key: cord-0980089-80rvi86v
authors: Donham, Cristine; Barron, Hillary A.; Alkhouri, Jourjina; Kumarath, Maya Changaran; Alejandro, Wesley; Menke, Erik; Kranzfelder, Petra
title: I will teach you here or there, I will try to teach you anywhere: perceived supports and barriers for emergency remote teaching during COVID-19 pandemic
date: 2021-06-22
journal: bioRxiv
DOI: 10.1101/2021.06.21.449058
sha: 4463e1aa1ad9433d76e91619f71e16106e9500f0
doc_id: 980089
cord_uid: 80rvi86v

Background Due to the COVID-19 pandemic, many universities moved to emergency remote teaching (ERT). This allowed institutions to continue their instruction despite not being in person. However, ERT is not without consequences. For example, students may have inadequate technological supports, such as reliable internet and computers. Students may also have poor learning environments at home and may need to find added employment to support their families. Additionally, there were consequences to faculty. It has been shown that female instructors are more disproportionately impacted in terms of mental health issues and increased domestic labor. This research aims to investigate instructors’ and students’ perceptions of their transition to ERT. Specifically, we wanted to: Identify supports and barriers during the transition to ERT Compare instructors experiences with the student experiences Explore these supports and barriers within the context of social presence, teaching presence, and/or cognitive presence during ERT as well as how these supports and barriers relate to scaffolding in emergency remote courses Design Using grounded theory techniques, we applied two-cycle, qualitative analyses to assess the instructor transcripts. In first-cycle analysis, we used open coding to develop initial ideas from the data. We then used second cycle coding to generate categories with definitions and criteria agreed upon during discussion-based consensus building. Finally, these categories and descriptions were used to code student survey data. Analyses/Interpretations Instructors identified twice as many barriers as supports in their teaching during the transition to ERT and identified casual and formal conversations with colleagues as valuable supports. Emerging categories for barriers consisted of academic integrity concerns as well as technological difficulties. Similarly, students identified more barriers than supports in their learning during the transition to ERT. More specifically, students described pre-existing course structure, classroom technology, and community as best supporting their learning. Barriers that challenged student learning included classroom environment, student availability, and student emotion and comfort. Contribution Together, this research will help us understand supports and barriers to teaching and learning during the transition to ERT. This understanding can help us better plan and prepare for future emergencies, particularly at MSIs, where improved communication and increased access to resources for both students and instructors are key.

In the middle of the Spring 2020 academic term, many institutions of higher education were forced to 66 move all instruction online. The term "pandemic pedagogy" was quickly coined as educators, many of whom had 67 never taught online or remotely, scrambled to come up with effective ways to teach their courses (Schwartzman, 68 2020) . Although moving to remote instruction can enable flexibility of teaching and learning (Daymont et al., 69 2011) , the speed at which instructors and students were expected to move to remote instruction was unprecedented. 70 Therefore, it is important to distinguish this quick transition to remote teaching, or emergency remote teaching 71 (ERT), from the traditional online teaching. Here, we use the term online teaching to refer to traditional online 72 teaching (i.e., teaching online during non-pandemic times), and have adopted Charles Hodges's definition of ERT 73 as a temporary shift of instructional delivery to an alternate delivery mode due to crisis circumstances (Hodges et 74 al., 2020) . ERT is characterized by improvised, quick solutions in less-than-ideal circumstances, and it was the best 75 solution most universities had to academic learning. This is different from traditional online teaching where 76 instructors are intentionally designing a course to be implemented and delivered online, a delivery mode that has 77 been studied for decades (Bender, 2012; Lewis & Abdul-Hamid, 2006; Oliver, 1999; Young, 2006) . There are 78 numerous research studies, theories, models, and evaluation criteria created for traditional online teaching (Oliver, 79 2000; Ouyang & Scharber, 2018; Shelton & Hayne, 2017) . Studies have shown that effective online learning stems 80 from careful instructional design, planning, and using a systematic model for the development (Branch & Kopcha, 81 2014 ). This careful design process was likely to be absent in most ERT shifts due to lack of time and experience 82 necessary for instructors to carefully design their course for online purposes. 83

In addition to a shortage of time and experience, the move to ERT introduced a variety of issues that 84 instructors and students had not faced during in-person teaching, such as lack of communication (Gelles et al., 85 2020) . Previous research on emergency teaching during Hurricane Katrina in 2005 showed that virtual student-to-86 student interactions and remote class dialogues created opportunities for students to provide mental and emotional 87 support for each other (Lorenzo, 2008) . Other issues with ERT include navigating the course in a new manner, 88

finding new ways to implement formative assessment, communicating with students in a fair and equitable 89 manner, monitoring academic integrity, and managing everything through a remote platform (Brooks & Grajek, 90 2020; Johnson et al., 2020) . Additionally, many students moved home and were expected to attend college from 91 is necessary for students to feel as they are dealing with real people, that they belong in some way to a group of 169 learners, and that they are involved in sharing, negotiating, arguing, and discussing (Wang, 2008) . In general, 170 online learning environments should be active, allow student to construct their own knowledge, make effective use 171 of collaborative and cooperative learning methods, and be meaningful to students while promoting social presence 172 and community (Ally, 2004) . 173

Finally, the third element, cognitive presence, is defined as the extent to which learners can construct and 174 confirm meaning through sustained reflection and discourse. It is rooted in social constructivism, which is a robust 175 theory proposing that people's learning is shaped by cultural context, conversation, and collaboration (Dewey, 176 1986; Vygotsky, 1978) . Cognitive presence has four components: 1) triggering event (e.g., students felt motivated 177 to explore content related questions); 2) exploration (e.g., students used a variety of information sources to explore 178 problems posed in the course); 3) integration (e.g., learning activities helped students construction 179 explanations/solutions); and 4) resolution (e.g., students applied the knowledge created in the course to their work 180 and non-class related activities). Of the three elements in COI, cognitive presence has been identified as the most 181 difficult to study as well as the most challenging to develop and sustain in online courses (Garrison & Cleveland-182 Innes, 2005) . This difficulty arises from the fact that cognitive presence contains inputs (the triggering event), 183 processes (exploration and integration), and outputs (resolution) that can be hard to measure or observe, whereas 184 the other two elements consist of processes that can be more easily observed (Garrison & Arbaugh, 2007) . 185

While the COI framework has been tested for online learning and K-12 teachers during ERT (Whittle et 186 al., 2020) , it has only recently been used for ERT in undergraduate STEM classrooms at a Minority-Serving 187 Institute (MSI). For example, Reinholz et al. used the COI framework to study how the nature of student 188 participation changed in moving from face-to-face to synchronous online learning environments at an Hispanic 189 serving institute (HSI) (Reinholz et al., 2020) (Wood et al., 1976 ) and many researchers 195 and educators have used the concept of scaffolding to describe instructor roles as more knowledgeable peers for 196 guiding student learning and development (Hammond, 2001; Stone, 1998; Wells, 1999) . Scaffolding has been 197 interpreted in a wide sense as "a form of support for development and learning" (Rasmussen, 2001, p570). 198 Alternatively, it can be used as an umbrella metaphor to describe the way that teachers supply students with the 199 tools necessary to learn (Jacobs, 2001) . 200

While there are different approaches in the literature on how scaffolding may or may not be intertwined 201 with Vygotsky's ZPD (Vygotsky, 1980) , we used Wells's argument that scaffolding can be a direct application and 202 operationalization of ZPD (Berk, 2003; Duchesne & McMaugh, 2018; Wells, 1999) . Wells identified three features 203 that characterize educational scaffolding: 1.) The essentially dialogic nature of the discourse in which knowledge is 204 co-constructed (dialog); 2.) the significance of the kind of activity in which knowing is embedded (activity); and 205 3.) the role of artifacts that mediate knowing (artifacts) (Wells, 1999, p.127) . Furthermore, the relationship 206 between classroom challenge and support is important in scaffolding (Hammond & Gibbons, 2005) . Hammond and 207 Gibbons (2005) found that highly supportive, but minimally challenging environments may be too easy to elicit 208 growth in knowledge, whereas experiences that are highly challenging but lack sufficient support will likely result 209 in failure. This becomes an important aspect of scaffolding to consider as we examine how instructors and students 210 experienced the switch to ERT. 211 212 Therefore, we had three main objectives: 213 Instructor interviews were carried out via Zoom between May 11 and May 27, 2020. The semi-structured 259 interviews ranged in length from 10-90 minutes, where the participants were asked six questions regarding 260 perceived supports and barriers, changes in interactions with students and instructors, pedagogical changes made or 261 planned, and potential future supports. This manuscript will focus only on perceived supports and barriers (Table  262 2). Instructors were not interviewed by anyone from their own department (i.e., biology faculty interviewed 263 chemistry faculty). Four of the authors conducted the 31 interviews. After the interviews were completed, they 264 were transcribed using a clean verbatim transcription service. 265 266 interactions with instructors and other students, and potential future supports. This manuscript will focus only on 275 perceived supports and barriers from the student surveys (Table 2) . 276 277

Instructor interview coding 279

Grounded theory techniques (Charmaz, 2006; Strauss & Corbin, 1998) Overall scheme depicting methods used. First cycle coding was completed for instructor interviews followed by second cycle coding where categories and themes were generated from initial codes. Following instructor data analysis, student survey data was coded by authors using categories generated previously. Following instructor and student analysis, an external expert panel was used to validate categories and descriptions. The color blue represents first or open cycle coding, the color red represents second cycle coding, the dark red represents deductive coding with student data, and the color yellow represents study validation. from grounded theory methods, we used two cycle qualitative analysis (Miles et al., 2018) to explore instructor 283 interview transcripts in a section-by-section fashion (Figure 3) . 284

285 First (open) cycle coding 286

First-cycle qualitative coding allows researchers to gain a comprehensive and integrated view of a dataset 287 (Miles et al., 2018) . It is intentionally cyclical, such that a code generated during the first cycle is not meant to be a 288 static assessment. Rather, fluidity is essential. As we engaged in first cycle analysis, we used open coding to look 289 holistically across all the data, and to identify repeating indicators of instructors' perceptions of teaching during the 290 COVID-19 pandemic. Strauss and Corbin (1998) describe open coding as a process of examining properties and 291 dimensions that exist within the data, allowing the researcher to identify unique and discrete aspects (Miles et al., 292 2018) . To do this, we looked across all instructor transcripts and began assigning codes that indicated how 293 instructors perceived supports and barriers while they conducted ERT. Prior to moving into the second cycle of analysis, we engaged in consensus-building with five researchers 299 (CD, EM, PK, WA, and HB). Consensus building in qualitative analysis is a critical measure of ensuring validity 300 and trustworthiness (Corbin & Strauss, 2014) . To address inter-coder consistency, we independently coded 301 approximately 10% of all transcripts, and then we used discussion-based consensus building to address 302 discrepancies in codes. Saldaña (2015) describes this process as interpretive convergence, specifically useful in 303 qualitative analysis where dynamic interpretations of data are paramount, as opposed to seeking statistical 304 significance in quantitative methodologies. In our efforts to converge toward common codes, we discussed both 305 our individual open codes as well as our analytic memos. Analytic memos in qualitative analysis serve as a 306 researcher's dialogue, both with themselves and each other, about what the codes mean to them as they are coding 307 (Charmaz, 2006; Corbin & Strauss, 2014; Miles et al., 2018; Stake, 2005) . Table 4 depicts open codes and analytic 308 memos during one of our consensus coding sessions. It shows five of the author's initial codes, generated 309 individually, and the consensus code generated as a group after discussion. 310 311 Second (axial) cycle analysis 316

The aim of second cycle analysis is to find linkages between the discrete parts that were earlier identified 317 in open coding (Miles et al., 2018) , in order to find "broader categories, themes, theories, and/or assertions" (p. 318 234). Essentially, this involves looking for similarities and differences across the previously identified properties 319 and dimensions of the dataset (Strauss & Corbin, 1998) . We engaged in this process of creating relational 320 categories through axial coding. Table 5 and Table 6 Four researchers (CD, EM, PK, and HB) worked together to make descriptions for supports and barriers 329 categories. The following categories and descriptions emerged from the data (Supplemental Table 1 ). Support 330 categories included: prior experience, timing, technology for remote teaching, community, help with technology, 331 socio-emotional factors, teacher beliefs, working from home, help with teaching, course attributes, student comfort 332 interacting online, and reducing cognitive load. Barrier categories included: communication difficulties, time 333 management, instructor teaching inexperience, instructor technology issues, teaching and learning resources, 334 student integrity, administrative issues, student presence and participation, student emotion and comfort, student 335 technical issues, assessment difficulties, instructor emotion, responsibility and workload, and instructional space. 336 337 Student deductive coding 338 A deductive coding approach was used to identify the various student support and barrier categories 339 generated from coding the instructor interviews (Table 2) . First, two researchers (PK and MCK) independently 340 coded the student responses using the 12 support categories (Table 9 ) or 14 barrier categories (Table 10) to get 341 initial codes (Table 7) . Each student response was one to three sentences in length and could contain more than one 342 category. 343 344 Table 7 . Individual coding with support example 345 ID Student response MCK initial code PK initial code Skyler Online lecture recordings allowed me to revisit past material to ensure I was understanding the taught concepts.

Preexisting course structure Classroom technology

Note: There were no analytic memos for this example 346 347

Next, PK and MCK met to discuss their categories until reaching 100% consensus (Table 8) . When coding 348 student responses, authors noticed that the working descriptions of some categories would only fit the instructor 349 perspective and needed to be redefined to be used for student coding. Authors then met to discuss which 350 descriptions should be changed and rewrote them so that the codes could be used for both students and instructors 351 alike. 352 353 PK MCK Need to modify "teaching resources" to "teaching and learning resources" as this describes not having access to learning resources Trouble with something before COVID, but can't access the resources that they used to. Maybe technical issues, lack of face-to-face support and lack of student resources 356

Validation 357

Following student coding using the categories and agreed upon descriptions, we brought the categories, 358 themes, and descriptions of categories to an expert feedback panel of five. The expert feedback panel was made-up 359 of STEM educators (both biology and chemistry), discipline-based education researchers (DBER), and learning 360 scientists at a research-intensive institution unrelated to the one in this study. This expertise allowed the panel to 361 provide valuable feedback on category descriptions. 362

The formatting for the feedback was organized in two parts. For the first part, the first author (CD) 363 presented the support instructor and student themes, categories and their descriptions along with examples 364 (Supplemental Table 1 ). Expert feedback panelists were then allowed to ask questions prior to a short content 365 validation. This content validation consisted of splitting the panelists into two groups and providing each group 366 with the same three support quotes representative of categories in each of the three themes (Supplemental Table 3  367 and 4). The panelists were given all the support categories and descriptions and were asked to match the quote with 368 a category and provide justification via notes. Additionally, authors CD, PK and JA were present in both groups to 369 take notes on feedback. After each group had finished choosing categories for the quotes, everyone met back in 370 one group and the answers were discussed. Comments and feedback are reported in Supplemental Table 3 and 4.  371 Following this discussion, the barriers, their descriptions, and themes were then presented by CD to the expert 372 feedback panel (Supplemental Table 2 ). The same two groups were then created, and panelists were again given 373 the list of barriers and their descriptions and asked to decide which category fit each quote best. After each group 374 had finished all comments and questions were recorded (Supplemental Table 5 and 6). Again, CD, PK, and JA 375 were present in both groups to take notes on feedback. 376

Following the expert feedback panel, all coders met and discussed the feedback. This discussion allowed 377 changes from version 2 to version 3 for categories and descriptions (Supplemental Table 7 The findings presented here are separated into the two primary areas of study, supports and barriers. 382

We identified 134 unique support codes which led to the development of 12 categories, and, ultimately, three 383 support themes: 1) tools and support for class content, 2) mental & emotional support, and 3) preexisting supports. 384 A total of 203 unique barrier codes were identified, which led to the development of 14 categories from which the 385 following three barrier themes emerged: 1) structures getting in the way, 2) spending more time and effort, and 3) 386 affective issues. 387 388 Instructor supports 389

Instructors described a variety of supports when discussing their switch to ERT. These supports ranged 390 from existing structures that made remote teaching easier to emotional support from colleagues. Here, we describe 391 how we went from transcript to themes ( During the open coding process CD identified supports that were both technical and moral, while EM 407 identified a support of hearing how others were handling the situation. During the consensus process, discussion 408 between the consensus coders focused on whether to differentiate between technical and moral support, how 409 specific to be with respect to technical support, and whether it was the people or the discussions that were the 410 actual support. This discussion resulted in two codes: After consensus coding all the supports, a total of 134 unique supports were identified. After reaching consensus on all the transcripts, the researchers individually collected the codes into 418 categories, then met to resolve any differences. This process led to the creation of 12 categories, shown in Table 9 . 419

These categories covered topics, such as access to technology, help with remote teaching pedagogy, access to a 420 supportive community, the comfort of working from home, and the timing of the switch to ERT. As On the surface both quotes seem similar, in that both Constance and Josephine identified people they could 435 talk to as a support. However, Constance's quote points to the support of having someone available to answer 436 specific, course-related questions about best practices, whereas Josephine's quote is more about the camaraderie of 437 people in similar situations and knowing that others are persevering under difficult circumstances. As a result of 438 these differences, Constance's quote was coded as 'help from more experienced colleagues', whereas Josephine's 439 quote was coded as 'community of colleagues'. These differences also led to the quotes being placed in different 440 categories. As Constance's quote was focused primarily on pedagogical help, it was placed in the 'help with 441 teaching' category, whereas Josephine's quote, which focused much more on the emotional support of others, was 442 placed in the 'community' category. 443

Instructor themes for supports 445

Once the codes were categorized, the researchers individually organized them into themes, and then met to 446 resolve any differences. This process led to the creation of three broad themes of: 1) tools and support for class 447 content, 2) mental/emotional support, and 3) preexisting supports, shown in Table 9 Danielle's quote was representative of the supports that fell into the 'preexisting supports' theme. These 477 were supports that were already in place that inadvertently made the transition to remote teaching easier for the 478 instructors. Other examples were instructors who had prior experience with online teaching, student familiarity 479 with streaming services, and courses previously delivered using a flipped modality. 480 481 During the open coding process CD identified a decrease in student participation as a barrier, while EM 497 identified a decrease in quality of student work as a barrier. During the consensus process, discussion between the 498 consensus coders focused on gauging participation and measuring quality, which resulted in barrier code of 499 'decrease in student assignment submissions.' After consensus coding all the barriers, a total of 203 unique barriers 500 were identified. 501 502

While there were many more barriers than supports coded, the barriers collapsed into a similar number of 504 categories (14 barrier categories, compared to 12 support categories), shown in Table 10 . These categories covered 505 topics such as issues with instructors and students using technology, increased responsibility, and workload, 506 decreased student motivation, and concerns about student integrity. As an example of how these barrier categories 507 came about, consider the two quotes, below: 508 509

With a pre-recorded video, that's impossible. Then, students won't even watch it. I was convinced 510 I shouldn't do asynchronous. Then, through the experience, I discovered that conviction was 511 reinforced, because the students told me. I asked them multiple times, and they told me, "This is 512 much better precisely because we can stop you and just ask you to explain the thing, again." ---513

Arturo 514 515

It's up to them to use their time the way they want to do it, but what we're finding is it's hard to 516

project with the statistics whether they're actually watching the videos or skimming through the 517 videos or how many times they're watching the videos. That's hard data to obtain. The only way to 518 get at it is to maybe get an assignment back from them-a lab or a quiz or whatever. You covered 519 some of the stuff on the videos and kind of points out, "These are some of your flaws that you're 520

not really looking at the videos." ---Roberto 521 522

Similar to the support quotes discussed above, both Arturo and Roberto identified barriers that, on the 523 surface, seem similar in that they both present difficulties with video lectures. However, Arturo's quote really 524 pointed to the difficulty of getting immediate feedback from students about specific topics, whereas Roberto's 525 quote was more about collecting data about whether students are really watching the videos. As a result of these 526 differences, Arturo's quote was coded as 'asynchronous teaching was not possible due to students not being able to 527 provide timely feedback about content understanding', while Roberto's quote was coded as 'assessing if students 528 are watching videos.' In addition, these differences led to the codes being placed in different categories. Because 529 Arturo's quote was about the potential difficulty of engaging and interacting with students via an asynchronous 530 modality, his code ended up in the 'instructional space' category and Roberto's quote, which focused much more 531 on difficulties measuring student interaction, ended up in the 'assessment difficulties' category. 532 533

Once the codes were categorized, we individually organized them into themes, and then met to resolve any 535 differences. This process led to the creation of three broad themes of: 1) structures getting in the way, 2) spending 536 more time and effort, and 3) affective issues, shown in ERT, ranging from course attributes to prior experiences that students had before the start of ERT (Table 9) . 578 However, 'course attribute's (38.6%), 'technology for remote teaching' (22.9%), and 'community' (11.4%) were 579 the three most frequent categories described by students as supporting their learning during ERT (Figure 4) . Here, 580

we describe how we went from student survey responses to categories for Claire's class. For example, one student 581 wrote that: 582 583

Homework assignments and discussion sections with my TA were extremely helpful, and the lab 584 sections also helped. ---Sammi 585 586

Homework assignments, discussions sections, and lab sections were all attributes of the course that had 587 existed prior to the transition to ERT. Therefore, Sammi's quote was placed into the category 'course attributes' 588 since the student discussed components of the course (e.g., homework assignments) that help them ease into the 589 transition to ERT. This was the most frequent support category described by students. Two other students, Ash and 590 Jax, wrote that: 591 592 I think the instructional videos provided helped us greatly in learning the content of the course. 593

Zoom calls helped us catch up and know his outlook on the plan for the coming week of the course.

---Ash 595 596

Zoom discussions, zoom lectures, recorded lectures, online tutoring services outside of school ---597

Jax 598 599

Ash stated that features of the Zoom platform aided their learning, which lead to their response being put 600 in the category' classroom technology.' In addition, Jax listed Zoom lectures, discussions, and other online sources 601 as resources that helped their learning, which lead to this response being placed in the category' classroom 602 technology.' Finally, another student wrote that: 603 604

Having online lectures actually felt like I was more connected to the class rather than in person, 605 so that really helped me out. ---Ekene 606 607

This quote was placed in the category 'community' because of the connection to the class that aided the 608 student and helped them sustain their learning. 609 610 611

Student barrier categories 612

Students described ten barriers when discussing what challenged their learning during ERT. These barriers 613 ranged from concepts such as instructor technology issues to instructional spaces (Table 10) . However, 614 'instructional space' (23.7%), 'student emotion and comfort' (21.1%), and 'student presence and participation' 615 (21.1%) were the most frequent categories students found challenging their learning during ERT ( Figure 5) . Here, 616

we described how we went from student survey responses to categories for students from Constance's class. For 617 example, two students wrote that: 618 619

That I felt there wasn't much learned as if I were in person. ---Nasim 620 621

Not being able to study in groups or ask my friend to help explain my questions. Also, there are 622 fewer physical demonstrations in classes and I tend to get more distracted at home ---Ollie 623 624

Nasim described feeling as if they were not learning as much remotely as in-person. Additionally, Ollie 625 described how the remote learning environment was difficult and distracting compared to the in-person learning 626 environment. Therefore, both quotes were put in the category' instructional space.' Additionally, two students 627 wrote that: 628 629

It was hard to focus at home because there were a couple of distractions. Also I kind of lost some 630 motivation through online learning. ---Xia 631 632

Productivity was a challenge of mine. I live in a household with younger siblings who are not only 633 rowdy but also need help with their online classes as well. ---Rio 634 635

There are many different reasons that might prevent students from attending or participating in course 636 activities or office hours. These factors can range from different emotional states to familial responsibilities and 637

can have a negative impact on students' learning. Because Xia and Rio described facing some of these difficulties 638 causing them to be unproductive, they were placed in the category' student presence and participation.' Lastly, one 639 student wrote that: 640 641

When we transitioned to [video only] lectures. I felt cut off and sort of isolated from the class. It 642 was starting to feel pretty lonely and discouraging. ---Mitra 643 644

Because Mitra wrote that this isolation was a barrier to their learning, and it resulted in them feeling lonely 645 and discouraged, it was listed as 'student emotion and comfort.' We placed it in that category because we felt it 646 could be included as a concern for student emotional well-being. 647

Instructor and student comparisons 650

To further investigate the student experience, we decided to compare the student survey data to their 651 instructors. We wanted to know if the students that we surveyed had a similar experience in their transition to ERT 652 as their instructors. It was fortunate that we had all four instructors of the students surveyed as part of our 653 instructor data. All the student data came from students with at least one course taught by an instructor that also 654 participated in our study, and we wanted to compare the supports and barriers identified by the students with these 655 four instructors. When assessing the instructor and student category data we found that there were more barriers 656 described by students than supports, like the instructors' results. The most frequent support categories applicable to 657 students were 'course attributes', 'technology for remote teaching', and 'community'. In comparison, the 658 instructors of those students' most frequent support categories were 'community', 'help with teaching,' and 'help 659 with technology.' We found that both instructors and students overlapped with mentioning the 'community' 660

The percentage of barriers from students. Each value is the total number of each category divided by the total number of all categories, 76. For example, 'instructional space' was mentioned 18 times and 18/76 = 23.7%. Note: no student used a single code more than once. If we look at Haik, the instructor, and one of their students, Mitra, we can see how they both described 672 their perspective colleagues as part of their supportive community. Instructor Haik discussed how his community 673 of support is other specialists in his area of research and Mitra discussed how their peers during synchronous 674 teaching provide support. 675

The categories' instructor teaching inexperience' and 'time management' were some of the most common 676 categories from instructors as barriers to their transition to ERT. However, neither of these categories were 677 mentioned by students, showing that even though there may be overlapping supports between students and 678 instructors, the barriers seem to differ more. One common barrier cited by instructors and students was the 679 category' instructional space', which we have described as "difficulty implementing or participating in teaching 680 and learning activities." 681 682

With breakout rooms, students weren't really-they needed to be introduced and encouraged, The instructor Claire and one of her students Sammi both described how 'instructional space' inhibited 692 their teaching or learning. Claire described how she found it difficult to get students to work in groups and find the 693 students and groups that needed her help. Sammi felt it was difficult to focus with the format that Claire was using, 694 as well as they felt it was impersonal and did not connected to their pace of learning. It seems that Claire and 695

Sammi both struggled with Zoom, especially the breakout room feature, as the ERT teaching platform. 696 1

We contextualized support and barrier categories within the COI framework by illustrating the positive and 699 negative influences on teaching presence, social presence, and cognitive presence (Figure 2 ). Our data shows the 700 importance of providing supports to enable instructors and students to maintain strong teaching, social, and 701 cognitive presence. Without these supports, there are potential impacts on work-life balance and mental health of 702 faculty and students. 703

Teaching presence 705

Teaching presence is defined as "the design, facilitation, and direction of cognitive and social processes 706 for the purpose of realizing personally meaningful and educational worthwhile learning outcomes" (Anderson et 707 al., 2001a) . One of the themes we considered a support for teaching presence was 'tools and support for class 708 content.' Having communication tools and the right technology was critical for both students' and instructor's 709 success during the transition to ERT. For example, several instructors mentioned how their early access to 710 technology allowed them to quickly transition from in-person to remote. This quick transition helped them 711 maintain their teaching presence. In contrast, the instructors that had difficulties with technology, such a delays 712 getting Zoom licenses or poor internet connection fell under the theme 'structures getting in the way.' This theme 713 of 'structures getting in the way' directly hindered instructors' ability to provide and maintain a consistent teaching 714 presence, thereby hindering student's ability to experience a strong teaching presence. 715

An example of a support category we considered an influence on teaching presence was the category 716 'prior experience.' We defined prior experience as "knowledge, skills, and experiences before the start of ERT." 717 For example, instructor Diane said: "We did an interview with a Skype scientist, and Zoom worked really well for 718 that, because I was already using Skype, so I just transitioned to that platform," which indicates her prior 719 knowledge with Skype was a contributing factor to her ability to quickly transition to ERT, and it allowed her to 720 maintain her teaching presence, largely unchanged, for that aspect of the course. Additionally, Diane's prior 721 knowledge was a support in the switch to ERT because it allowed her to maintain meaningful learning outcomes 722 for her students. 723 Figure 5 . Scaffolding in learning contexts for students. Predicted student outcomes. The top left quadrant, 'demands too high; failure likely' is when barriers or challenges are not met with equal support, leading to a likely failure for the student. The lower left quadrant 'low motivation; boredom and behavior/problems likely' is generally seen when challenges and supports are both too little. The lower right quadrant 'comfortable/easy; little learning likely' is when challenges are too low but supports are high. The last quadrant, the upper right is when challenges or barriers and supports are equal leading to an 'extension of learning and capability' Adapted from Hammond & Gibbons (2005 

Social presence is "the ability of learners to project their personal characteristics into the community of 758 inquiry, thereby presenting themselves as 'real people'". The 'community' category within the theme of 'mental 759 and emotional support' was something both students and instructors mentioned as important in their transition to 760 ERT. Instructors discussed things like having a support group to allow them to learn from others and having 761 friends to vent to about their frustrations. Having a sense of community provided a basis for social presence in that 762 in our study we considered instructors as learners as well as students. Instructors in ERT were learning how to 763 teach remotely in a pandemic; therefore, instructors' social presence was also something important to consider in 764 our analysis. 765

Another category that aligned with social presence was 'student comfort interacting online', which was 766 defined as "using communication tools and modalities familiar to students. This barrier to ERT was categorized as 'student presence and participation,' which was defined as the "inability of 783 students to attend or participate in class, office hours, or course activities." We argued that lack of student presence 784 and participation decreases student's social presence. 785

Cognitive presence 787 Similarly, instructors described 203 unique barriers that we collected into 14 categories, best represented by the 814 themes 1) Tools and support for class; 2) Spending more time and effort; and 3) Affective issues. 815

At the same time, 69 undergraduate students in STEM classes were surveyed about supports and barriers 816 that they had experienced during the ERT transition. The student survey responses were analyzed by inductive 817 coding, using the codes identified during the instructor analysis. The most important support categories identified 818 by students were 'course attributes', followed by 'technology for remote teaching' and 'community'. The biggest 819 barriers identified by students fell into the 'instructional space' category, followed by the 'student emotion and 820 comfort' category and the 'student presence and participation' category. 821

Considering these supports and barriers in the COI framework, we found that some faculty found ways to 822 maintain a strong teaching presence and a healthy work-life balance, but most of these supports were local and 823 existed prior to the transition. Despite institutional efforts most of the broad, large-scale supports that faculty 824 identified were logistical in nature and primarily helped the faculty to maintain a strong teaching presence. While 825 faculty may not be familiar with the COI framework, they all recognized the importance of maintaining a strong 826 teaching presence and facilitating a strong social presence for students. However, most of the faculty identified 827 several barriers that prevented them from maintain and facilitating teaching and social presences. In addition, most 828 instructors had trouble maintaining a healthy work-life balance and found the transition to be extremely stressful. 829

All of this suggests that either 1) the supports provided by the institution were not meeting the instructor and 830 student needs or 2) the instructors and students were unaware of institutional supports. None of the instructors or 831 students identified institutional supports that helped them maintain a healthy work-life balance nor supported them 832 in maintaining a strong social presence during ERT. Following this, recommendations for future emergencies for 833 universities could include either more supports for instructors and students, such as greater access to free 834 technology, both software and hardware as well as increased access to reliable internet bandwidth, or, finding 835 better ways to disseminate where and how to get these supports. Additionally, maintaining a healthy work-life 836 balance is vital for everyone. Increased workload was cited by many students and instructors as a significant 837 struggle they faced. It is crucial that universities find ways to help alleviate this added strain, whether it be through 838 hiring additional staff to create things like laboratory instructional videos or simply lowering expectations so that 839 instructors and students that have family obligations directly stemming from the pandemic can attend to them. 840

Additional time for transitions for preparation of new online course material or training for students to become 841 more familiar with remote platforms would also be helpful. 842

Limitations and future directions 844

We acknowledge that there are several factors that limit our study; however, these limitations provide 845 opportunities for future studies. First, we conducted a convenience sample at only one MSI, UC Merced, so there is 846 limited generalizability. Also, we did not employ a systemic approach to ensure even distribution of faculty and 847 students across STEM disciplines at our institutions, so there were more chemistry and biology participants based 848 on the departments of the co-authors. In particular, the students surveyed were from only four of the 31 instructors 849 interviewed. As the transition to ERT during the COVID-19 pandemic was a one-time event, we will not be able to 850 collect more data from other MSIs or within UC Merced to see if our patterns persist in different study contexts or 851 with more participants. It might be possible to work with other MSI's that conducted similar work and compare 852 our survey data for patterns. Second, this study focuses on self-report survey and interview data that is often 853 perceived as less objective compared to well-developed and validated classroom observation protocols (AAAS 854 2013). A risk of surveys and interviews is that the participants inaccurately self-report their teaching and learning 855 experiences, limiting the conclusions that can be made from these data. Our future work aims to triangulate the 856 self-report survey and interview data with classroom observations and student learning gains data (i.e., concept 857 inventories) to better understand classrooms interactions and student learning. Third, this study only described the 858 perceived supports and barriers through the transition to ERT, not the continuation of it. Therefore, we collected 859 instructor interview data at the end of the Fall 2020 semester to better understand what served as supports and 860 barriers for instructors during the continuation of ERT. Taken together, we found there are more barriers than 861 supports and that increased resources and communication between students, TAs, staff, faculty, and administration 862 could help mitigate future emergency remote hardships and experiences. 

Designing quality e-learning environments for emergency remote 889 teaching in coronavirus crisis

Foundations of educational theory for online learning. Theory and practice of online learning

Assessing teaching presence in a computer 899 conferencing context

Developing a community of inquiry instrument: Testing a measure of the community of inquiry framework 903 using a multi-institutional sample. The internet and higher education

Discussion-based online teaching to enhance student learning: Theory, practice and 906 assessment

Child development 6th ed

Emergency remote teaching in a time of global crisis due to CoronaVirus 911 pandemic

Instructional design models

Faculty readiness to begin fully remote teaching

COVID-19 reveals gaps in the Valley's public health system, but 919 improvement is possible

20 Years of the Community of Inquiry Framework

Constructing grounded theory: A practical guide through qualitative analysis. sage

The coronavirus pandemic could shut down schools for months, leaving some students hungry 927 and far behind their peers

Impact of COVID-19 Pandemic on College Student Mental Health and Wellness

Basics of qualitative research: Techniques and procedures for developing 937 grounded theory

Grounded theory research: Procedures, canons, and evaluative criteria. 940 Qualitative sociology

Discipline-based education research: Understanding and improving learning in 943 undergraduate science and engineering

Social 946 presence enhances student performance in an online geology course but depends on instructor facilitation

Deciding between traditional and online formats: Exploring the 950 role of learning advantages, flexibility, and compensatory adaptation

Experience and education. The Educational Forum

How women are getting squeezed by the pandemic. The New York Times

Educational psychology for learning and teaching

Instructor and student responses to COVID 19 emergency remote 960 learning: A preliminary investigation of ten undergraduate animal sciences courses

E-learning in the 21st century: A community of inquiry framework for research and 964 practice

Critical inquiry in a text-based environment: Computer 967 conferencing in higher education. The internet and higher education

Critical inquiry in a text-based environment: Computer 970 conferencing in higher education. The internet and higher education

The first decade of the community of inquiry framework: A 973 retrospective. The internet and higher education

Researching the community of inquiry framework: Review, issues, and 976 future directions. The internet and higher education

Facilitating Cognitive Presence in Online Learning: 979 Interaction Is Not Enough

Exploring causal relationships among teaching, 983 cognitive and social presence: Student perceptions of the community of inquiry framework. The internet 984 and higher education

Compassionate flexibility and self-987 discipline: Student adaptation to emergency remote teaching in an integrated engineering energy course 988 during COVID-19

Scaffolding: Teaching and learning in language and literacy education

What is scaffolding. Teachers' voices

The difference between emergency remote 995 teaching and online learning

Virtual and Traditional Feedback-Seeking Behaviors: Underlying 998

Competitive Attitudes and Consequent Grade Performance

Emergency remote teaching 1002 and students' academic performance in higher education during the COVID-19 pandemic: A case study

Providing the scaffold: A model for early childhood/primary teacher preparation

US Faculty and Administrators' Experiences and Approaches in 1009 the Early Weeks of the COVID-19 Pandemic

Demographics of physics education research

Design considerations in emergency remote teaching during the COVID-19 pandemic: A 1015 human-centered approach

The impact of the COVID-19 epidemic on mental 1018 health of undergraduate students in New Jersey, cross-sectional study

A Structural Equation Model of Predictors of Online Learners' Engagement 1023 and Satisfaction

Implementing effective online teaching practices: Voices of exemplary 1026 faculty

Using reflections and questioning to engage and challenge online graduate learners in 1029 education

The Sloan semester

Investigating 1035 student engagement in blended learning settings using experience sampling and structural equation 1036 modeling. The Internet and Higher Education

Qualitative data analysis: A methods sourcebook

Evaluating online teaching and learning

Exploring strategies for online teaching and learning

Adapting the TPACK framework for online teaching within higher education

To Understand is to Invent: The Future of Education. Penguin Books

A Pandemic Crash Course: 1053 Learning to Teach Equitably in Synchronous Online Classes

/06/01/). Social presence in relation to students' 1056 satisfaction and learning in the online environment: A meta-analysis

Supporting the mental health of preservice teachers in COVID-19 through trauma-informed 1060 educational practices and adaptive formative assessment tools

Pandemic Pedagogy Facebook Group 1064

Resilient Instructional Strategies: Helping Students Cope and Thrive in Crisis

Measures of quality in online education: An investigation of the community of 1069 inquiry model and the net generation

Developing an instrument for evidence-based peer review of faculty online 1072 teaching

Exploring the Impact of the Students Assessing Teaching and Learning 1075 Program

Students' engagement characteristics predict success and completion of online 1078 courses

The metaphor of scaffolding: Its utility for the field of learning disabilities

Basics of qualitative research techniques. Citeseer. 1086 1087

Fall 2020 workforce diversity

Mind in society: the development of higher psychological processes

Mind in society: The development of higher psychological processes

A generic model for guiding the integration of ICT into teaching and learning. Innovations in 1100 education and teaching international

Scaffolding critical thinking in the zone of proximal development. 1103 Higher Education Research & Development

Dialogic inquiry: Towards a socio-cultural practice and theory of education

Should teachers be trained in emergency remote teaching? Lessons learned from the COVID-19 1109 pandemic

Emergency remote teaching environment: a conceptual 1112 framework for responsive online teaching in crises

The role of tutoring in problem solving

Student views of effective online teaching in higher education. The American Journal of 1118 Distance Education

Cognitive presence is defined as "the extent to which learners can construct and confirm meaning through 788 sustained reflection and discourse." The support theme 'mental and emotional support' was important because 789 poor mental health and distractions impacted motivation to explore course content, the first component of cognitive 790presence. For example, the support category 'work from home' allowed for personal support from spouses and 791 pets, didn't clearly indicate that there was an impact on teaching or social presence, but it clearly had an impact on 792 instructors' ability to teach during ERT. 793For example, instructor Martin described how working from home allowed him to feel supported, "And, of 794 course, informal support in mental issues, well, my spouse and my pets were really imperative to keep me sane." 795Despite this category not pertaining to teaching or social presence, it is still particularly important in describing 796 experiences during the transition to ERT. Mental health, critical to both instructor and student success, has been 797 shown to play a significant role in the lives of many during the COVID-19 pandemic (Copeland et al., 2021; 798 Kecojevic et al., 2020; Roman, 2020). 799Many instructors found it difficult to manage a proper work-life balance as shown in the barrier category 800 'instructor emotion,' which was defined as "frustration, fatigue, and guilt due to remote delivery." Instructor Claire 801 described, "There were multiple days where, if I did shower, it was at 3 in the afternoon and I was sobbing in the 802 shower because I just couldn't see it getting any better, and still try to maintain quality." Claire's struggle to 803 maintain a separation between work and home life, coupled with her feelings of hopelessness, did not align with 804 teaching or social presence, but might have impacted how she engaged with ERT. 805 806 CONCLUSIONS 807In this article, we have examined the perceived supports and barriers that affected instructors and students 808 during the rapid transition to ERT resulting from the COVID-19 pandemic. A total of 31 STEM instructors were 809 interviewed about supports and barriers that they had experienced during the ERT transition, and interview 810 transcripts were analyzed via two-cycle quantitative analysis process drawing from grounded theory methods. This 811 process led to the identification of 134 unique supports that were collected into 12 categories, best represented by 812 the themes 1) Tools and support for class content; 2) Mental/emotional support; and 3) Preexisting support. 813