Evaluating uses of Learning Technology

Download Report

Transcript Evaluating uses of Learning Technology

Evaluating uses of
Learning Technology
Martin Oliver,
MST/London Knowledge Lab
Overview
Some general definitions and history
What is it we’re discussing, exactly?
Issues for evaluating learning technology
What problems do people face when doing this?
Data and distance
What can we find, and what does it tell us?
Tools to support people (very brief)
What resources exist to support people?
The ELT toolkit project (A detailed look at a tool)
How can tools like these make a difference?
What can we learn from them?
Conclusions
Prelude
This first section’s really about orientation
Think about your experiences of evaluation
If you can think of a positive experience:
What was it that made it good?
…even if you can’t think of one:
What was it about your experiences that were
bad?
Capture these and revisit them later
What do I mean by
evaluation?
A contested term
Judgements about the value (benefits) and
worth (costs) of something
A way of describing something (ethnographic)
What evaluators do…
What I don’t mean
Entirely personal judgements like reviews (or
checklists), without data collection
Assessing student learning
…entirely valid things, but not what I’m talking
about
A brief history of educational
evaluation
The beginning…
A tradition that grew from measurement theory
Firmly rooted in the experimental method
Educational interventions as things applied to
populations
…and the backlash (c. 1970s)
Alternative traditions rejected this approach
Illuminative, ethnograpic, naturalistic
approaches arose
Sought to re-define what counted as ‘valid’
evaluation, but politically weaker
This pattern still evident
US legislation endorses controlled experiments
A paradigm war?
Certainly two opposed traditions
Battle lines seem drawn around methods
…but Hammersley suggests it’s more
about philosophy
Logical positivists looking for stable,
controllable interventions - even if they’re doing
qualitative grounded theory
Relativists looking to interpret what’s happening
and make recommendations based on personal
judgements
Points out that many people are actually
eclectic, rather than hidebound (“principled”?)
A third way
Patton: utilization-focused evaluation
Most commissioned reports are never read
A good evaluation isn’t one that’s
methodologically rigorous - it’s one that helps
people act (make decisions)
Principle of designing for “intended use by
intended users”
No good working for people who won’t act choose a different audience
No good working for those who are powerless
to act - instead, influence those with power
(There are ethical issues here…)
Repositioning evaluation
Evaluation positioned as a social, political
activity, not as value-free science
Emphasis on rhetoric - persuading an audience
(understood not as a ‘type’ or ‘role’ but as a list
of names)
‘Stakeholders’ - whose voice will be included in
this process? What authority, if any, will it have?
(Will they just provide data, or help frame the
study, interpret, present, etc.?)
Recognition that the evaluator has a stake in
the process too - reflects on their credibility and
integrity
Repositioning evaluation
Also: treating evaluation as an educational
intervention - “process use”
What can those involved learn? From
negotiating its scope, gathering data,
contributing to analysis, debating
interpretations…?
Looking at opportunities for feedback - ongoing
interventions for improvement (action), not just
summative judgement
(cf. Feedback in learning)
Creating opportunities for dialogue between
‘stakeholders’ - socially constructed
understandings, involving multiple perspectives
Revisiting your experiences
Review the list from the start
To what extent are the “good” features
consistent with utilization-focused
evaluation?
Are the “bad” experiences linked to a
particular approach?
Are any of you unwitting utilizationfocused sympathisers?!
End of the first part
Overview of relevant issues from
educational evaluation
Intended to provide a common ground
and insights into the specific problems
of learning technology
Also directly relevant to any other
evaluation you do
Next, on to our specific concern…
So what about Learning
Technology?
This section will look at the things that are
distinctive about evaluation in relation to
learning technology
Evaluation within this context echoes wider
shifts in educational evaluation
…although usually a few years later…
Same contestation between paradigms
Evaluating Learning Technology
Some novel features
Large number of practitioner-researchers
No formal training as evaluators
‘Common sense’ evaluators - no theoretical
foundation to their work
Large number of funded projects are told to
evaluate their work
(esp. after TLTP phase I - lots of
development, no information about its
value!)
Many have to evaluate the project that pays
their salary, sometimes with an external ‘check’
Evaluating Learning Technology
Same power-laden confrontation between
paradigms
Qualitative, interpretative perspective
common in action research/practitioner
studies
Quantitative, positivistic perspective
championed by policymakers and
Evidence-Based Practice
Say “what works”, not explain why (“easy
answers” for funders)
Hierarchy of evidence – from RCTs to
qualitative methods and GOBSAT
Evaluating Learning Technology
Particular problems with comparative
studies
A difficulty for all educational evaluation
If one condition is believed to be better, can you
justify withholding it, ethically?
If outcomes are affected by teaching, changing
outcomes changes what was learnt – so can’t
use same assessment as no longer appropriate
(Constructive Alignment)
Particular difficulty for Learning Technology
What is “E-Learning”, “Blended Learning”, etc
anyhow?
Evaluating Learning Technology
“Is e-learning better than traditional forms of
learning?”
What we mean by e-learning today isn’t what we meant a
year ago
It’s not the type, it’s the specific instance – is this well
designed?
It’s not what it is, it’s what you do with it – are you using it
well?
What exactly is “traditional” learning? Do we really want
to assume this is a stable point of comparison?
“No significant difference” phenomenon – tends to be
different, not “better”, unless study is designed to
measure what e-learning does well
“Are books better than other resources for
learning?”
Evaluating Learning Technology
The difficulty of attribution
One innovation amongst many: postcompulsory education is riddled with new
initiatives; what’s the root of any particular
change?
False negatives: students learn differently but
cover this up in order to perform on ‘normal’
tests (learn new and old forms of knowledge)
False positives: technology a symptom of a
wider change in attitudes or practice (is
technology the symptom or cause of widening
access in post-compulsory education, or are
both symptoms of something else?)
Coincidence: groups may just be different (even
if randomly created, but particularly if cohorts)
Evaluating Learning Technology
What kind of comparisons can we draw?
People can and do undertake comparative studies
Tend to compare preferences for one intervention or
another
Sometimes measure group performance against
some invariant test (e.g. standardised exam) – point
of reference
Typically try to ‘control out’ things interpretative
researchers find interesting (influence of teacher,
etc.)
Can be informative and useful if you’re treating
this as a case study – an insight into specific
performance, rather than building a general law
The problem of impact
A common request:
“Establish the impact of this new form of
teaching”
Remarkably hard to answer
A brief exercise (2-3 minutes):
List ways in which introducing a new form of
technology can have an impact
Consider a variety of roles/people
What kinds of evidence might you look for in
each case?
The problem of impact
The TLTP III EFFECTS project
National initiative to see whether accreditation
and staff support was an incentive to lecturers
to adopt new technology
Required evidence of selecting, planning,
implementing, evaluating, disseminating
technology use
Now established as a SEDA award (PDF-ELT)
External evaluation: establish the impact of
the project
How could we make sense of this?
The problem of impact
Multi-layered model
What was the impact on learners?
What was the impact on lecturers?
What was the impact on the
organisation?
What was the impact nationally?
The problem of impact
Initial plan was different evaluators for
different levels
Lecturers evaluate impact on learners
Local project team evaluate impact on lecturers
and organisation
Project evaluators evaluate national impact
Worked ok, but…
Lecturers hadn’t the time and found this hard
‘Evaluation’ the weakest outcome in
assessment across all sites
A picture with lots of holes
The problem of impact
As an example: impact on staff
Evidence of change in teaching practice
(observation, documentation)
Evidence of changed role
(Promotion, involvement in committees)
Evidence of change in career direction
(Publishing educational research)
Evidence of change in attitudes
(Self report, changed use of discourse)
Detailed picture, but not amenable to
widespread study; relied on snapshots
The problem of impact
To summarise:
Easy to find evidence of change, hard to
establish what it means for our study
Hard to draw out general conclusions when the
‘focus’ is so ill-defined
Much can be said that’s relevant to local
practice – evidence of some kind of impact is
almost unavoidable (but is it the right kind…?)
Need a position to interpret this against
and good rhetoric to present this to others
http://ifets.ieee.org/periodical/vol_3_2002/v_3_
2002.html
End of the second part
This section’s focused on the specifics of
evaluating learning technology
Highlighted issues of design and
interpretation facing studies in this area
Raised a number of general concerns,
including disposition and approach of those
asked to evaluate
Next section will look at the issues of
gathering data
Data and distance
The previous section has explained why it’s
hard to establish impact
This section will look at some of the
problems of getting hold of data at all
Issues arise in relation to the point of new
technologies
Many introduced to increase flexibility
Many claim to support new forms of learning
What can we gather as evidence and what can
we infer from it?
Data and distance
“Informating” (Zuboff)
Computerised activities make things
explicit and generate information
Such information is, potentially, data
Discussion archives, use logs, site hits,
etc.
A helpful source of information?
Ready for processing - in electronic form
Only a partial account of what took place
Data and distance
Can we see what we need to?
Chris Jones: ethnographic study of a course
with online collaboration
Incident of ‘cheating’ observed face-to-face
Calls into doubt the veracity of ‘easy’ data
‘Distance Pedagogics’ (Peters)
Lecturing at a distance (e.g. video link) is still
lecturing and needs no new pedagogy
Private study may be on-campus but provides
flexibility and choice to students and raises
issues familiar to traditional distance educators
Similar issues here: public or private?
Whose context?
Data and distance
Importance of triangulation
Each part reveals an element of the
wider picture
Interpreting multiple sources of data
reassures and (potentially) explicates
However…
Ongoing problem: most of what’s
important with learning is private, so how
can we learn about this?
Data and distance
Old methods
Travel for observations - time intensive; fine for
cases, but less good for general conclusions
Travel for interviews, or interview by phone
Surveys etc
All are opportunities, but access becomes
a serious issue (and no guarantees data
will be provided)
Data and Distance
Old methods in new formats
Online survey - higher response rates, but
caution about missing out those least happy
with technology (usually an important group)
Online interviews/focus groups - readilycaptured data, but different skills and pace
required; more thoughtful, less spontaneous; if
open (rather than selected), vocal minority an
issue
However, mostly self-report; what else can be
accessed?
Data and distance
New(er) methods
Traces - hit logs: ‘dumb’ data (about access,
not use or intention) that needs interpretation
Discussion archives - access to exchanges that
are fleeting in traditional settings
Can raise ethical issues
‘spyware’ for data
status of comments (as permanent, as data) in
online discussions - data protection act
Can be easy but inappropriate to gather lots of
data
Summary of part three
General issues
Getting the data you need is harder, as it’s
private and distributed
Online data collection methods harder to
control, which may raise questions about
interpretation
Questions of interpretation raised: ethical
justifications, relationship to context(s)
Next section will look at tools designed to
help with this
Supporting people who
evaluate learning technology
This section builds upon a problem
mentioned in passing in EFFECTS
evaluation
A pressure to evaluate technology use
Lots of conceptual and methodological
issues
No support or training for teachers
Supporting people who evaluate
learning technology
The response
A plethora of tools to ‘fix’ this
Assumption that a sensible teacher with
a bit of information and guidance will get
through ok
Relevance to this:
Raise awareness of tools like this
Highlight ‘good’ (sanctioned) approaches
to practice
Tools to support evaluation
Existing tools support data analysis
SPSS, NVivo, etc.
…but only if you know what they should be
used for
Another collection of tools focuses on
evaluation design
LTDI Evaluation cookbook
TLT Flashlight project (US)
MEDA Evaluation tool for training software
ELT Toolkit
Different kinds of tool
MEDA - a handbook of questions
Flashlight - a database of question
Works on the assumption that “very different
educators need to ask similar questions”, using
surveys
LTDI - a cookbook of methods
Assumes educators can choose a topic but
might need help with methods
ELT toolkit - tries to do both
http://www.elt.ac.uk/materials.htm#evaldiss
The evolution of the ELT
toolkit
The brief: produce something that will
help practitioners evaluate in spite of
the difficulties
The initial idea: a structured
walkthrough supporting study design
Selection of methodology
Selection of methods (guided by
methodology)
Selection of data analysis methods
The evolution of the ELT
toolkit
It didn’t work
First study: couldn’t even get to using the
toolkit
Participant didn’t know what they wanted to
know
Goals kept shifting
Discussion led to re-framing… endlessly
Patton’s “process use”
Being involved in evaluation design was
educating for the participant but this didn’t
actually help them carry out a study!
The evolution of the ELT
toolkit
Revision of the tool
Introduced new steps to address context
Identification of stakeholders (individuals)
List their concerns and turn these into
questions
(Existing three steps: methodology, data
collection methods, data analysis methods)
Approaches to communicate findings to
audience(s)
Explicitly framing this as a social process
Similar to Draper/TILT’s “inner” and “outer”
steps for evaluation design
The evolution of the ELT
toolkit
Developed a paper-based version of the
tool
Tested it
It worked fairly well, but took a long time
Received funding from the JISC
Implemented as the online Evaluation of
Learning and Media Toolkit
(Incorporated another tool for curriculum design
- badly)
http://www.ltss.bris.ac.uk/jcalt/
Expanded functionality
The evolution of the ELT
toolkit
Three main sections
Evaluation planner
Evaluation adviser
Evaluation presenter
Plus things we didn’t want lost
The methodology section JISC didn’t like
References and links
Within each section three types of activity
Tell us something (free text entry) - usually
context
Make an open choice (list of suggestions with
‘other: please specify’ option)
Enter data and a model recommends things
The evolution of the ELT
toolkit
Data entered pulled together in a final
report
Printed off as 2-4 page evaluation plan
Summarises decisions made
Captures contextual information
Presents this in an ordered way
Amenable to sharing with others (managers,
funders, research assistants) or using as an
outline workplan
Simple idea; huge success with users!
Further development: option to share plans
An example
Forming questions had been a
problem
Developed an activity to support
question framing
Start with concerns
Rephrase these as a series of questions
Combine or contrast different types of
questions
Pick the one that seems most useful
An example
Concern: student learning
Exploratory questions
What do students learn? Who learns best with
this resource? How do students use it? What
do they think about it?
Comparative questions
Do some students use it differently to others?
Does this group perform better on tests than a
group that doesn’t use it?
Measurement questions
How long do they use it for? Do they use it all?
If students do better on tests, how much by?
How many complaints where there about it?
An example
Negative questions
What was wrong with it? Why did students
dislike it? Did it hinder their learning? Who
found it hard to use? What problems did it
cause for people?
And then, combined questions
How much better did these students do than
those? Why did some students like it more than
others? What led to problems arising for
students?
And finally - select the one question (ok, at
most, three) that your study will seek to
answer
Evaluating the Evaluation
Toolkit
Had to do it to ourselves…
Used the toolkit to do our evaluation plan
Implemented it
Published in JCAL 18 (2), 2002
What we learnt
The usability of the tool was fairly poor
Took about 3-4.5 hours to do a full plan
Experts thought this was wonderful
Novices thought this was far too long
Editing shared plans one way to reduce this
Evaluating the Evaluation
Toolkit
A success?
It did its job - even complete novices produced
credible plans
In addition - experts prompted to think about
methods they hadn’t previously used
Suggests people were learning from this as
well as ‘just’ designing studies
However…
“Comfort zone” - if novices wanted to do a
survey and it wasn’t recommended they’d override the list - needs to be “challenging enough”
No evidence of impact on practice - did anyone
implement the plan? (Longitudinal studies)
Summary of part three
Given the complexities of evaluating
learning technology, it’s no surprise people
need support
Tools are regularly developed as a way of
providing this
They can be interesting in their own right
Impact of ELT Toolkit on design
However, hard to judge their own impact
But also interesting as a representation of
“good” practice
Summary of part four
Toolkit designed as a stand-alone resource
…although works best when introduced in a workshop,
preferably with peer discussion
Whether used or not, useful as a way of
highlighting issues and suggesting structure
Provides outline of a decision making process
(stakeholders, questions, data collection, analysis,
presentation) – useful for planning, but also for staff
support, training, etc.
Highlights complexities – each activity relates to an area
worth thinking about; sensitises to issues
Can be educational to look at tools like these,
whether or not you follow them
Where does this leave us?
A recap
An overview of the issues is evaluating learning
technology
Overview of themes from wider educational evaluation
Specific issues facing evaluation of learning technology
In addition
Examples of tools that might prove useful
A detailed look at a particular tool – the ELT toolkit – to
illustrate how it can help (at least with design; no
evidence for practice)
Something that places control of evaluation back in the
hands of practitioners
Where does this leave us?
My hope… any of the following:
Support you in being better informed
participants, commissioners or readers of
evaluation studies
If you’re studying this yourself, provided
ideas about how to do it
Sources of evidence of impact
Methods in relation to ‘absent’ students
Sounded notes of caution about how to
interpret and present studies
Causes, comparisons, interpretations etc