The Assessment Report
Software Engineering Group
09/12/2008
1. Student Learning Outcome (SLO)
Be able to apply the principles and practices
for software design and development in the construction of software systems of
varying complexity. (SLO11)
2. Assessment Plan
The First Iteration
·
Develop performance
criteria – by 31 March 2008
·
Verify performance
criteria for the assessment– by 16 May 2008
·
Assess artifacts – by 15
June 2008
·
Analyze the results of
the assessment – by 15 July 2008
·
Create an assessment
report – by 15 August 2008
The Second Iteration
·
Modify class objectives and/or
the assessment method based on the result of the first iteration – by 15 October
2008
·
Update performance
criteria, if necessary – by 31 October 2008
·
Assess artifacts – by 19
December 2008
·
Analyze the results of
the assessment – by 9 January 2009
·
Create an assessment
report – by 15 January 2009
3. Method(s) of Assessment
Performance criteria were used to assess artifacts already produced by
COMP380/L and COMP480/L students as part of their coursework. The criteria are
as follows:
The scale for satisfying each criterion: 4 (Exceptional or Grade A), 3
(Good or Grade B), 2 (Satisfactory or Grade C), 1 (Poor or Grade D), 0
(Unacceptable or Grade F).
A.
Performance Criteria for COMP380/L
1)
Assessment criteria for analysis (use case
descriptions and diagram)
Criterion ID |
Criterion |
Measurement concerns |
CR-1 |
Identify
system responsibilities |
Use
cases need to cover requirement and specifications. |
CR-2 |
Develop
system responsibilities |
Development
of each use case based on the recommended format that includes use case
properties, primary scenario and alternative scenario. Each use case should
be complete and correct. |
2)
Assessment criteria for design (class diagram)
Criterion ID |
Criterion |
Measurement concerns |
CR-3 |
Identify
classes |
Population
of classes and associations from use case descriptions. The classes chosen
should be high cohesive, low coupling, more sufficient, less complex, and
complete. |
CR-4 |
Identify
relationships between classes |
Class
relationships should be chosen appropriately using the different types of
connections such as regular association, generalization, aggregation,
composition, and dependency. |
CR-5 |
Identify
and allocate attributes |
Details
of a class are represented with attributes. Attributes can be simple
primitive types (integer, floating point numbers, etc.) or relationships to
other objects. Each attribute can be described in detail. |
CR-6 |
Identify
and allocate operations |
Operations
specify how to invoke a particular behavior. Each operation can be described
in detail. |
3)
Assessment criteria for design
(Sequence Diagram)
Criterion ID |
Criterion |
Measurement concerns |
CR-7 |
Identify
structural elements |
The
sequence diagram can model interaction at use case. Where the interaction is
modeling a use case, the starting point is the classes that participate in
“use case realization.” |
CR-8 |
Draw
sequence diagrams |
Draw
sequence diagrams for the use case descriptions. Each sequence diagram should
be displayed with a frame, life lines, messages, constraints, conditions,
etc. In particular, proper messages types (methods calls, sending a signal,
creating an instance, destroying and object, confirmation, etc.) should be
used depending on the situations. If necessary, proper fragments (“alt,”
“opt,” “break,” “critical,” etc.) should be used. |
B.
Performance Criteria for COMP480/L
1)
Assessment Criteria for requirement
and specification document
Criterion ID |
Criterion |
Measurement concerns |
CR-1 |
Feasible |
The
requirement is doable and can be accomplished within budget and schedule. |
CR-2 |
Concise |
The
requirement is stated simply. |
CR-3 |
Unambiguous |
The
requirement can be interpreted in only one way. |
CR-4 |
Consistent |
It
is not in conflict with other requirements. |
CR-5 |
Verifiable |
Implementation
of the requirement in the system can be proved. |
CR-6 |
Allocated |
The
requirement is assigned to a component of the desired system. |
CR-7 |
Assigned
a unique identifier |
Each
requirement shall have a unique identifying number. |
CR-8 |
Devoid
of escape clauses |
Language
should not be speculative or general (i.e., avoid wording such as “usually,”
“generally,” “often,” “normally,” and “typically”). |
2)
Assessment criteria for design
Criterion ID |
Criterion |
Measurement concerns |
CR-9 |
Feasible
and implementable |
Design
should be made to be successfully implemented. |
CR-10 |
Consistency
with requirement |
Design
should be consistent with the requirement document. |
CR-11 |
Enough
information |
Design
should include enough information for implementing the system. |
4. Scoring Rubrics
Before conducting the assessment test, scoring
rubrics were established as follows:
5. Results of the Assessment
5.1 COMP380/L
Among 10 group
projects, three projects (30%) fell into “Good,” five projects (50%) into “Fair,”
and two projects (20%) into “Poor.” Table 1 shows the details.
Table 1. The assessment results of COMP380/L
based on teams.
CR1 and CR2 turned
out to be “Good,” CR3, CR5, CR6 and CR7 were “Fair,” and CR4 and CR8 were “Poor,”
as shown in Table 2.
Table 2. The assessment results of COMP380/L
based on the evaluation criteria.
5.2 COMP480/L
Among 4 group
projects, two fell into “Fair” and two into “Poor.” Table 3 shows the details.
Table 3. The assessment results of COMP480/L
based on teams.
CR7 turned out to
be “Good,” CR7 and CR8 were “Fair,” and the rest were “Poor,” as shown in Table
4.
Table 4. The assessment results of COMP480/L
based on the evaluation criteria.
6. Analysis of the Assessment Results
Reflecting the
2007 assessment results and actions, rubrics were specified in great details
for this assessment and design activities were stressed in the previous classes.
See the appendix for details.
6.1 COMP380/L
As described on page
1, the assessment criteria for COMP380 consist of three sub-groups: developing
use case descriptions, class diagrams, and developing sequence diagrams. Each
sub-group is refined with a set of criterion.
The first
sub-group criteria evaluate the student capabilities on identifying and
developing system responsibilities, particularly functional requirements. The deliverables
are use case diagrams and their associated use case descriptions. Overall, the
results of the assessment criteria for analysis are good.
The second
sub-group criteria assess the student capabilities on developing class diagrams
that describe the structure of the system. The results were relatively fair. Students
were able to identify classes and develop their attributes and methods. Students,
however, seemed to face some degree of difficulties in fully and fluently developing
class relationships by using the different types such as generalization,
aggregation, composition, and dependency.
The third
sub-group criteria review the student capabilities on developing sequence
diagrams that help software developers understand system behaviors. A sequence
diagrams consists of a set of objects and interactions between the objects.
Students were good at identifying relevant objects but they needed to specify
the interactions in more details with proper message types and proper fragments.
In summary, compared
to the 2007 assessment result, the 2008 assessment result shows some degree of improvement
but the design skills need to be improved.
6.2 COMP480/L
<<Bob,
kindly take care of this section.>>
7.
Recommendations
for Actions/Changes
7.1. Should this assessment
activity be repeated? If so, when?
Yes, we need to repeat this assessment in Fall
2008 to see if there has been improvement in the results.
7.2.
Should changes be made in the way this assessment was done? If so, describe the changes.
Our current assessment method is a formal and indirect approach. We, the
software engineering group, carefully planned and carried out the assessment
activities. We also indirectly assessed the student learning outcome by
assessing the student project deliverables.
We now consider a formal and direct approach. Although indirect
approaches are efficient and effective on evaluating certain student learning
outcomes, we have a sense that this SLO can be more accurately assessed by a
direct manner. Therefore, we plan to create an assessment test so that we can
see whether or not students understand and apply the principles of software
design.
7.3. Should there be any changes in
curriculum? If so, describe recommended
changes.
No.
7.4. Should any other changes be
made?
No
Appendix
Summary Assessment Report – SLO11
Software
Engineering Group
9/7/2007
Student Learning Outcome: Be able to apply the principles and practices for
software design and development to real world problems. |
||||||
Performance Criteria |
Strategies (How and Where Taught) |
Assessment Method(s) |
Context for Assessment (How and Where Measured) |
Time of Data Collection |
Assessment Coordinator |
Persons of Responsible for Evaluation of Results |
Develop a software requirements and specification
(SRS) document that addresses a reasonable set of requirements in sufficient
details. |
COMP 380/L COMP 480/L |
Faculty evaluation using rubric |
Comp 380/L Comp 480/L |
Spring 2007 |
George Wang |
George Wang Diane Schwartz Robert Lingard |
Develop a software design document (SDD) that
includes appropriate designs and properly reflects the SRS document. |
COMP 380/L COMP 480/L |
Faculty evaluation using rubric |
Comp 380/L Comp 480/L |
Spring 2007 |
George Wang |
George Wang Diane Schwartz Robert Lingard |
Construct code that should follow coding
conventions such as names, comments, layout, etc. |
COMP 380/L COMP 480/L |
Faculty evaluation using rubric |
Comp 380/L Comp 480/L |
Spring 2007 |
George Wang |
George Wang Diane Schwartz Robert Lingard |
Results 5/22/2007: There is a
mismatch between design document and implementation. Design documents tend to
not updated reflecting implementation. One of the reasons is that the design
documents were created before implementation was started and were not updated
when implementation was finished. |
||||||
Actions 5/22/2007: A formal
assessment will be done. We will organize a panel in Spring 2008 whose
members are former COMP380/L or COMP480/L instructors. Rubrics will be more
specific and the project deliverables of all COMP380/L sessions will be
reviewed. Meanwhile, in COMP380/L course objectives, design activities will
be stressed. The basic design activities will be also addressed in
lower-level course courses such as COMP110, COMP182 and COMP282. |
||||||
Second-Cycle Results: None so far. |