librarypunk - 003 - LearningAnalytics.js

Episode Date: February 21, 2021

We're joined by Dorothea Salo to talk about learning analytics! Learn more about the Data Doubles project at https://datadoubles.org Add your pay: https://bit.ly/libsal  THE BIO BUTTON: https://www.chronicle.com/article/the-surveilled-student    

Rush Limbaugh claims to be pro-life, but dies anyway. Welcome to Library Punk. Hey, hello. My name is Justin, and this is what it would sound like if we're a beautiful woman. My voice mod is working, right? Because I can have to hear myself. I can totally hear you. Yeah, that's exactly what women sound like. That's how femininity works. See, the fun thing about being a trans T.M. is I did speech therapy, and so I can do that without a voice mom.

I can do mail-to-mail transition. Oh, yeah. Let's see how that goes. Hang on. I just called this one IRA. So I spent 16 years in the provisional IRA. They said we're going to put Margaret Thatcher's statue on a 10-foot length, and I said, you know, I reckon I can piss 11 feet.

I can do that, too. Hi, I'm the governor, Jesse Ventura, of the state of Minnesota, and I'm tired of these texts, these taxes, and you should vote for me this November in the elections for governor. I'm just even sure. Is he still running for stuff? No, but I'm pretty sure he has a YouTube presence. Mm-hmm.

Oh, definitely. Well, we can all make ourselves sound like Margaret Thatcher, because she's dead and can't talk anymore. Hang on, I think I have one for her. Hello, I'm against mine workers. I've never watched it. I have no idea how to do any of that. So welcome. My name is Justin. I am a scholarly communications librarian. My pronouns are he and him. Hi, I'm Sadie. I am an IT administrator at a public library. My pronouns are she, her, or are they, them?

I'm Jay. I am a metadata librarian, and my pronouns are he, him, or they, them. And I'm Dorothea. My pronouns are she, her, and I am a distinguished faculty associate in the Information School at the University of Wisconsin at Madison. Hi, and I'm Carrie. I'm a health sciences librarian in Milwaukee, and my pronouns are she, her. Yeah. So we have a guest. Our first guest, our first first guest. It's exciting. That's how we know we're trying to be

legit. Yeah. Legitimacy is overrated. I've managed to avoid it so far. Fair enough. Well, no, we're like super legit. We're on, is it circulating ideas? Like, he put us on his list of other library podcasts, I think. Wow.

Yeah, we're making the big leads now. Yeah, I guess. and I've got my own voice changer and my own sound board because the Sincaster broke the other one. Oh, no. I've got some fun stuff. This one's for you, Jay. The new power box. I don't even watch Game Grumps, but the first time someone showed that to me, I was like, yeah, it's so good.

And then, yeah, I went through the stuff that Carrie sent me. And that's on a hot button now. Yeah, anytime we mention like ALA or Elsevier or Freeze Peach, we do the... That's called a ham horn officially. I don't know if you know that, but the official terminology is a ham horn for those metadata enthusiasts out there. The day I learned. The only soundboard with a controlled vocabulary. That's right.

Library punk. That'd be fun, though, actually, like... Like, is there a taxonomy of sound effects? I'm curious now. Actually, yeah, there is because, like, there are, like, different kinds of screams. There's the Wilhelm scream, which is, like, a very official type of scream. So, yeah, there's, like, a taxonomy of screams and shit like that. So, uh, I mean, I took film classes.

I know about the Wilhelm scream. I didn't. And, yeah, I didn't take those kind of film classes. I took the ones where, like, wrote about, wrote papers about impotence in post-colonial Africa and stuff like that. So, yeah, really useful. That's the only paper I ever got an A-plus on. I got to write a paper about Toshiro Mufine's dick because I was examining post-war masculinity in Stray Dog and Rachamone.

So I just got to say Phalas a whole bunch, and I got it A. It was great. Does he hang hog in that movie? I mean, he and Rashamon. But then Stray Dog, he's a cop. and someone steals his gun. So, oh, no. That's a very convenient topic to write a paper on.

Oh, yeah. No, totally didn't do that for any reason. Mm-hmm. Yeah. So learning analytics. Yeah. Yeah, how about them? How about those learning analytics?

So, like, I mean, like, that's a good enough transition. Like, so, like, film class to learning analytics. Learning. Yeah. I mean, like, we learned stuff about penises in college, and now people want to monetize that. Exactly. Yeah. Well, I wanted to warm myself with the bio button because I was writing about it all week in the group chat while I had no power.

And it was like, all I could do was really read articles. The internet wasn't working that well. So until I ran out of battery too and then ran out to gas. I still want to have gas. Anyway, Texas. Insert fart joke. I don't have that drop. We are a refined podcast.

Excuse you. I have wine and everything. Absolutely. That I get from a snooty subscription service. Like, ho-ho. One of those assholes. I would have done that if they did it for whiskey. Jay, are you in a wine club?

No, I get, it's one of the things where it's like, a media property I enjoy told me to. And I was like, oh, that sounds like a good idea. It's Bright Sellers. When I was in grad school in Illinois, I actually got Wink. That was way cheaper. And then I moved to Utah where you can't even look at alcohol half the time. Right.

Oh, hi, Arthur. My cat for, you know, just jumped up. Hop up. Okay, that's a good site. Yeah. Bright Sellers is pretty good. They're a little more pricey. Okay, we don't have those endorsements yet.

So hold out for that. Yeah. I'm not have to believe all that. We're not the wine-sponsored library podcast. Yeah. But anyway, that's, yeah. It seems like that should already exist. I want the snack box subscription.

I want to chill snacks. I'd chill snacks with you. Yeah, like what would we sell our souls for? Yeah. I would take mattresses, too. Like, I want snacks and mattresses. Mm-hmm. Sox would be good.

Oh, yeah, socks. Well, I mean, I'm really particular about socks, so I wouldn't chill for just any sock. But, like, I want a princess and a pee situation where I have a stack of mattresses. Like, I've always wanted that. Live your dreams. Yeah. Maybe a podcast will get me there. Okay.

Just a picture of the princess and the pee, and it's just like women think it's okay to live like this. So the buy button. Oakland University Which is not in Oakland California. It is in Grand Rapids, Michigan For those who you don't know. That's silly. Yeah. Let me double check on that, but I'm almost positive

It's in Michigan somewhere. So there's like two Kansas cities or something? No, okay. No, this is nothing like the two Kansas cities. As a person who is from Kansas City, this is nothing like the two Kansas cities. Oh. Yeah, I saw your Twitter thread.

That was a fun. Oh, yeah. So Oakland University is in Oakland County, Michigan. Oh, it's by Detroit. Never mind. Okay. Because it's where Troy is, which is right outside of Detroit. It's also where the Detroit Zoo is.

Anyway. Yeah. So that's where Oakland University is, which is in Oakland, Michigan, not to be mistaken for Oakland, California, which is the most well-known to Oakland. The Kansas City thing is that the real Kansas City is in Missouri, but the two cities are right next to each other, even though everyone thinks that you're from Kansas.

Anyway, that's a controversial opinion. They should just rename that. Yeah, that place where the four states meet, they should just rename that Kansas City, so there are four of them. Yeah, the four corners down in the southwest. I'm pretty sure. Yeah, dude, let's like, yeah,

let's do another, like, freaking white colonialist project and rename. I'm pretty sure that's all native territory, too. Just name everything Kansas City. Why not? Have some consistency for once. People from Kansas City would love that. And as a person from Kansas City, I hate that. So Oakland University, which is in, I wasn't paying attention.

Detroit. It's in the state of Detroit. Yeah. The great state of Detroit, go lines. So they have to wear a coin-sized bio button attached to their chest with medical adhesive. So, like, you can't even, like, you can't even put it on a lanyard. You've got to glue it onto your tits every day. Like, what if you're allergic to the adhesive?

What, okay, so here's, okay. It's just super glue properly. I have an actual, like, bone deformity around my chesticular area, and it, like, it causes, like, and there's, like, two different forms of it, and one, can cause your chest to like poke out like a pigeon. Oh yeah, I've heard that. And the other causes you to be sunken chested. And I have like, yeah.

So like I have a snack tray in the middle of my body. Yeah. Convenient in a way. Yeah, it really is. What's the largest big gulp you can fit in it? I cannot fit a big gulp in there. It's really like I can fit like a handful of almonds in there. It's like really good for trail mix and stuff like that.

but it's not like a cup holder. I have been able to fit a can of beer in there. Nice. Like an eight-ounceer, but like not a tall voice. Like one of these skinny ones? No. A skinny one is like, I think like a good eight-ounceer. Like a just a regular size PVR or a ham's fits in there.

No, okay. Let's eat and digger. Yeah, like a thicker can. Because the skinny ones, like it's like a balance issue. Hmm. Yeah. Anyway, weird stuff. worry about my body.

We're learning so much today. The medical adhesive. Yeah. Please don't. The medical adhesive is not good. Yeah. I was like, I can't wear a heart rate. I can't wear a heart strap chest monitor because they're really uncomfortable because

of how my bone deformity works. But, you know, this does actually say something about some of the assumptions underlying these analytics programs, they kind of assume that all people are the same. Yeah. Yeah, it assumes you're able-bodied. Like, yeah, you don't, like, because I mean, there are people with, like, lots of other bone deformities that are way worse than mine. Just like skin allergies or.

Yeah. Oh, yeah. Yeah, if you have eczem or psoriasis or something like that. Or like, I'm one of those people that, like, adhesive. Yeah, if you break out. I don't know if I'm allergic to it, but it just, like, irritates my skin a lot. It's terrible. Yeah, I mean, that's.

Yeah. I mean, and the article didn't even go into that that much, even though it didn't. People are, um, people are alike enough that a one size fits all solution, if you can call it that, um, solves everybody's problems and everybody's happy world without end. I'm in. And I just, what is wrong with people? I mean, we could be here all day on that one, but I'm just saying. Yeah. It's, so the, the whole thing is for COVID. and it checks your temperature, respiratory rate, heart rate. So that's supposed to indicate if you have COVID in some way or a cold or the flu or, you know. Well, as somebody with anxiety, you know, if you're keeping track of my heart rate and my respiratory rate, I'm going to look like I'm high all of the time.

I'm also on not one, but two stimulants and also Wilbutrin and also testosterone. So my heart rate is like I should probably be dead. So like, and also like this, to go back to like the privacy element a little bit, not analytics, but like this sort of like is how, you know, is this actually necessary for the conditions we're in right now? Like what kind of tradeoffs are we making to because of the coronavirus situation? Like I'm still kind of not sure how I feel about like contact tracing and stuff because I'm like, privacy. No, don't even look at me. Don't put a thing on my phone. But also the sort of like we're in a pandemic.

Like it's hate it. Yeah. I'm also of the mind of like I feel like there are easier ways to be safe. You know, we have like everyone else is doing safe things easier. Like you don't have to do this big techno-dispopian thing. And I think maybe that's where. some of this comes from. There's, you know, my question for a lot of these measures is how well are they tailored

to the situation that we find ourselves in? At my shop, the way it works is that if you're in the dorms, you're getting tested, if you want to go into a library, or library, any campus building, you have to have had a negative test within the last eight days, basically the last week. but it's specifically a COVID test. Right. So it's nicely tailored. This bio button thing is picking up all kinds of information that it's COVID irrelevant.

It's not collecting any kind of relevant data to the problem at hand. Right. Way too broad. Yeah. Even though that's our excuse for getting it. Yep. Yeah, someone sold them on it. But the thing is like it's, I wasn't clear on how it knew if you had COVID.

that's not really explained fever well like it's supposed to go off if you're within eight feet of someone else with COVID and it uses Bluetooth to do that which Bluetooth is you know famously secure and so but how does it know if the other person near you has COVID they report it yeah because isn't that just the thing that like your iPhone can do where you if you're basically it can basically it probably has like a little memory stick in it and it's like probably storing some memory in there like you just We just need location data for that.

Yeah, exactly. Well, so I was reading about this a little bit, and because the different state contact tracing apps use kind of all over the place things. And when it comes to privacy, the Bluetooth version is actually better, if I'm remembering correctly, because it doesn't use GPS data,

so it can't actually dial you down to where you were at that moment. It just uses like a random string to pick up who is around you to remember who's around you, and it doesn't necessarily keep any information about the person that was around you either. So, yeah, Bluetooth as a protocol is not, you know, the world's most secure, but most people have their Bluetooth on their phones all of their time anyways. So, yeah. Yeah, the bio button is way, way overkill for that.

I do really like how the original student that they interviewed was like, what if I, you know, what if I go to a Black Lives Matter protest that turns violent, like am I getting punished for being there, starts this petition to, you know, make it not mandatory, but voluntary, and wakes up to right-wingers being like, this is communism. We're not going to wear it. No, the petition said that the college was turning communist. That reporting was really bizarre. That was like some really sloppy, I mean, it's the, this a chronicle.

But like, like, I mean, full disclosure. But I mean, like, I feel like that. Like, this is communist. Like, it's like, in what way? Like, just, like, I, Marx wrote nothing about this. And communism will have to stand in line for bread, the person says, in a bread line in America. Yeah. Yeah, like the point that failure of capitalism, this is socialism, this is communism.

Yeah. So I guess the leader of the team was like a philosophy professor, which was really funny to me. Because like if we just teach ethics, this stuff won't happen. It's funny what the humanities does. Mm-hmm. Yeah, because humanity is what's going to save everyone, right? Yeah, totally. Information literacy.

Yay! Yeah, I'm going to wrap about it, and that'll save everyone. No. No. No. Hang on, I got to be... Oh, hell yeah, Carrie, break it down. My name is eight-mile.

I got the framework. Watch my brainwork. All you other librarians. Can't handle all these slick licks that I'm spitting out. Watch me. Do my Prisma frameworks. Watch my flow chart. Watch me blow darts.

step. The story's the greatest thing that ever happened to me this week. That was horrifying. So bad. I went on YouTube to see if she'd already done one and I looked up library rap and I saw some racist-ass rap. Right. Because it's like it's not inherently racist to be a white person who's rapping. And all of these are like, my name is something and I'm here to say, like the educational rap style. Yeah, but they were wearing like, cringing back. Oh God. Just fully know. I know I I shouldn't be rapping. And I did that. And I'm aware. And I'm sorry, everyone. Go weed white fragility and then, you know, call us in the morning. It'll be okay.

I've got, I'm no longer talking to white people in one earbud, and I've got white fragility and the other earbud every night when I go to sleep. I've watched a few YouTube. I think I'm good. Yeah. Yeah. You can, can you please edit out that rap and everything that fell out up because I I don't think that was appropriate in any universe. No. Okay.

If you think it's funny and it'll pass, that's a, I trust your judgment. It'll be funny. I have no idea how the levels will balance out, but we'll see that. There was another thing in this article that, gosh,

what was it? Elypsis Health, which was an app that uses machine learning to flag speaking for depression. So you have to like answer a bunch of questions. It's just, Minlo was the university and this was like a Silicon Valley

university, I think. Yeah. And so they're just letting a startup run their health and wellness. Of course. Because that never goes wrong, Philadelphia. Oh, or who was that lady who made the blood thing? Oh, oh yeah.

Yeah, because I just watched that documentary is so good. Um, yes, it is quite good. Um, nanotainers. I like to, sometimes I cosplay as her just around the house because I can. I have the turtlenecks. I have the lipstick. Yeah.

It's, we'll get into this more because I'm, I think I'm going to skip the Lumen, well, I'll save the Lumin learning bit, but it's kind of like the first question I have for Dorothea, which is, you know, it says at the end of the, the article, maybe accreditation requires a certain level of exam. They're talking about exam proctoring. And as the students just want to know why. And I want to know, like, is that the case based because I was in your webinar last fall and I rewatched it? And I remember that students want to know why things are happening. But like, I'm not entirely sure that that's the whole story. Sorry, can you rephrase the question? I'm not sure what I'm not sure what I'm.

I'm answering here. Yeah, students just want to know why there's more exam security and then they would be okay with it. If you're talking about anything coming out of data doubles, all of the data doubles work so far has been pre-pandemic. All right, so we didn't ask about anything as specific as exam proctoring because, hey, that had not happened yet. So taking my data doubles hat off, just being a person who hangs around on Reddit and the discussion forums and my classes, especially my introductory information security class, are quite active. So what I'm hearing is not only do students want to know why this is being done to them. They want to understand why it has to be this intrusive. They want to know, aren't there any other ways to teach?

What they don't often ask, and this is something that I found, at least, in the first phase of data doubles, is that they don't think to ask, where else is the data going? what are the strictures on can it be sold? Can it be shared? Who gets access to it? How long does it persist? All of the questions that I is an extremely paranoid librarian would ask doesn't really occur to most students,

which personally I find deeply unfortunate, but here we are. Like they don't even know that those are questions to ask. Right. Let alone remember to ask them. Something we, all of us who were doing interviews for the first phase of data doubles heard from our respondents several times was, wow, you know what? I never really thought about these questions before. And there I am trying to keep my, trying to keep my poker face on. My proper research interviewer poker face.

It was difficult sometimes. Yeah. So once they kind of had that, I guess. I mean, what was your research method? So did your research method allow for, I guess, exploring that a little more? Or was it just kind of an aha moment of like, I didn't even consider that question anymore? Were you able to dig into that a little more or no? Sometimes, yeah.

The phase one of data doubles was structured interviews. So we had a spiel. We had a set list of questions. to ask folks. But we were allowed if we heard something interesting to probe. And that's pretty normal

for structured interviewing when you do hear something interesting. You're allowed to go a little deeper. I guess what I would say about that is, all right, I'm going to channel one of my co-investigators. For a number of respondents, what seemed to happen

was that during the interview itself, there their understanding would evolve, the things that they were saying would grow more nuanced. A number of them were just suddenly interested in a way that they weren't necessarily coming into the interview. So it was a fascinating process. It really was.

Yeah, I was reading in the preprint about kind of the methodology was, gosh, I've got it highlighted here, but the social constructivist framework, which is just by asking you're going to make them think about these questions. So I imagine that a lot of times you, you just have to let them work it out and just kind of sit there and wait for them to ask the follow-up question. And, you know, as is pretty normal with these things, they all had our contact information, or at least contact information for the person who was interviewing them. And I may, I don't

completely remember. But one or two at least said that they were interested in taking my information security and privacy course once they'd gotten through the interview. So nice little bonus for me. That's good at least. That's kind of cool, like, how that opens up a window for folks once they start to learn about that kind of thing. Because I think that's one of those things that, like, I think you start to see with, you know, some of the, especially like even just kind of covering that bio button story or whatever,

students are interested in these things and they are curious and they notice, they notice shit like that. Like as soon as you kind of tip them off to that kind of thing, they're curious. They want to know. Like all the dots connect suddenly. They start to kind of connect those things and they want to dig down deeper. And I, this was something that kind of, I mean, this was kind of a generalizing generational thing that I heard from a, I was in a student life training or something like, or, oh yeah, I do this kind of, I do university service with a living learning community. But like they were, they were doing an in-service thing and they were talking about how, especially this like, you know, Gen Z or

whatever you want to call them, they're really interested in particularly like digging down and like getting to the root of things and really fully understanding the whole scope of something, which I thought was really interesting because, you know, there's some variation on like, you know, what kind of generalizations we make about generations. But I thought that was a really interesting perspective that like, you know, so often we read literature that makes a lot of assumptions about digital natives and stuff like that. And I'm just like, no, these folks are pretty tuned in. And they and I think that was something that I think spoke some truth to, I think,

what you're observing and what I've observed a little bit to in some of the classes that I've taught. Last summer, I premiered a new assignment in introduction to infasek. I basically had them find out whatever they could about a particular kind of data gathered about students on campus. And it was the first time this never happened to me before that pretty much every single assignment that was turned in went way over the page limit. because they wanted to tell me what they were finding out. They wanted to tell me what they looked for, the information that they expected would be there, but was not what they couldn't find out.

And they totally wanted to tell me how they felt about it. Good. Yeah. I'm not objecting. Like, hell yeah. Good for them. Yeah. Yeah, I think definitely a lot of it, at least coming from my side of things,

is a lot of things people just aren't aware enough to ask questions because a couple of months ago, I did like a basic security presentation for my library. And I had a lot of people afterwards being like, wow, I didn't even know that about, you know, passwords or, you know, and it wasn't anything, it wasn't anything like really revelatory. It was, you know, how to how to use pass phrases instead of passwords and, you know, how to spot spam emails and stuff like that. So it was, I think as soon as people know, regardless of generation, people are going to start thinking about it.

And one of the things about it is that so often it is framed, like, and this goes sort of back to like our privacy discussion we had. Of course it does, yeah. Is that like we so often frame this as a very individualistic concern. You know, what are they gathering about me? How can I stop them gathering it about me? And what we rarely do is we never really frame it as like, how do you make sure the network of people you interact with are, you know, doing these things?

If you're a professor, what are you doing to help your students to this kind of thing? Like, I feel like that's an angle that's often left out of privacy discussions, especially like in universities where every single thing a student looks at is, or anybody it looks at is gathering information about them. Like even, so we have like Alma Primo where I am and like I can totally just go in and see like with the top hundred or whatever search terms are by month, by year, by user group, faculty, staff, undergrad, grad, like all of these different user groups. And I can just like go in and look at that. It won't tell me like it was this user specifically. but it's like that's data that's being collected on like in my university community and I bet no one besides like me even knows that that's being collected.

So yeah. So a funny thing about that. Yeah. Funny thing about that. Most in both the interview phase and the survey phase of the data doubles research, students were pretty. pretty strong in their beliefs that, for example, the library was keeping track of everything that they checked out, like sort of a library permanent record, if you will, they expected that. Oh, yeah. People, every job I've ever had where I've worked circulation, people expect that you

have a permanent record of what they've, what they've checked out. And I'm like, no, why don't like we do that? Oh, yeah. No, it's not like that. Yeah. So, you know, we have a little bit, I think, of a, of a communications, a public relations problem around, I mean, it's not the only problem. We have, goodness knows. But a PR problem around our privacy commitments, people have no idea. Mm-hmm. Mm-hmm. And they certainly don't know why.

Yeah. Yeah. And, like, another issue with it is like, so I've been getting interested in, like, user experience. testing and the different methods for that. And it's sort of like, when I was looking at the notes for this episode, I was like, well, collecting some of this information is actually really useful from like a user experience testing standpoint or for like a curriculum development or just like improvements, like seeing how are people interacting with your university. How are people succeeding in a class or something where it's like, oh, this is actually

useful information most, you know, sometimes. Sometimes they get irrelevant stuff that doesn't matter because it's like, you know, just collect everything you can and that's bad. But it's sort of like, like there's even the, like the user experience testing where it's like you track people's eye movements and everything. You ask them like, I'm very interested in like how different user groups or different disciplines search.

And it's like, that's pretty invasive actually. It is quite invasive, but it doesn't happen without their knowledge and consent. True, yeah. It makes a huge difference, all right? And again, this is something that came up in data doubles. A lot of students were like, we're nodding their heads and saying, yeah, I think we're okay with this, but gosh, I really feel it you should tell me first. Like where the practice itself isn't the problem, it's just people haven't even been informed. Yes.

And what exactly is being collected? It's like, oh, I'm okay with the university or whomst ever knowing this, but they don't need to know this. So why are they collecting this along with it? Yeah. Yeah. When I did user testing, like, I went into a room and was like on camera. So it wasn't, you know, people knew what was going on. But when I was working and I controlled basically every system at my last university, I put in just a very basic like,

free web tracker and you can track where people click, where people move, and all this browser data that was extremely easy to collect. I'm interested though, going back to your course

where the students did assignments on information gathered on them, did any of them do anything in like IT, like the university network campus sort of thing? Last summer, no. This spring, I have some students looking into a couple of things that are related to that.

I think it's a function of ignorance, again, how many students realize that logging on to the Wi-Fi on their campus basically geolocates them to a considerable level of precision. Like, I know what building you're in. I probably know what floor you're on. Yeah, I mean, if you just look at, you know, which access point they're connected to, that gives you a very small area where they could be standing. And you can probably even track their path if you've got like a mesh network of access points. You can watch them walk from one access point to the next and just know exactly where they're going if you're looking at it in real time. And it's 100% identified because.

because you had to log on to get on the Wi-Fi in the first place. I mean, public libraries have that, too. Then, like, this isn't just an academia problem. No, no. I mean, none of our libraries are really huge, but the ones that are, the one that is big enough to have multiple access points, if I wanted to, I could watch a Mac address of a cell phone or a laptop walk around,

you know, if we were open. Of course, we're not open. But I could watch that if I wanted, and nobody would not be any of the wiser. And then you turn on wire shark. Yeah, exactly. And start packet capturing from the entire wireless network. Exactly.

You know the back address. You can see where this person is going on the web. It's so easy. I teach my undergrads to do it. Yep. And actually, I just had this come up at work not too long ago. We redid a whole bunch of our network stuff at a couple of our branches. And we were using an outside vendor for it.

because they had the particular knowledge of our, like, geographical area, sort of how the network was laid out. And when they set up the access points, they set them up. So they were hopping on the same network as all of our public computers and, you know, getting IP addresses that way and stuff. And I just, like, I had a tiny panic attack because, like, our network pulls so much data for our devices. I don't want to be pulling that same data for patrons devices, you know. And I was just like immediately changed it over. So it's less, a lot less trackable. Changed the mode of the access points.

And when I talked to them, they were like, oh, well, you know, we were thinking because that way you could you could block a bad actor a lot easier if you did it this way, as opposed to the way that you just changed it to do. And I'm like, okay, yeah, I get that we will need to block bad actors if it happens. but I don't know if that's really worth compromising the privacy of literally everybody who hops on our Wi-Fi network. Can I ask a question? I had something to say about that. Yeah, go ahead.

Yeah, so I'm relatively new to learning analytics. It's something I hear, but, and it's not that I haven't done, like, instruction or anything, but that's not what I studied or anything. It's like, yeah, I know, you know some, like, infosex stuff, but, like, the whole learning analytics, thing I'm fairly new to. Is learning analytics just like data gathering and privacy invasion but with the purpose of making like courses

better or something or just like making services better? Like is it kind of privacy invasion with that's the rationale. Okay. That's the rationale. Yeah. I mean, you know, it's big data for education. So it's not pretty much the same

discourse around it as big data everywhere else. Oh, so it's now just like a buzzword where it's like we're an educational place and we're data gathering and that's learning analytics and then you can do anything with it? Our big data is learning analytics, yes, exactly. Oh, okay, so it's not even like, we are definitely

using this for this. It's like, no, we're just collecting it, but because we're a college, it's learning analytics now, wink. And I mean, you know, the same pardon my language, bullshit rationalization or what is, what is,

it is, it is surveillance. I'm using the S word. It is surveillance. Right. Yeah. Hamhorned surveillance. The same stuff that you see everywhere else. Smart cities, right?

Predicted policing, all of that crap. Yeah, it's skinner. If we just know more about them, they'll be safer, they'll be better, everything will be wonderful. I can't. No, thank you. no thank you techno paternalism. Right. You can see yourself right out of there.

And I mean, yeah, I've, I, I, uh, I, uh, I read the, the Shoshana Zuboff book, surveillance capitalism. So, you know, I'm, I'm, I'm part of that cult. So, uh, you know, I am all about, I think the bit about the bit about the bit that really hit me hard was when she talked about, um, Walden 2 and how, so many of these tech companies are so kind of, I guess, driven by this idea of behavioral predictives. And, you know, I think that's part of what I see so much in these learning analytics

conversations is like, if only we can like predict your behavior, then we can save you money, then we can graduate you faster than we can, but there's so much money to be made in it too. Or like the admissions ones where they can look and be like, we can tell whether or not you'll succeed or something. so that basis if whether we admit you to our college. Oh, no. It's more insidious than that.

Oh, yeah. It's even more insidious than that. A number of, go to the Chronicle, go to inside higher ed, you'll find this one. I forget where exactly got caught doing this, but they were deciding whether students, you know, a student's level of interest in matriculating at that particular university. based on how long they spent on the website. Oh, wow. And they're like so many people.

Oh, yeah, I remember that one. Right. There's a term of art in research methods and statistics, availability bias, which is that you're biased toward the data that you can actually gather, even when it is totally bullshitting you. And that's, I think, what's happening in an awful lot of,

Big data generally in learning analytics specifically. These are the data that we can gather, so we're going to, you know, milk everything we can out of stuff. Dorothea, you'll get a real kick out of this. The school in question looks like it was UW Stout. Yeah. That would explain why I'd heard about it. Yeah. Yeah, no.

I was in like a committee meeting, like on a campus, like a college level, like a university committee service, yay tenure. And it was like before the meeting started and someone in the committee was talking about like, oh yeah, we're thinking about, you know, getting this like AI analytics thing that will, you know, determine whether or not we should admit someone like whether or not they'll succeed or something. And I just like was horrified. And I was like, no. And like they're like, and I was like, are you kidding me? What? And it's like, and I started talking about like, like, oh my God, like that there's no way that's going to end up good. Just like the ethics alone. And like, like, I was like, are you kidding. And like, and like, and like, I was like, like, and I was like, like, and like, like, the bias, like algorithm bias. And the person was like, well, like, you know, there's a data science, computer science person here. She probably disagrees with you and can assure you at safe as stuff. I was like, okay. Like, you know, I'm not saying that like teaching the humanities TM is going to save us,

but there actually is a difference in how some disciplines are taught with like, are you actually being trained to think about people? when you're doing this. Like that doesn't mean you will, but is that mindset even exposed to you when you're designing these things? When I was taking IT classes, I was like astounded.

I worked in libraries before I went into IT, so I already kind of had that mindset. But when I'm taking these classes, there's no what about the people behind it. It's literally just about the systems and what they can and can't do, nothing about why you would want to do it that way other than outside of just network security kind of stuff, which, as I'm sure we're all learning, is often a word for, you know, security is actually a buzzword for privacy invasion in a lot of ways.

That is one thing I will give librarianship and, like, at least the program I went to is that it did try to frame everything as like, and we're doing this for people. That doesn't mean that librarianship is actually doing that, but they pretend to teach you that. in your classes. Pretend. They do make a... I teach you that. Yeah, they make a pretty good effort at that, but I mean, there's also like a little...

There's some softness going on there. And some like, you know, vocational all white saviorism is a huge part of it a lot of the time. Yeah. But yeah, no, that's... Often when I hear these discussions about, like, big data or, like, data collecting, especially when it's like an education

or librarianship. It's like, was no one actually thinking about the people they're collecting data about? Like, even if their intentions are good, like, you know, let's say they're not like, you know, capitalist demons. What if they're like, no, honestly, getting this data will make this, you know, in this software will help you, like, do this, like, in that sort of techno-utopia thing where, like, they actually have good intentions. but because they aren't necessarily thinking about the more or less like there's a real human person who you're getting information about when you do this, where it's like you sort of lump everyone into the user, TM, the patron, the student, and ignore that it's like these hundreds, thousands of individual people. It just always sort of blows my mind because I'm like, you know, if you actually thought about this a little bit, this would actually be really cool and might actually be super helpful, but right now it's

terrifying. The other thing I want to point out here is that the evidence base, I mean, we've been doing learning analytics for a few years now. There's been more than enough time to show that it's demonstrated

actual benefits, and it largely has not... Really? Right. The big success story is from Georgia. and the the the Georgia State one

right the wise and where wherefores are really really muddy there because they didn't just do the big data they also threw a bunch of money at a bunch of undergraduate advisors like human beings okay

so Occam's razor it's not the analytics yeah they really bury that one in there. Yeah, yeah. But like... Sunday. Oh, yeah. They really do. They really make a big stink about the analytics, but like, if you dig through enough, then you kind of

start to see that there are people there. Yeah. But that finally makes sense with the Georgia thing. Yeah. But aside from that, aside from that, right? Huge hype about big data thing, nudging, whatever, dealing with summer melds, whatever. None of it works. So what the hell are we invading people's privacy for if this stuff

doesn't even work? Well, and it's also expensive too. Right. Where are we spending our money? Yeah. Could we spend our money on people instead of machines, please? Mm-hmm. Mm-hmm. Yeah, I would really like that. That's the thing. I skipped over the Lumen Learning bit because that's more my area of you know the

argument this is David Wiley's company and he works a lot with OER and Lumen Learning sells courseware which has you know machine learning feedback that sort of thing

and his whole argument is well OER in itself is pretty much neutral in terms of student success which is that's not what I've been reading for a lot of years yeah it's I mean we can

put that to the side for now. Okay. And, and then so, but his big argument is that it's the, the adaptive courseware that really makes the learning improvements. But, you know, if, if that's not the case, plus also it makes money and also it keeps OER sustainable. That's the argument, sustainability. I have in here an italics very serious person. So that's kind of our running, that's our running enemy on this show is very serious people. You ever think there's a tiny bit of motivated reasoning, the guy

from Lumen Learning going, hey, here is how we can make Lumen Learning more sustainable. Oh, and that he canceled, like, he took his ball and went home because people got mad that he invited a bunch of publishers to an open education panel keynote. I remember that. I was there.

It was very funny. He told nobody before it happened. So, So he just said, this is the last one during the keynote. And he's like, it's up to you guys now. Wow. And then I was in the planning for Open at 20.

I have another question. So based on the sort of like, you know, we have all of these big data and like learning analytics and stuff. And oh, because evidence based, right? That's the whole reason we're doing. Right. And then it shows that actually it hasn't made that difference that. you know, we even might may have made these changes or something and it still just doesn't matter. What does that say about then like, because even with like things like user experience testing,

you're getting such a small percentage of the population who might be using your discovery layer, for instance. And then like you're making changes of like, you know, how easy is it for a student to figure out how to sign into their library count or something? Because that's still data you're collecting. And so if we're getting this huge big data, getting information on everyone and it's not helping, is smaller user experience testing where it's like smaller amounts of data? Is that helping? Like, just collecting data about anyone help actually?

Or is this something we tell ourselves to make ourselves like think that we're doing a good thing and doing right by people? I think that's part of it, right? We like to feel that we're not just sitting here. we're doing something. And there's fashions in that. But I think some of it does go back to method. Okay. I think it was Jacob Nielsen, who was like, you know, test on three people. That's all you need. You will get plenty of actionable feedback from like three people to find plenty of stuff that you need to fix. Then you fix it and you get another three people. Right. You don't necessarily need large.

numbers to learn something about a system as relatively uncomplex as a website, not even the entire website, but a part of a website, but you can test methods, right? Their entire fields of endeavor that are basically trying to figure out whether and how certain research methods work. With big data, We're kind of undergoing testing in the field, if you will. We're all just kind of ready fire aim about all this. You removes the methods from it, kind of. Right. Yeah.

Going, who was it? It was Mr. Longtail, Chris Anderson, whoever the hell that was, who was just like, oh, well, okay, this is the end of the scientific method. It doesn't matter anymore because of data. That's the biggest person I've ever heard. I may have the attribution wrong, by the way. I would have to look it up. But anyway, so we're starting to get the feedback about how well this stuff works.

And what we've been learning is it's incredibly biased. Yeah. When you start a big data project, you are basically baking bias right into it. Predictions are garbage. There was, and this was another local project, well, partly. It was a multi-site thing. But anyway, they did a contest, basically, with a huge data set on American families, basically asking people to predict, okay, how well do these kids in these families do? Right.

And so lots and lots of fun, big data, number crunching, many, many different models. And what turned out to be the best predictor was basically a simple regression. That was all you needed to do. Right. We know how to do regressions for centuries. too, you don't need all of the big iron. And in fact, it doesn't work. So this is not a message that's terribly palatable to big data, data science communities.

But I have to think the day of reckoning is coming, right? You can't ignore the really terrible track record that's being built up. And I will say, around people-centered big data projects. Right. You want to throw big data models at predicting what a quasar is going to do. You want to throw it at climatology. I have no problems with that. Yeah, that's fantastic. Like, that's great.

Yeah. Like, yeah, when you can, like, you know, throw a bunch of data into a machine and have it spit out, you know, like science. Like, when it's, you know, like, it just gives you what you give it, right? Yeah. Yeah. When people don't, aren't at stake like that. Mm-hmm. Nope.

Yeah, it's like, and I think, you know, it's not necessarily people framing it as like algorithmic bias or AI bias, but as these stories of like, oh, we made a Facebook chatbot or something that turned into a Nazi in like two minutes. Like those kinds of new stories that kind of get, right, that kind of get widespread appeal. I think that might be helpful for people to start seeing like, oh, like just because, you know, code is code. when you put human beings and human behavior and whatnot in it, that sort of throws out whatever sterile, neutral idea about code you have. Because people are right and like code is code and whatever you give it, it will just do it what you tell it to do. But it's like the way you write it and then the data that you actually give it. And so, yeah. Machine learning is pattern matching.

All right. It's finding patterns. And machine learning has no ethics. So it cannot tell the difference between, oh, this is an interesting pattern that we like and we should maybe follow up on. And this is a pattern of racism. Right. This is a pattern caused by transmisia, right? Machine learning can't do that.

And too many people are adopting machine learning without taking on the associated responsibility of making those judgments themselves. And even if you put people back into the equation, that doesn't necessarily mean that they're going to be able to pick on that either. Absolutely, because a lot of these predictions are incredibly inscrutable. Yeah, a lot of algorithms are very, very black boxy. like, you know, their intellectual property for Google or whatever. So, of course, they're not going to release the code. So therefore, you know, it can't be vetted by anybody who has something more ethical in mind. Google who's fired how many AISists at this point?

Oh, my God. I feel like every other day I think of that. I have the great opportunity to see Dr. Sophia Noble. And, yeah, I'm starting her book now. It's been on my list for a very long time. But, yeah, and that just reminds me of, like, her research. research with just the Google like search. Absolutely.

Yep. Machine learning is going to see that, you know, the word Asian is associated with porn. And that's the conclusion it's going to draw. It's not going to be able to say, hey, maybe there might be something racist about this. Because it doesn't know what porn means. It doesn't know what it means. And like, that's another reason why like actually like semantic web and link data is important. Because you can sort of be like, this is what this word means actually.

I mean, it's not going to solve that problem, but I know just like libraries as a general have been so slow to even adopt the concept of like, maybe we should tell our machines what we actually mean when we say this is what a subject. Yeah, given more of context. Yeah. Yeah, like context is so important with these things. There's also that book, is it like algorithmic bias in library discovery that came out? Like I have that read the. Yeah, I have that checked out and I've read the book. like online article about it.

I need to read it because I handle our discovery layer where I am. But like it's a infuriated. Oh, totally. Yeah. Oh, trust me, I'm already infuriated just because I'm like, oh, I know there's something bad. But it's like, I'm going to say, my ADHD is happening. Yeah, I forgot what I was going to say. But anyway, yeah, that's a good book. People should probably read because I haven't read it yet, but I'm assuming it's,

it's great. Maybe I'll think of what I was going to say in a second here. Let me just think. And then y'all can go talking one, I think. I was going to, because earlier Jay was asking about learning analytics and institutional analytics, those lines kind of are blurring because when I was, they were originally defined differently, whereas institution analytics would be more like the advising, like, you've got to see in this course. so we're going to send you to a counselor.

And I also had a question in there if you knew about the Georgia State, was that also targeted financial assistance? Because it said it was mostly low-income students. So did they also do like small grants, I wonder? I don't recall offhand. I know there have been small grant programs at various institutions. I don't know if Georgia State was one of them. I wouldn't be surprised.

And I wouldn't be surprised if that worked. Well, I've heard they're pretty successful when you do, just when you have a student in like the like the second to last semester just giving them a thousand dollars it increases their odds of graduating significantly which makes perfect sense you know you have to invest in things that whole thing reminds me of the old joke about the person looking for his watch under a lamp post because that's where the light is even though it's not where he lost the watch availability bias again right we're gathering all of this click data about what people are doing in the learning management system or whatever and we're

not looking at their financial situation. So if we only predicate our interventions on the data that we can gather from the learning management system, we're missing a lot. Yeah. It's an over-reliance on what's available rather than focusing on or like what can be easily exploited or something. You know, maybe, I don't know if that's the right term, but I think that maybe is the right direction. Yes.

I remember what I was going to say now. So I, like I said, I handle our discovery layer, and I am but a lowly English BA. You know, I took courses in like spooky literature and gay literature. And like, you know, I look at math and I go, even though I'm quite good at math, actually. But I've been having to teach myself, and by teaching myself, I mean, you know, go into like the web development. developer mode on other institutions that have like Primo VE and like look at like, oh, this is what their job is like their custom JavaScript file looks like. And this is what their CSS file looks like. And so many of them have like the Google Analytics put in. And like I know, yeah, like I know my library website as a whole has Google Analytics put into it. I haven't put Google Analytics into Primo. God no. But it just sort of makes me think because it's like, you know, knowing how many people have visited. the website and how they're interacting with the website, that can actually be important information, but there are other options that aren't Google analytics. There's like free open source,

like programs like that that you have more control over. And it's like, especially at my institution, and I know my institution is not the only one where this is a problem. The university is like a budget thing, got rid of library IT. Campus IT now. And it's like, so it's like we don't have control over our own website anymore. And so it's like, you know, libraries, like, we have this over reliance on third parties. Now, like on vendors, on, you know, more like systematic or consensual or academic IT and not being able to control our own stuff because we're not given the funding or we're not given the actual like workforce to be able to do that stuff ourselves. Because if we could do it ourselves, we might actually be able to do it.

it in a more conscious way, but because they don't give us the resources for it, we have to resort to, well, you know, campus IT does this and we don't have a lot of control over it. We can like talk to them really nice and they might think about it. But then that's a decision that will affect the entire university and not just the library anymore. So like, you know, because of budgetary crisis and whatnot, it's like how much are we forced to sacrifice? Like what is taken away from us because of that. And like the whole data collection thing is definitely a part of it. Like it got, it's worse now because we were thinking about moving to like a free open source

non-Google alternative. And then they're like, psych, you don't have IT anymore because the Huron report. Yeah. You now have to file a ticket for any kind of help. Even if you want to just go upstairs and talk to someone, you have to. file ticket. Yeah, because they're housed in the library. Yep, no, so I've got a file ticket. Go through

the help desk. Yeah. Another place this is coming up and I encourage everybody to keep an eye on it is a single sign on authentication for library e resources. Oh yeah, I've heard about that one. Like J-score like forces you to like have an account now or something. That's

one tactic that they're taking but you also need to watch the seamless access people. Yep. Okay. All right. Be watching. them real closely. They basically want everybody on Shibboleth, Samel's single sign on. And the way

that they are designing, I guess I will say, the standards and the tech around this, pretty much guarantees that vendors will know a lot more about what identifiable individual

patrons are searching for, looking at, etc. Yeah, this is what we talked about on the privacy episode. Oh, excellent. Good. Yeah, I was going to say. Yeah. Rock on.

Because we were talking about the Springer integration, because it's basically what you were talking about, the single sign on, is they want to know if there's leakage happening. So if you're on Researchgate, they want to know what university you're with anyway because they want to say, oh, look, this is our content on Researchgate. And then they show you the statistics. because they want to remain relevant because they know that we will unsubscribe if usage numbers just go too low. That's one of the statistics we gather that actually matter, which is if no one's using it,

we will cancel the service. There's another aspect to this, though. Well, there's many other aspects to it, but the one that I want to bring up right now is in the shadow of a young man who is no longer with us named Aaron Swartz. Right. Yeah. Yeah. We, if, you know, if seamless access goes through, I confidently predict many more lawsuits against both universities and individual patrons.

Oh, yeah. Because the proxy server will not be in the way. Yep. Right? Mm-hmm. The vendor will know exactly who did what. Mm-hmm. Yeah.

Yeah. Yeah. And then something else I brought up back on the privacy episode. how and like library freedom project like shout out to them every single episode we do we should have like a little like oh j mentions library freedom like they have like a vendor oh we need a ring the singing freedom yeah like they have like a vendor scorecard you can look at for like library vendor and like i said like oh um like when i when i have taught instruction i go oh by the way if you have like privacy badger or something that blocks trackers or third party cookies or something that it might break this database. So if you're having a problem using it, you might actually have to go in and, like,

adjust your settings or something. Because it is collecting info. And, like, I know that personally because I've had that problem where I've gone into databases, and it's just, like, not working the way I want it to because of whatever settings I have turned on. Or even just, like, it won't work without this certain, oh, God, I don't know, like, some, like, benign setting you can turn on or off.

it just won't work without that. And students don't know this. And it's like, we're the ones basically doing product demonstration for vendors. Yep. And we're not telling our students or patrons. Because like in a public library, like I'm assuming you have classes. Guys, I've never worked in a public library. Because public libraries have databases too, right?

Yeah. Yeah. You do. We do. And I'm assuming there's some sort of like instruction, even if it's one-on-one in a reference interaction. It's like those patrons are also getting data collected about them. And they access through proxy server and they log in using their library account.

And so it's like, you know, not only is that an invasion of privacy, but it's like the types of searches they're doing and like the more like, you know, quote, learning analytics because that's the good way, right, like data that they're collecting about it. With just like our library software. You can also look at like the type of product that, you know, this affects a variety, this would affect a variety of products. So in addition to like a scholarly academic database, you can look at electronic

resources that provide different types of information. So you could even look at like a citation management tool like Mendley, for example, which else of your owns now. Yeah, Taco Bell owns that. which is a demolition man reference. I love demolition man. It's so good. So good, the three shells. I'm a connoisse here of the three shells.

And yeah, and so it's one of those situations where like, you know, when you look at other electronic resources, in addition to just like our straight-up academic databases and what they're doing with intellectual property, you know, you can even look at expanding into public libraries and things. thinking about other types of electronic resources that this would be vulnerable to. So looking at like people, I remember when I worked in public libraries, one of the things that we had access to was like Chilton manuals online and things like that. Or like, oh, if you have like access to like LinkedIn learning or something like that, that's another platform that's going to, that's exploiting a lot of data out of people. LinkedIn learning.

Yeah, we have that too. Yeah. So ours is actually run through our IT department. oddly enough. Mm-hmm. So, yeah, library problem. It's like,

when it's in a public library, is it still called learning analytics, or is it just big data then? I don't know. It's called customer relationship management. Oh, okay. So it's all the same thing.

It's just got, you know, whatever context appropriate. Yeah. Gotcha. Yeah, I got out of public libraries before that really crept in.

Yeah, that happened at my, my previous library, too, is, like, I think we had been using Orange Boy. I don't know if any of you... Yeah, yeah.

I'll assume it's bad. It's bad. It's bad. And we were transitioning to a different service that was worse. I forget what it was called now, but... O-P-L-E-Ys? No. That's one of them.

They're other. Yeah. But part of it was like, you know, the big data kind of collection. We were a couple of us in the IT department. I was fortunate enough to work with another library. He was a librarian turned IT. And we were both just like, why would we ever want to make this decision? And we were basically told, well, it'll help us in marketing in the long run. And we need that to be able to pass levies. So we recognize that it's kind of not necessarily compromise. That wasn't the word they used, but that was basically what they said. We're willing to

compromise to be able to continue to get funding in the future by having all of this marketing information on our patrons. It's like where you're forced into that situation, even though you don't agree with it. Exactly. Like how all of the public libraries, it's like even though the ALA, you know, air horn, but like there were, you know, as a profession, we're very against like the internet filtering. But so many, I guess, like districts, cities, like whatever, funding bodies of public libraries will say, like, if you don't put filters on your public computers, we won't give you money. Oh, yeah.

No, it's a federal program called E-Rate, and I've dealt with it a whole bunch where basically- Yeah, that's part of, yeah, IMLS. Yeah. Yeah, if you don't use, if you're not CEPA compliant with your filtering, you can't get the, well, like, 75% or 70% refund for your internet. connection costs, basically. So yeah, it's basically flagrail. Yeah, like the neoliberalization of everything, basically. Like I know. Yeah, I think that was a

Janet Reno case. As I recall, I think I learned about that one in library school. I don't ask me what I remain during that part of my life. But as I recall, the case to have to filter internet and public libraries in order to get funding and it's only for people under 18 because the kids will look at

the porn exactly you can't have kids knowing about sex and also like the the ability to get around that by you get around people looking at porn with acceptable use

policies yeah And that was, when I worked in public libraries, that was a cudgel. We wielded pretty aggressively. Yeah. Fun tangent. So when I was in grad school, very first class semester, I took one of my friends for our final, you know, poster project we had to do.

She did hers on just like the history of like pornography in libraries and how all of the sort of discussion about it used to be like, do we collect it? How do we collect it? had a weed store, preserve it. And then as soon as the internet happened, it's all just about filtering. And it's like the discussion about everything else stopped. Yeah. I mean, actually like some of that still does happen a little bit on the collection development side. Like you'll still see some articles around that and like how do you, if you do collect it like where do you keep it? And how do you like there was a big debate like when the Madonna book sex came out? Oh, I bet. Yeah. And like where do you,

And like, I remember one of my old professors said, oh, yeah, we just, like, kept it in a reference. Yeah. Because it was an oversight. Yeah. The good books are the ones that say C librarian on them. Yeah. Yeah, exactly. They're non-circulating.

Like, it's just, like, make sure everyone can see it. Makes it accessible. It also comes up in the context of e-book approval plans or e-book bundle. There have been a number of e-book deals with various public libraries, public library consortia, that went south because nobody thought to ask, okay, how much erotica is in this particular bundle? And local sensibilities were not respected. I mean, that's the only reason I even, like, use my local library Libby account as I'm like, what kind of romance novels do they have that I like to read? Yeah. And the answer is none.

but when I lived in like Salt Lake City they were a little better because it's a bigger I live in not a big place anymore. Oh yeah. There's not enough money for those. I mean, they're fucking expensive. Yeah. I live in Western Washington. I went out of my way to get a King County library card because they have one of the biggest overdrive collections. Oh, I think you have to pay for that.

Oh, no, I didn't. Okay, good. Reciprocal agreement because. Oh, no. Yeah. Yeah. So I won't know, I don't know if I should admit this, but I cheat. So I have, I have three library, three active library cards. My boyfriend is the same thing. Yeah, my boyfriend has one from like where he used to live and just like he hasn't told them. So because my parents live in Kansas City, they have like, my mom has like three reciprocal cards and I can use her old address to maintain, like,

like three reciprocal cards there because they have a weird library districting system because of essentially what I will call racism. Yeah. Yeah. Yeah. Library jandering. I worked for both public library systems there too. And I will say that a lot of that is rooted in racism.

Oh, yeah. If you're ever listening to this major public library systems of Kansas City, I am paying attention. and I don't care I'm never working there again No skim off my back And also my local Milwaukee public library Which I finally returned a book that I had

Checked out for two years so I can use their services again So good for me Librarians are the worst at returning books I was just about to say we never return shit We're the worst We are like no I know this works is mine Like, I have, like, yeah, I have like a $115 late fee before we got rid of them at my, at my work. At least I can't blame mine on COVID.

I don't think I owe it, actually. I'll pay it someday. I'll pay it eventually. Like, don't worry. Just not right now. Yeah. Just not like, you know, I don't feel like it. Like at that point, the replacement fee is cheaper than the late fee.

Just buy them the new buck. Yeah. I mean, they have them. Yeah. Yeah, so it's like, what harm is done? Yeah, they have the things I returned. Exactly. It was some, it was some documentaries, actually, so they might not have been that much.

It might have been actually more expensive to replace them because academic licenses. The, like, public performance rights that get bundled when you buy things. Yeah. Yeah, it's so, I'm like, I might be a metadata person, but I am like a huge copyright nerd. Like, I just finished up the new. the first wave of the New Hampshire, like, copyright first responders. And, like, I took a copyright class in, like, grad school and everything. And, like, yeah, public performance rights, which I actually learned about as an undergrad because I did Rocky Horror.

And so every fall, we had to pay $500 to Criterion, whatever, not like Criterion Collection, but, like, Criterion is also the name of a place you could buy performance rights for. And, yeah, Rocky Horror. And it's like we had to be like, well, we're only performing it once. Yes, we are charging admission. It's $5. Like all of this stuff. And like, yeah, it was a like the whole reason we did any fundraising and paid, you know, had membership fees for a little sci-fi fantasy club was to pay for Rocky Horror every year,

every year. So that's like the whole reason I know about that kind of stuff. So fun. I miss doing it. Anyway. We're at an hour and a half and I want to do this. this kind of ideal world segment where we kind of imagine utopia. So so much of this is like caught up in being very serious people, right?

Like we're working within the system we're in. But in an ideal world. I'm not capable of that just so you know. Well, we all get caught up in it sometimes. Maybe you don't. I've never been chill about anything in my entire life. Very serious. I am persona non grata in the entire state.

of Minnesota, released among its librarians. What about among Governor Jesse Ventura? I think Jesse Ventura's got a soft spot for you because you protect my data. All right, again. My cat is staring at me because I'm laughing.

He did a big stretch. Oh, right behind me. Okay. Yeah. So in an ideal world where in luxury, gay space automated communism. What would learning analytics still exist? And if so, what would they look like?

They would certainly not exist in their present form. Where they exist, they would be in the form that we're familiar with and familiar with the ethics of controlled, consented experimentation. We're like some of the data are actually useful and we'll actually go towards a good purpose. Well, and we don't get what we're seeing, particularly in library learning analytics, of what I call no notification, no consent specials, right? Where data is being hoovered up and sprayed all over the place. And students are not told. Students can't opt out. And it's just incredibly irresponsible and it needs to stop.

Yeah, like I totally should put that. on the Primo Lib Guide I'm making that's like, oh, hey, by the way. Yeah, I think that would be like a front and center, like, first day opt-in. Like, I think the only way that they could ever be okay would be an entirely opt-in system. Because I see my search terms all the time because I test things with the same search term every single life. So ghosts is one of the top. Yeah. Yeah.

And then, like, all the instruction library terms, like web of science. I think this also speaks to like so many of our linked data systems. Like how do we design our systems of linked data to still be convenient to a user, but not take advantage of them? Like provide context without being invasive and stuff. Yeah. And none of this like you can you can agree to our cookies or you can leave the website stuff. Right. That's not consent.

No, yeah, exactly. And EBSCO is very bad about that. There was a whole lot. There was a great article that was like how the web got ruined or something and like someone showed it on their phone and it had the little accept, like how the web became unreadable, I think is what it was. And someone like took a screenshot of it on their phone because it had a little like, you have to accept all these cookies banners. They like took up half the screen or something. Here's a thing.

That's like a bad adaptation to like the GDPR stuff, right? Yeah. That's just like the web's bad adaptation. And I think that's really unfortunate. And I think if we had like a push for better adaptations to laws that are designed to protect our data, you know, things could be a little better. Like if you had a browser that could set up like a cookies profile for you that would, you know, that a website could interpret somehow or other like, please accept these types of cookies for me or whatever. Like, you know, something like that would be a better adaptation than every. website being set up like that.

Okay, Firefox, go work on that. Because the whole point of privacy, because the whole point of privacy is not anonymity, unless that's what you want. Privacy is just controlling what people know about you and how and when. Because like, if I want to share that information, that's still me controlling my privacy. But it's like, no, I'm actually fine with this website or the service taking this information from me.

I can opt into this. Often I don't. But it's like if I were okay with that, it's like, yeah, that's still. me having privacy is having that option. Like if you're an exhibitionist stripper, you don't do it under your own name. Mm-hmm. And that's still you can't deal with your privacy.

Yeah, exactly. Yeah, privacy is not anonymity. That's a thing people often confuse. The other thing I think we need not to let go of is just remembering the power differentials, particularly with students. Right? So, first day opt in, that would be better than where we are now.

But what brand new fresh person, right, on their very first day on campus? Yeah, yeah, yeah, yeah. It's going to say no to an authority who wants them to basically give up on their data. Yeah, they're going to think their greatest based on it. And I also think this, like, you know, I always think about this a lot in terms of, like, student loans. Yeah. And how you come to understand your student loan. Like, like, I didn't understand that until I had to pay it.

Like, you know, who enters that kind of scenario? And I think, you know, maybe. and I would always like I like the idea of opt-in but like ideally none since you know it doesn't fucking do anything as we're finding out there's only so much you can mitigate that power yeah right yeah so what we need and what is very much not in evidence in learning analytic circles or library learning analytic circles right now it's a little bit of self-restraint maybe we don't collect every piece of data we could. Maybe we don't keep it indefinitely.

Maybe we don't, you know, slice and dice it down to, you know, the tiniest little demographics because, hey, re-identification, it's a thing, right? Maybe we restrain ourselves. And I think part of that, too, it's like it needs to be audited, too. Like, you can't forget it and forget it. You got to come back every couple. of years or so and re-review that and make sure that it's still actually

you're actually keeping the information still meets your purpose. Otherwise you might as well just keep it indefinitely. Yeah. It's like I see is it Becky Yuse? Is that how you say her? It's like one of those things I better yeah. She posts all the time it's like data and info can only leak if you have it.

Yeah, exactly. I love her. I love her too. Shut out. Also, I just want to say for a second. I'm a little mad at Justin for misleading us into like entertaining the thought that like interlearning analytics could be good for a second. Like what kind of cop bullshit is that? I didn't. I just wanted to be a cop.

Yeah, that was some cop bullshit. I just want to know it as an ideal form of something that can work. Well, because like I'm someone, it's like I think, like I'm a metadata discovery person, but most of my like all my class. were about it, but my actual graduate assistantships in college, it's like I worked in the reference department. One of my bosses was Lisa Hinchliff. I did impolet stuff with her. My first job at Utah, and even now, it's like I'm faculty, so I have to therefore be a subject liaison and do instruction when it's requested of me. And because I'm also like a metadata person, like it gives me this perspective of it's like, oh, I know what the back end looks like, and I know what the front end looks

like and like how to like, oh, if I do this on this end, this is how it affects it on this end. And oh, because I know the back end, I can teach people how to search whatever. And so my whole shtick is just like, okay, how can I use, like, how can I do research, like even human subjects research, like interviews, user experience testing, whatever, to improve, like how we do metadata and back in stuff. because I'm like, that's actually pretty useful information because so often it's just librarians, we just make it, like, especially in tech services, we're just doing it, you know, because like, oh, well, this is the way I like to do it and this is the way it's easy for me. And sometimes we don't necessarily think about, well, is this making it better for patrons? And how do you learn that, right? It's just also, because it's like at a certain point, I'm like, actually, I would like to know all of that stuff, actually. It would make me make it, help me make it better for people. But it's like, If we can't get it in any way that's sort of ethical or good or actually helpful, then it's like, well, I guess I'll just sort of try to guess because I would rather it be a little subpar than to be invasive. And thank you for that.

Yeah. Yeah. Because I think I know what's good for people because I do have that sort of both side perspective. Because so often metadata people don't. And most of the reference people I know, I say metadata around them. And they're like, ah. Like, I attended catalog class money years ago and I hated it. And they don't even like engage with it.

But because I've like actively done both at the same time, it's like, I actually understand that connection pretty well. And so I'm like, actually, me knowing a lot of this would help. But if I can't get it, you know, if it has to be a big data learning analytics way, then like, no way. I'm taking a stand. I'll make Primo the way I want it. And if it sucks, then whatever.

But yeah, especially with subject headings. Like that's a big thing. Like I would love to get like some mass information about like how people actually search for things. And that should help us shape like not just like we shouldn't just relay on literary warrant. Remember that aggregated information is for the most part vastly less dangerous. If you're just looking at the aggregation of the search terms, the aggregation of the hits, I don't care. Have a ball. Yeah. All right.

It's when you start drilling down into identities and demographics that I start to get really, really willing. User groups. Like if I were just looking at like the popular search strings that Primo collected. Sure, rock on. But then I can break that down by user group, whether they're faculty, staff, undergrad, grad, what campus their like student ID is from. Or if they were a guest, you know, like what country their IP address from. I can't be like, I can't look at this IP address searched for this, but I can learn a lot of information about the person. As long as the user group stay pretty big. Yeah.

That's fine. Yeah. It's okay. Okay, because I'm like, oh, I can see. I mean, most of the top searches are just the instruction library. To be honest, because they just, we all just use it. Yeah. And then I do ghosts when I test things.

So I'm like, oh, that's me. There are tricks you can do. Like, if your staff machines have static IP. addresses, you basically rip that net block out of the data. Yeah, so I guess it's like for librarians when we're, or like anybody who's actually trying to improve these things where we actually

might need to collect data. What are some, I guess, guidelines of collecting learning analytics TM or collecting data to improve things? Like what are some guidelines you as like an infosec person actually teaches this and knows more than all of us? Like what are recommendations that you would have?

Collect only what you need when you need it. Permanent, you know, just collecting for the sake of it, don't do that. That just always gets you in trouble. De-identify as best you can understand that de-identification is not the same thing as anonymization. Which people would understand that. Taking my name out of something does not mean you can't tell it to me. Yep. I ran into that in my thesis.

Good. I'm glad. Yep. you know I'm with Becky data minimization is the big thing and you know

be realistic about what the data can do for you and the methods I'm assuming and the methods understand that this is only ever a very partial view of

your universe and behave accordingly and never forget that your data are people your data are people. Treat them with respect. Yeah. I think that was something we talked about a little bit in the privacy episode was we collect data that is kind of useless to us all the time. And that was kind of where

I was going with the asking like what ideal learning analytics would look like is, is there a need for this constant ongoing question? Or should you just do the three people, then three people, then three people? Like some occasional U.X testing. I actually really like Asking them first. Yeah, it's like, I really actually like the, I watched a Luna Learns webinar or something about like using user experience testing to max like to make your primo instance better or whatever. And I forget which university was it, but they did UX testing and then developed like the UX personas. Where it's like, you know, of the population we study like undergrads tended to be like this. So we actually made a person.

So like to help us think about these people as people and like in general. So that one, they didn't have to do extensive UX testing as often, but it helped them to not reduce everyone down to the user TM. And even within that, to not just be like, this is an undergraduate, but to be like, this is a person. And they're, you know, a first year, like they're a first generation student. And they live on campus, but they might have a kid or something like to actually look at, like, create like a person instead of a demographic. And I was like, oh, that's great. I had never heard of that before. I definitely want to do that kind of thing if that's actually useful and good.

Or am I being misled? The persona-driven techniques have been questioned. Let me say that. So I think you do want to do your research behind that. No, I mean, you know, it's the stuff that you would expect. People developing personas that are basically a little more than stereotypes. Yeah. They're not evidence-driven.

They're based on people's internal assumptions. And that typically doesn't end well. Right. I mean, there's bias in everything. Right. Yeah. So do your reading on that. And I will just point out that the original book that introduced the persona concept, as

UX currently knows it, has this incredibly able-less title. I apologize for this title. Alan Cooper, the inmates are running the asylum. I'm sorry! Oh. Oh. Yeah, so. That's pretty sad.

Yeah, that's pretty sad. I mean, like, I know we're, you know, in, you know, I'm not trying to be a like, well, back then, you know, because not every. But when was this written? The 90s, I think. Oh, okay. Not saying that makes it better, but I'm like, oh, okay. No, yeah. Right.

I mean, no. have been 2021. Yeah. Like, okay, where the hell was the publishing

house? Like, I'm sure some people were taking offense to that too then, but...

I am sure. Yeah. But, so I'm not saying don't do it. I'm saying, do it with your eyes open.

Right, because I feel like using analytics, you know, where you're just collecting what you need and you're remembering, forcing yourself

to remember this is people. Like, like, if you have to mass collect data, to make your services for those very people better, like it's never going to be perfect.

And I think that's another thing. It's like, we're like, well, we can't make this perfect. And we've already have all, like, then we might as well just give up. Like, it's never going to be perfect. But that doesn't mean like we shouldn't try, I guess, even though we know, you know, at the end of the day, there's probably something we're going to fuck up. But, yeah.

Yeah. Well, I think to summarize, you know, because we were just saying it would be great to do all this in a limited way. But earlier in the conversation, we said, you know, institutional analytics and learning analytics and IT tracking and wireless tracking, it all can be used to de-identify people. And it can all be used. And we don't know who the malicious actor is because if we say,

we're not going to send this to the, like Jay, you can't see the IP address, but I bet someone somewhere can't. Totally. Like the internet service provider or something. And they can be subpoenaed, which makes them the threat. Yeah, my dad used to work at an internet company and all the time, because I'm like, oh, yeah, like, I'm really interested in, like, you know, digital privacy and whatnot. And he's like, oh, yeah, no.

I mean, it's been a few decades since he worked in that realm. But he's like, oh, yeah, I could see literally every single webpage, every single customer of this internet company was going to. Because he, like, managed the servers and stuff. It's just like you can see everything. Well, but we and libraries are accustomed to that. Yeah. Because when the information universe

of people was basically the physical book, that's when we developed the Code of Ethics and we said, hey, confidentiality, we can know, but we can also keep our mouths shut. The FBI has not been here, baby. Yeah. It was a pretty low bar. Yeah.

And yet. Yeah. Is it a disage bar? Yeah. And we go back to episode two. I'm going to pretend I know what you're talking about. Yeah. I'm going to pretend I'm a horse girl.

Yeah. The decalotage bar. I've been told I have horse girl energy. Cool. It feels like a robot horse. Like the one horse lady who runs on all fours. Is that what people say? that? Like you gallop places?

I don't know. I'm putting this in the group chat. You should definitely I know what you're talking about, but... I'm sending you the video. You should compose a drone piece about this, Carrie. I actually am right now. It's like h-or.S-e. Yes. Like horse whisper

Twitter account. And oh, by the way, for we'll listen to the podcast. You don't know. like Carrie made our theme song. Oh, yeah. Yeah. Yeah. You should go, you know, listen to her music. It's great.

Oh, thanks. If you like cool synth drone stuff, which I do. Thanks. It's a niche. Yeah. Yeah, it's like, oh, hell yeah. I'm going to go listen to Elian Radegua. Do dishes.

Basically, yeah. Exactly. Yeah. I listen to podcasts when I do dishes, like library punk. Uh-ho. Uh-huh. plug in Arthur staring at my screen and he's ashamed

Absolutely without shame Oh you be getting a breadloaf Arthur Huh Well I think we can close out I did want to point out the Library salaries list I just want to plug that for people to go in Put in your salary

I made a bit leave for it So it's bit.ly slash L-I-B-S-A-L-S-A-L-L-SAL, all lowercase. Is this a big data operation? Yeah, I mean, use... It's temporarily contributed data. Yeah, I know. Yeah. I know it's totally ethical.

I'm actually on board. It's for the workers, so... Yeah, I'm totally for salary sharing. If I'm not on here, I should. And you can use Incognito browser to make sure it's not picking up who you are. If you're, like, logged in and... Chrome all the time like I am. I think incognito only stops so much because of certain things that PHP can do. And your IP still knows what you're doing. Exactly. I learned this when I went to

night school for web development. Oh, nice. Yeah, I went to night school. Yeah, I was really proud of that one. Good for you. Yeah, community college. It was good time. Community college was awesome. I have to earn a whole ass second master's degree to get tenure. So, you know, in hell. I went to our local community college to pick up more than I knew at the time of the technology pieces of introsect. Yeah, it's great. I got my IT degree from a community college. I was like a part-time employed librarian and went and did some web dev stuff and it's really, really valuable and great. Also, they had a culinary school at the community community.

college where I did my night school and it was they had the best food. I think one of the entries on the salary list is me before I transitioned. I think. Jay is just destroying the gender was that the woman voice filter that you

just put on? No, so so to get a little and luckily my gender like my speech therapist she was like just because it says it's like we don't have to do everything. But people who are socialized, raised as females, T.M. T.M. T.M. T.M. Tend to talk more in the front of their face and let the voice resonate more in their nose and their mouth and in the front of their face and talk up like this. Whereas people socialize as men tend to bring it more back

in the center of the mouth or the throat. It's not quite down here. But it's like, even if the pitch doesn't change, it's more towards the center of the mouth instead of the front of the mouth. Huh. Yeah. So I can just like, because I used to, if like, if I, if I, I don't think my voice has changed that much. But if I listen to old recordings and myself, it's totally more in the front here and the pitch is higher. But, yeah, I trained for like six months to get a, the res. Because it's like, I didn't have to train pitch because testosterone does that. Right, right.

I just had to change residents and gestures and affect and stuff. And my teacher was an opera singer, so that was fun. Oh, yeah. It sounds an awful lot like singing voice training. I've been singing in choir since I was like, well, amazing. Yeah, and like because I knew music, she's like, oh, you understand what I mean when I say this. Yeah, yeah. Because there's like voice terminology for that and I don't know what you. But I also work with, um, part of my subject liaison is with a speech and, uh, speech pathology and communication disorders. And so like some of them will do like research on, you know, like, uh, like, uh, uh, gender. affirming voice therapy and things like that. So like, you know, we've done some fun exploration

of articles and stuff. So, you know, I know some of the literature. I know how to find those articles. You got the, the deeds. I don't know the actual science behind it. So it's fun to, know what the actual science is, though, because I didn't know that. So that'll help me next time I get a research question on that. Oh, ho. The more you know. Yeah, indeed. Just teaching yourself everything as you need to learn it. That's right. Absolutely.

Like with learning analytics, you only take what you need. Oh, hey. Oh, ho. Full picture. Oh, ho. Full circle. So, Dorothea, do you want to plug your Twitter or anything or your research with data data doubles?

Talked enough about data doubles. You can find out more at datadoubles.org. I will plug an article I have coming out fairly clean. She's got an article about to drop. Oh, next month, I hope. You article just dropped. Nice. Physical equivalent privacy.

The thesis of the article is that it shouldn't matter to your privacy, whether you access a piece of information in physical form or online. Yet it does. And that's kind of a problem. Yeah. It totally does. Awesome. I'm totally entering my salary information right now.

Good for you. Nice. Yeah. Because I thought I'd done this before. Well, then I'll close it out. So, Dorothy, thanks so much for coming here and explaining all of your research so that I didn't have to write as many notes this week. Oh, this was great. This was my first podcast.

Super fun. Thanks, y'all. Oh, really? Oh, wow. You are great. Thanks for putting up with us. Oh, you are wonderful. Good night, everybody.

All right. Good night.