After screening using the inclusion criteria, a final 16 studies were selected for systematic analysis. These sources are listed in Table 1, which contains a concise description of the studies’ characteristics: location, learning context, type of data, sample size, research approach, and attainment measures.
The included studies were systematically analyzed by making use of two coding schemes: one on the study level and one on the results of those studies (Lipsey & Wilson, 2001). In the first coding scheme, general study information (such as publication year, study design, and methods), sample characteristics (gender and age), educational context specifics (such as the tertiary institution, its region, and the student cohort), and methodological characteristics (such as the type of comparison groups, type of data collected, analysis used, and methodological concerns) were coded for each study. In the second coding scheme capturing the results of the studies, we gathered information about the conclusions reached regarding the impact of LC on attainment, attendance, and students’ perceptions about the LC. The double-coding of the 17 articles was conducted separately by the two authors and then compared. There were no disagreements to be resolved.
In the following sections, we explore the themes reported in the identified studies to answer the research questions.
What are students’ perceptions of lecture capture and its uptake?
Students see immense value in LC because of the flexibility it provides, this being the most obvious and widely accepted perception of LC (e.g., Hall et al., 2020; Meehan & McCallig, 2019; Trenholm et al., 2019; Yoon & Sneddon, 2011; Yoon et al., 2014). Because the lecture information can be accessed at any time, LC availability is perceived to facilitate a better study/work/life balance and provide equitable access to content for students who have other commitments (Hall et al., 2020). Frequently, it is seen as a safety net (Loch et al., 2016).
From an academic perspective, students see LC as a positive influence because of its utility as a study tool to re-watch content (e.g., Hall et al., 2020; Yoon & Sneddon, 2011). Gouia-Zarrad and Gunn (2018) survey of engineering mathematics students found that 93% of students perceived LC as a necessary support. Clearly, the majority of students find use in technology. Moreover, it is argued that LC is especially useful in mathematics, as discussed by Loch et al. (2016), who reasoned that as mathematics is a hierarchical subject (knowledge building on previous knowledge), students can use LC as a tool to avoid falling behind.
Unsurprisingly, high utility value perceptions translate into high uptake. In Jones et al. (2018) study of an undergraduate statistics class, 88% of students accessed LC at least once. The proportion reported in Yoon and Sneddon (2011) is even higher, with 95.5% of students watching at least one recorded mathematics lecture. The purpose of LC is that students who would be discouraged after falling behind now have the tools to catch up. On the other hand, LC makes students more likely to fall behind because of this sense of security, which may lead to increased cramming near key assessments.
However, even after attending the lecture, some students still often find value in watching the recording (Hall et al., 2020; Yoon & Sneddon, 2011; Yoon et al., 2014). For example, in a study of two first-year mathematics courses at the University of Auckland, Yoon and Sneddon (2011) analyzed student viewing patterns through a self-report questionnaire. They found that a large proportion of students reported pausing their LC viewing “for examples and ideas, toggling back and forth within the lecture or only watching relevant parts” (p. 439). This behavior was especially prominent for non-native speakers, who are potentially more susceptible to falling behind because of language comprehension difficulties. In a recent study of first- and final-year mathematics students’ experiences at two Australian universities, participants reported that “they were rarely able to fully understand content by attending lectures in person and found that the ability to pause and rewind the recordings mitigated these issues. These capabilities were particularly valued by participants who described themselves as “slow learners” or “not that good at maths”” (Hall et al., 2020, p. TBA).
However, this positivity towards LC should not be confused with a desire to replace lectures. Wood et al. (2018) reported that physics/mathematics students still saw value in attending the lecture over watching the recording because of the social contact and ability to ask questions. Similarly, Loch et al. (2016) found that nearly all students surveyed did not want a reduction in contact hours (e.g., lectures, tutorials). Perhaps it is the safety net LC provides that gives it universal praise from students, yet the evidence asserting its value when directly compared to traditional teaching is missing.
A minority of students do see LC as a substitute for attendance. Based on an analysis of student self-reports in a large first-year mathematics course at the University of Auckland, Yoon and Sneddon (2011) revealed that 30% of students viewed LC as a replacement for live lecture attendance. Similar perceptions were reported by Khan (2013): in a large statistics course at the University of Western Australia, 23–25% of students identified LC as a primary reason for missing lectures. Of note, Docherty et al. (2019) investigation on first-year mathematics students at the University of Edinburgh found that only 9% of students who missed class went on to watch the lecture. The performance aspect of this relationship is discussed more in the attainment section below.
Does lecture capture have an effect on attendance?
A concern for staff is the perceived impact that LC has on attendance, with the fear that students will use LC as a substitute for rather than a supplement to face-to-face lectures (Loch et al., 2016; Wood et al., 2018), yet the current research on the relationship between LC and attendance is mixed.
Of the 16 relevant studies this review identified, 10 did not consider attendance, none reported a positive impact, three sources reported insignificant/no change, and four reported a negative impact on attendance (Table 2). Attendance was often not considered, perhaps because (1) it is difficult to record non-compulsory attendance, (2) there is little desire to investigate it because of the prevailing staff opinion that LC decreases attendance, and (3) attainment was the main focus of the studies.
Table 1 Studies selected for the systematic review
In all studies that explicitly reported a drop in attendance, the total attendance reductions were in the range of 23–30%. However, none of these studies were specifically designed to establish a causal relationship. All four studies used student self-report data and/or lecturer observations to reach a conclusion (from the University of Western Australian by Khan (2013), from the University of Auckland by Yoon and Sneddon (2011), from Germany by Zimmermann et al. (2013), and from a multi-site study of two mathematics for engineering courses at the University of South Australia and Loughborough University, UK, by Trenholm et al. (2019)). For example, as seen in both Zimmermann et al. (2013) and Yoon and Sneddon (2011), 30% of students stated that using LC was a substitute for attending class. Statistically analyzing group differences in students’ questionnaire responses at the start and the end of the semester, Trenholm et al. (2019) observed that individuals who watched lecture recordings attended significantly fewer lectures than their peers who rarely watched videos.
This tendency for a large group of students to see LC usage as an “either-or” instead of a “both-and” appears common, as explicated in studies like Yoon et al. (2014) and Trenholm et al. (2019). Students in a multivariate calculus course at Loughborough University exhibited this behavior, as reported in a study by Inglis et al. (2011). The researchers investigated ways in which students used learning resources in a typical blended learning environment with an option to use LC instead of attending lectures. In their analysis, which included tracking the lecture attendance of 534 students (via swiping their library cards on entry) and recording of LC access, the data were entered into a hierarchical cluster analysis in order to classify the typical behavior of students. To their surprise, the result revealed that none of the identified behavioral clusters involved students making heavy use of more than one resource, meaning that students who relied on watching videos for content rarely came to live lectures and vice versa. The finding led to the authors’ suggestion that what they observed was “blended teaching” (p. 500) as opposed to blended learning. Le et al. (2010) concluded similar findings: less than 10% of students both attended and watched more than half of the lectures online. In the same vein is a finding from another large-scale study of student usage patterns of LC and live lectures by Howard et al. (2018) at University College Dublin. In a Maths for Business course, students were given the choice of going over the course contents via LC (presented as short online mini-lectures), live lectures, or a combination of both. Researchers collected quantitative data on each student’s resource usage (attendance at live lectures and usage of online videos) for the entire class of 522 students and employed model-based clustering to identify four distinct resource usage patterns with lectures and/or videos. Remarkably, the largest cluster (N = 313, 60%) consisted of students who chose to cover the course material through videos and did not make much use of live lectures as a resource.
The contextual factors, such as institutional characteristics and location, could play a significant role. Reasonably, the more commuter-oriented a university is, the more likely a student is to choose LC over attendance because of the opportunity cost in time. The same logic would hold for any negative perception of attending campus (e.g., cost of campus food, weather); thus, different institutions may face more extreme attendance decreases than what is listed above because of their individual circumstances. But rationally, in all cases, attendance would likely decrease (at least marginally). However, great care should be taken in extrapolating these attendance reduction figures to other institutions. Attendance being a multi-variable function, it would be unwise to expect identical results. The figures above are only presented as an illustration of what appears to be the norm.
The method of recording attendance that each study used may influence these data. This review supposed that self-reporting bias would be a factor, projecting that surveys may show a more positive image of attendance than the reality (Chester et al., 2011), but taking attendance slips may also artificially improve attendance; this is discussed in more detail in the “Methodological concerns” section. In a similar vein, perhaps studies that investigated the initial implementation of LC and others that investigated already-established LC may show differences in attendance conclusions (students may form non-attendance habits over time). But there are not enough attendance data to investigate this question in this review.
Assuming attendance is affected, should it matter if LC is a perfect substitute for attendance? Attainment, not attendance, is the main measure of learning outcomes of higher education. This question is discussed in the next section.
What is the relationship between lecture capture usage and student attainment?
The most contentious aspect of LC is the impact its implementation has on student attainment (Nordmann et al., 2019). Reasonably, LC can provide repeat learning opportunities for students to engage in revision and better understand the content. As explained by Zimmermann et al. (2013), mathematics proofs are often not understood the first time around, and hence, LC could be a worthwhile tool for mathematics in particular. However, it may also provide a shortcut for less motivated students.
While recognizing the limitations regarding the use of students’ grades to reflect the changes in their mathematical cognition, we use the term “attainment” in the most commonly understood way, as educational achievement in the course of study, which is usually measured by student performance on assessment components (e.g., Pournara et al., 2015).
Of the 16 studies investigated in this review, six did not report on attainment, one concluded the neutral impact of LC (with the exception of students who were not following up on their intentions to watch LC after missing a lecture—this practice was strongly associated with poor grades; Yoon & Sneddon, 2011), one concluded a positive relationship, and nine concluded a negative relationship between LC and attainment (Table 3). This lends some rudimentary evidence that the blanket LC policies that are being rolled out at many tertiary institutions across all faculties may be premature.
Table 2 Summary of conclusions about the effect of LC on attendance Table 3 Summary of findings: relationship between LC and attainment Of the eight studies that reported a negative impact on attainment, all found that regular substitution of live lecture attendance with LC was associated with lower achievement (Howard et al., 2018; Inglis et al., 2011; Le et al., 2010; Meehan & McCallig, 2019; Mullamphy, 2011; Sorensen, 2015; Trenholm et al., 2019; Zimmermann et al., 2013). Consistent with the literature review of the previous decade, Trenholm et al. (2019) concluded that there is a significant negative correlation between final course grade and LC views (ρ = − 0.443, p = 0.014).
However, these studies did see benefits to groups of students who used LC supplementarily (Meehan & McCallig, 2019), but once a student used LC as their primary learning tool, they underperformed compared to their peers attending lectures. Sorensen (2015), examining data from 12 semesters at the University of South Dakota, found that the failure rates in the second year calculus courses were significantly higher in the semesters when the LC was provided; however, Zimmermann et al. (2013), reporting on the incorporation of the LC into two courses (Introduction to Elementary Geometry and Application-oriented Mathematics) for pre-service teachers in Germany, pointed out that “[t]he failure rate in the exams of both investigated courses is about 23%, which is relatively low for mathematics courses at German universities.” (p. 154). Arguably, in Zimmermann et al.’s case, such a comparison cannot be taken as objective evidence in the absence of data from the same courses for pre-service teachers without the LC's provision. This is because the educational context and demographic characteristics cannot be ruled out as primary factors in the observed relationships and, therefore, should be adequately controlled for in all reported associations.
Only one study reported a potential positive impact of LC on attainment. Based on qualitative analysis of in-depth interviews of students, Wood et al. (2018) reasoned that LC increased attainment because of its usefulness when circumstances stopped a student attending class, preventing a student from falling behind. LC’s utility in re-watching content is a reason why student attainment may improve. Additionally, the authors posited that the use of LC could compensate for disadvantages of attending information-heavy live lectures, as all students reported that having to multitask in taking notes while listening to the lecturer effectively was cognitively demanding. In spite of this plausible reasoning based on students’ opinions, it is worth noting that the study was not designed to empirically ascertain the impact of LC on attainment.
Some studies found that students who were already performing poorly were the most affected by the introduction of LC (Trenholm et al., 2019). Interestingly, this negative relationship between attainment and LC often appears in disconnect with the perception some students have of the utility they gain from LC (Trenholm et al., 2019). Gouia-Zarrad and Gunn (2018) found that students perceived increased performance and satisfaction with a course when LC was implemented despite a weak negative relationship between LC usage and performance, with low-achieving students using LC the most. Similarly, Howard et al. (2018) found not only that the group of students who relied on watching LC and not attending lectures (the largest cluster, N = 313, 60%) achieved the lowest grades in the course, but also that a “portion of students in this cluster strongly believed videos are superior to lectures in maximising their learning for the time available owing to a more concise format with less repetition, flexible use, efficient and faster pace” (p. 542). At James Cook University (Australia), in a survey of a mathematics for engineering class, 85% of students felt that LC did not have a negative effect on their academic performance (Mullamphy, 2011).
A previously mentioned study by Inglis et al. (2011) examined students’ patterns of usage with lectures, online videos, and the university mathematics support centre in three similar mathematics courses at Loughborough University (n = 534). For each student, the following was recorded: attendance at live lectures (via swiping the students’ library cards), the number of times they viewed online lectures (via logfiles on their Virtual Learning Environment server), and their number of visits to the mathematics support centre. However, it could not be confirmed that the visits to the mathematics support center were in relation to the courses in the study (as students could be seeking support for other problems). On performing a hierarchical cluster analysis, the authors identified four clusters of students: those who primarily attended live lectures (N = 214), those who primarily accessed the online lectures (N = 70), those who primarily used the mathematics support center (N = 60), and those who made little use of any resources (N = 185). It was shown that the students in the different clusters adopted significantly different strategies for their academic study and that, across all participants, the most important predictors of examination success were incoming diagnostic test achievement and attendance of live lectures during the course.
In a later study at University College Dublin, Meehan and McCallig (2019), on performing a cluster analysis, found that the students in the predominantly LC cluster (N = 30) achieved worse when compared to the students in the predominantly live lectures cluster (N = 22) (55.26% versus 57.69% on the final exam). Importantly, this was despite the inverse difference for the two clusters on the measure of prior achievement recorded in the Irish school system: 47.50 points (out of 100) for the predominantly LC cluster versus 43.21 points for the predominantly live lectures cluster.
As investigated by Le et al. (2010), a potential explanation for the association between poor performance and LC is not only that less motivated students use LC but that the technology facilitates more of a surface learning approach. Le et al. (2010) study of two calculus courses found that the highest usage of the ‘pause feature’ in LC was associated with the lowest attainment. Interestingly this contradicts a result from a previous identical study of a psychology course (Bassili, 2006). Le et al. (2010) reasoned that mathematics requires less rote memorisation and more understanding of concepts; usage of the pause feature may indicate memorisation. These findings were not controlled for prior achievement, so they may instead highlight students who could not keep up with the speed of the lecture and thus were more likely to perform poorly. However, Trenholm et al. (2019), who controlled for prior attainment, also found evidence to support Le et al. (2010) claim through the usage of the learning approaches questionnaire R-SPQ-2F. Regular LC users scored significantly higher in the “surface approach” category of the questionnaire relative to less frequent LC users. This lends support to the hypothesis that the relationship between LC usage and low attainment is not only because low-achieving/less motivated students tend to use LC more but because of some intrinsic quality of LC itself. As summarized by Trenholm et al. (2019), “regular RLV [recorded lecture videos] use, overall, maybe depressing the quality of student learning” (p. 13).
For some, there is no apparent reason why LC is not a functional substitute for a classroom in terms of the audiovisual content, but perhaps the lack of an academic environment leads students to be more easily distracted when using LC, thus lowering the quality of their cognitive engagement. This report hypothesizes that this is one of the principal reasons a negative correlation is often found between LC and attainment. LC may be a functional substitute for a highly motivated student, but removing a strict academic environment (i.e., social pressure from the classroom environment) allows a less motivated student to get the impression of learning even when they are not cognitively engaged. Trenholm et al. (2019) also reasoned in support of this hypothesis, suggesting an explanation for why LC may be an inferior learning method: because there is a lack of two-way interactivity (even in courses with little student input), as there is a “tacit acknowledgement of the ‘other’ in the room” (p. 14).
A perceived reason why LC may increase attainment—that LC encourages students to spend less time note-taking and more time focusing/engaging in lectures (Wood et al. 2018)—appears to be questionable. As mentioned in the previous section, only a minority of students attend lectures and watch LC; thus, this benefit of LC appears weak. In this review, no current studies that investigated this phenomenon in a mathematics course were found. Another behavioral change this review could not find research on was the social interactions students have before/after lectures. The introduction of LC may reduce incentives for students to form academic relationships with each other (e.g., talking about and checking someone else’s notes for a missed class), and because of this fewer student-to-student discussions may take place, reducing learning opportunities.