
In this episode of EDUCAST 3000, hosts Ryan Lufkin and Melissa Loble dive into the fascinating world of psychometrics with Instructure experts Russell Ligon and Alexandra Lee. They share their personal paths into the field, reflect on how assessment practices have evolved over time, and unpack why psychometrics is key to building fair, effective assessments for all learners. From designing balanced systems that support diverse learning styles to leveraging the latest in technology and AI, their conversation explores how assessments can truly reflect student growth. The group digs into the nitty-gritty of psychometrics — validity, effectiveness, and the human side of testing — and looks ahead at what’s next in educational assessment. Along the way, they highlight the importance of creating a positive culture around assessment that fosters meaningful learning and growth.
Takeaways:
- Failure can be a powerful teacher in the learning process.
- The shift from centralized to decentralized accountability models in education.
- Assessments must be tailored to state-specific learning standards.
- Psychometricians play a crucial role in ensuring equitable assessments.
- Balanced assessment systems provide a comprehensive view of student learning.
- Assessments can serve multiple purposes: of learning, for learning, and as learning.
- Effective assessments require ongoing feedback and adjustment. Assessments require time and come with opportunity costs.
- Validity is essential for assessments to be effective.
- Trust in assessments is crucial for educators.
- Positive cultures around assessment can improve experiences.
- Balancing assessment and instruction is vital.
- Authentic evaluation of student growth is needed.
What is Educast 3000?
Ah, education…a world filled with mysterious marvels. From K12 to Higher Ed, educational change and innovation are everywhere. And with that comes a few lessons, too.
Each episode, EduCast3000 hosts, Melissa Loble and Ryan Lufkin, will break down the fourth wall and reflect on what’s happening in education – the good, the bad, and, in some cases, the just plain chaotic. This is the most transformative time in the history of education, so if you’re passionate about the educational system and want some timely and honest commentary on what’s happening in the industry, this is your show.
Subscribe wherever you listen to your podcasts and join the conversation! If you have a question, comment, or topic to add, drop us a line using your favorite social media platform.
-
Psychometrics Reloaded
Welcome to Educast three thousand. It's the most transformative time in the history of education.
So join us as we break down the fourth wall and reflect on what's happening, the good, the bad, and even the chaotic. Here's your hosts, Melissa Lobel and Ryan Lufkin.
Hey there. Welcome to Educast three thousand. I'm your co host, Ryan Lufkin.
And I'm your other co host, Melissa Lobel. And we are joined today not with just one guest, but two guests, and actually two really special guests. These are some of our colleagues here at Instructure that are specifically focused on thinking about how to unpack research and understand indicators for student performance. Or if we wanna think more broadly, I will say assessment perhaps, but it's so much more than that.
And we're excited to pick their brains. There's a lot happening today in the world, as well as a lot of opportunity for how we as educators think differently about how we understand student success, learner success, and how we help each other drive to that and meet our outcomes. So with no more further ado, please welcome Russell Ligon. He's a psychometrician here at Instructure alongside Alexander Lee, who's a senior researcher here at Instructure.
Welcome both.
Thank you for inviting me. I'm happy to be here.
Yeah. Super excited to be here and talking about how do we measure learning.
Love it. Before we jump into today's topics, let's, we'd love to hear a little bit more of your background and how you actually ended up in the world of assessment. So, Russell, let's start it with you.
Sure. So my background is actually as a biologist.
I have a PhD in evolutionary biology. But about five years ago, I decided to make a career transition to one that would give me more flexibility for, you know, personal and family reasons. And I switched into a role as a data scientist. I still wanted to work in a field in which I was kind of excited about contributing and giving back something that I could be proud of at the end of the day.
And so I was focused on kind of green energy, biotechnology, and education. Luckily for me, I ended up at a small company called Learning OVations, which was an education startup focused on improving childhood literacy. So as I was hired as a data scientist, but, as many of you know, at a startup, you have to wear many hats. And I got the opportunity to develop, become familiar with, and then put into practice kind of psychometrics.
Basically, tools that would help me understand the assessments that we were using to better provide information to teachers to make kind of customized tailored instructional choices for their readers. And so that's how I got into assessment education, and psychometric.
That's an incredible background. Also, by the way, I've always said that psychometrician is, like, the best job title ever. It's like It is. Something about it. I'm a psychometrician.
It is. I I love it.
And, but I always have to, like, double check that I'm spelling it right whenever I Yeah.
Al, how about you?
Yeah. I have a sort of winding path, but maybe more related to education than Russell's.
So I, you know, really became interested in education research due to my own experiences as a classroom teacher. I was really lucky to get to teach in lots of different places. I taught in Thailand, taught in Singapore.
I moved back to the states and taught in the rural Mississippi Delta, and then I also taught in my hometown of Denver, Colorado.
And when I was teaching high school English in Mississippi, I really became interested in learning more about student motivation and how I, as a teacher, could better motivate my students. And that, you know, ignited a curiosity in me that eventually led me to go back to school to pursue a PhD in educational psychology and educational technology at Michigan State University.
And we are recording this on the first day of March Madness, so I have to say go green.
Yes. Those are the startings. Shortly. Yes.
Yes.
And maybe I will I'll kick myself later, but I'm By the time this goes the way.
Will go the answer. Yeah.
Yes.
After completing my PhD, I became really interested in doing applied research in education. So similar to Russell, I wanted to do work that matters and was directly impacting teachers and students in a positive way. So I started looking for jobs, you know, in industry, which is moving at a fast pace and being, you know, technology that's being used by teachers and students. And that's what led me to my current role as a researcher here at Instructure.
And I feel really lucky to get to do work that's meaningful to me and that I can relate to on a personal level having been a teacher. And so really enjoy the work I get to do here focused on assessment research and efficacy studies and those sorts of things. And assessment research is really just near and dear to my heart since I taught in a state tested subject, and I saw firsthand just the stress that teachers and students are under in those state tested subjects. And I really wanted to make sure, as a teacher, I was helping my students be successful on those end of course assessments so that they could go on to do great things, go on to college, go into career paths, apprenticeships, all of those different pathways.
I love that, Alan. I can relate to the state tested subject. I started my career also in the in the classroom. I taught high school in New York City.
And there was not just one assessment, but two. You chose your level in which you were assessed that I and and as a new teacher, I remember the stress around having that be not only state assessed, but that was, like, that was the impact on students because that was pathing them. I had ninth graders. That was pathing them already that early.
And, like, it's just this added in comparison to the classes that I taught that weren't state assessment courses, it had such a distinction on sort of my experiences as well as I know the students, which kinda leads me to the next question we love to ask our guests is, like, a favorite learning moment or teaching moment. So it could be one where you were teaching your your students. It could be one where you were learning something yourself. It could be something you've observed, something that you've worked on with all of the organizations that you've helped do research.
Would you mind starting and sharing with us a favorite learning moment?
Yeah. So, I think that this learning moment to me that comes to mind is really related to how I try to approach teaching, and it was really impactful for me. So a learning moment that comes to mind for me was not my favorite at the time it happened, but looking back now as, you know, a middle aged lady, I can really appreciate how this was such an important moment for me and really transformed how I approach studying in school for the better.
So the moment is, you know, picturing me as a freshman in high school, and I had my first experience failing a test. And the test was on Greek mythology in English class and required me to study and memorize a lot of detailed information about gods and goddesses. I think a lot of folks have taken a similar test to that when they were in high school.
And the problem for me was that at that point in my life, I just didn't know how to study. And my favorite teacher, miss Scott, at George Washington High School in Denver, she taught me how to study. And the way she taught me how to study was by giving me my first f, and I needed to have that happen to take it seriously.
And the thing was that she didn't just give me an f. She also gave me an opportunity to retake the test and to work with her to get help on learning how to study and prepare better. And so the combination of her having really high expectations of me and not sort of passing me on that test when I didn't deserve to be passed on the first time and giving me an opportunity to learn and grow an important skill, studying for an exam effectively, is something that I'll be forever grateful to her for. And, and I really am grateful that I got an f on a test in my freshman year of high school.
Oh, man. I think we can all relate to that story. And so often when we ask our guests these questions, we get, you know, the, I remember this because it really inspired me to be this. And it's like, no. No. No. There's, like, there's points in our life where it's it's it's a wake up call or a, you know, a shake.
I think that's I actually love every time we ask this question on our podcast.
I think of a different experience in my life. But how what an amazing teacher to use that opportunity to actually change the underlying skill of studying, right, and focusing on that. That's amazing.
Yeah. Alright. Russell, you're up.
So so Al and I did not coordinate, these Sure. Answers, but but I was thinking, you know, of some learning moments, because I know this is a thing that you love to ask your guests, and I'm happy to share it as well. My learning moment that I wanna share is also one related to failure and learning how to study. Except mine happened when I was in college.
It was a bit later. Essentially, like, I was a good student and I got into a good college, but I did so without having again, having had that moment that Al had as a freshman in high school where I didn't really know how to study and then kind of build knowledge as opposed to gather facts. And so I went to Pomona College, and, I was a bio major. And basically, I scraped by with, c pluses in the first year as a freshman in chemistry, general chemistry.
And then as a sophomore, I was in organic chemistry and was continuing to kind of be towards the lower part of the class in terms of performance. And I had a professor, Cynthia Salasie, and she took those of us who were having trouble, and she created a study group for us. And she was spending her own time, and this is kind of before I even knew about office hours and how you could go ask for help from professors. And she was working through these problems with us and showing us how we needed to be able to put what we were learning in the classroom and in the lab into practice, kind of understanding molecular structure and, you know, complex organic chemistry in a way that was different from simple fact or knowledge regurgitation, which was had been kind of something that I was fine with at that point, but it wasn't sufficient to succeed.
And so the time that she took to help those of us kind of learn how to take information and use it to create, you know, kind of knowledge and information to be able to build on that was something that I am very grateful for. And I did by the second semester of organic chemistry. I was able to get a b minus. Nice.
And so that's the grade I'm most proud of.
Impact. Yes.
In all of my college, you know, the slow upward trajectory from I think it was exactly c minus to c to c plus to b minus. And so thank you, professor Salassi, for taking that time to to, again, teach in a different way, but the same you know, it's analogous to ALs. How to learn, how to use information, and build on that.
Yeah.
And that's a classic course that's we call it a weeder course or however you wanna I mean, I know so many people that didn't have that opportunity and never made it through that course. And then you stayed in science, which is just rad. I love that. But, like, it it kept you in a field that you are really passionate about.
That's so cool. So I mentioned at the beginning, this is a topic that I'm super or an area of education that I'm super passionate about, but I'm not sure all of our listeners have some of maybe the same foundation that maybe Ryan and I do. So just because we have all sorts of people listening to this podcast, Al, if we can start with you and perhaps we'll focus assessment. We'll use the word assessment, but let's focus it on student performance and indicators in that.
So sort of where are we today as compared to, like, ten years ago? Maybe give us, like, a baseline or grounding so that we can build our conversation on that and make sure all the listeners are thinking in the same in the same way as we all are.
Yeah. And I like thinking about assessment as a broad way that we understand what students have learned, and I like to think about learning as including things like social emotional learning, soft skills, hard skills in the careers, and then, of course, learning standards and academic types of learning. And so when I think about, you know, where we are today and where we were in our history in ten years ago, a somewhat recent past. One thing that comes to mind to me is how federal policy in the US has really shifted and how we're holding, you know, teachers and schools accountable for helping students learn. And so I'm really interested in the policy changes and accountability models and how these have shifted away from a centralized model and towards more of a decentralized model over time and really shifting power and decision making back towards the states.
And since we do have lots of folks listening to this, I wanted to sort of walk through a couple key historical moments related to this. So, you know, when we look back at the trends over the last ten years, but really scope going a little further back to twenty years ago, we can really see how this shift from centralized accountability models towards more of a decentralized model has taken hold. And so a little over twenty years ago, the no child left behind legislation was passed.
And that really, I think, had great intentions of wanting to make sure all students are learning, but but really centralized accountability at the federal level. And then we saw, you know, sort of a shifting away from federal accountability models towards the states with the Every Student Succeeds Act passing now within about ten years ago. And with the Every Student Succeeds Act, you know, there was more of a role in states defining their own accountability models. We saw assessments now broadening out to include more differences in learning standards and other competencies.
And then today, in twenty twenty five, you know, we can see this move, you know, increasing even more and really decentralizing education decision making towards the states. So more to come. I think this is the trend, you know, I kind of see looking back at our history of accountability models. And and accountability models are important because it has implications for assessment. And that's why I'm really interested in it.
And so, you know, these differences in accountability models from state to state have resulted in a greater need for assessments to be tailored and customized to each state and each state's differing standards and learning priorities.
And in addition to tailoring the assessments to the unique learning standards of each state, there are also increasing differences between states and other outcomes being measured for accountability.
Recently, some examples have been coming across my desk where states are including a greater emphasis on career skills, portfolios of learning, and really assessing the whole learner beyond just these academic standards. And so I think as someone who does assessment research, I'm really interested in thinking about how this variability state to state impacts assessment. And how do we design good assessments to sort of rise to the occasion of this variability?
Yeah. This is this is really important right now as we're thinking about how do we create outcome oriented and equitable learning environments. It's funny. Every I talk a lot about AI and that shift towards more skill based, more outcome oriented is kind of omnipresent in every conversation. So, Russell, where does the work of a psychometrician fit into that broader conversation?
That's a great question. So one of the, you know, fantastic kind of toolkits that a psychometrician or an assessment research team has available to them, and it's called item response theory. And it's a class of analyses basically that let you understand the items and students simultaneously. The items is kind of a general term.
Quest it can include specific questions, but also kind of broader, like, fill in the blank type things. This item response theory, it basically lets you see items in the real world. And so, you can use these tools to identify places where individual items or assessments at a broader level are behaving differently than expected for different subgroups of assessment takers, so students. And so when you talk about equitable learning, you want to be able to evaluate or assess those students equitably, fairly as well.
And so a psychometrician can help ensure that the items that we include in assessments are free of or have at least minimal bias. And through repeated cycles of evaluation and data cleaning, essentially, through these psychometric tools, we can do our best to minimize the impacts that the items that we're creating and putting in front of students to evaluate mastery on topic A or B are items that are essentially free of bias that we inherently impute into them when we create them. No one can help it. Right?
It's just we have our cultural, historical backgrounds and that implies I was gonna say it's not really one size fits all.
Right? Like Al was pointing out, you've got these regional differences, very, like, global differences. We've got to account for those.
Yeah. And so, you know, that's a big part of it is kind of this, like, nuts and bolts. We wanna we can evaluate using psychometric tools, the assessments to have a quantitative measure of the degree to which they are or are not performing differently. The other thing is we want to be able to deliver those assessments in a way that makes sense to all learners. And so technology, not necessarily AI in this case, but technology, we want to make sure that our assessments are interpretable and fairly delivered and scaled and scored quickly so that teachers can then use that information rapidly to make decisions. Because that's, you know, that's the end goal is to we're not assessing students for the sake of assessment. We're doing it in many ways.
Making it actionable. Yeah. Exactly. Cool.
Yeah. I'd love to build on that. And so, Russell, you've shared how to think about the role of assessment creation and delivery and the importance of the science behind how ensuring that we remove bias from and other things from from assessment. And, Al, you've chatted already about the history and what do what do we need to be thinking about from a legislation or policy perspective.
As you look forward, Al, to thinking about how you'd advise districts or states in what they need to know or think about assessment, how how do we combine these two worlds and perhaps give them some practical ideas around how should they think be thinking about assessment in their state, in their district, How even maybe should a teacher be thinking about assessment?
So that, again, we can all see, regardless of whether you're teaching a second grader or a lifelong learner, we can all help people progress and achieve the outcomes that they're hoping to from their learning.
Yeah. So when I think about, you know, what schools, districts, teachers, instructors in higher ed should be thinking about, I really think it's important to consider a balanced assessment system as a good approach.
And so no single assessment, even if they've got really strong items on them confirmed by item response theory, is going to adequately measure and provide all the information that different stakeholders need to have to make informed instructional decisions.
Assessments are designed with different purposes in mind. And so, you know, you really do need to think about balancing different forms of assessment with different purposes as, you know, the way to approach assessment and having it be sort of an ongoing process. When I did pull a paper, because I'm a researcher, off the National Academy of Education website, they recently published a report around reimagining, balanced assessment systems. And how they are defining a balanced assessment system is one that is intentionally designed to provide feedback to students and information for teachers to support ambitious instructional and learning opportunities.
And so when I read that definition, you know, to me, a question arises, which is, well, how can educators make sure they're strategic and intentional and how they're using assessment to learn what they need to from it, which then would enable them to make sound instructional decisions and really increase student learning?
And so, you know, how how I kind of like to think about a balanced assessment system is thinking about assessment as as being of learning, for learning, and as learning. And that to me is an intuitive way to kind of conceptualize what this looks like in the real world and how these assessments have different purposes. I wanted to just kinda walk through those three ideas.
So first, thinking about an assessment of learning, I think this is what we most often think of when we think of assessment is an assessment that's summative in nature and is used to sort of evaluate students. So when I was talking at the beginning about my learning moment and getting an f on a test, that was an assessment of learning.
Another example of that sort of assessment would be state assessments at the end of the year. So that end of course assessment, I help my students get prepared for when I was teaching English too.
And these assessments, you know, are useful. They're useful data points, especially at the school or district level, to identify trends or the state level or the national level to sort of think about what are the trends we're seeing and what students are learning and not learning and and what are macro changes we can make to help them learn more. So think about curriculum changes, professional learning initiatives, those sorts of things. The next sort of idea around assessment is thinking about using assessments for learning.
And so these assessments would be, you know, sort of quicker but more readily available assessments that aren't summative. They're more formative, and they're really being used on a very regular basis, even daily basis, so that teachers can get kind of real time feedback on what their students know and do not know and so that they can readily adjust their instruction and, you know, pretty quickly in real time. And, you know, and that those more assessments those formative assessments are really helpful for teachers in particular, and students can also get feedback from them. And then a final way to think about a balanced assessment system are assessments as a learning tool in and of themselves, and I think this is one that sometimes, like, blows people's minds.
But, you know, research in cognitive psychology has found that when students are taking assessments, they have increased memory of new information. They're more likely to identify misconceptions and learning gaps themselves, and that's especially true when they're given feedback. And so when assessments can be used for learning, that's another important component to consider is making sure that there's a feedback cycle that students are a part of when they're taking assessments. And that, you know, sort of helps them develop their self regulated learning processes, metacognition, and will ultimately set them up to be highly skilled learners for the rest of their lives.
And so I think, you know, that's sort of the advice I would give to folks when they're thinking about how to design an assessment is it's not gonna be a one size fits all approach, but you really wanna think about different forms of assessment, giving you different pieces of information. And then that's kinda helping fill out a whole puzzle that'll really allow you to make strong decisions around instruction and interventions.
Yeah. So as we kind of evolve into these new areas of assessment and try to do them at scale, this is one of those areas where technology really plays an important role. Right? Russell, what role does technology play in good assessment design and delivery?
Yeah. That's such a good question. And and building off of Al's point, it depends kind of on the goal of a particular assessment. And so to build on an analogy or to build an analogy that's been used probably before, I really like thinking about these assessments for learning, kind of these formative assessments that provide information to teachers and students about where they are in their learning journey as kind of measurements of where you are on the map.
So if you have a goal that you're trying to reach, taking an assessment lets you understand your position on your map relative to that end goal. But in this case, in this analogy anyway, you're not just simply pressing a a button on your GPS and getting your location. You have to take some measurements and you have to get your bearings, measure, you know, different landmarks to get your position. So it takes time because just like in the real world of assessment, assessments take time.
And although some assessments can be used as learning via those processes that Al mentioned, they also are often taking away from instructional time, right? So they provide valuable information, but it doesn't come for free, right? It comes at a cost. There's a time.
And so one of the key ways that technology can improve assessments and improve how assessments are part of a balanced pedagogical plan is by speeding them up. Now, you want efficiency, not just speed. So you wanna still get good quality information from the assessments, but you don't want it to take an hour and a half out of your day. And so one of the ways the technology has been used effectively within the realm of assessments, especially informative assessments, is through computer adaptive testing.
Computer adaptive testing takes into account individual student learning trajectories. And ideally, it takes into account, like, a large bank of items that can be customized based on that student's estimated kind of ability at the start of the assessment. And then it incorporates their answers to more quickly hone in on their true ability or mastery of a given subject or topic.
So that's a way that's a key way that But it'll be almost impossible to do that without the technology to kind of access the Absolutely.
It's one hundred percent impossible without, you know, that technology. And it's also tailored. Right? So you can get high precision information about both your highest and lowest performing students in a Goodman classroom from the same delivery, assuming that the technology was in place, the item banks in place, and that the information which comes from psychometrics is there to inform how that test is delivered. So that to me is, like, a super important way that technology and assessment can be married moving forward. I mean and have been. I don't mean to imply that this is not widely used.
I would say another one that is key is integrating these, student information systems so that the teachers, the district administrators have access to kind of, like, holistic sets of information about not just performance on assessment a in class b, but also attendance and performance in previous years and other key important parts of of who is The life metrics.
Right? That that I'll throw them back in student success.
Yeah. Right. Yeah. Yeah. That's exactly right. So and Al mentioned this as well, you know, no assessment's gonna do everything.
So you need to a balanced set of assessment strategies. And, yeah, we want holistic information about students, and that's what students want as well in many cases.
Awesome. Yeah. I think one of the things too that technology has surfaced up perhaps is there's so much more assessment out on the market.
Products, content, services, you name it. Right? There's it's I feel like there's more and more of that. Some of it doing really impactful things like you've described, Russell, and then, you know, some of it you wonder. So I'm curious, Al, how do we know a particular assessment or approach or technology? How do we know it's effective, or or where does the evidence play into this?
Yeah. It's a really important question.
And I think it's one everybody, you know, needs to have a working knowledge of so that they can be informed consumers in the market. And, you know, in my view, for an assessment to be effective, it has to give you the information you expect it to give you. You really need to be able to trust that the assessment's giving you valid and reliable information on what students know and do not know. If you can't trust that it's giving you the best information possible about what your students know, then you can't use it to inform instruction.
Or you might be trusting it, and you shouldn't be. And you're making decisions that are going to potentially harm students in the worst case scenario or not help them. Right? And so making sure that you're able to evaluate an assessment at a high level to see if it's valid and if it's measuring what it says it is is really important, I think, as consumers and teachers.
And everyone should have a little bit of a working understanding of how to look at validity evidence of an assessment.
And so, you know, as a researcher in Russell as a psychometrician, you know, we look towards a couple sources to figure out, you know, how we show validity of an assessment, how we establish that it's valid. And so there was a researcher, Samuel Messick, who really provides a framework that we use a lot in our work, and he identified key sources of validity evidence.
And these sources of validity evidence are also what underpin the standards of test development.
And these are the standards that assessment developers should be using to be aligned with research based best practices. They're updated by leading education research associations, American Education Research Association, National Council on Measurement in Education, and then the American Psychological Association make these standards. And they're really building off Messick's work around sources of validity evidence. And so I'm not gonna go through all of them in detail, but I wanted to kinda hit, like, a couple top ones I think folks can really look for. And so one source of validity evidence is looking at the test content itself. And so to my earlier point, you know, the test should be aligned closely with the standards that it's seeking to measure. And so making sure that, you know, the questions are aligned with your state blueprints or your instructional pacing, that's an important source of of making sure that it's gonna be a valid assessment.
Another thing that you can sort of look towards is looking at the characteristics of the items on the assessment. And so this is, you know, Russell's bread and butter as a psychometrician.
But you should be able to, like, look at some documentation and see, oh, there's been some sort of item analysis done to validate the items that are included on my assessment, and that's an important component.
Another thing you can look at is, is this assessment if it's on state learning standards, is it correlated with other assessments that are measuring the same thing? So a best practice, to look towards with looking for correlation studies or predictive validity studies, or that's showing that, a benchmark assessment, for example, is highly correlated with a state assessment, if because they should be measuring the same thing, and they should be doing that comparably well. And then a final source that is my personal favorite source of validity evidence. So I am a nerd about this stuff, but I love this idea. And it's the it's showing that a a test is valid because it has positive consequences on student learning.
And so this does sort of make sense. You know, assessment providers are saying, if you use our assessment, you're going to be able to make better instructional decisions and drive higher student outcomes.
Well, let's test that. Let's see if it's a valid assessment. If it's not valid, it's not going to make those informed decisions possible and won't lead to higher student outcomes.
And so, you know, research studies can be designed to really look at is the use of the assessment tied to higher student achievement.
And that's another way of looking at the validity of the assessment. And then I like that one in particular because I think it's important to think, you know, at the end of the day, is this helping students? And if it's not, then we shouldn't be doing it. So Yeah.
And it helps us weed through you know, I think everybody has intentions of being good actors in the space or I'm gonna hope for. I'm gonna glass half full that one. But it helps us weed through what's more, again, hope versus what's real or what's marketing versus what's actually happening on the ground, which is why I love that one as well. And just for our listeners, I'll remind you again, we'll get a link. Both Al and Russell have already mentioned a number of different research studies and resources. We'll make sure there's a robust collection of links along with the show notes. That way, if you wanna dig in deeper to any of these areas, something really, you know, struck your fancy, you will able to be able to sort of follow the trail that they were sharing.
Yeah. Well, and it wouldn't be the Educast three thousand podcast if we didn't talk about AI at some point. This and so much of the conversation around AI has been focused on academic integrity, right, and how it's really undermine traditional assessment models. Russell, what role does AI play in assessment moving forward? How do we actually leverage these powerful tools?
Yeah. I think there's an inherent assumption that the data you get from an assessment, from the perspective of a psychometrician, is generated by a student. Right?
So with that assumption How do we make sure it is actually generated by Yeah. Let's let's make sure that that it's actually a kid taking the test or a lifelong learner and not a a boss a lot with educators.
Right? That it's gonna be students or AI submitting homework that's graded by AI and no learning happens in that loop in between. Right?
Yeah. That's a little bit of a horrifying, you know, worst case scenario from my perspective because to me, Al and I were talking about this in kind of advance of of this opportunity to chat with you guys. You know, learning is such a fundamentally, like, human process.
And, of course, there are statistical elements that have been extracted and built incorporated into these models that produce incredible content, but still a joy of life for many people, myself included, is learning. And it is this kind of one of, like, the production of art. Yes. AI can do art, but it doesn't take away the joy and the beauty of human made processes.
So from kind of just stepping away from that, like, broader perspective in the role of or potential roles of AI and assessment, one of the key areas, of course, is content generation as you as you mentioned. Right? There is great potential for AI to generate varied yet topically relevant questions, items that could be used to evaluate, you know, student learning. But you knew the but was coming.
Oh.
You have to trust it. And so especially depending on the kind of assessment you're talking about, any content that's generated by someone who is not a content expert, who is not doesn't have that deep background, and who is kind of more degrees removed from the actual teaching of that topic in the context in which it's being used, you run the risk of decreasing that assessment validity. Al was mentioning, you know, you have to be able for an assessment to be valid. Someone has to be able to look at that assessment and say, oh, yeah.
That's measuring what it's set out. This is a third grade math assessment. I can tell that because I've been a third grade math teacher. Not me personally, but, you know, in this scenario.
So it has the potential to streamline some of those processes, those item creation processes.
But at the end of the day, you know, if you're gonna depending on the use case of the assessment, it's still really important to have people involved in the Even in the loop aspect.
Right?
For sure. You have to have and because assessments have to be trusted for them to be used, there may be a mandate for a given district to use a particular tool that they purchase. Right? But if the teachers are getting results that they don't trust because the test has low validity that you were talking about, Melissa, it hasn't been empirically demonstrated either quantitatively or even qualitatively.
If there's a disconnect where the items when the teacher looks over a student's shoulder and says, what is question even asking?
You know, why is that on this assessment? They're gonna do what the district mandates, but they're not gonna use that information in the way that informs their instruction, and rightly so if they're if they don't trust it. Right? So there is definitely a role for AI in kind of streamlining a lot of these processes.
But at the end of the day and this is true across all content in my perspective and probably yours too, Brian. Like, there are many places where it can streamline, but there's still quality checks by humans. And where is that trade off when it's still actually, like, a net gain in productivity?
You know, one of the conversations that's probably up more recently is the idea of AI drift. Right? That idea that initially, you might get some really great answers, and it might be really consistent. Three months down the road, you might find that it's drifting off, and we don't really know why. But we've gotta have that kind of constant human loop maintenance on it. And so as we're thinking about leveraging these tools even to scale these really powerful approaches, we gotta make sure we got humans plugged in to to make sure it's not wandering off into the forest.
Yeah. A hundred percent. And that's part of best, like, assessment practices as well regardless of AI is is continuous evaluation and refinement. And so I talked about, like, the ability of psychometric tools to be able to identify biased items. We can also identify high performing and low performing items where a high performing item is one that gives you lots of information, and it gives you information about the difficulty. And using that information lets you build kind of an assessment based on a blueprint based on IRT, but you have to continue to evaluate because student populations change over time.
Well, and Al's point about, like, assessment as learning, I think was really interesting because I think in a lot of times when you as I've used AI tools to explore topics that I'm writing about or talking about, I'm learning a lot about it in really powerful ways even as I'm, like, preparing to you know, if my test or my assessment is the actual presentation I'm giving. Right? I'm learning through the process. So it strikes me that there's some opportunity there to really use AI again with with the proper checks and guardrails as almost a part of the learning process.
Yeah. I'm particularly this is such an interesting topic. I'm particularly concerned about this in the lifelong learning space or the continuing education space. I think in k twelve in the US and in other countries, the equivalent primary and secondary, there are more checks and balances, some more structure, some more formality to that.
I still think we have we have gaps. Right? And but there's more opportunity to bring in the human. I recently and Alan and I were chatting about this briefly.
I recently went through some I'm a scuba diver, and I did some scuba diving instruction. And I took an assessment alongside some e learning that I was doing in order to to learn a new skill. And it was very clear that assessment was built by AI, and there were questions that it was pulling that made no sense and had no alignment actually to the content. And so I go over this with, like, my professional instructor, and he looks at that, and he's like, this doesn't make sense.
And in my mind, and he's like, well, who wrote this? And I'm thinking, I know who wrote that. That was probably because they don't have that same checks and balances or practice. Right?
Like, as we start to expand, this is why we wanted to do this episode was let's give some folks some foundation for how do you think about assessment so that regardless of where in the learning life cycle you are, you can have that kind of the right kind of impact, but you can use it in the right ways so that it is actually driving outcomes. So I'm I'm thinking sorry about that that side note, but it just made me think of that. I'm like, oh, I have an immediate example of where that recently happened to me. So let's end on future predictions, if that's okay.
Well, our last question will be about the future. Al, where's the future going when we think about assessment, particularly in the framework that you've shared?
Yeah. I mean, I think if I could kinda wave my magic wand and have something happen in the future, what I'd really love to see is positive cultures around assessment.
I think that assessment, a lot of times, makes people feel really negative or they've had really bad experiences with it. And I think it's really important to create and foster positive cultures around assessment and around how it can be used. And when I think about how that can happen, culture is a big thing. It's hard to change. And so I like to think about it at sort of a systems level and really thinking about the learner at the center of this ecosystem with different layers on top of it. And, you know, education is a part of a complex dynamic system where there's lots of different relationships that are really important to consider.
And when we're wanting to change cultures around assessment, we need to think about all of these interlocking relationships at different levels of this big ecosystem in as influencing what the student experiences at the end of the day. And I really think it's important to kind of acknowledge, like, the complexity of making culture change towards a more positive culture around assessment.
And I, you know, am a motivation researcher, and so I've thought a lot about, you know, mindsets and beliefs around assessment and how that's really important for what folks experience around it. And so I really think thinking about, you know, how the mindsets of district leaders influence the mindsets of teachers that then influence the mindsets of students and how we need to kinda consider all of those interlocking parts and relationships really strategically to build positive cultures of assessment.
And, you know, one personal example that comes to mind around this idea was that when I was teaching, I tried to essentially be an open book with my students, and acknowledge my mistakes. And then I sort of realized in doing that that I was modeling for them how you learn and grow and improve over time, and that all feedback is important to helping you master something new and really leaning into that. And so when we would do assessments, I would try to model for my students what I was seeing at the class level in the data about my teaching and modeling for them. Oh, here are some things I'm not doing very well on.
Like, you know, we're like, I'm seeing in the data that these are some gaps in what you're learning, and that's a bit on me as the teacher. Right? But then also having them individually look at their own data and have those same realizations. And I think having that sort of feedback loop where we're all sharing how we're learning and growing and using the assessment information to provide feedback on that process is really critical.
And I think really emphasizing, at the end of the day, we all have the same goal.
Like, our goal is for you to be successful, and the administrators to the teachers to the students all need to be excited about this goal that you're working towards together. And so I think when I think about the future of assessment, I'd love to kinda put all the bad history of teaching to the test, drill and kill all that stuff away, and think about how can we embrace assessment with a growth mindset, a mastery orientation, and a positive culture around it.
So Love it.
Love it. Russell, how about the future for you?
I have a micro and a macro. So my micro is about the kind of opportunity cost that comes with assessment. So the idea that when you're giving an assessment, you are not actively teaching. And so aside from the assessment as learning component, which would be fantastic as a way to kind of shift mindsets about assessment, design, and creation, where that can be a key additional outcome.
So basically finding that balance, which only comes through using kind of high quality, valid, reliable assessments. So making sure that the assessments that are used to inform instruction and make educational decisions are giving the best information in the shortest amount of time, because there's only so much time in a day, especially true if you're a teacher. And then the macro is kind of a broader, just big picture, like, how are we going to develop ways to more authentically evaluate student growth and mastery. And so, traditional psychometric tools and assessments are fantastic, but they provide information of a very particular and and somewhat limited scope.
And so finding kind of that true match between what a teacher is seeing in a classroom when he or she can see a student in all these different contexts and building assessment or assessment strategies that better capture that.
Yeah. I feel like we need to wrap up with, like, a super friend style theme song, psychometricians.
Like, I love this. I love what you guys are doing. This is amazing. Thanks for being on the show to share this with us.
Thank you so much for having us.
Yeah. Happy to be here.
Yeah. So much inspiring work here. Thank you. And as our future evolves, we're having you back. Absolutely. We've got all sorts of topics parking lot in here that we wanna talk more about. But thank you again for being here.
Yeah. Thanks for having us.
Thanks for listening to this episode of Educast three thousand. Don't forget to like, subscribe, and drop us a review on your favorite podcast player so you don't miss an episode. If you have a topic you'd like us to explore more, please email us at instructure cast at instructure dot com, or you can drop us a line on any of the socials. You can find more contact info in the show notes. Thanks for listening, and we'll catch you on the next episode of Educast three thousand.