c r e s s t / u c l a center for the study of evaluation national center for research on evaluation,...
Post on 22-Dec-2015
213 views
TRANSCRIPT
C R E S S T / U C L A
Center for the Study of EvaluationNational Center for Research on Evaluation, Standards, and Student Testing
(CRESST)
New Models of Technology Sensitive Evaluation: Giving Up Old Program Evaluation
Ideas
Eva L. Baker and Joan L. Herman
SRIFebruary 25, 2000
C R E S S T / U C L A
Goals of Presentation
Outline Purposes and Challenges of Technology Evaluation
Describe Present Limitations of Technology Evaluation
Suggest Improvements
C R E S S T / U C L A
Purposes of Technology Evaluation
Soothe Anxiety
Justify Expenditures
Judge Impact
Identify Short-Falls
Improve Outcomes
Shore Up Managers’ Images
Demonstrate Technology Use in Evaluation
C R E S S T / U C L A
Limitations of Current Approaches to the
Evaluation of Technology
Conception of Evaluation
Designs
Measures
Validity of Results
That About Covers It
C R E S S T / U C L A
Limitations: Concept of Evaluation
The Scholarly Overlay of “Formative” and “Summative” Evaluation Makes Little Sense in Education in General and No Sense in Technology Implementations
Focus on “Value Added” Using New Outcomes Instead of Limiting Measures to Lowest Common Denominator
Evaluation Should Match Cycle-Time of Technology, e.g., No Five-Year Studies
C R E S S T / U C L A
Limitations: Designs
Existing Designs Are Usually Messy, Volunteer Studies of Available Classrooms
Randomized Treatment Allocations Are Possible, But Compensation from Home and Other Environments as Well as Pressures for Equal Access Make Them Impractical in the Long Run Without Creative Strategies
Treatments Need to Be Reconceptualized in Terms of Control and Uniformity
C R E S S T / U C L A
Limitations: Design/Procedures
Need for Collective Bargaining Agreements for Participation in Evaluation—Data Provision, Types of Information, Ability to Monitor Children and Adults
Human Subjects and Informed Consent
C R E S S T / U C L A
Limitations: Measures of Technology Effects
Opinion, Implementation, Smile-Tests
Student Performance Measures Insensitive to Implementations Mismatch Between Desired Goals and Measures Standardized Measures Mom and Pop Measures Lacking Technical and Credible
Qualities Masked by “Standards-Based” Rhetoric
C R E S S T / U C L A
Families of Cognitive Demands
Self-RegulationCommunication
ContentUnderstanding
ProblemSolving
Teamwork and
CollaborationLearning
C R E S S T / U C L A
Cross-Walk to Guide the Simultaneous Design of Assessment
and Instruction
Cognitive Models (Task Specification, Scoring Guides) Become Implemented Different Subject Matters
Domain-Independent and Domain-Dependent Components
Used for Design and/or Administration and/or Scoring
C R E S S T / U C L A
Next Generation: Authoring Systems for Multiple Purposes
Not an Item Bank
Capture Content, Task, Process
Instant Scoring and Feedback
Expert-Based
Beyond the Screen
C R E S S T / U C L A
Limitations: Validity of Results
Source: Inappropriate Measures, Unclear Assignment, Treatment Vagaries
Even in the Best Case: Generalizing to What? By the Time We Complete a Study, the Treatments Are Obsolete
C R E S S T / U C L A
Suggestions for Improvement
Defining the Objectives of Evaluation
Fixing Measures
Addressing the Concept of Evaluation
C R E S S T / U C L A
Distributed EvaluationCharacteristics and Functions
Conceptually Congruent with Distribution and Decentralization
Provides Information for Users and Program Managers
Allows Flexibility in Implementation Measures, e.g., Measures of Engagement, While Raising the Standards for Validity
C R E S S T / U C L A
An Indicators Approach
Flexibility
Longitudinal
Data Capture
Disaggregation and Aggregation
Data Export/Import
Local Engagement and Feedback
C R E S S T / U C L A
An Indicators Approach
Report Generation to Appropriate Constituencies
Updatable
Operational
Creates the Right “Management, High Tech” Patina
C R E S S T / U C L A
Quality School Portfolio Longitudinal
Database
Standards-Based
Multi-Purpose
Multi-User
Multiple Occasions
Local Goals
Automated Reports
C R E S S T / U C L A
Key Attributes of Distributed Evaluation
Measures: Fixed and Flexible
Some Common Performance Measures
And “Indicator” Mentality for Outcome Measures from Archival Sources
Common and Site Specific Implementation Measures
Fixed and Flexible Data Collection Schedule
Feedback Is a Feature in Every Design