to select the correct answer to a question from an array of possible
answers and those items that pose a question and ask the student to supply
the correct answer. The former of these include true–false tests, matching
tests, and three types of multiple choice tests. In all tests, careful planning,
and attention to detail, provides the best measurement product.
By reading and studying this chapter, you should acquire the competency to
do the following:
• Link instruction with the use of formative evaluations in the classroom.
• Provide the documentation needed to support report card grades and
conference with parents.
• Devise a curriculum map for a course or grade level.
• Build a table of specifications for an achievement test.
• Explain the six levels of cognition as identified by Benjamin S. Bloom.
• Describe how to construct high-quality multiple choice items.
• List the major pitfalls in the writing of matching-type items.
• Know the advantages of using well-written true–false items.
TESTING AND TEACHING
Good teachers love to teach. If there is anything about the job of being a
teacher that is disliked, it is testing and evaluating. One of the first tasks given
to graduate assistants by their mentor professors is grading. Likewise, teacher
166– –PART III TESTING OF STUDENTS BY CLASSROOM TEACHERS
Learning Objectives

aids are frequently asked to read and comment on student work for the
teacher. Some high schools hire paraprofessionals to read essays written in
English classes. Unfortunately, this general dislike of the evaluation process
carries over to the mandated statewide assessments. The vast majority of edu-
cators see them as a waste of resources and time that would be better spent
on instruction. Yet, in this era of accountability, no educator can afford to
ignore educational measurement. As Popham (1999) observed, all the rules
changed when the test scores from schools appeared in the local newspapers.
Testing and Instruction
The first step in teaching is the identification of the goals for the instruc-
tional process. These goals should be expressed in terms of what each child will
be able to do or know as a result of the instructional process. Once these goals
have been identified and sequentially ordered, instructional activities can begin.
Good teaching involves ongoing monitoring of the learners and the con-
stant tweaking of the instructional process. During every instructional hour
good teachers are continually integrating questions into their instruction;
they are always looking for those who don’t understand or are confused. This
ongoing questioning represents an informal approach to assessment. It is
informal in that it is idiosyncratic and designed for that particular moment.
Informal data collection related to a child can also occur in conversation with
other teachers and aids, through observation of the child, by an analysis of
errors the child makes, by a careful reading of the homework the child com-
pletes, and by asking questions of the child’s parents.
FORMATIVE TESTS
Classroom quizzes and tests serve a purpose that is similar to informal assess-
ments. These measures are referred to as formative. The name formative is
used because they are designed to inform instructional practice. The best
instructional practices involve the use of frequent quizzes designed to inform
both the teacher and the learners of achievement and acquired new skills
and of areas of learning difficulty. This provides documentation of the effec-
tiveness of the teaching and/or indicates areas needing reteaching
(Ainsworth & Viegut, 2006).
Standardized tests and statewide assessments are designed to measure
after instruction has occurred. These measures are used to quantify how
much was learned by each child. This type of assessment is a summative
Chapter 6 Planning and Constructing Classroom Tests– –167

measure, as it is given to summarize what has happened. All high-stakes tests
and final examinations in secondary schools and colleges fall into this group.
Application of Formative Measures
The type of measurement that is most useful to the teaching process
occurs while learning is still underway. These tests are formative in nature, as
they can inform the instructional process. Most classroom tests and all
quizzes can be formative. The formative testing process makes it possible for
a teacher to determine whether students hold a misconception in common
or if they have learned and achieved what was expected. This can be as
simple as doing a brief analysis of the items from a classroom quiz. After com-
mon learning problems have been identified, it is possible for the classroom
teacher to reteach the topic, and the instructional outcome will be improved.
The key concept in formative evaluations is that the classroom assess-
ments must support the instructional process. This involves much more than
deciding on a child’s grade. Formative assessment data make it possible to
modify and improve the instructional process while it is occurring. In addi-
tion, formative evaluations also provide a quality-control system for what is
happening in the classroom (Leahy, Lyon, Thompson, & Wiliam, 2005). Each
time a teacher asks a student a question, a datum has been collected. When
such data are used correctly the nature of student learning can be under-
stood, and the quality of instruction can be improved.
Thus, the learning cycle includes (a) instruction, (b) formative and informal
assessment, (c) feedback for the teacher and feedback for the students, (d) mod-
ification of instruction and modification of the behavior of the learner. Each
of these four steps uses data generated by the assessment. The information
provided by the assessment can be as simple as identifying and rewarding a
student’s good work. It can also involve helping other students to create a learn-
ing scaffolding similar to that of the model student (Good & Brophy, 1995).
Formative assessments also can help the instructor in setting the correct
pace for learning. This pacing task needs to be accurately set and monitored.
When instruction moves too slowly students become bored and disconnected.
When the pace of learning is too fast many students may be left behind.
The cognitive depth that students need to employ to learn new material
can also be monitored through the use of formative assessment strategies. By
asking the students a few questions before instruction begins, each of which
is written to a different cognitive level, and asking the questions again during
the instructional process, it is possible for the teacher to ascertain if students
are employing the optimal learning strategies (Carroll, 1963).

Student Response Pads
A new approach to formative assessment uses an interactive computer
system to give teachers real-time formative assessment data during instruc-
tion. This new technology has been called “classroom clickers” (Duncan,
2005). To initiate this system, schools first invest in hardware that resembles
remote controls for each child (see Photo 6.1). These student response pads
provide a way for each student to communicate directly with the teacher’s
computer. During instruction the teacher can ask or in some way present a
question and students are able to supply immediate answers using their
response pads. This is much like being able to call on every child in the class
at the same time. The pulse of the class can be taken by the teacher at any
moment and the result seen on a handheld computer screen. In the wireless
environment needed for the student response technology, the data can be
outputted onto a wireless PDA. A number of textbook publishers have begun
to supply classroom response systems as an incentive for schools to purchase
their textbooks and curriculum materials.
Case in Point (6a)
The response pad system has been widely demonstrated by several television
quiz shows recently. One example was the original ABC network show Who
Wants to Be a Millionaire. The show is now syndicated by Disney’s Buena Vista
to most TV markets in the United States. One of the “lifelines,” provided by the
game rules, allows the contestant to poll the studio audience for their opinion
on the multiple choice questions being asked. A bar graph representing the
audience’s choices appears on the screen. This polling of the audience is com-
pleted by a response pad system.
In the classroom the teacher can periodically provide multiple choice
questions for the class to answer. The pattern of answers will demonstrate how
many students understand the concept being taught. By writing lesson plans
with several such marker questions imbedded each day, the teacher can keep
close tabs on the learning that is happening in the classroom. The new gener-
ation of clickers does not require a “clean line of sight” to operate and facili-
tate more give and take in the normal classroom setting (Guy, 2007). The total
number of response pad units in the schools and colleges of the United States
was estimated to total 8 million in 2008. This reflects the falling costs of the
systems. It is now possible to purchase classroom clickers for less than $100
per unit (Cavanaugh, 2006).
For more
information, see
“Considerations on
Point” at
www.sagepub.com/
wrightstudy

Quizzes and tests can also be administered using this system. Another
advantage of the use of response pads for classroom instruction is that there
is evidence that their use improves student memory and retention of instruc-
tional material (Chan, McDermott, & Roediger, 2006). The multiple choice
test questions can be presented on paper, but a better model involves the
presentation of the items using a PowerPoint format. The response pads have
coded signals to the computer identifying each child. The system is able to
download student test data into a Microsoft program that will summarize and
analyze and save the data in the teacher’s grade book.
Special Needs Children
A second function of the formative evaluation process that occurs within
the classroom is the early identification of children at risk for failure. The
need for vigilance on the part of primary-grade teachers has never been
greater. The fact that all third-graders must take a high-stakes test, which in
many states could force them to repeat the year, makes it critical that educa-
tors watch for children who are struggling to keep up with their classmates.
The early identification of children who are at risk for failure can make it
possible to initiate remediation and developmental instruction before the
children become lost in the system.
Photo 6.1 Students Use Response Pads
SOURCE: From Eduware Inc. www.eduware.com. Reprinted with permission from Bill Stevens,
president of Eduware Inc.

Learning Standards
One method for improving a school’s average scores on state assessment
tests is to map out the state’s standards and design local methods to measure
and track the progress of children toward achieving those learning standards
(Guskey, 2005). In this effort, formative assessments can be linked directly
with the learning standards and objectives approved by the state and
included in the high-stakes assessment. This is more than a simple paraphras-
ing of the state’s exemplar test items, involving an analysis of what the learn-
ing standard requires and what students are able to actually do. This analysis
must include the underlying skills implied by the standard. Classroom mea-
sures designed to assess student performance against the published stan-
dards can be thought of as providing benchmarks for progress toward
proficiency (Herman & Baker, 2005).
A number of school districts have begun to employ formative testing
to reduce the gap between minority and Anglo-White children. The Fairfax,
Virginia, schools have adopted the use of a series of formative mini-tests
during the school year. These formative tests monitor progress toward the
achievement of the standards for learning that are measured by that state’s
assessment tests. Fairfax’s mini-tests are designed to point out problematic
vocabulary and undeveloped specialized skills that are required for the state’s
testing program (Glod, 2006).
Test Preparation
Another reason for introducing both formative and summative testing in
the early grades involves the need to familiarize children with the testing
format. The contrived environment that is needed for the administration of
Case in Point (6b)
Computers have recently become significantly easier for children with dis-
abilities to use. President Clinton signed into law the Communications
Technology Amendments to the Rehabilitation Act in 1998. That law has a set
of mandates requiring all computers provide adaptations for users who have
a disability (P.L. 105-220, Title 29 U.S.C. 794d § Sec.508). These adaptations
include keyboard and mouse modifications for the orthopedically dis-
abled, varying font sizes for the screen, adjustable sound levels, and speech-
recognition software.
For more
information, see
Point” at
www.sagepub.com/
wrightstudy

high-stakes assessments can be disconcerting to children unfamiliar with the
testing process. Young children are comfortable helping each other and shar-
ing answers, behaviors antithetical to high-stakes testing. Young children also
expect that their teachers will provide them with answers and help them
when they experience difficulty. The regular use of tests and quizzes by
primary-grade teachers can introduce children to the reality of the summa-
tive testing that they will face starting in third grade.
Parental Reports
Another function of formative classroom evaluations is to provide the
documentation needed to submit a report card grade (see Chapter 10) and
to prepare for parent–teacher conferences. Parents need and deserve to
see work samples from their children during conferences with teachers.
Work samples that were part of formative evaluations during the school
term can be organized and presented as part of a folio of each child’s work
and progress. The simple recitation of grades from a roll book will not work
for most parents who want to see exactly what problems their children are
encountering and what successes they have had. The classroom response
system provides printouts depicting graphic data designed to demonstrate
both the absolute and relative progress of individual students. It can pin-
point areas of difficulty as well as areas of success. This type of evidence is
difficult to debate and can lead to productive and cooperative meetings
with parents.
Case in Point (6c)
Not missing an opportunity, in 2006 the Educational Test Service purchased a
test consulting company that specializes in formative assessment systems. The
entry of ETS into the formative testing field is resulting in a rapid expansion
of the availability of new formative classroom assessment approaches. One
new product from ETS was the development and sale of access to a bank of
thousands of formative assessment items. These items are designed to allow
schools to develop their own practice tests for the state-mandated assessment.
In 2007–2008, the ETS division on formative assessments became the central
player in a national research study on the implementation of formative assess-
ments (Cech, 2007).
For more
information, see
Point” at
www.sagepub.com/
wrightstudy

SUMMATIVE EVALUATIONS
Children face summative evaluations of their work a number of times each
school year. At the end of each high school year, students sit for the dreaded
final examinations in their various courses. Elementary and middle-school
teachers provide students with tests after each instructional unit. And the
state now has a series of mandated achievement tests that children must face
during the spring of at least seven of the school years.
Classroom Achievement Tests
A day rarely passes in the lives of school children without their being
tested. By the time a student graduates from high school it is likely he or she
has taken 2,000 tests, quizzes, and other written evaluations. Research has
shown that teachers at all grade levels do not have a high level of confidence
in their ability to develop tests of the highest quality (Stiggins & Bridgeford,
1985). The frequent use of classroom quizzes and tests has been demonstrated
to be a motivating factor for most children and may even be a direct cause of
learning (Slavin, 1994; Tuckman, 2003).1
The motivational effect is maximized
if the time between taking the test and getting a grade or score report is short.
The use of unannounced tests increases both student stress and parent
angst (Guskey, 2002; Partin, 2005). Yet, pop-quizzes produce a small achieve-
ment advantage and are a favorite technique for ongoing formative assess-
ments by many teachers. By using this strategy, secondary students are forced
to always read and prepare before the class, whereas a quiz may or may not
happen. Therefore, the teacher makes his or her class the primary one on the
student’s homework schedule and places his or her assignments ahead the
homework assignments made by all the other teachers. In deciding to employ
this method, it should be remembered that few opinions and attitudes are as
stable over time as those about teachers. One major dimension in teacher
evaluations is the perceived “fairness” of classroom tests and evaluations.
Case in Point (6d)
This is an easy truism to verify. If you think back to your school days, it is easy
to remember the best and worst teachers you experienced as a student. One
factor of the poor teachers may have been a capricious evaluation and
For more
information, see
Point” at
www.sagepub.com/
wrightstudy

Classroom Test Validity
All teachers can build valid and reliable classroom measurements. (A full
discussion of these two dimensions is provided in Chapters 4 and 5.) One
core issue in the validity of a classroom test is the match between the con-
tent that was taught and the questions on the test. When there is a mismatch
between the two, students and their parents may be expected to complain. A
mismatch is also a sign that the test does not have content validity. Two tools
available to teachers to enhance the content validity of classroom tests are
curriculum maps and a table of specifications (test blueprints).
The careful design of classroom tests can also contribute to improving a
school’s scores on the state-mandated assessments. This is possible when the
classroom tests are each linked to the state’s approved standards for learn-
ing. That set of endorsed learning standards should be at the heart of each
school’s curriculum. In other words, public schools have an obligation to
develop and teach a curriculum aligned to the state standards for learning.
This makes it possible for classroom teachers to develop classroom assess-
ments that are an indication of how well the students are achieving the state-
required learning standards. Each school should have a central map of its
curriculum and how it links to the state standards.
The Mohonasen Central School District (NY) developed curriculum
maps that show the sequence and details of what is to be taught in English,
mathematics, science, and social studies for each school year.2
These maps
make it possible to develop a series of formative mini-tests that track all
children as they progress toward the goal of being graded “proficient” on the
statewide mandated assessment (Glod, 2006). Student performance on the
grading system. This area, evaluation, was also noted by Michael Scriven
(1997) as one of particular concern when students evaluate their instructors.
One new, and highly contentious, direction for student evaluations of
teaching can be seen on a number of college campuses today. Each year there
are more campuses that have online teacher evaluation systems in place.
Centralized online systems can reduce the subtle influence that can be exerted
by professors who distribute paper-and-pencil forms for student use in the
evaluation of the teaching they experienced. These may have been in answer
to the movement that was started by student groups on some campuses to pro-
vide an independent, but publicly accessible, faculty evaluation system (e.g.,
Universities of Utah and Mississippi). It is only a matter of time before such
systems appear on high school campuses.
Case in Point (6d) (Continued)

mini-tests can inform the teacher of the need for individualized instruction
designed to enhance the learning for children who have not acquired a nec-
essary piece of the curriculum.
Curriculum Mapping
A curriculum map can be thought of as the school’s master plan for
learning. It should be a publicly available document that captures the scope
and sequence of the learning objectives and activities in each subject and in
each grade. It is more than a curriculum guide in that it also specifies the
skills that students will develop and presents a timeline for instruction in
each curriculum area. The map also specifies major educational activities
(e.g., field trips and guest speakers) and specifies what assessments are
needed for each grade level and classroom.
Curriculum mapping begins with the state learning standards. A team of
teachers and curriculum specialists then translates the standards into educa-
tional goals and objectives that specify the skills and content to be learned.
All subjects and grade levels need to be mapped, even those for which there
are no state learning standards.
To review a school district curriculum map for the Spotsylvania, VA,
School District see http://205.174.118.254/cmaps/ and from Litchfield, AZ, see
http://www.lesd.k12.az.us/curriculum_site/cmap.htm.
The major advantage of curriculum mapping is in providing guidance for
the teachers and in the facilitation of instructional planning. Another advan-
tage is that the maps make it possible for teachers to know exactly what was
taught the previous year and what will be expected of the children during the
ensuing year. Also, parents can quickly see what is being taught and when
that instruction is to happen during the school year. For this reason, posting
the curriculum map on the school’s Web page is a recommended strategy.
Table of Specifications
The construction of valid classroom measures requires planning the test
along two dimensions. The first dimension is central to content validity, and
the second is the level of cognition the test items will require. These two axes
can be plotted together in a test construction blueprint known as a table of
specifications (see Table 6.1).
On the vertical axis of the table are specific content objectives for the
weather unit. When a classroom teacher builds a test, this axis should exactly
reflect the areas of the curriculum that were taught. This process of creating

Objectives
Identify cloud forms
and explain their
formation
Emphasis = 40%
(16 items)
Understand weather
instruments
Emphasis = 10%
(4 items)
Read and interpret 2
scales on a
thermometers
Emphasis = 10%
(4 items)
Describe and
interpret 3 types of
weather fronts
Emphasis = 15%
(6 items)
Explain the “water
cycle”
Emphasis = 15%
(6 items)
Explain tropical
storm system
Emphasis = 10%
(4 items)
Emphasis =
# of Items =
# of Items =
# of Items =
# of Items =
# of Items =
# of Items =
Knowledge
(40%,
16 items)
7
2
1
3
2
1
Comprehension
& Applications
(40%,
16 items)
6
1
2
2
3
2
Analysis, Synthesis,
& Evaluation
(20%, 8 items)
3
1
1
1
1
1
Table 6.1 A Hypothetical Table of Specifications for a Science Achievement Test of a Weather
Unit for Fourth-Grade Children
Level of Competence Required to Answer the Questions
TOTAL TEST ITEMS = 40

a table of specifications prevents the test from overemphasizing those areas
for which it is easy to write questions and underemphasizing those areas
where test items are difficult to construct. In organizing the table of specifi-
cations, the teacher should assign a weight for each content area. This weight
serves as a guide in determining the number of items needed for each part
of the test. The weight can take the form of a simple percentage, as is shown
on Table 6.1. Each of these content areas for which questions are being writ-
ten should match the instructional objectives specified on the school’s cur-
riculum map for the grade level.
Cognitive Requirements
The second axis (horizontal) of the table of specifications indicates the
cognitive level that items on this test should require. The question of the
level of cognition implied by any learning objective, or needed to respond to
any test question, was clarified over 50 years ago. In 1956, Benjamin S. Bloom
and other members of a committee of the American Psychological
Association provided a taxonomy, or scale of cognition. This scale defines six
levels of thought, which are further divided into 19 subcategories. These
levels of cognition can be used to classify the level of thinking required by an
instructional objective or to answer a test question.
The lowest level of thinking is that which Bloom’s committee referred to
as “knowledge,” and the most complex was “evaluation” (see Figure 6.1).
It is the lower level of cognition (knowledge) that dominates almost all
teacher-made tests. Two doctoral dissertations from the 1990s demonstrate
this point. Rosemary Castelli (1994) examined the syllabuses and final exam-
inations from 18 freshman and sophomore liberal arts courses at a suburban
community college. The courses included those from the humanities,
science, mathematics, and social sciences. She found that a preponderance
of syllabuses for the courses included course objectives requiring students to
use higher order thinking skills. However, the final examinations for all but
one of those college classes were dominated by questions requiring the low-
est level of cognition (i.e., knowledge). The exception to the rule was an
introduction to philosophy class in which the instructor asked examination
questions requiring analysis and synthesis.
Maureen Finley (1995) followed up the Castelli study with a project that
examined the upper-level courses, including the Advanced Placement
classes, of a suburban high school. Here the curriculum guides for the classes
had many learning objectives written requiring higher levels of cognition, but
the questions on the midyear examinations did not.

There are a number of reasons why teachers do not write questions
requiring students to employ higher order thinking skills. Perhaps the most
telling are that such questions are difficult to write and take much longer to
grade. Typically, these higher order questions require the use of extended
answers and an essay (constructed response) format.
Bloom also attempted, but never finished, a taxonomy for educators to
use for the classification of the affective domain. That taxonomy was written
by David Krathwohl and includes the following broad titles: Receiving,
Responding, Valuing, Organizing and Conceptualizing, and, at the most
advanced level, Characterizing by Value or Value Concept (Krathwohl, Bloom,
& Masia, 1964). Another uncompleted effort by Benjamin Bloom was to
define a psycho-motor domain. That partial work includes a starting point of
Imitation, followed by Manipulation, Precision, Articulation, and, at the high-
est level, Naturalization (Atherton, 2005).
Cognitive Level
Knowledge
Comprehension
Application
Analysis
Synthesis
Evaluation
Assessment Verbs
Quote directly, who, what, where,
when, list, define
Interpret, estimate, discuss,
translate, summarize
Solve, calculate, examine,
demonstrate, apply,
differentiate, explain
Sequence and order, select,
compare, argue, connect
Deductive and inferential
thinking, combinatorial, create,
plan, invent, design
Summarize, assess, judge,
evaluate, defend, convince
Capabilities
Memorization of facts, dates, lists, and events
Ability to recall details of subject matter and the
vocabulary of the field
Ability to transfer knowledge to other contexts
Capability of grasping meaning and predicting
consequences, ability to interpret and translate
Problem solving skills and the ability to employ
theories and concepts to new tasks
Ability to see latent meanings and patterns in
information. Capable of ordering parts to see
components and commonalities
Using the components from analysis to create new
concepts or conclusions. Ability to hypothesize
outcomes and to combine concepts into new ideas.
Assess theories and evaluate constructs and
concepts. Compare competing models and make
discriminating judgments.
Figure 6.1 Bloom’s Taxonomy
SOURCE: From A taxonomy for learning, teaching, and assessing: A revision of Bloom’s taxonomy of educational objectives
(1st ed.), by Lorin W. Anderson and David R. Krathwohl. Boston: Allyn & Bacon. Copyright 2001 by Pearson Education.
Reprinted by permission of the publisher.

Cognitive Taxonomies
Bloom’s original model has been supplemented by several recent modi-
fications. A less complex model was developed for the teachers of North
Carolina in 1989 (Northwest Regional Educational Laboratory, 1989). This
model divides the domain of cognition into five levels. Starting at the lowest
level, these are recall, analysis, comparison, inference, and evaluation.
Norm Webb, of the Wisconsin Center for Education Research, developed
a cognitive taxonomy designed to help align state standards for student learn-
ing with statewide assessments. That taxonomy has four levels: (1) recall and
reproduction, (2) skills and concepts, (3) strategic thinking, and (4) extended
thinking (Webb, 2007).
A student of Benjamin S. Bloom, Lorin Anderson, led a group of cognitive
psychologists in the development of a new model for the original taxonomy by
Bloom. Their work provided a two-dimensional model (Anderson & Krathwohl,
2001). This system provided one dimension representing the level of knowledge
required for the material that was learned, and the other dimension presented six
levels of cognitive process needed to perform the mental task (see Figure 6.2).
In addition to this new cognitive taxonomy, Anderson and Krathwohl
also proposed four levels of knowledge that could be learned: factual, con-
ceptual, procedural, and meta-cognitive. Their system then plotted the six
levels of cognitive process that may be used to answer test questions with the
four levels of knowledge, thereby producing a grand total of 24 distinct levels
of cognition (four knowledge levels multiplied by six cognitive processes).
Thus, the analysis of any test item can be carried out in two dimensions cov-
ering 24 distinctly different item types.
Bloom’s Taxonomy
Knowledge
Comprehension
Application
Analysis
Synthesis
Evaluation
Anderson’s Taxonomy
Remembering
Understanding
Applying
Analyzing
Evaluating
Creating
Figure 6.2 Comparison of Bloom’s and Anderson’s Taxonomies

TESTING FORMATS
Once the table of specifications is complete, the next step is to write the ques-
tions that will be a part of the new instrument. The first choice is whether to
employ an open-book or closed-book approach. This choice is very much
linked to the educational philosophy of the teacher and to the learning objec-
tives for the course.3
The obvious advantage to employing an open-book test
format is that students are put on notice that the test will require analytical
answers and that the memorization of lists of facts is not the best approach for
test preparation. Another option is the use of an oral-examination system
(Lunz & Bashook, 2007). This method for student assessment is often used in
graduate and professional schools; also, at the undergraduate level, oral exam-
inations are often employed in assessments of languages and communication.
Beyond the open-book test format is the take-home examination format.
Take-home examinations combine elements of a performance evaluation
with a structure similar to a traditional examination. This approach is a good
method to employ when the expectations for learning are at the highest level
and require many hours of analysis to answer. The emergence of virtual
schools and online university courses has reignited research interest in take-
home examinations (Rakes, 2005–2006).
Conventional wisdom notwithstanding regarding classroom courses and
real schools, there is no research evidence that open-book examinations
lower test anxiety, reduce cheating, or improve the quality of learning
(Cassady, 2001; Kalish, 1958). Lacking any clear evidence of an advantage for
the use of open-book and take-home examinations, most teachers at all levels
prefer to employ a closed-book, in-class test format.
Item Format
Once the format decision is made, the teacher is then ready to chose
between a “selected response” or “constructed response” format, or possi-
bly a combination of both (Popham, 1999). Constructed responses are the
answers created by the student to questions such as completion, short
answer items, essays, and compositions. The latter two can also be consid-
ered to be performance items (see Chapter 9).
Selected Response
These test items require the student to select the best answer from sev-
eral provided on the test. The three major formats for selected items are

true–false, matching, and various permutations of the multiple choice ques-
tion. There is a general set of test construction rules that apply to all three
forms. First, the number of questions should make it possible for more than
90% of all students to complete the test. Girls tend to work at a slightly slower
pace on multiple choice format tests than do boys, and the range of scores
shows less variance among girls than it does among boys (Longstaffe &
Bradfield, 2005). This may reflect the greater tendency for risk-taking by boys
(Ramos, 1996). For this reason, girls tend not to do as well on this type of
test. The test length should not present an obstacle to those who take a more
reflective approach to answering the questions.
Also, each question should have a single focus and use language that
is unambiguous and available to all students. The directions should be clear
and succinct and provide insight as to how the test will be scored. Students
who are impulsive in their tempo tend to miss the directions and begin tak-
ing the test as soon as it is in their hands. All test directions should be printed
on the test and also read aloud by the teacher. This assures that everyone is
following the correct test directions.
True–False Questions
The first of the three principle formats of selected answer questions is
true–false. This well-known item type has distinct advantages over other
forms of test questions. These items are incredibly easy to score and provide
an easy to administer objective measurement of student achievement.
Students can answer true–false items at a faster rate than any other format
of question. This makes it possible to ask many more questions in a given
time period.
Figure 6.3 Classic Peanuts Cartoon With Charlie Brown and Snoopy
SOURCE: Peanuts: © United Feature Syndicate, Inc.

True–false quiz and test items can serve an instructional role when they are
used to motivate students during review sessions. By using a worksheet featur-
ing true–false items, a teacher can identify misconceptions and areas of confu-
sion. Computer-savvy teachers can post true–false questions on their Web page
prior to an examination as a way to provide students with a home study tool.
Chief among the disadvantages of using true–false items is the difficulty
in writing good questions. Additional problems include the tendency to test
at the lowest cognitive level, the ease for student collaboration (cheating),
and the large amount of variance that is a function of random guessing. If a
student never saw the questions, and only worked with the answer sheet of
a true–false format test, the most likely grade would be 50% correct.
A well written true–false item is unequivocally true or false, without any
ambiguity. Ambiguity is seen in words such as sometimes, usually, and typi-
cally. These words tip off the test taker that the best guess for the answer is
“true.” Also, in writing such unambiguous items, avoid using exact specification
terms or defining words such as always, never, not, only, and must, which tip
off the test taker that the best guess for an answer to the question is “false.”
Here are some examples of bad items:
It never rains in Southern California. (F)
Water always boils at 1,000° C. (F)
There are occasions when rain falls in the Mojave Desert. (T)
Here they are again but with better wording:
1. Southern California receives less annual rainfall than the northern
part of the state. (T)
2. At standard pressure and temperature (STP) water boils at 1,000° C. (F)
3. Las Vegas, NV, receives less annual rainfall than does the Mojave
Desert. (F)
Good true–false test items can be written for any subject area.4
1. When the atmospheric pressure falls, water boils at a temperature
greater than 100° C.
2. The petitioning employee won in the case (Hill v. C. A. Parsons [1972]
1 Ch. 305).
3. If a
→
→
is a linear combination of b
→
→
and v
→
→
, then b
→
→
is a linear combina-
tion of v
→
→
and a
→
→
.

4. The Socratic educational ideal is based on the view that discovering
the truth involves critical thinking.
5. The validity of a test determines the maximum level of reliability the
test can have.
6. Under International S.T.P. conditions PV = nRT.
7. Compound interest accumulates as FV = P(1 – r)n
One variation of the true–false test has students indicate how certain
they are that a question is true or false. Under this system, students indicate
T + 3 if they have no doubt if a question is true. The scale extends from F –
3 for being secure in the belief that the item is false, F – 2 for holding a mod-
erate belief in the item being false, and F – 1 for having a high degree of
uncertainty about whether the item is false, to T + 1 for being slightly positive
the item is true, T + 2 for holding a moderate degree of certainty, and T + 3 for
being very positive of a true answer.
Grading of each item is done on a seven-point scale (–3, –2, –1, 0, +1, +2,
+3). If the item is skipped and not answered, then it is scored zero. If the item
was really true but marked F + 1, the student receives a score of –1 for the
question. If the student marked that “true” item T + 2, he or she would be
scored +2 for the question, etc.
This complex variation of a true–false test is generally not well received by
the majority of students. Students who are risk takers seem to prefer this
option. The advantage for the instructor of using this approach is that it results
in a less skewed distribution of scores than is typical of true–false tests.
Matching Questions
Another selected format for test items that is frequently used in the mid-
dle grades is the matching question. This format provides a way to ask ques-
tions that requires cognition that is at the comprehension, application, and
analysis levels. The questions can be answered quickly, making it possible to
cover a wide area of curriculum in a relatively short test. The matching ques-
tions can be easy to mark and are less threatening to most students.
The problem with matching items is that the probability of any part
of the matching question being correct is influenced by the other sections
of the item. If there are five stimulus items to match against seven alterna-
tives, the chances of guessing one correctly is 1:7; however, if the student
knows the answer to four of the five stimuli, then the chances of guessing
the last stimuli correctly becomes 1:3. These items are difficult to write, as all
stimuli and alternatives must be drawn from the same subset of the knowledge

domain. This difficulty may make mixing the areas on the question attractive
for the teacher. Matching items are always homogenous in topic coverage. If
the various question parts represent two or more distinctly different areas of
knowledge, then the item becomes very easy to guess. For example:
Match the items listed in the first column with those in the second column.
Not all items in column two are needed to answer this question.
A. Potassium _____ Electricity
B. Lavoisier _____ Combustion
C. Table salt _____ NaCl
D. Ozone _____ NaCO3
E. Franklin _____ Magnetism
_____ K
_____ O3
In the earlier grades teachers sometimes have children answer matching
questions by drawing lines connecting the stimuli and appropriate alterna-
tives. While children like this type of activity, it can present a nightmare of
lines and erasures for the teacher to grade. Another common error in con-
structing this type of test item is to put too many stimuli and alternatives into
a matching item. This can create confusion for those children with poor read-
ing and short-term memory skills. The optimal length is five or fewer stimuli
and seven or fewer alternatives. The format should have all the longer stems
on the same side (the left) and the shorter answers listed on the other. This
reduces the need to constantly reread a long list.
It is also important to avoid grammatical clues between stimuli and alter-
natives. Another example showing common errors of grammatical agreement
is shown below. This next question also illustrates another item writing error.
Read each item carefully and match the states on the left column with their
official state animal.5
STATE ANIMALS
1. The state animals of A. Bison
Pennsylvania are:
2. The state animal of B. American buffalo
Oklahoma is a:
3. The state animal of C. White-tailed deer & Great Dane
Kansas is an:

In this example, a little knowledge of English grammar provides the cor-
rect answers. Also, the answers American buffalo and bison are too similar
to be used as alternatives in the same question. A better choice would be to
include a state with a dissimilar state animal, such as Alabama, where the
state animal is the racking horse.
Multiple Choice Questions (Type A)
Select the best answer from the choices listed below the following question.
Which is true of multiple choice questions (MCQ)?
A. The use of MCQs for testing is widespread throughout education.
B. MCQ items in a test require that the student only recognize the cor-
rect answer.
C. MCQ items can be scored in an objective and reliable way.
D. Writing a high-quality test with MCQs is a labor-intensive task.
E. All the above6
Multiple choice items like the one above are made of several parts. The
question statement is known as the stem of the item. The various answers are
the alternatives, the correct answer is the keyed response, and the wrong
options are the distracters. There are three major formats for MCQ items.
The type of item demonstrated above is the standard “type A” item.
The item stem should provide either a complete question or present an
incomplete sentence. The former is better for younger children. In reality,
most MCQ stems written as incomplete sentences can easily be rewritten
as complete questions. The stem should pose only one problem that can be
read and clearly understood. The length of the stem should not require
students to reread it several times to understand and remember the task
before they read the alternatives.
In writing the stem, avoid including negatives. The simple insertion of
the word not into a question can magnify the item’s difficulty and flummox
impulsive students, who may never see the word. See the following two pos-
sible sample stems from a question on a social studies test about the Civil
War. (Asterisk denotes correct answer.)
From the following list of cities, which was a state capital for a Confederate
state in 1862?

Or this alternative stem:
Which of the following cities was not a state capital for a Union state in 1862?
A. Columbia*
B. Sacramento
C. Augusta
D. Springfield
Poor stems leave the student wondering what is expected. The following
is an example of such a murky item stem.
Newton’s Second Law . . .
A. electricity
B. bodies at rest
C. time
D. movement*
Here’s a better item:
Newton’s Second Law explains the relationship between mass and accelera-
tion in terms of what property?
A. weight
B. force*
C. length
D. light
The alternatives include one correct (keyed) answer and two to four dis-
tracters (wrong answers). All the alternatives should be plausible and related
to the same topic. All too often a teacher writing a multiple choice item starts
with a correct answer then writes a stem for the item. Next, the item author
casts about for the distracters. The result is often an odd collection of unre-
lated nouns that can appear to be the product of free-association.
The best choices for distracters are errors that the students have made
previously on homework or misconceptions that came up during class. When
the teacher is writing distracters it is not the time to become whimsical or

attempt comedy. The unexpected humor can interfere with the focus and
attention of the students to the task. Research has shown that optimal item
reliability occurs when there are three options for each item. One option
is the correct answer (keyed answer) and the other two plausible alterna-
tives (Rodriguez, 2005). However, most published tests provide four or five
answer options on each question.
Four other guidelines for writing MCQs include ideas about “all and
none,” length, sequence, and grammar. Try to always avoid the use of “all of
the above” or “none of the above” as an alternative. Both of these alternatives
lead to classroom “lawyering,” as students make a case after the test is
returned that there is really a better answer out there somewhere (none of
the above) or that there is something about each alternative that is partially
correct (all of the above). Also, the option of “all of the above” is logically
incompatible with test directions to “select the one best answer.” There are
times when there are only three options. The next item is an example.
We are all familiar with the observation that the sun rises in the eastern sky
and sets in the west each day. What is the explanation for this phenomenon?
A. The earth rotates on its axis from west to east.*
B. The earth rotates on its axis from east to west.
C. The earth doesn’t move and the sun rotates about it.
In this example, “all or none of the above” would not work as a fourth
option.
One way to telegraph the students which of the alternatives is correct is
to make it longer than the other options. Testwise, students know that “when
all else fails, select the longest answer.” This occurs as the question author
tends to overwrite the keyed answer to guarantee it is indeed correct. As in
the case of matching items, it is important not to alert students to the correct
answer by providing grammatical clues. Also the location of the keyed answer
should be random. There is a tendency to place the correct answer in posi-
tion A of the alternatives as it is the first answer that the teacher writes.
In the animal kingdom a frog is classified as an_________
A. amphibian*
B. reptile
C. marsupial
D. fish

The following hypothetical examination is one in which you can get every
item correct knowing the rules for making good multiple choice questions.
Please read the following questions and select the best answer to each question.
1. The purpose of cluss in furmplaling is to remove?
A. cluss-prags*
B. tremalls
C. cloughs
D. plumots
2. Trassing is true when?
A. luso tresses the vom
B. the viskal flens, if the viskal is conwil or scrtil*
C. the belga frulls
D. diesless kils easily
3. The sigia frequently overfesks the tralsum because of what?
A. all siglass are mellious
B. siglas are always votial
C. the tralsum is usually tarious*
D. no trelsa are feskaole
4. The fribbled breg will minter best with an?
A. derst
B. marst
C. sartar
D. ignu*
5. Among the reasons for tristal doss are?
A. when the doss foged the foths tristaled*
B. the kredges roted with the orats
C. few rekobs accepted in sluth unless foths are present
D. most of the polats were thronced not tristaled
6. Which of the following (is, are) always present when trossalls are
being gruven?
A. rint and vost
B. vost*
C. shum and vost
D. vost and plone

7. The minter function of the ignu is most effectively carried out in con-
nection with?
A. a razama taliq
B. the thrusting bding
C. the fribbled breg*
D. a frally rnoz
8. ???
A.
B.
C.
D. *
Key to the answers:
1. The word cluss is in the stem and in the answer.
2. The conditional “if” gives it away; also it is the longest answer.
3. The word usually is the giveaway here.
4. Here the key is found in the agreement of “an” with “ignu.”
5. The answer is in the stem.
6. The word vost is in all other answers.
7. The answer in question 4.
8. The pattern of the other correct answers is A, B, C, D, A, B, C; ergo, D.
Each multiple choice test item is one that can be guessed. When a
student knows one or two alternatives are wrong, it is possible to improve
the chance of being correct when guessing. Yet, every item can be guessed
even if the student has no idea which answer is correct. In that unfortunate
case, the likelihood that the student will guess the correct answer is a func-
tion of the number of answer alternatives. If there are four alternatives, the
chance of a correct, albeit blind, guess is 1:4. Should this occur on every item,
the unlucky student will likely score near this “chance level,” or about 25%.
If there are five alternatives, the chance of a correct guess is 1:5. Knowing this
fact, a number of test publishers have a correction formula used to correct
for guessing (Lord & Novick, 1968).
This is done by counting the number of wrong answers on the test and
subtracting those that were not attempted. The number remaining is the
number attempted but answered wrong. This value is then divided by the
number of alternatives. This provides an estimation of the number of items

correctly guessed on the test. This value (the number answered wrong
divided by the number of alternatives) is then subtracted from the total
number of questions answered correctly. This is then reported as the raw
score “corrected for guessing.” Because boys are more likely to take blind
guesses than are girls, this approach may produce a differential effect by gen-
der (Ben-Shakhar & Sinai, 1991).
All students deserve advanced warning if this scoring method (correc-
tion for guessing) is used. When the test is scored using a correction for
guessing, the best strategy for test takers to employ is to skip any item where
they have no knowledge about any of the alternatives. If one or more of the
alternatives can be eliminated, then it will pay the student to make an “edu-
cated guess” between the two alternatives that are possible.
Case in Point (6e)
The publishers of most textbooks provide teachers with online banks or CD-
ROMs with test questions as well as printed books of test items. Occasionally
these item banks are even organized by instructional objectives and item diffi-
culty. While there are no absolutes, the quality of these items is generally poor.
One reason for this is that the authors of the texts rarely write their own item
manuals or item banks. A large number of full-time graduate assistants have sup-
plemented their incomes by writing the questions to accompany new books.7
For more
information, see
Point” at
www.sagepub.com/
wrightstudy
The second and third formats for multiple choice items are type K and
type R items. Both of these are easily adapted for writing questions requiring
higher order thinking skills. These items are extensively used on the qualify-
ing examinations of various medical and other professional licensing boards.
Multiple Choice Questions (Type R)
This format for multiple choice items is sometimes referred to as an
extension of the well-known matching question. This format requires that
test takers have good reading comprehension skills. The format is used in
qualifying examinations in medical disciplines. It also has many applications
when there is a large body of high-consensus material that can be used in
item development. This includes most science disciplines and a number of
areas in the social sciences.

Extended-matching items are multiple choice items organized into sets
that use one list of options for all items in the set. A well-constructed
extended-matching set includes four components:
1. A theme, or the subject area being assessed by the items.
2. An option list, or a list of all the potential alternative answers.
3. A lead-in statement that provides a common task to complete based
on the question posed by each item stem.
4. A list of two or more item stems (questions) that can be answered by
the options.
Theme: Microbiology
Options:
A. Adenovirus
B. Aspergillus fumigatus
C. Bacillus anthracis
D. Candida albicans
E. Chlamydia psittaci
F. Coccidioides immitis
G. Coronavirus
H. Corynebacterium diphtheriae
I. Coxiella burnetii
J. Coxsackievirus
K. Epstein–Barr virus
L. Haemophilus influenzae
M. Histoplasma capsulatum
N. Mycobacterium tuberculosis
O. Mycoplasma pneumoniae
P. Neisseria gonorrhoeae
Q. Neisseria meningitides

R. Pneumocystis carinii
S. Rhinovirus
T. Streptococcus pneumoniae
U. Streptococcus pyogenes (Group A)
Lead-in:
For each patient who has presented with fever, select the pathogen most
likely to have caused his or her illness. Select the one most likely pathogen.
Stems:
1. A 7-year-old girl has a high fever and a sore throat. There is pharyngeal
redness, a swollen right tonsil with creamy exudates, and painful right
submandibular lymphadenopathy. Throat culture on blood agar yields
numerous small hemolytic colonies that are inhibited by bacitracin.
Answer: U
2. For the past week, an 18-year-old man has had fever, sore throat, and
malaise with bilaterally enlarged tonsils, tonsillar exudates, diffuse cer-
vical lymphadenopathy, and splenomegaly. There is lymphocytosis
with atypical lymphocytes. The patient tests positive for heterophil
antibodies.
Answer: K
Multiple Choice Questions (Type K)
Type K multiple choice items allow the test developer to ask questions
with more than one correct answer. This format of question was used on
early forms of the SAT. In the recent past this item type was used primarily for
questions on medical board qualification tests. The use of these items has
generally been supplanted by the use of R-type multiple choice items.
Students who are asked to answer this type of question must have a good
level of reading comprehension and a good memory. The stem that sets up
the question must be memorized and then compared with each of the com-
plex list of possible answers.
Type K questions provide a set of directions that are repeated at the top
of every page of the test where these items appear. This multiple choice for-
mat also has a stem that is expressed as a problem for the test taker to solve.
The third part of the question provides the answer alternatives.
DIRECTIONS
For the following questions, one or more of the alternatives given are
correct. After deciding which alternatives are correct, record your selection
on the answer sheet according to the following key:

Mark A if alternatives 1, 2, and 3 only are correct.
Mark B if alternatives 1 and 3 only are correct.
Mark C if alternatives 2 and 4 only are correct.
Mark D if alternative 4 only is correct.
Mark E if all four alternatives are correct.
SUMMARY OF DIRECTIONS
A B C D E
1, 2, 3 correct 1, 3 correct 2, 4 correct 4 correct All are correct
STEM
A 27-year-old woman complains of diplopia and difficulty maintaining her
balance while walking. These symptoms began suddenly yesterday morning
while she was bathing. During the past 2 years she has also had several day-
long episodes of blurred vision in one or both eyes. She remembers that one
of these episodes also began while she was bathing. In addition to a thorough
clinical examination, an appropriate initial diagnostic workup would include
the following:
1. cortical evoked response studies
2. basilar arteriogram
3. cerebrospinal fluid protein
4. bone marrow biopsy
Correct answer: 2
Summary
The development and proper use of high-quality tests is a part of good teach-
ing. Ongoing quizzes and the use of classroom clickers are part of formative
evaluations, while achievement tests are a form of summative assessment.
Teachers can use tests to identify learning problems, individualize instruc-
tion, prepare children for future high-stakes tests, and collect data to share
with parents.
The best classroom achievement measures are designed following a
plan that includes a two-dimensional blueprint. The blueprint provides both

content and cognitive level guidance for the teacher and can result in a
valid measure.
Select items, such as multiple choice questions, provide an opportunity
to test a wide swath from the curriculum that students have studied. The
writing of quality items requires time and practice on the part of the teacher.
While students cannot bluff on a multiple choice test, they can guess the cor-
rect answer by being test wise and understanding a few simple rules.
1. Examine the use of tests in a school with which you are familiar. Do
the teachers of this school use tests in a formative way, or are all tests
used as summative achievement statements?
2. What points would you make in your presentation if you were asked
to provide an inservice lecture on the advantage of formative testing
in the schools?
3. Write a multiple choice test item based on this chapter of the textbook
that requires test takers to employ Bloom’s cognitive level of either
“application or analysis” to answer. Then rewrite the same question
so that it requires Bloom’s level of either “synthesis” or “evaluation”
to answer.
4. Write an outline of a presentation you would make for high school
students on how to do well on multiple choice tests. Include in your
outline issues of wording, guessing, and grammar.
5. Assume that you are asked to explain to your school’s parent organi-
zation why the purchase of interactive classroom communication sys-
tems (clickers) would be a good idea that will benefit the students.
Educational Assessment on the Web
Log on to the Web-based student study site at www.sagepub.com/
wrightstudy for additional Web sources and study resources.
Discussion Questions
Student Study Site

NOTES
1. The level of motivation to do well on a classroom test (low stakes) is related to
many factors, including the gender of the children (Eklöf, 2007). Boys tend to
have less motivation but a higher level of confidence in their knowledge base
than do girls.
2. See www.mohonasen.org/03curriculum/curriculumhome.htm.
3. Some professors of educational research who emphasize application and have
little need for students to recall facts use a take-home examination model. This
format can reduce student anxiety in complex subject areas while still determin-
ing if the student can handle complex tasks.
4. Answers: 1. T, 2. T, 3. F, 4. T, 5. F, 6. T, 7. F
5. Answers: 1. C, 2. B, 3. A
6. Yes, I know it is bad form using “all the above.”
7. All of the examination items supplied to faculty teaching this course were
written by the text’s author.

PLANNING AND CONSTRUCTING.pdf

More Related Content

Similar to PLANNING AND CONSTRUCTING.pdf (20)

More from International advisers (20)

Recently uploaded (20)

PLANNING AND CONSTRUCTING.pdf