Examining the Reliability and Validity of Teacher Candidate Evaluation Instruments

afternoon, everyone. Thank you so much
for joining us. We’re excited to be
hosting this webinar today. I’m Marc Brodersen, a senior
researcher from the Central Regional Educational
Laboratory or REL Central at Marzano Research. We’re very happy to have
everyone with us today. Today, we’re going to provide
an overview of the need for developing standards for
reliable and valid teacher candidate evaluation
instruments. We’re also going to
hear from leadership from several educator
preparation programs in North Dakota about their experiences
in developing a teacher candidate evaluation
instrument and the steps they’ve taken to examine its
reliability and validity. We’re very excited to
have representatives from the Council for the
Accreditation of Educator Preparation and the North
Dakota Association of Colleges for Teacher Education. As I mentioned, this webinar
is being hosted by REL Central. REL Central is one of
10 regional educational laboratories that is funded
by the Institute of Education Sciences in the U.S.
Department of Education. REL Central serves a
seven-state region, including Colorado, Wyoming,
Kansas, Nebraska, North and South Dakota, and Missouri. As with all of the
RELs, we are charged with providing
technical assistance and conducting
research on priority topics for stakeholders in
our region and nationwide. All of our work is directed
by our stakeholders and organized under research
partnerships or alliances. This webinar is being
conducted under the umbrella of our Educator Pipeline
Research Alliance, which currently has
projects focusing on educator preparation,
evaluation, and mobility. We very much hope
that this presentation is of interest to everyone
that is participating today. Now, I’m going to hand
this over to Steve Meyer from REL Central, who will
introduce our presenters. STEPHEN MEYER: Hey. Good morning, everyone. This is Steve Meyer. I’m with REL Central
and RMC Research. I’d like to welcome you
all to today’s webinar and am pleased to introduce
our producers– our presenters, excuse me. First, we’ll hear from Gary
Railsback of the Council for Accreditation of Educator
Preparation or CAEP, which as most of you I
expect to know is the national organization that
accredits educator preparation programs. Gary is a vice
president at CAEP, where he has oversight of
the accreditation process and the accreditation team. He has a wide
range of experience with accreditation through his
work at previous institutions and at CAEP as a trained site– lead site visitor and inquiry
brief lead site visitor. He’s performed site
visits in 15 states and was a reviewer with
the California Commission on Teacher Credentialing and
the Oregon Teacher Standards and Practices Commission. Gary is going to tell us
about the CAEP guidelines for establishing the validity
and reliability of assessments created by educator
preparation programs and some of the challenges that
programs have faced in meeting CAEP guidelines. Following Gary, we’re going to
hear from three presenters who each represent an
educator preparation program in North Dakota. So, Sarah Anderson is
an associate professor and accreditation coordinator
in the Division of Education at Mayville State University. She’s a former high
school special educator and is now teaching graduate
and undergraduate pedagogical courses. Her research interests
include teacher appraisal for continual improvement, poor
instruction progress monitoring for RTI, and effective
instruction and interventions. Stacy Duffield is a
professor in the School of Education at North
Dakota State University. She teaches graduate and
undergraduate teacher preparation courses. Her main research
interests include teacher preparation
and assessment, and middle-level education. She’s been involved
with the work of the Network for Excellence
in Teaching since 2010, helping to further
the development of valid and
reliable instruments for teacher preparation. And last, Alan Olson is a
professor and assessment coordinator in the School of
Education and Graduate Studies at Valley City State
University in North Dakota. He teaches graduate and
undergraduate preparation coursework and
he’s been involved with common metric
assessment efforts through the Network of
Excellence in Teaching, as well as the North Dakota
Association of Colleges of Teacher Education. And each of these
presenters is very familiar with the CAEP Standards,
having served as CAEP reviewers over the years. And, for the past few years,
Sarah, Stacy, and Alan have been working as part of a
collaboration of the 12 member institutions of the North
Dakota Association of Colleges of Teacher Education, or NDACTE,
to develop a Student Teacher Observation Tool. And so, they’re going to
tell us about the process they used to develop the
tool and how they examined its reliability and validity. All right, in the
next slide, I’m going to talk about
the webinar objectives. These are our
objectives for today. The webinar is really
designed to familiarize you with the CAEP requirements
for demonstrating reliability and validity of teacher
candidate evaluation instruments and focusing on
those developed by educator preparation programs. And then we’ll also be
talking about approaches for examining and supporting
the reliability and validity of these instruments. All right, next slide. This is just a set
of resources that have informed the presentations
today and some things that you may want
to refer to later. There’s a link to
the slides available, I believe, in the chat box. And this is also something that
is downloadable via the IES website. So just a quick overview of
what we’re featuring here. The CAEP Standards themselves
are the first document that specify the expectations
for CAEP accreditation. Next is the CAEP
evidence guide, which has guidance around the
use of data and evidence in educator preparation and also
in the accreditation process. And there’s also
guidance in there related to data quality,
data collection, and data analysis and specific
guidance related to the topic we’re
talking about today, which is the reliability and validity
of candidate evaluation instruments. Next on the list is a report
about a teacher evaluation rubric that’s used in Texas. And, like the candidate
evaluation tools that we’re discussing
today, this is a rubric-based
evaluation tool. And the report that
we’ve linked here illustrates how evidence
of reliability and validity may be generated. And then, the last two
bullets on this slide list a couple of
resources that were developed by our presenters
from North Dakota. These are reports
of studies that have been used to
examine the reliability and validity of the Student
Teacher Observation Tool, which we’ll hear about
in a little bit later. All right. So next, we’d just like to
take a couple of minutes to open up the Q&A box to hear
from you all –this question. So we’re interested, just to
kind of kick off our thinking about this discussion,
in why you all think reliability and
validity are important when evaluating teacher candidates? So, I invite you
all to use the Q&A box to weigh in with a
response to this question. And we’ll take a
couple minutes– maybe about a minute or
so– to allow you to do so. So again, if you hover at
the bottom of your screen, a black bar will
appear and you’ll have the opportunity to click on
the Q&A box and your question. OK, great. So we’ve got a
few thoughts here. So one response– let
me just kind of convey what people have said. So, “Why is reliability
and validity important when evaluating
teacher candidates?” So one is for quality
assurance to know that we’re assessing the right things. So that’s the idea
of validity really is that you’re sort of measuring
what you intend to measure. It’s important to ensure that
we have a clear understanding of how we define
“ready to teach” and ensure that we’re
measuring that and only that. So again, like
focusing the content of what you’re measuring to
be what you intend to measure. Validity is a potentially
important predictor of effectiveness
after graduation. Reliability and validity
help ensure that everyone is measured the same way. And then knowing it’s
information that helps know that you’re measuring
candidates consistently. It helps identify teachers who
are most likely to be actually effective in the classroom. So those are just some of
the comments we’re getting, I think that are all in the
spirit of how we collectively think about reliability
and validity. So, equity is another one. Making sure the instrument
is measuring the construct accurately. So I think I’m
seeing that we’re all very much on the same page
about where this comes from. When we hear from Gary
Railsback from CAEP, we’re going to hear a lot
of these priorities echoed in his presentation. And I think that’s where,
in any measurement, you know, we’re trying to make
sure the measures are doing what we set out for them to do. So with that, I think
I will turn this over to Gary from CAEP. And as I said, he’s going
to talk about the guidelines for establishing the reliability
and validity of assessments created by educator
prep programs and some of the challenges they faced. And we’ll be hitting more on
the importance of these topics in his talk as well. So let me turn
that over to Gary. And Gary Railsback from CAEP. GARY RAILSBACK: So thank
you all for participating, and especially to Marc and
Steve for setting this up. Validity and
reliability are really important as EPPs (educator
preparation programs) make recommendations
for states for who’s going to get a
credential and who’s going to be in a classroom. So I just wanted to start off
with some general reminders. When EPPs start working
on their self-study, we ask them to think about
these kind of assessments– performance assessments– in
really two broad categories. One is a proprietary assessment. So that would be one that’s
been developed by somebody else. Maybe it was purchased by
an individual or the state. And the validity and
reliability should have already been done on that, but we don’t
ask people to provide that. The self-study just
says “if available.” So what we’re going
to focus on today are really the larger
category, where you feel like those proprietary
assessments don’t cover all of the standards and you
want to do some things that are customized for your campus. They’re usually
developed by faculty. So we’ve got some examples here
of proprietary edTPA, Danielson Framework. And we’re going to
spend most of our time on the latter category– EPP-created assessments. So I get lots of
calls about this, and people want to know
what is the magic number. There is no minimum. There’s not really a maximum,
but having been a site visitor, I would keep that
number single-digit. A lot of EPPs only
have two to three. So especially on the
initial programs, we’re not looking
for you to develop, like the old NDACTE policies,
which said six to eight. The ones that are most common
are a clinical practice observation, sometimes
dispositions, possibly a unit plan. It could be a portfolio. Those are the ones that
are really most common. Starting fall of
’19, we’re also going to be moving into
doing site visits for advanced-level programs. And if you’re not
aware of that yet, I’d encourage you to
look at our website to see if your programs
actually fall into the scope. But essentially what people are
doing at the advanced level is they’re developing some
instruments that might be related to Standard A.1.1, which
are the six broad professional skills, maybe content-specific. And then if it’s in a
full advanced program and it has clinical practice,
maybe something for there. They could also be doing
something in dispositions. So as we think
about validity, we want to make sure that
whatever assessment we’re using is actually measuring what
it’s supposed to measure– that it’s not measuring
something else and then we’re using it for
some other purpose. So I did cite here out of
our handbook, our definition. So “it’s an operation, a test,
or some other kind of measure that we’re really sure
that’s what it’s measuring.” And so to help people
think about that, CAEP developed an
instrument that’s called the CAEP
Evaluation Framework. Sometimes, it’s called
the CAEP Evaluation Tool. Sometimes, it has a
whole bunch of words in the title beyond
that, but essentially we’re going to go through that
and kind of look at what CAEP is looking for. The way the rubric is
put, the left column is below sufficiency,
the middle is sufficient, and the right is for
the eager beavers that want to get the highest
grade, but it’s not required. So that’s above sufficiency. So when we look at the
CAEP Evaluation Framework, there’s seven different
sections of it. And section four
and five are the two we’re going to look at for
validity and reliability. So for validity
below the sufficient, we’re finding places that there
was no plan at all to establish any kind of validity. And we accept multiple
kinds of validity, and this will be talked about
later in the presentation as North Dakota
gives their example. But we need to
have some validity. Is it content validity? Whatever. We want to make sure that
the instrument before it was actually used was piloted. And we want to make sure that
it was done by more than one or two people. So the common way
this used to be done is two or three faculty
teaching in an area might develop an instrument,
and then they just run with it. And so we’re trying
to move people to the middle of the
sufficiency and involving some external stakeholders. And then the last
one is to develop some steps to do this research. And if it’s on the
below sufficiency, there wasn’t evidence
of that in the site– the self-study, or any
other materials given. So now as we move toward
the middle of the column here in the sufficient, we
find that there was a plan. We know how the EPP
developed validity. We know what kind of
validity it’s looking at. Most people do content validity,
but that’s not a requirement. If it was new or revised,
there was some kind of a pilot, and that’s clearly delineated
in the description. The EPP details its
current process or plans for analyzing and
interpreting the results. And then whatever they
did follows generally accepted research methods. Now in the right column,
some people look at that and think, well, that’s
what I have to get. And that’s just saying this
is like way beyond or above, so you don’t have
to do these things. One of them is reporting
the validity coefficient using certain types
of validity, which are a little more complex,
like predictive validity. But again, those are
above sufficiency and they’re not required. OK, so we’d like to
take a question here. R. MARC BRODERSEN:
All right, yeah. So we’re going to open up
the Q&A box again and so we’d like to hear from you. What do you think are
some challenges you think EPPs face when developing
teacher candidate evaluation instruments and examining
the reliability and validity? OK, so we’re starting to
get a few responses here. So some of the challenges
that are being raised here– one common– is that
time constraints can be a constraint. GARY RAILSBACK: Absolutely. R. MARC BRODERSEN:
We have another as having sufficient
number of faculty to help with the process. Another challenge might be
to define the constructs that are to be measured. And I think related to that,
we have another comment here about measurement
without the content and maybe how those
two can be aligned. We have another one that’s
funding and time for field testing. And I think we’ll be addressing
a lot of these issues as we move forward
in the presentation. How about we’ll do
this last one here. “Getting stakeholders from
the public schools involved might be a challenge.” So thank you everyone
for your comments. And we’re going to hear
some more from Gary. GARY RAILSBACK: All righty. So what we’re going
to go through now are some examples of
EPP-created assessments that have been used in
the accreditation process on both initial and advanced. So again, we’re looking at
that evaluation framework. And on the left side
on below sufficiency, one of the things that we look
at is curricular validity. It “refers to the
extent to which the content of the assessment
matches the objective of a specific program.” So you know, as a
site reviewer, I saw a lot of places trying
to give course grades. And there’s just not a lot of
strength in a course grade. So, I would encourage people
not to use course grades. Even if it’s a specific course– that’s somewhat problematic. End-of-course
assessments might be a portfolio, it might be
a term paper, or whatever, but there’s no discussion
about validity and reliability. So those are some
examples that come out on the below sufficiency. Some other examples–
face validity. It might be a
dispositional data. A lot of EPPs develop a
dispositions instrument where they ask
candidates, or faculty, or the clinical faculty
out in the schools to rate a candidate. And it might be very qualitative
or it might have no analysis. Another one is
candidate interviews, which is a great thing to do. But they’re often done
without an instrument and there’s no analysis. Same thing for portfolios. In the middle column,
the ones that usually reach the sufficiency. Most people are doing
content validities to make sure that
whatever is in your rubric that you’re assigning and
you’re using for assessment. And the common ones– lesson plans. A lot of teacher ed programs
have a unit plan, teacher work samples, portfolio assessments,
observation, capstone thesis. Or the better one would
be a problem-based project that would be in partnership
with a school district. Over on the right
side again, this is above sufficiency– so not
required, but it’s predicting. And that might be doing
something in preservice, trying to predict whether
that person is going to be a qualified teacher
a year or two down the road. And then comparisons
of candidates in education programs
with either other higher ed institutions, or it
could be even across campus if it’s a content course. So for us, validity
is that it can be supported through
evidence of the following. There’s some agreement
among reviewers that the description or
the narrative about it is measuring the same thing. There is expert
validation of performance, or looking at artifacts,
expert validation of the items in an
assessment or rating form, and then the measure’s
ability to predict performance in a future, and
that’s predictive, which is on the right side. So there’s a number
of different ways that EPPs are going about
developing content validity for some of these
common instruments that we’ve talked about. My first couple
of weeks at CAEP, I got some interesting
emails from people. And one email said, does CAEP
require the Lawshe method? And I said, “no.” And then a week or so
later, I got another one. And somebody said,
CAEP does not allow you to use the Lawshe method. And I don’t know where
these rumors started. We have used the Lawshe
method at our CAEP conferences and a lot of other
places, but we don’t require you to use a
certain kind of methodology. The Lawshe method
has been used a lot because it asks you
to go out and have external stakeholders evaluating
whether this is beneficial, but it’s really not. In the Lawshe article, the
little equation there– if algebra scares
you, that’s just the equation that’s used to
conduct the content validity ratio. So, and I actually did this
at Azusa Pacific before I left last year. You take some
P12-based educators. If it’s content,
maybe faculty members, P12 administrators, leaders,
candidates, partners, parent advisory boards. In other words,
you’re getting people that are familiar
with what you’re doing and you’re asking them. And then you can do
this in a survey. You can send it
out electronically. And you’re just asking
people, is this task that we’re asking
somebody to do, is it essential for whatever the
task is of the job– classroom teaching, school counselor,
school administrator, whatever. Is it essential, is it
useful, but not essential, or not necessary? And then you just determine
how many people that you asked. And if you have a
lot of panelists, anything beyond 50
percent is most likely going to get you a
content validity ratio. So you don’t have to go out
and hire a psychometrician or have somebody from your
psych department do this. It’s just going out and
working with your advisors, your advisory council, or
whatever that might be. So our definition of
reliability is “the degree to which test scores for
a group of test takers are consistent over repeated
applications of a measurement and are inferred to be
dependable and repeatable for an individual test taker.” Now, if EPPs are developing
two to three assessments, one person may not
take it more than once. So we’re not saying
that you have to do it, but as you look at it
over the three cycles, you might be able to tell
a little bit about that. The next slide talks
about reliability. And so again, we’re going
back to the CAEP Evaluation Framework. On the left side, what
makes it below sufficiency is the “plan to
establish reliability does not inform reviewers
whether reliability is being investigated or how.” So the reviewers don’t really
know what was going on. The “described steps do
not meet accepted research standards for reliability
or that there’s actually just no evidence.” What most people
are doing is, when we go to the middle
column, is they’re doing inter-rater reliability. So let’s say it’s an
observation instrument. They’re having a candidate
observed by a classroom teacher and a faculty member. And they’re looking
at those two scores to determine what the
inter-rater reliability is. It can be supported by multiple
raters of the same event. And it doesn’t have to be live. One of them could
have been on video. We’re looking for stability
or consistency of ratings over time and evidence
of internal consistency of those measures. So here’s a little example
that we put together to kind of think about this
supporting the reliability. First of all, you start
with content expert– some kind of feedback
on that instrument. Then you get some feedback. You clarify it. You may change it a little bit. You get content experts,
developers of the measure to review and seek feedback. Then you pilot and
you examine that data. So common challenges
EPPs’ staff face in meeting the CAEP Standards. So if you look at
the CAEP Standards, the component
we’re looking at is 5.2, where it asks that
your instruments are valid and reliable. So most of the
time, as I’ve said, there’s going to
be two or three. And they might be for Standard
1 about content or pedagogy. They might be in Standard
2 about something about clinical practice. They might be in Standard 3
about dispositions or something that’s non-academic. And so what our teams do is they
look at all those instruments– two or three of them. And then they say, do
the majority of them– the new handbook starting fall
of ’18 moved to 75 percent– do 75 percent of them meet
the sufficiency criteria? So if you go back to my
slides in that middle section, that’s really what
we’re looking for. And what we find
in most cases is they haven’t done
content validity, or it was done by
two or three people, or we don’t really
know how they did it. And as far as the
inter-rater reliability, which is the most
common method, we don’t find much
evidence of that. And that’s a change from both
NCATE and TEAC to actually say, you’re making a
consequential decision. How can you demonstrate that? The other thing I’d say
about Standard 5 of the five standards, it’s the one
that most institutions had trouble meeting. OK, I’m going to turn it over
to the North Dakota group. STEPHEN MEYER: Great. Thanks so much, Gary. So I think we’d like to
take a few minutes here to see if folks have questions
related to Gary’s presentation before we move on
to North Dakota. So again, I’ll invite
you all to use the Q&A box to submit some questions. So one question, Gary,
was around the last thing that you said about Standard 5. Apparently, there
was an audio problem for at least one person. GARY RAILSBACK: OK. So Standard 5 is the
quality assurance standard that ask essentially
for an EPP to have partners– stakeholders. You define what that is. And having them involved with
creating your instruments, evaluating, looking
at data, and then using that for program
improvement in the future. So that kind of cycle that the
regional accreditors are also pushing EPPs to work toward– that’s the one that
more institutions have gotten either stipulations
with or on probation because they don’t fully
meet that standard. And the big part of
that is component 5.2, which looks at validity and
reliability of the instruments that you’re using in your
quality assurance system. STEPHEN MEYER: OK, great. Thanks. Another question is around
whether there is an expectation that assessments used for
operational assessment will need to have
reliability or validity? GARY RAILSBACK: So
operations like on the unit? STEPHEN MEYER: That’s
a good question. Maybe I’ll ask the questioner
to clarify that question. GARY RAILSBACK: Yeah, I
need some more clarification on that. STEPHEN MEYER: OK, so another
question in the meantime. How is CAEP building consistency
among chairs and teams to look for and rate
validity and reliability? GARY RAILSBACK: Well,
that’s a very good question. So I’ve been working
with training volunteers the last two summers. And we’re now doing
it regionally. So I did one in West Virginia,
and one in Tennessee this fall, and in a couple of weeks,
I’m going to Mississippi. And so that’s what we’re
working with new trainees. The CAEP Evaluation
Framework has been out there, I think, since about 2015. So it’s not really new. But getting those
500 volunteers that were trained several
years ago back in, we haven’t had a
chance to do that yet. And that’s why we do
a lot of webinars. We try to get people
to go to CAEPCon. But even if it doesn’t
happen at the team level, it goes to the
accreditation council. And I’ve also been doing
the training for them to make sure that there
is inter-rater reliability between the individuals
and the different panels. And that’s really where
the major decision is made. Again, we’re looking
for meeting all– if you look at that
evaluation framework, there are five sections on
a performance assessment, and then Section 6 and
7 are about surveys. So if you just look
at the first five, we’re looking for
a summary statement that the EPP-created assessments
are at that essential category. It’s not, well, you got one
wrong or you got two wrong. It’s a summary statement. STEPHEN MEYER: Great. OK, thanks, Gary. Another question
is around what is the expected inter-rate
reliability coefficient? Can you say a little
more about that? GARY RAILSBACK: Yeah. So that actually hasn’t been
in the evaluation framework, but it is in the handbook. And the suggestion is 0.8
but it’s one of five factors. So it’s not like if
you don’t have 0.8, then it’s going to
sink the whole ship. That’s one of five factors. STEPHEN MEYER: Great. OK. And I think just
one last question. So this is an
individual who says their program has three
distinct pathways that use three different
clinical field evaluations. Would they be expected to
come together and create one evaluation tool
if they plan to use the clinical experience
for a tool for validity and reliability? GARY RAILSBACK: That’s
a good question. So if you’re talking about an
initial program that has three pathways– and I’m certainly
familiar with that– that’s a choice you need
to make if you want to have three different instruments. What CAEP says is any
performance assessment you’re using to make the
case your program meets the CAEP Standards needs
to be valid and reliable. So you can either do
each one of those three and make sure they’re
all valid and reliable or you could take the best of
them, put them together in one, and make sure that it was. But that would be
a campus choice. You decide if those programs
are really different, then that’s a decision
that you get to make. STEPHEN MEYER: Great. All right, thanks. Thanks, Gary. And thanks, everyone,
for submitting questions. We are going to shift
over to our presenters from North Dakota– Dr. Stacy Anderson– or sorry,
Sarah Anderson, Stacy Duffield, and Alan Olson. They will be talking about
a candidate evaluation tool that they collaboratively
developed called the Student Teacher Observation Tool. And before we get into
their conversation, we’re going to take one more
moment to have a quick Q&A to just kind of kick off this
conversation with a discussion of “What do you think are
the pros and cons of educator preparation programs
staff developing their own instruments versus
purchasing or adapting an existing instrument?” So as [AUDIO OUT]thatthat are
[AUDIO OUT] administered, some states have tools that
are administered statewide. But I just wanted to– kind of before we get into this
conversation of North Dakota’s work– to think collectively about
what some of the pros and cons are of doing local
development of an instrument versus adopting something
that’s already out there. So please consult
your Q&A box once more and share your thoughts. OK, so one comment is that a
locally-developed educator prep program developed
instrument can be tailored to address a local mission. So there may be local
priorities or emphases that you can better reflect
on a locally-developed tool. It’s more customized. One advantage is that there
are more opportunities to test the reliability and
validity in a local context. So that’s certainly
an important aspect is that you can kind of
demonstrate reliability and validity for the
candidates who you’re serving. And if you have a
locally-developed tool, the opportunity to do
that may be better. There is also the
potential of less costs with a locally-developed tool. A downside– here,
we have a comment– is that you have to actually
go through the process yourself of evaluating
the tool’s reliability and validity. You know, you can’t
draw upon some existing study that may be out there
for a widely-available tool. Of course, you can
tailor the tool to exactly what you
want to measure. OK, a couple other comments. You know, one
advantage, which we’re going to hear about from
our North Dakota colleagues, is that a locally-developed
instrument may give you the opportunity to partner
with other institutions in your locale to put together
something that kind of makes sense for the candidates who
you serve in a particular area. So if you’re looking to meet
standards for a given locale, like a state or a
district, and you’re trying to meet
certain priorities, you have the opportunity to
partner with other folks. And there are other
advantages of partnership that we’ll hear about. So maybe I’ll just
mention one more. One of the cons of adopting a
widely-administered instrument like a commercial
instrument is it may not align particularly well
with your state standards or your program values. So thanks a lot for
sharing all those things. There are lots of factors to
weigh in making this decision about developing something
locally versus , you know, adopting something
that’s already out there. So I’m excited to
turn over control to our North Dakota colleagues. Alan Olsen is going to
start out this conversation. And they’re going to
talk about their work in developing the Student
Teacher Observation Tool, and also the
work they’ve done to demonstrate its
reliability and validity. So over to you, Alan. ALAN OLSON: As representatives
from the North Dakota Association of Colleges
for Teacher Education, we’d like to share
our experiences about creating a valid and
reliable assessment instrument. My name is Dr. Alan Olson from
Valley City State University. I’ll tell you about some of the
benefits and challenges, pros and cons that we experienced
while we developed an EPP-created assessment,
and a little bit about our collaborative process. Dr. Sarah Anderson, from
Mayville State University, will discuss the
process that we used for developing the instrument. And then Dr. Stacy Duffield,
from North Dakota State University, will explain
the NDACTE efforts to establish reliability and
validity of the assessment instrument. Our experience began when North
Dakota EPPs were transitioning from NDACTE to CAEP. And we had a common need. We had assessments,
but they didn’t all have actionable descriptors
and they weren’t quite at the depth we needed to have. And so the need was to develop
valid, reliable, practical assessments in
order to gather data that we could analyze
for program improvement and meet our
accreditation standards. We wanted quality assessments,
so we could get quality data. And we all needed to make
those types of improvements. So the idea for
collaboration came about because it was mutually
beneficial for all of us. We all had a similar need
and we could share expertise among our colleagues, across the
campuses to fulfill that need. We’re going to talk about
sharing among the institutions, but even within our
own institution then we have to share
on our own campuses and find colleagues
with different talents and different expertise to
help us if it’s only our EPP. But in this case, we’ll be
talking a little bit more about collaborating
across institutions. The concept of having
common assessments is something that we came
about because we wanted to meet the CAEP sufficient levels. And we were used
to the work that was being done by the
Network for Excellence in Teaching Efforts. It was funded by the Bush
Foundation in Minnesota. And our State of North Dakota
received permission from the NExT– this Network for
Excellence in Teaching– for their exit survey, completer
survey, and employer surveys. And we liked this idea of
some common assessments that institutions didn’t
have to use in our state, but could use in
our state in order to have valid and
reliable instruments. And we believed that we could
develop a student teaching instrument that would meet
the CAEP sufficient levels and become a quality assessment
that each of us could use. We pursued a grant
from the AACTE, and that was the beginning
of our collaborative process. Some of the benefits
and challenges that we found in
collaborating with each other. Some of the benefits– getting a variety
of perspectives really was helpful to us. And often, we had ideas
from multiple stakeholders that ended up making
a difference for us. And the idea that one
of us can have an idea but with the help
of others, sometimes that idea can be much
greater than anything we’d have thought of
individually on our own. We involved our stakeholders to
increase the potential of skill and expertise for research, or
assessment, or field experience that could help us. And also it gave us
a common language. When we were talking to each
other across universities. We had the same
assessment, and so we had the same types of data, so
we could talk apples and apples with each other– compared
to the days when we all had our own separate student
teaching instruments. And even the same topic of
assessment, the questions might be different,
or planning lessons, the questions
might be different. And now, we had a
common language, common assessments that we
could share with each other. It improved our
communication and networking for future collaboration. And it also– we talked about
having mutually beneficial needs and how this
was going to help us. And it did help all of
us feel the need for what we needed to do for assessment. It also helped us have
some common assessments, so that if somebody
was going to have– a cooperating teacher was
going to have a student teacher from Valley
City one year, and Mayville the
next year, and NDSU the next year, and the
University of North Dakota the next year, they’d still
have the same instrument. And principals could be used
to seeing our similar completer surveys, employer surveys. And our cooperating
teachers could be used to similar student
teaching instruments. And that was helpful for
our cooperating teachers and our state as well. And then the collaboration
of using–working with other partners helped share resources
and expertise as well. Some of the challenges
would be that it does take more time and effort
to get together, figure out schedules, and get to
meet is a challenge. And there has to be
some concessions. So some loss of autonomy was
a challenge for collaboration. Some of our pros and cons
of having an EPP…created assessment, and the ones that
were discussed that Steve was able to gather in the
chat– they were excellent– about tailoring the
instrument to be more exact, or having a local instrument
to meet a local mission, or being able to be
more customized– those were all great points
that people mentioned. For us, we talked about how
we worked with standards, to begin with, but then we
could proceed with a little bit more freedom, like the
InTASC (Interstate Teacher Assessment and Support
Consortium) wording. That is pretty lengthy. And all the InTASC items
gets to be pretty lengthy. And we had a little more
freedom and autonomy to make our own decisions
about how we would use those. And then a greater opportunity
to develop assessments that were practical
and meaningful to our EPPs and our state. And items could be aligned or
complement other assessment instruments. We had an exit survey, a
completer survey, an employer survey that we liked. We wanted to have some
alignment with not only InTASC, but with our other
surveys so we could look across instruments
and across perspectives of cooperating teachers, student
teachers, first-year teachers, and employers of our
first-year teachers. And then also an
assessment that can be validated to the population
using the instrument. And I think the other comments
that came in from people on the chat were excellent. Some of the cons– increased time and effort, more
responsibility for the process, and some of us would
lack confidence in our validity or reliability
statistical analysis type things. And that’s where we need
to seek some expertise, or as Gary mentioned, something
like the Lawshe method isn’t quite as
complicated to find ways that we can find validity and
reliability that maybe aren’t quite as complicated or
to seek some expertise from some partners to help us. And then the creation process
can cost money and take time. We’ve been able to share about
our student teaching assessment instrument at some different
NDACTE and CAEP conferences before and at the state level. And we’ve had some people from
other states have interest in our assessment tool. And this is an idea of
some of the other states that people from
different institutions have shared some interest
in the observation tool that we created to be valid and
reliable for use in our state. SARAH ANDERSON: All right,
thank you, Al (Alan). So this is Sarah Anderson
from Mayville State University chiming in. And I’m going to share a little
bit more about our process and outline what
that looks like. So our decision to collaborate
on the common performance evaluation was really
prompted, as Al mentioned, by that chapter support grant
from the American College Association of Colleges
for Teacher Education. We were awarded that chapter
grant way back in June of 2015. And subsequently at our
state chapter monthly meeting that September, representatives
from our NDACTE organization then made the
collaborative decision to select what would become our
fourth statewide common metric. So our discussion then
led to the selection of our preservice teacher
performance measure that would again accompany our exit
survey transition to teaching, which is administered one
year post- graduation, as well as our employer surveys. So this decision
then really became a part of the discussion
that was occurring across a number of our
educator preparation programs with their school partners. As about 5 percent
of our local EPPs were in conversation with three
of our largest school districts in our state about how
could we simplify processes, how could we make field
experience and student teaching placements more
of a common process. How could we identify what
some of those best practices were and really make it easier
for our cooperating teachers that are working with
teacher candidates coming in from a variety
of institutions? That really made
this work very timely when we were thinking
about our partner schools and what their
needs were as well. So I volunteered to
lead that work along with six of my committee
colleagues from NDACTE. We ensured that we
had representation on that subcommittee across the
different types of institutions in our state that included
research institutions, regional, private, as well as
our tribal teacher preparation programs. So with that, we were able to
take a look at that purpose statement of why were we
doing this in the first place. And Al explained that very well,
that we did have this need. It was evident. We had a very clear statement
from that CAEP sufficiency rubric that told us that we
need to look at our validity and reliability in a
very different way. And so that purpose was
pretty clear at the beginning of what our end goal
needed to look like. So the steps then to start
gathering the information from the tools
really first began with taking a look at
what was already in place. We, at the 12 universities, did
have all of our own performance assessments, which did serve
the purpose of evaluating preservice teaching skills. Each of our tools did have
alignment, to begin with, within task and
our CAEP standards, but as Al mentioned, none
of really which completely met those sufficiency
requirements for validity and reliability. So we really did start
with some good basis for developing a common tool. Those tools, as
we collected them from all of our
partner institutions, had a variety of rating
levels, whether that was from a one to
four or one to three. We had both quantitative
and qualitative formats that were in place, again
varying levels of actionability of the descriptors
in those statements. And a number of
elements for rating, ranging from very small numbers
to very intensive numbers. So as our institutions then
were working through how are we going to move forward, not only
did we take our own instruments and take a look at
those, but we also took a look at our P12
partners to see what was happening in our schools. We really wanted to make sure
that a preservice evaluation instrument was also
aligned with what was happening with teacher
supervision and evaluation in our state. And so, across our
institutions, we found that the most common ways
teacher evaluation was being conducted was through the
Marzano Teacher Evaluation Model, the Danielson
Framework for Teaching, and the Marshall Teacher
Evaluation Rubrics. So these were brought
to our subcommittee also as part of our starting
point for instrument development. So the committee also then took
a look at other literature, doing essentially a
literature review, to see if there were
other developed models beyond these ones that we could
use also to inform our process. So, as we gathered
each of these tools, we pulled them together,
and then came back to our basis of our
InTASC Standards, which when we took a look at the
standard’s progressions, actually had 166 individual
constructs to measure the skills of teaching, which
if we’re going to put that into the hands of a
cooperating teacher, is not exactly well-received
in terms of their time, energy, and effort. And so we knew that we
needed to make this feasible, also looking at who would
be filling this out. We also then specifically
looked at the CAEP Standards and took a look at what the STOT
would be used for as evidence within our programs. So, through that process, we
acknowledged that we really would be addressing
our candidate knowledge skills and dispositions
of Standard 1.1. We would be collecting evidence
for state and SPA reports in Standard 1.3. We would also be
addressing our partnerships with our P12 schools as
explained in CAEP Standard 2. We also really
needed to investigate how the tool would provide
evidence for standard three and selectivity during
preparation and at completion. And so we knew that we would
be developing this tool with those end goals in mind. We also wanted to ensure
that was in the final product that we were able to address
the crosscutting themes of diversity and technology. And so, those the
instrument was developed to make sure that we could
address each of those items. So coming back to some of
the steps, our subcommittee, then, as we pull together
all of these resources, which was really in that development
phase, we had to, of course, collaborate in-person. That was one of
our greatest assets was the ability to come
together, because of funds by that chapter grant from AACTE
to do the work face to face. And that did have a huge
taxation on resources in time for us, as we full know what
a faculty’s workload can be in a typical week or month. So we had about
four work sessions where we were face to face three
to four hours and one full day. We had a number of
web-based meetings and smaller work in pairs
on each of those 10 impacts, standard, and the associated
items on the thought. We had a number of
times where drafts were sent out to committee
members for review at each institution and, again,
to touch base on expertise in certain areas. That brought us to a
period of refinement. So as we move through
with the refinement, the seventh draft was
actually the first draft that we brought to the
whole NDACTE statewide group for review and
institutional feedback. We also, at that point,
gathered some information from CAEP site visit reviewers. We also brought our seventh
draft to the CAEP staff for some additional guidance,
and ensuring that we were on the right track. And we ended up with
12 drafts before we were ready for an
initial pilot, which I think speaks to the
complexity of the task that we were given and also to
the quality of the end result. So we really did
consider at this point the use of the half point
scale, whether or not those would be included. Those were based on
Marzano’s Proficiency Scales. And we did take a look during
some pilot exploratory factors if our cooperating
teachers were choosing to use the half points. And they really did. A large number of them chose
to use those half points based on that Marzano
Proficiency Scale. So we ended up with
a seven-point scale. We had multiple discussions
about the performance levels and the labels on what
those scores would be– again, referring back
to our 12 institutions. And as Al mentioned, we had
to make some concessions. At our institutions, we might
be using levels and labeled differently than the
statewide tool to look at, and it was one of
the pieces that we did have to compromise
sometimes that our institutions on that we were changing
those indicators. We also, at this point,
consider the use of “does not language” within
the actionable descriptors. And we ensure that the order
was from the distinguished category or a Level
four that was listed first next to the actual item. You can also see in this example
from InTASC Standard, one with our two elements that are
being evaluated that we also made sure that we had very clear
connection back to which InTASC Standard was being measured. That then led us to the next
portion of our work, which was in our first pilot. And so, that first pilot was
a voluntary participation of an exploratory factor
analysis in May 2016. We, again, leveraged our funding
from the AACTE grant and North Dakota State University. And DSU conducted that analysis. And my colleague, Stacy, will
describe little bit later on here the pilot and
validation process, and what those
results look like. From that first exploratory
factor analysis, we look towards some
instrument refinement and in what would end up
being our final version. And we took those results,
the recommendations– we took those, especially,
the double-barreled items– we had a lot of items that
had multiple indicators being measured. We worked through
those together. We looked at the
language, and how they were loading
on alternate factors to direct them more towards what
they were intended to measure. That refinement process
really worked specifically with the verbiage of those
actionable descriptors. And, again, a collective
and collaborative process to reach that end goal. We also at this point,
with the first pilot had some stakeholder feedback
through our Institutional Advisory Board through
graduate students in the educational
leadership program. And we also held a
presentation at our department of Public Instruction
State Conference where we could solicit
feedback from across the state from administrators and
teachers in the field. We brought that back to
our NDACTE committee. And that resulted
in six more drafts. Those recommendations
continued to come in. And pilot two, then, the second
exploratory factor analysis occurred in December of 2016. This was completed with our
18th version of the stock. We administered that to
a cooperating teachers working with student teachers
from 11 of our 12 institutions. And results showed some pretty
impressive commonalities, which we’ll share with
you in a few moments here. So, from that second
pilot, we ended up coming to a final version, which
we know is draft 20. And this is what you
will find available on the NDACTE website. And that is available for
download at this time. And simultaneously, to
implementing the 20th draft, which was the first
to go statewide, we were also the recipient of
a second AACTE chapter support grant, which would help
fund the development of inter-rater reliability
training modules, that again could be used statewide. So the first time that the STOT
went statewide in North Dakota was the 2017/18 academic year. We really wanted to
ensure that there was institutional flexibility
in the administration. What systems it
would be put into to collect that data
breach institution could be used with whatever
system was in place. And again, how that
institution chose to use data for decisions
within their own programs was maintained. So, for example, at
Mayville State University, we operate with the
Taskstream Operational System. And all of these items can be
put into that system for us to use. We use it formidably during
the clinical experience with acceptable
targets for performance that are set by our own faculty. And we use that for making
continuance decisions within our program
for our candidates. We also use it summatively
during the student teaching experience. And we have that
completed by the students for self-evaluation
reflection, as well as by the cooperating teacher
and the university supervisors so that we do have those
multiple perspectives. So each institution is
left with that flexibility of how they will use the
STOT within their frames of reference. So the Inter-rater
Reliability Training Module is now available. So again, at the
ndacte.org website, that was released
in October of 2018. You also have that thought
and validity information that was mentioned at the beginning. And you have links to at the
start of this presentation. That inter-rater
reliability training involves panels of experts
from across three of our North Dakota institutions. Again, some teaching
videos that were used to practice
evaluating and using the STOT and some training,
as well as on bias, and what that looks like
when ratings are being given. We also utilize some examples
from student capstone portfolios of some of the items
that were used for training. So, at this stage,
we are actually working on the confirmatory
factor analysis. We had enough changes between
the first and the second pilot that the confirmation occurred
during our first full use of statewide data. So the process for doing that is
that here at North Dakota State University the analysis is run. We take that
information and results from our own
institutional system and we place that into
a common data template. And that data is then
all processed together. So we have both
an aggregate score from state of North
Dakota and we also then have our institutional
numbers, which for smaller regional
institutions, this is one of the greatest
advantages is that we can have a bigger N to work
with in terms of doing some of the statistical
work, as well. So NDSU has really served as
the lead research university conducting that work. And we are at the
point now, as we will be seeing those
results very shortly to do a second round of revision
and refinement of that tool, as we ourselves engage in the
continuous improvement process. And the exciting part of all
this work and the collaboration that’s occurred
amongst our institution is that this confirmatory
analysis work of the STOT is actually occurring alongside
our fifth statewide assessment, which is the
disposition, evaluation. And an exploratory
factor analysis has been completed with that,
and we are piloting that across the state this semester. R. MARC BRODERSEN:
All right, thank you so much for all
that great information. Before we move forward and
talk about the approaches that have been taken to
look at the reliability and validity of the STOT,
we want to open up a poll and get your thoughts
about, can an instrument be unreliable but valid? So, if you have a chance, you
should see the poll results up in front of you. And so, just take a moment,
and then we will move forward. OK, so they’re
starting to slow down. What we have so far is 51
percent say –of respondents say, “Yes, the instrument
can be unreliable but valid;” 38 percent say, “No, it cannot
be unreliable and valid;” and 11 percent say
that “It depends.” Can we go to the
next slide, please? So the answer is no. For a test to be valid,
it must be reliable. So, as we discussed, there’s
different definitions or approaches for validity. But, basically, validity
is, are you measuring what you think your measuring. Or are the conclusions based
upon this data, the assumptions you’re making about this,
can those be supported? If a test is unreliable and that
it generates different results when you use it, the
items inside of it are not consistent
with each other. You do not know exactly what
it is that you’re measuring. It might be a measure
of a teacher content knowledge in an area. But if the test is
unreliable, it’s also picking up something else. So generally
speaking, a test must be reliable as a
requirement for validity. So thank you everyone
for your responses. And I think we’re going
to learn a little bit more about reliability
and validity as it was addressed with the STOT. STACY DUFFIELD: So
I’m Stacy Duffield. I’m from North Dakota
State University. And I’m going to talk just a
little bit about the process that we underwent
to ensure validity and reliability of the Student
Teacher Observation Tool. And I want to repeat
what Sarah said. Just kind of getting
us started that we did have some grant
support that enabled us to do a lot of this work that
may not have been possible had we not had some of
that funding support from AACTE for
our chapter grant. And we also had, through our
experiences with the network for excellence in teaching,
had a protocol and a great deal of experience in doing this
work through the development and validation of the
networks common metrics. And so we did have a bit
of an advantage in network. So I’m going to walk through
the five types, the five things that we did for our validity
and reliability process. And so I think when we
look at this process, it’s really important to start
at the beginning of the STOT development process. And Sarah just went through
a great deal of that. But I’m going to parallel
that with the work we did with the validity
and reliability work. We really follow it, I guess,
the Wiggins and McTighe, Understanding by Design process,
where we began with the end in mind. And we knew that we wanted to
end with a valid and reliable instrument. And so with that, we
recognized that we needed to be very mindful of
what needed to happen along the way to ensure that
we had a high quality instrument through that
development process. It wasn’t something we
think about at the end. So we work toward that
quality with three types ability that included content,
face, and construct validity. So, as Gary pointed out,
we did exceed what CAEP would require for sufficiency. We also wanted to ensure full
of internal and inter-rater reliability. And I’m just going
to walk through kind of an overview, I guess,
of how we did that work to ensure validity and
reliability for the STOT. So first, face validity– face validity doesn’t mean that
an instrument really measures what it’s supposed to measure. But rather that in the
judgment of the raters, it appears to do so. So it’s a fairly low bar. But, at the same time, that’s
the place you need to start. So, while it’s the
least sophisticated, it’s so important because it
engages users and asks them if the instrument seems to
meet its intended purpose. We established face validity
through that pilot and feedback process that Sarah described
with the cooperating teachers and university supervisors. We also, as part of that, had
added an open-ended feedback item to the pilot at the end. And we invited the university
supervisors and cooperating teachers to give us
qualitative feedback as they were filling it out. So, just kind of in summary,
while face validity is useful and it’s important as a
beginning part of the process when you’re developing
an instrument, it’s just not enough. And so we went
forward from there. So with content validity with
our next type of validity, we looked at– and while it’s
similar to face validity, it really does take a
more formal approach and often uses some
lower level statistics, like Gary talked about
with the Lawshe method. It does involve
experts in the field. And those experts
judge the questions on how well they cover the
material or the content that you want to assess. We wanted to make sure that the
STOT items actually measured what it was we wanted
them to measure. And so, when I’m composing
the items, as Sarah described, we began with the
InTASC Standards, and those other measures that
were being used in the field. And we aligned to
STOT with previously validated instruments, the next
surveys that were mentioned. Content validity
is often measured by relying on the
knowledge of people who are familiar with the construct. And these subject matter
experts are provided with access to the tool. And they’re asked
to provide feedback and we did those processes. But then the next step
is really important. And that was a formal systematic
analysis of that feedback that we received. And it informed our decisions,
as Sarah talked about, with the different versions
that as the team went through, as we developed, what we would
share with our stakeholders. So with construct
validity, again this does exceed what CAEP
would require necessarily, but we knew we had the funding. We knew we had invested
people in the process. And we knew that we did have
the resources and expertise. And so we did want
to make sure that we did get that internal validity. So construct validity
is “the extent to which the instrument
captures a specific theoretical construct or trait.” And we did this through
a factor analysis. The initial exploratory
factor analysis was then used for revision. But because the revisions that
we did were fairly substantial, we weren’t able to
move forward with just a CFA or a confirmatory
factor analysis. We really did need
to run a completely separated exploratory
factor analysis (EFA). So, in the end we’ve
had, the instrument has undergone two
exploratory factor analyses. And we’re trying to get
our hypothetical construct to wash out in the factors. And I’m just going to tell you
a little bit about the work that we did. We began with four hypothesized
factors, as Sarah said, we use the InTASC Standards. Our hypothesized factors
were the categories that InTASC uses, so the
learner and learning content knowledge,
instructional practice, and professional responsibility. And we wrote our items
with those constructs in mind, wanting
in the end for them to bear out in the
factor analysis. So the first internal
factor analysis, the first EFA that we did, we
only ended up with two factors. So the first three constructs–
learning and learning, content knowledge, and
instructional practice– all putting together in one
large unwieldy construct. And we had a second
factor that did represent the professional
responsibility. And so, Sarah said, there
was a pretty lengthy revision process. And we want back to
work on the items and found that we had
some double barreling. We had some unclear wording. We found that we were using
words that were probably making people think about
instructional strategies instead of, perhaps,
content knowledge. And so we just really
went back to work and thought about
how the wording was impacting the way people
were interpreting the items. And we do have those
validity and reliability analysis for view out
on the NDACTE website. So if you are one of those
statistical people that would really like to read
about the oblique rotation– oblique solution and the KMO
(Kaiser-Meyer-Olkin) and all of things, you are very welcome
to go out and take a look at that. But I won’t go
that deep for this. But you can take a
look at the chart that we have on
the screen and you can see that we
actually ended up with some really nice
numbers, with 0.35 kind of being the threshold. You can see that
we’re well above that. All right, so the next slide. The second pilot
exploratory factor analysis we had taken the
assessment instrument down to 34 items, which we
thought was pretty manageable for a cooperating teacher. We also did some
work with timing, like how long would it take a
cooperating teacher to complete this, and really
thought carefully about wanting to make
sure that we got– everybody was completing
the instrument, and that we were getting good data. And they weren’t
just rushing through. And so we were pretty mindful
about all those sorts of things as we were creating
the instrument. What you’re seeing
on the screen right now are example results
from the second pilot. And in addition to
the more sophisticated statistical analysis doing the
exploratory factor analysis, we also just did some
basics with frequencies with the responses. I’m looking at means
and standard deviation because it helped
us to see things– like, was our scale working? Were people able to use
those half point items? What did they think of those? And in some of our
feedback, and some of them are qualitative data that
we gathered along the way, we heard from those cooperating
teachers and university supervisors that
those half points were critically important. And what we found
is that we were getting more accurate
ratings because when someone was between scores– say, it was a little bit
better than a proficient level, but they weren’t quite
distinguished– people were choosing the top
rating because they wanted to acknowledge that. And so we were ending up
with inflated ratings. And when we were able to
give that half- point that– above a proficient
could be recognized without going to
the next category and creating those
inflated ratings. And so you can kind of see
here how they broke down. You can see, especially
with a 3.5 and a 2.5, those half ratings
were heavily used. And so then looking at
reliability, what was just pointed out in the question
before this section, you have to have reliability if
you’re going to have validity. And so we did run an
internal reliability using a Cronbach’s Alpha, which
is probably something that is pretty familiar to most of us. And we did get
very good numbers. You can see that they’re
all 0.9 and above. And in some cases, what
you’ll hear is that above 0.9 is too high. But that’s relative. And it’s in context. And so we went back to look. And it told us that we may
have some repetitiveness in the instrument. But going back and
looking, we really found that we had trimmed
it down to 34 items. And that every item we had in
there really was necessary. And so we didn’t
feel that looking at the items,
themselves, that we were worried about being above 0.9. We were happy with
the results and felt we needed to keep the
items that were there. And then finally, Sarah
talked a little bit about inter-rater reliability. And we know that you can
have a great instrument. It could have incredible
internal validity and reliability. But if it isn’t
being used properly, and you don’t have
inter-rater reliability, your data isn’t worth much. So your interpretations aren’t
going to be high- quality. And so it was very important
to us to take that next step and work on inter-rater
reliability. And we have the
second chapter grant and were able to work
with the expert panels. So we found videos
that were teachers teaching– kind of everyday
teachers from classrooms. And we would just
use short excerpt to focus the viewing on
particular indicators, making accuracy of the
reading more likely. We used expert panels. And we set a minimum of at least
five members in each panel. And the person that
was on the expert panel needed to be an expert at the
level of the classroom featured in the video, and
for secondary, needed to have some expertise in
that content area, as well. We used a rating process. And in fact, this whole
process, we really need to give a lot of
credit to Erica Brownstein from Ohio State and
the work that she did. We really borrowed heavily
from the process she used. We spoke to her. We went to her sessions
at different conferences and really found that useful. And the rating process
that we used, first, we had the expert panel
rate independently. So they watched the video. And they rated without any
influence from anyone else, and then determined
what evidence they were using for that rating. And so we had them jot down
notes on a recorded document. They’re thinking around
why they gave that rating. Then the group spoke
together about the ratings and presented their
rationale to each other. And after that conversation,
then the raters re-rated the video and
arrived at a consensus rating based on the evidence. And so that became the
foundation for the training modules that you’ll see
out on our NDACTE website. And you’ll see that you have
the expert rating panel. And we have the videos in there. And then along with
what Sarah described at the beginning, just
some basic training on removing bias
from their rating. And so finally, I just
wanted to share this slide. This is just a sample of
the independent rating and how it was collected. In this particular case
with this expert panel, we used Qualtrics,
a survey platform and loaded the
video up on a link. And the independent
raters sat by themselves before coming to the meeting. And they did their
independent rating and entered their feedback. And so when they gave
different levels of feedback, they entered their
evidence so that we could see that beforehand. And it was all
recorded and documented so that we have that kind of
systematic and formal approach to doing this inter-rater
reliability training. And then we met together
to do the consensus work, and we had a facilitator
for each of those groups that did the documentation
and collected that evidence. So we have all of that very
much like a research project, to be honest. And that ends the end of
what I want to talk about. But if you have any more
questions because it was just a survey, you can
absolutely reach out to us or ask questions
right now and we’ll try to do more clarification
or add more detail. And we just thank you
for this opportunity to share our experience
and, perhaps, give you some ideas for some
of the work you might want to do on your own campuses
and with collaborators across your state. R. MARC BRODERSEN: All right,
well, thank you so much. It’s all very interesting. So we are near the
end of our webinar. However, we do want
to open this up if there is any
remaining questions. We’re trying to address
those as we could, but we received a lot. So it’s hard. We can’t address
every single one. But if there’s anything
that has not been addressed or you have questions for
any of our presenters, please use the Q&A box. And we’ll try to address a few
before we have to close out. So I have one question
here is, “Have you considered doing a
consequential validity study?” STACY DUFFIELD: We
have not yet, actually. Next on the docket
is we want to do with some predictive
validity, and probably with the next
survey instruments. But we have not put that on
our list of things to do yet. But if you use the STOT, and
you want to partner with us, shoot us an email and we would
definitely be interested. R. MARC BRODERSEN: And
I think that addresses another question. We have one person
that was asking if this instrument is available
for other institutions to use? SARAH ANDERSON: Yes, it is
available on the ndacte.org website. And we just ask that you
give copyright back to NDACTE and that is indicated on
the instrument itself. So it is available
and ready for use. Again, we just
appreciate hearing if you are using
it so that we can begin to kind of keep track of
where it is being implemented. R. MARC BRODERSEN:
And I think you might have already
addressed this, but another question
we have is, “You mentioned student teachers
use the STOT to self-evaluate as a learning tool. Have you, or you considering
looking at validity studies for those self-evaluations?” STACY DUFFIELD: We have not yet. We started out with just
the cooperating teachers and university supervisors. But again, that’s
certainly something that we can put
on our to-do list. And if Joshua wants
to partner, we are very interested in
doing this kind of work. R. MARC BRODERSEN: Great. SARAH ANDERSON: And at
the institutional level, a lot of what we’re
working with is the STOT is the formative process. And so the self-evaluation
component for a student, we do use that data to take a
look at how it relates to the university
supervisor, as well as the cooperating
teacher’s rating. But the final ratings in terms
of continuance in the program and what we’re really looking
for in terms of providing data on our completion
for our candidates is really on the supervisor,
their cooperating teacher, not the student. Which is why we brought
that up as a reflective opportunity for them and to
get to them to make comparison and then set some
growth goals, as well. R. MARC BRODERSEN: OK,
maybe one more here. “Can the STOT be used for
the first and third years of full-time teaching at a
self-evaluation and supervisor evaluation by the
building principal for measuring student impact?” STACY DUFFIELD: There are a
couple questions in there. So it’s not on the
InTASC Standards, which are professional
teaching standards not just for pre-service
but for in-service, as well. So the intention is
there for the instrument to be used for teachers. We have not validated it
on in-service teachers. But I know that there
are institutions in North Dakota that are using
it for their CAEP Standard 4 with their in-service teachers. And we– need to get enough numbers. You really want
100 or more if you want to do a good
internal factor analysis. And so we’re just waiting
for some of those numbers. And then the second question is
about measuring student impact. It really isn’t that
sort of an instrument. It’s an observation tool. And it doesn’t measure
student learning… directly. R. MARC BRODERSEN:
All right, well thank you very much for that. I think with that,
we’ll come to an end. So first, I want to give
a big thank you to Gary, to Stacy, Sarah, and Alan. Thank you so much. It’s been a pleasure working
with you all to prepare for this presentation. So with that, thank you
all so much for joining us. And we really
appreciate your time.

Leave a Reply

Your email address will not be published. Required fields are marked *