Building Capacity for Collaborative Action: Can Collective Impact Help Advance Urban Forestry?
A Look Ahead for the Multi-State Collaborative to Advance Learning ...
Transcript of A Look Ahead for the Multi-State Collaborative to Advance Learning ...
A Look Ahead for the Multi-State Collaborative to Advance Learning
Outcomes Assessment/VALUE Initiative
Terrel L. Rhodes, Vice President for Quality, Curriculum and Assessment and Executive Director of VALUE, AAC&U
Julie Carnahan, Vice President, State Higher Education Executive Officers
Kate McConnell, Senior Director of Research and Assessment, AAC&U
October 17, 2016
Why Are We Doing This Work?
2
Purpose and Vision for the Multi-State Collaborative
3
Change the dialogue currently focused on:
To…
Access and Completion
Quality and Success
A Collaborative Vision for Student Learning
Valid Assessment of Learning in Undergraduate Education
Launched in 2007
Campus-based
Authentic assessment of student work
Privileges role/importance of faculty as authors of assignments and arbiters of quality
An advocate for state policy leadership
A liaison between states and the federal government,
A vehicle for learning from and collaborating with peers,
A source of information and analysis on educational and public policy issues.
Emphasis on success
Taking the Vision to Scale in Twelve States
5
MSC
Participants:
CT, HI, IN,
KY, MA, ME,
MO, MN, OR,
RI, TX and
UT
Plus – GLCA
& MNCp
Steering Committee: State
pt. persons from each state +
reps. from SHEEO & AACU.
Institution Point Persons: Person
from each campus in each state.
-Root assessment of learning in authentic work & the expertise of faculty
-Establish benchmarks for essential learning outcomes
-Develop transparency of shared standards of learning to assist w/ transfer
Goals:
VA
The Current VALUE/Multi-State Collaborative Project
• Purpose
– Sea Change in Assessment
– Reliability
– Validity
– Local value
– Policy debate = learning
MinnesotaCollaborative
(MN)
Great Lakes Colleges Association (GLCA)
Multi-State Collaborative
(MSC)
Multi-State Collaborative (MSC) to Advance
Learning Outcomes Assessment
Demonstration Year
7
MSC Demonstration Year by the Numbers (2015-16)
• MSC states: Connecticut, Hawaii, Indiana, Kentucky, Maine, Massachusetts, Minnesota, Missouri, Oregon, Rhode Island, Texas, Utah
• 79 public institutions uploaded artifacts
• By sector:– 37 four-year, including 8 research institutions– 42 two-year
These results are not generalizable across participating states or the nation in any way. Please use appropriately.
8
MSC Demonstration Year by the Numbers
• 10,948 pieces of student work were submitted [number of pieces of work approximates number of student participants]
– Students had to be 75% of the way to completion of institutional degree requirements
– 3,031 artifacts scored twice (28%) in order to measure inter-rater reliability
• 1,573 assignments were submitted[number of assignments approximates number of faculty participants]
• More than 190 scorers
• Three rubrics – critical thinking, quantitative literacy, written communication and option of civic engagement
These results are not generalizable across participating states or the nation in any way. Please use appropriately.
9
VALUE Rubric Approach Assumptions
• Learning is a process that occurs over time
• Student work is best representation of student motivated learning
• Focus on what student does in terms of key dimensions of learning outcomes
• Faculty and educator expert judgment
• Results are useful and actionable for improvement of learning (and accountability)
10
Inherent Challenge for VALUE: Navigating Methodological Complexity
12
LEARNING OUTCOMES
• State and Sector Coverage of Outcomes• Mean Scores by Learning Outcome• Score Distribution by Criterion• State Comparisons for Overall Mean Scores
VALUE & Validity
90%Participating faculty believed that the VALUE rubric = Useful tool for evaluating student work quality
VALUE & Validity in the Pilot Year
>80%
Sufficient range
Descriptors were understandable
Descriptors were relevant for making
judgments about levels of learning
VALUE Rubrics
16
~75% Dimensions encompassed the core meaning of the learning outcome
VALUE Rubrics
17
DEMONSTRATION YEAR PRELIMINARY DATA
18
CRITICAL THINKING – 2 YEAR
0%
10%
20%
30%
40%
50%
19
CRITICAL THINKING – 4 YEAR
0%
10%
20%
30%
40%
50%
21
QUANTITATIVE LITERACY – 2 YEAR
0%
10%
20%
30%
40%
50%
22
QUANTITATIVE LITERACY – 4 YEAR
0%
10%
20%
30%
40%
50%
23
WRITTEN COMMUNICATION – 2 YEAR
0%
10%
20%
30%
40%
50%
24
WRITTEN COMMUNICATION – 4 YEAR
0%
10%
20%
30%
40%
50%
• Methodological
• Philosophical
• Pedagogical
Nature, Implications of Complexity
Quantitative Terms Qualitative Terms
From “Qualitative Analysis on Stage: Making the Research Process More Public” by Vincent A. Anfara, Jr., Kathleen M. Brown, & Terry L. Mangione, Educational Researcher, October 2002 vol. 31 no. 7 28-38.
Validity Credibility, Transferability
Reliability Dependability, Confirmability
VALUE
• Comparing the validity & reliability of the VALUE process to standardized tests will always be an “apples” to “oranges” proposition.
• That said, establishing the validity (credibility & transferability) and reliability (trustworthiness & confirmability) of the VALUE process in its own right is a key priority for AAC&U.
Nature, Implications of Complexity
Purpose = Discuss Reliability in Relation to Inherent Complexity of VALUE
Scores (rubrics)
AssignmentsScorers
Order of Operations for Discussion:1. Reliability vis-à-vis Demonstration Year Data
• Discussion of preliminary analyses• Discussion of “in the weeds” plans for
investigating reliability, generalizability more deeply
2. Reliability & MSC’s Multiple Moving Parts• Scores• Scorers• Assignments
• % agreement and Intraclass Correlation Coefficient (ICC) test to assess agreement beyond chance
• ICC often used to look at ordered categorical data (data that “behaves” as ordinal data) –like the VALUE rubrics
• ICC (1) – used when do not have consistent/fully crossed raters for all artifacts
• Most conservative/smallest of ICC tests
Preliminary Pass at Interrater Reliability
Commonly Used Cutoffs for Qualitative Ratings of Agreement Based on ICC Values:
<.40 = Poor
.40-.59 = Fair
.60-.74 = Good
.75-1.0 = Excellent
What does the ICC value mean in very “lay” terms?
An ICC(1) Average Measures Score of .449 means that 44.9% of the variance of the mean of the raters is “real” – not simply attributable to chance.
Start w/ the Big Picture
ICC(1) for Average Total Scores
.59 .63 .62
QuantitativeLiteracy
Written Communication
Critical Thinking
Civic Engagement
.62
Scores (rubrics)
AssignmentsScorers
• More robust interrogation of reliability data• Fall 2016 study with Gary Pike (IUPUI) more
thorough consideration of IRR, appropriate tests (e.g., Gwet’s AC1)
• Broader generalizability study
• Look at literal source of scores – rubrics themselves
• Consider the possibility… not all criterion, performance levels are created equal from a reliability perspective
Scores (Rubrics)
• Mine the existing data to identify outliers, more “expert” scorers, disciplinary differences
• Enhanced training (in what ways?)
• Enhanced criteria for participation (how stringent?)
• Different approach to scoring (e.g., scoring to a “4”, etc.)
Scorers
• Mine the existing data (part of our planned work for 2016-2017) – what should we be looking at?
• Improve assignment design – how?
• Establish – and enforce? – assignment design parameters. What would that look like?
Assignments
A Careful Balancing
Act
Methodological Philosophical/
Pedagogical
Lessons Learned- Overall
40
Lessons Learned
• Increased focus on campus capacity
• Assessment as a high impact practice
• Importance of assignment for demonstration of learning
• Faculty development and collaboration and engagement for learning
• Increased attention to equity – based on data
41
Lessons Learned - MSC Structure, Organization, Leadership, Process
• SHEEO agency – coordinating role, financial support, outreach to legislators, governor, accreditors, K-12, business leaders, public
• Institutional leadership – Provost, CAO, Assessment & IR Directors, Faculty, champions for the work
• Statewide Assessment Council – voluntary association of experts, inter-institutional communication, technical support, collegial support
• Statewide & campus convenings, assignment design work shops; norming calibration training
42
Lessons Learned – Campus Capacity and Engagement
43
Level Learning Outcome9-State
Avg Score Connecticut Avg Score
4-Y
ear
In
stit
uti
on
s Critical Thinking 2.01 1.99
Quantitative Reasoning 2.33 2.12
Written Communication 2.53 2.53
*Rubrics – 4 point scale
Reflections on the Pilot Year
Learning Outcome & Faculty
Faculty (N)
Number of Artifacts
First Year
Soph Junior SeniorCCSU Total
MSC Total
Critical Thinking (33 Majors)
12 16 21 58 130 225 119
Quantitative Reasoning (19 majors)
5 0 6 29 82 117 78
Written Communication (28 Majors)
13 13 19 62 97 191 87
Grand Total(45 majors – 75% of majors)(27 faculty – 45% of dept)
27 29 46 140 318 533 283
Reflections on the Pilot Year
Questions and Follow-Up
Julie Carnahan, [email protected]
Kate McConnell, AAC&[email protected]
Terrel L. Rhodes, AAC&[email protected]
46