739 Is the Library One-Shot Effective? A Meta- Analytic Study Dani Brecher Cook* The one-shot instruction session is a dominant mode of teaching in academic libraries. While many conference presentations and articles about methods have been shared, there is little consensus about whether a single library session promotes student learning about information literacy topics. This meta-analysis gathers studies that employ quantitative measures of student learning in an attempt to determine if the one-shot is an effective modality for learning. Results indicate the need for a more critical look at the grouping of “one-shot” as a methodology and the need for further robust research on acquisition of student learning outcomes in the one-shot context. In academic libraries, the most pervasive method of providing information literacy instruction is the so-called “one-shot,” where students in a given course attend a single session facilitated by a librarian during their academic term to learn about conducting library research (Nicholson, 2016). Since the one-shot is so broadly used in the academic library context, there is a clear need for a meta-analytic review of the literature, so that library practitioners can make informed choices about how to spend their time, resources, and efforts in supporting undergraduate student learning. The aim of this meta-analysis is to attempt to answer the research question: Does the academic library one-shot result in improved information literacy knowledge and/ or skills of undergraduate students? So-called “bibliographic instruction” emerged in the 1980s; and, by the 1990s, it was firmly entrenched as a key component of academic librarians’ work (Martin & Jacobson, 1995). Almost as early, the one-shot began to receive criticism as an effective teaching tool (Gavin, 1995). As Bowles-Terry and Donovan put it, “one-shot instruction sessions were born out of necessity,” yet they have continued to be the dominant mode of work for instruction librarians (Bowles- Terry & Donovan, 2016, 137; for more recent statistics, please see Hsiesh, Dawson & Yang, 2021, 3). While many librarians may prefer other ways of engaging with students, the constraints on library instructional engagement remain. Due to the prevalence of one-shot sessions (which en- compass a wide variety of pedagogical approaches), there is a significant and divisive literature about the overall goals of the one-shot model of information literacy instruction, as well as the method’s efficacy. Can meaningful learning occur in a single librarian-mediated session? Some practitioners claim that this method is “preposterous for many librarians” (Walker and Pearce, 2014), while others assert that the one-shot is a valuable tool for student learning (one example: * Dani Brecher Cook is Associate University Librarian, Learning and User Experience at the UC San Diego Li- brary, email: danicook@ucsd.edu. © 2022 Dani Cook, Attribution-NonCommercial (https://creativecommons.org/ licenses/by-nc/4.0/) CC BY-NC. mailto:danicook@ucsd.edu https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ 740 College & Research Libraries September 2022 Wang, 2016). In her recent editorial, Pagowsky (2021) claims that the one-shot is “transactional and keeps us in cycles of ineffectiveness.” The literature ranges from self-reports of student attitudes and behaviors toward the library, to pre- and post-testing of specific knowledge, to authentic assessments1 of information literacy in student work at the end of a term. While the profession was called upon to robustly assess learning outcomes in instruction as early as Davinson (1984) and Barclay (1993) and the literature has grown large since then, there continues to be little agreement on which outcomes are meaningful measures, best practices for experimental design, and assessment instruments. The category of the “one-shot” includes a diverse range of pedagogical approaches, goals, and outcome measures, all bound together by the common qualities of being a single standalone, time-limited (often 45- or 60-minute, though may be as short as 15 minutes) session. Within the academic library context, the one-shot is meant to deliver information literacy knowledge, which can be measured and understood in a variety of ways. Sobel & Sugimoto (2012) found that instruction librarians use a wide variety of assessment tools and outcome measures, with a focus on access and resource selection. Between 2000 and 2016, many studies used the ACRL Information Literacy Competency Standards (2000) to determine outcome variables (for in- stance: Hsieh & Holden, 2010; Rosenblatt, 2010), and other studies continue to use specialized subject matter information literacy standards from ACRL (such as Tran et al., 2018). After the adoption of the ACRL Framework for Information Literacy in 2016, the literature moved away from Standards-based outcomes assessment to concept-driven outcomes (examples: Hurley & Potter, 2017; Tomaszeski, 2021). Yet other studies employ entirely different frameworks for information literacy assessment, such as the Research Readiness-Focused Assessment (Wang, 2016), or focus on the diversity of source types used (Howard et al., 2014). Despite the wide variety of outcome measures and assessment types, these studies are often framed as measur- ing the efficacy of the one-shot in teaching information literacy–related outcomes. Within the literature examining one-shot library instruction, there is a broad range of quality of studies as well. Early assessment literature from the 2000s has been criticized for its reliance on convenience sampling, as well as a lack of emphasis on long-term retention of skills (Spievak & Hayes-Bohanan, 2013). However, more recent literature tends to be more expertly designed and executed, though longitudinal studies remain scarce. Additionally, most of the literature focuses on early-career undergraduate populations, such as introductory writing courses or college-preparedness courses, with very few studies examining graduate populations. Methods Meta-analysis was selected as the most relevant method for conducting the work of examining the aggregate impact (or not) of the one-shot. A meta-analysis is a formal review that collects relevant studies that examine the same topic (but may have different methods or specific outcomes) and synthesizes their results using statistical methods. Relevant studies were identified by searching LISTA (Library, Information Science, and Technology Abstracts), Library Literature and Information Science Full Text (H.W. Wilson), ERIC, and Education Source in October 2021. The search strategy combined the terms (“one shot” OR “one-shot” OR “single-session” OR “bibliographic instruction”) AND (assess* OR evaluat* OR measur* OR test*) AND (undergraduate OR college OR university) AND (librar* OR information literacy), and was not restricted by date range. The bibliographies of studies that met the search criteria were mined to identify additional studies. Is the Library One-Shot Effective? 741 Case studies, quasi-experimental studies, and experimental studies were all considered for this meta-analysis. To be included in the meta-analysis, studies had to test the educational effectiveness of a one-shot library session, meaning that researchers examined a dependent variable that measured knowledge acquisition gained from the one-shot experience. For the purposes of this meta-analysis, the researcher did not include studies that examined GPA or course grade as outcome variables, as these are controversial metrics with many potential confounding variables (examples: Fisher, 2018; Robertshaw & Asher, 2019). Studies that mea- sured affect, confidence, or anxiety were excluded. Studies that investigated the effectiveness of online modules, multisession instruction, or term-long courses were excluded. Only pub- lications from journals were considered for inclusion in the meta-analysis. Prior to making this decision, recent conference proceedings from LOEX and ACRL were examined. Eligible participant populations were undergraduate students at community colleges, four-year col- leges, and universities. Studies were not limited by publication date. Due to the researcher’s language limitations, only studies in English were included. Only peer-reviewed journal publications were considered for inclusion. The researcher reached out to a small group of authors to ask for additional information about their studies to determine if they could be included. While there is an emerging best practice in the social sciences to include two inde- pendent reviewers in conducting a meta-analytic review (Siddaway, Wood & Hedges, 2019), the present meta-analysis was conducted by a single reviewer. As shown in figure 1, 3,058 potentially relevant citations were retrieved from the structured literature search. Titles and abstracts were reviewed to determine the relevance of the article, as well as to preliminarily identify whether the article was a research study. Of those, 66 articles were sub- ject to in-depth reviews, in- cluding examination of the methods and data reported in the results and discus- sion. To be included in the meta-analysis, study design needed to be experimental or quasi-experimental and include a measure that indi- cated difference due to the one-shot intervention. Nine of those articles included relevant data, such as a t-sta- tistic, chi-squared statistic, or mean and standard deviation FIGURE 1 Flow Diagram of the Meta-analytic Review Process 742 College & Research Libraries September 2022 for the researcher to calculate an effect size estimate. Articles that did not include the size of the population, n, were not included. Two other articles were identified by citation mining. After identifying the 11 articles for the present meta-analysis, the researcher entered data extraction elements in an Excel spreadsheet. Some articles reported more than one study or variable that could be used (such as M.E. Cohen et al., 2016). Elements included were inde- pendent variable, dependent variable, study size n, degrees of freedom, one-tailed t-value, p-value, and effect size r. For articles that reported t-statistics, the effect size r was calculated using Cohen’s equation: (Cohen, 1965). Only two articles in the meta-analysis reported chi-squared statistics, and the effect size was calculated using: (Rosenthal, 1991). The articles were also coded to represent population characteristics (in other words, major or general education, class standing), institution type, length of the one-shot session, and experimental design (that is, authentic assessments versus pre- and post-testing) to conduct moderator analyses. Please see appendix for the codebook. The total n of these studies is 1,572. From 11 identified studies, 16 effect sizes were able to be calculated. Fisher’s transformation (zr) was also calculated by the researcher. A summary table is avail- able in figure 2. FIGURE 2 Summary Table of Articles Included in Meta-analysis Article Dependent Variable n df t Effect Size r z r M.E. Cohen et al. 2016 Score on information literacy (IL) content quiz 64 67 –6.94 0.646698 0.7696 M.E. Cohen et al. 2016 Score on IL content quiz 24 23 –1.51 0.30032 0.3099 Howard et al. 2014 Number of sources used 207 205 0.94 0.06551 0.0656 Howard et al. 2014 Simpson Diversity Index 199 197 1.2 0.08519 0.0854 Howard et al. 2014 Usage of library sources 207 205 3.48 0.2362 0.2407 Hurst and Leonard 2007 Number of source types used 184 182 –5.35 0.3686 0.3868 Lantzy 2016 Score on IL content quiz 31 31 9.46 0.8618 1.3003 Lantzy 2016 Score on IL content quiz 26 24 4.69 0.6915 0.8509 Martin 2008 Type of information resource used 200 1 N/A (=1.612) 0.0898 0.0900 Mery, Newby, and Peng 2012 Score on IL content quiz 32 62 2.6465 0.3186 0.3301 Is the Library One-Shot Effective? 743 All data were analyzed using Jamovi 1.6 and the MAJOR meta-analysis package for Jamovi. Results Of the 11 papers under examination in this meta-analysis, five conducted paired sample t- tests, four conducted independent-samples t-tests, and two conducted a chi-squared test for their statistical analyses. When researchers did not report the type of t-test conducted, unless it was explicitly stated that individual scores were paired, the present researcher assumed an independent samples t-test. Only one study was truly experimental, taking a sample from a population and divid- ing it into treatment and control groups (that is, Howard et al., 2014). The majority of the studies (n = 8) in the analysis were quasi-experimental, and used a pre-test/post-test design. Most study populations included convenience samples to some degree, which may impact the outcomes of the studies. For example, we could imagine that instructors who opt their courses into a study on educational effectiveness may care more about teaching, and thus be better teachers; or that students who agree to take an optional class and assessment may care more about learning. All 11 of the studies in this meta-analysis yielded p-values in the expected, positive direction, indicating that, at the least, one-shot library instruction does not seem to damage student learning. Heterogeneity testing indicates that the students are significantly hetero- geneous (τ = 0.326, p< .001). This means that there is significant variation in the results of the various studies, which is expected due to the wide variety of methods and outcome measures in these studies. In a fixed effects model, the overall effect size r is 0.268. In a random effects model (K = 16), z = 4.53, p <.001, 95% CI: (.229, .577).2 Converting Fisher’s z back to r for ease of interpretation, r = .383. As can be seen in the forest plot in figure 3, half of the studies’ 95% confidence interval includes 0. When a confidence interval includes zero, that indicates that there is a chance that there is no treatment effect and, in this case, that the one-shot makes no difference. Examining the moderator variables, there does not appear to be a specific charac- teristic that unifies the studies that have confidence intervals that do not include zero—they all share characteristics with other studies that do include zero. FIGURE 2 Summary Table of Articles Included in Meta-analysis Article Dependent Variable n df t Effect Size r z r Portmann and Roush 2004 Score on IL content quiz about source usage 38 37 –0.055 0.0090 0.0090 Spievak and Hayes- Bohanan 2013 Selection of a reliable source 119 1 N/A (=3.85) 0.1508 0.1519 Tewell 2014 Score on IL content quiz 90 89 1.536 0.1607 0.1621 Walker and Pearce 2014 Score on IL content quiz 69 68 –6.134 0.5968 0.6882 Walker and Pearce 2014 Score on IL content quiz 69 68 –6.754 0.6336 0.7475 Wilhite 2004 Score on IL content quiz 13 12 –2.3 0.5531 0.6229 744 College & Research Libraries September 2022 FIGURE 3 Forest Plot for the Random Effects Model FIGURE 4 Funnel Plot Is the Library One-Shot Effective? 745 One study (Lantzy, 2016.1) may be overly influential, as can be seen in the funnel plot in figure 4. That single study creates asymmetry in the funnel plot. A symmetrical funnel plot would indicate that precision of the studies increases as sample sizes get larger. This funnel plot indicates that the literature may be missing smaller studies. It is likely that the asymmetry in this funnel plot is due to the strong heterogeneity of the studies under review. Moderator variables of interest were also considered. Institution type does appear to have an effect (p = .007), with midsize public institutions demonstrating the largest effect sizes. The researcher also investigated if the effect measured is a meaningful moderator variable. Does looking at a specific outcome (such as the number of web sources cited) have a different effect size than a more diffuse variable (for instance, change in information literacy score)? Compar- ing Howard, Nicholas, Hayes & Appelt (2014) with Walker & Pearce (2014) and Portmann & Roush (2004) indicates that the effect size is significantly different at the p < .05 level for these two groups (t1 = 17.0, p = .04). The analysis showed that change in information literacy score on a standardized test was more apt to be changed than the number of web sources cited. For the practitioner, this may indicate that measures of learning that require students to apply their knowledge in new contexts may be more difficult to attain than answering test questions specifically aimed at information literacy concepts taught in the class. Finally, a publication bias analysis was performed. Publication bias examines whether studies that have significant results are more likely to be published than those that show null results. As librarianship is a heavily practitioner-focused field, there may be several factors at play here: Practitioners may be more likely to contribute to the gray literature (though a review of the recent proceedings of the LOEX conference yielded no additional candidates for inclusion); practitioners may not be incentivized to publish when their methods do produce a statistically significant result; or there is not an incentive for practitioners to conduct robust assessment of one-shot instruction sessions. The Begg and Mazumdar Rank Correlation was rejected as an approach due to the limited number of studies available (Begg & Mazumdar, 1994). However, the Rosenthal Method for Fail Safe N finds that approximately 874 null- result studies would be required to have a significant effect (Rosenthal, 1979), meaning that a relatively large number of studies would have to have not been included to drive the overall effect size to zero. Discussion While it is not possible to draw causal inferences from the studies included in the meta-analysis, as most of the studies were quasi-experimental in design, large amounts of evidence could begin to make a case for causality. Since only 11 studies were considered in this meta-analysis, the researcher hesitates to draw any kind of causal conclusion; the wide range of the 95% con- fidence interval could also be narrowed by the inclusion of more studies. The heterogeneity of the methods is a benefit of the studies meta-analyzed in this article, as it allows us to feel more confident in generalizing the findings (Rosenthal, 1991, 129). The overall effect size of the one-shot intervention on measures of learning in these 11 studies is approximately .383 (this varies depending on fixed or random effect models). Ac- cording to Cohen (1977), this is a medium effect size, meaning that it is likely to be noticeable. The 95% confidence interval goes as low as .226, meaning that it is possible that the effect size may be much smaller. Empirically, this meta-analysis indicates that there is some positive effect of the one-shot instruction session. That being said, it is important to note that in every 746 College & Research Libraries September 2022 study but one, outcomes were measured directly after the library instruction session. The literature reviewed for this meta-analysis, then, cannot comment on how efficacious one-shot instruction might be over time. Additional investigation and more robust studies are needed. It should also be noted that effect sizes were universally small (r < .2) in studies where an authentic assessment was employed. This is an important feature, as it may indicate that one-shot instruction has limited effectiveness in actual skill-building, as opposed to being able to answer factual questions on a quiz. For interest, the researcher performed meta-analytic processes on two smaller data sets: one consisting of the articles that employed information literacy-related tests, and the other of the articles that used an authentic assessment. For the studies that employed tests, a random effects model yielded (K = 10), z = 4.71, p <.001, 95% CI: (.335, .814). For studies that employed authentic assessments, such as analysis of source selection, a random effects model yielded (K = 6), z = 3.29, p <.001, 95% CI: (.069, .271). This indicates that one-shots that are targeted at specific skills to be measured on a test are more likely to have an effect than those that ask students to perform authentic tasks. As most re- search requires students to perform novel searches, evaluation, and synthesis, it should be considered if the one-shot adequately builds those skills. This meta-analysis does not review the literature of other instructional models, such as embedding librarians in courses, developing comprehensive online learning modules, or scaffolding information literacy instruction over an entire course in collaboration with fac- ulty instructors. Meta-analyses of these literatures would also be warranted, and effect sizes compared. In the present meta-analysis, two moderator variables were investigated: The type of dependent variable measured, and the type of institution where the study was conducted. Both variables were significant. There was a significant difference between specific variables examined and more general outcomes, which indicates that the field may benefit from clearly defining desired learning outcomes for one-shot library instruction, and measuring them across multiple studies. One potential moderator that would be interesting to explore, though not possible with the present literature, is the teaching experience of the librarian. Perhaps the most important lesson learned from conducting this meta-analysis, though, is not about the effect of one-shot instruction on student learning, but rather about the state of the library instruction literature (or perhaps the preparation that librarians receive in re- search methods). Of the 3,058 articles initially identified in the structured literature search, only 66 (or ~2%) were studies of educational effectiveness. Of that 66, only 9 (or 13.6%, or 0.2% of the original set) reported enough data and conducted sufficient statistical analyses to be included in the present meta-analysis. Even within the 11 articles ultimately included in the meta-analysis, there were small problems with data reported; one article misreported the degrees of freedom in a chi-squared test, causing the researcher to independently redo the analysis. Another article transposed greater-than and less-than signs, so that when they reported that a finding was not statistically significant and had a small t-value, they claimed p < .05. These transcription errors cause some concern for the validity of the data presented in these articles; but, because the literature is so sparse, the researcher did opt to include them in the meta-analysis. Two studies were rejected when percentages of students reported resulted in fractions of humans, indicating that missing values were not adequately reported. The field would benefit greatly from additional studies that clearly and accurately report statistical tests, including precise p-values, effect sizes, and confidence intervals. The vast majority of Is the Library One-Shot Effective? 747 studies in this field report descriptive statistics (primarily mean values and frequency tables) and a smaller number report percentages that fall into specific categories (“relative risk”), but more meaningful and informative analyses could and should be performed. In addition, well-designed experimental studies of the effect of the one-shot intervention would enrich the field and introduce the possibility of drawing causal inferences. The preponderance of the case study makes it challenging to trust the results of the meta-analysis. Given the continued prevalence of the one-shot in library instruction, the field should consider an increased focus on robust assessments of their efficacy. This may require addi- tional statistical or methodological training and support for library practitioners. Further, the results of the present meta-analysis should encourage the field to reconsider the classification of the one-shot as a “method.” The one-shot is not a monolith, but instead encompasses a wide variety of outcome variables, pedagogical strategies, timing, and populations, and be- ing able to compare studies that are more alike would yield more meaningful and actionable results for library practitioners. This study indicates that there is likely more to unpack in terms of what differentiates an effective one-shot from an ineffective one. Replication studies at different institutions may also be warranted. In conclusion, the researcher hopes that this meta-analysis spurs additional investigation into the field and that a more directional meta- analysis may be possible in the future. 748 College & Research Libraries September 2022 APPENDIX Codebook Variable Possible Values Definition Student Category Community college students; undergraduate students The population studied, as identified by the original study authors. Discipline English; education; business; sociology; unknown/general The type of course that the one-shot was associated with, as described by original study authors. Institution Type Community college; small liberal arts college; mid-sized public; large public The type of college or university where the study took place, as identified by the original study authors. Experimental Design Pre- and post-test; quasi- experimental; survey; portfolio assessment The type of study that was conducted, as reported by the original study authors. Some studies used more than one design and were coded as such. Length of Session 45 minutes; 50 minutes; 60 minutes; 75 minutes; 90 minutes; not described The length of the one-shot session where outcomes were being measured. Not all studies included the length and were marked as such. Analysis Type Chi-squared; independent sample t-test; paired sample t-test The type of analysis conducted on the data collected to report on the outcome variable. Dependent Variable Type Test; authentic assessment Describes the type of measure that was used to describe the change in the dependent variable. These were selected and coded by the present researcher. Notes 1. According to Wiggins (1989), the creator of this phrase, authentic assessment is a “true test” of student learning, where students are asked to demonstrate application of knowledge in an exemplary and nonstandard- ized task. Examples include essays, portfolios, and research projects. 2. Fixed effects models assume there is a single central tendency and can only be generalized to other popu- lations of the same ilk. Random effects models generally have a smaller effect size but are more generalizable and amenable to heterogeneity of studies. This is the case in this analysis. References American Library Association. (2000). ACRL standards: Information literacy competency standards for higher education. College & Research Libraries News, 61(3). https://doi.org/10.5860/crln.61.3.207 Association of College & Research Libraries. (2015). Framework for Information Literacy for Higher Education. https://www.ala.org/acrl/standards/ilframework Barclay, D. (1993). Evaluating library instruction: Doing the best with what you have. Reference Quarterly, 33(2): 195–202. Begg, C. B., & Mazumdar, M. (1994). Operating characteristics of a rank correlation test for publication bias. https://doi.org/10.5860/crln.61.3.207 https://www.ala.org/acrl/standards/ilframework Is the Library One-Shot Effective? 749 Biometrics 50(4): 1088–1101. https://doi.org/10.2307/2533446 Bowles-Terry, M., & Donovan, C. (2016). Serving notice on the one-shot: Changing roles for instruction librar- ians. International Information & Library Review 48(2): 137–142. https://doi.org/10.1080/10572317.2016.1176457 Cohen, J. (1965). Some statistical issues in psychological research. In Wolman, B.B. (Ed.), Handbook of clinical psychology. McGraw-Hill. Cohen, J. (1977). Statistical power analysis for the behavioral sciences. Academic Press. *Cohen, M.E., Poggiali, J., Lehner-Quam, A., Wright, R., & West, R.K. (2016). Flipping the classroom in busi- ness and education one-shot sessions: A research study. Journal of Information Literacy 10(2): 40. https://doi. org/10.11645/10.2.2127 Davinson, D. (1984). Never mind the quality, feel the width. Reference Librarian 3(10), 29–37. https://doi.org/10.1300/ J120v03n10_04 Fisher, Z. (2018). Who succeeds in higher education? Questioning the connection between academic libraries and student success. In Proceedings of the 2018 CARL Conference. Redwood City, CA. http://conf2018.carl-acrl. org/conference-proceedings/ Gavin, C. (1994). Guiding students along the information highway: Librarians collaborating with composition instructors. Journal of Teaching Writing, 13(1 & 2), 225–236. *Howard, K., Nicholas, T., Hayes, T., & Appelt, C.W. (2014). Evaluating one-shot library sessions: Impact on the quality and diversity of student source use. Community & Junior College Libraries 20(1–2): 27–38. https://doi. org/10.1080/02763915.2014.1009749 Hsieh, M.L., Dawson, P.H., & Yang, S.Q. (2021). The ACRL Framework successes and challenges since 2016: A survey. Journal of Academic Librarianship, 47(2). https://doi.org/10.1016/j.acalib.2020.102306 Hsieh, M.L., & Holden, H.A. (2010). The effectiveness of a university’s single-session information literacy instruc- tion. Reference Services Review, 38(3), 458–473. https://doi.org/10.1108/00907321011070937 Hurley, D.A., & Potter, R. (2017). Teaching with the Framework: A Cephalonian approach. Reference Services Review, 45(1), 117–130. https://doi.org/10.1108/RSR-07-2016-0044 *Hurst, S. & Leonard. J. (2007). Garbage in, garbage out: The effect of library instruction on the quality of stu- dents’ term papers.” Electronic Journal of Academic and Special Librarianship, 8(1). http://southernlibrarianship. icaap.org/content/v08n01/hurst_s01.htm The jamovi project (2021). jamovi. (Version 1.6) [Computer Software]. Retrieved from https://www.jamovi.org *Lantzy, T. (2016). Health literacy education: The impact of synchronous instruction. Reference Services Review, 44(2): 100–121. https://doi.org/10.1108/RSR-02-2016-0007 *Martin, J. (2008). The information seeking behavior of undergraduate education majors: Does library instruction play a role? Evidence Based Library and Information Practice, 3(4). https://doi.org/10.18438/B8HK7X Martin, L.M. & Jacobson, T.E. (1995). Reflections on maturity: Introduction to library instruction revisited: Bib- liographic instruction comes of age. Reference Librarian, 24(51–52): 5–13. https://doi.org/10.1300/J120v24n51_02 *Mery, Y., Newby, J., & Peng, K. (2012). Why one-shot information literacy sessions are not the future of instruc- tion: A case for online credit courses. College & Research Libraries, 73(4): 366–377. https://doi.org/10.5860/crl-271 Nicholson, K.P. (2016). “Taking back” information literacy: Time and the one-shot in the neoliberal university. In N. Pagowsky & K. McElroy (Eds.), Critical library pedagogy handbook (Vol. 1, pp. 25–39). ACRL Press. Pagowsky, N. (2021). The contested one-shot: Deconstructing power structures to imagine new futures. College & Research Libraries, 82(3). https://doi.org/10.5860/crl.82.3.300 *Portmann, C.A., & Roush, A.J. (2004). Assessing the effects of library instruction. Journal of Academic Librarian- ship, 30(6): 461–465. R Core Team. (2020). R: A language and environment for statistical computing. (Version 4.0) [Computer software]. Retrieved from https://cran.r-project.org (R packages retrieved from MRAN snapshot 2020-08-24). Robertshaw, M.B., & Asher, A. (2019). Unethical numbers? A meta-analysis of library learning analytics studies. Library Trends, 68(1): 76–101. Rosenblatt, S. (2010). They can find it but they don’t know what to do with it: Describing the use of scholarly lit- erature by undergraduate students. Journal of Information Literacy, 4(2), 50–61. https://doi.org/10.11645/4.2.1486 Rosenthal, R. (1979). The file drawer problem and tolerance for null results. Psychological Bulletin, 86(3): 638–641. Rosenthal, R. (1991). Meta-analytic procedures for social research. SAGE Publications, Inc. Siddaway, A.P., Wood, A.M., & Hedges, L.V. (2019). How to do a systematic review: A best practice guide for conducting and reporting narrative reviews, meta-analyses, and meta-syntheses. Annual Review of Psychol- ogy, 70(1), 747–770. https://doi.org/10.1146/annurev-psych-010418-102803 Sobel, K., & Sugimoto, C.R. (2012). Assessment of learning during library instruction: Practices, prevalence, and preparation. Journal of Academic Librarianship, 38(4), 191–204. https://doi.org/10.1016/j.acalib.2012.04.004 *Spievak, E.R., & Hayes-Bohanan, P. (2013). Just enough of a good thing: Indications of long-term efficacy https://doi.org/10.2307/2533446 https://doi.org/10.1080/10572317.2016.1176457 https://doi.org/10.11645/10.2.2127 https://doi.org/10.11645/10.2.2127 https://doi.org/10.1300/J120v03n10_04 https://doi.org/10.1300/J120v03n10_04 http://conf2018.carl-acrl.org/conference-proceedings/ http://conf2018.carl-acrl.org/conference-proceedings/ https://doi.org/10.1080/02763915.2014.1009749 https://doi.org/10.1080/02763915.2014.1009749 https://doi.org/10.1016/j.acalib.2020.102306 https://doi.org/10.1016/j.acalib.2020.102306 https://doi.org/10.1108/00907321011070937 https://doi.org/10.1108/RSR-07-2016-0044 http://southernlibrarianship.icaap.org/content/v08n01/hurst_s01.htm http://southernlibrarianship.icaap.org/content/v08n01/hurst_s01.htm https://www.jamovi.org https://doi.org/10.1108/RSR-02-2016-0007 https://doi.org/10.18438/B8HK7X https://doi.org/10.1300/J120v24n51_02 https://doi.org/10.5860/crl-271 https://doi.org/10.5860/crl.82.3.300 https://cran.r-project.org https://doi.org/10.11645/4.2.1486 https://doi.org/10.1146/annurev-psych-010418-102803 https://doi.org/10.1016/j.acalib.2012.04.004 750 College & Research Libraries September 2022 in one-shot library instruction. Journal of Academic Librarianship, 39(6): 488–499. https://doi.org/10.1016/j. acalib.2013.08.013 *Tewell, E.C. (2014). Tying television comedies to information literacy: A mixed-methods investigation. Journal of Academic Librarianship 40(2): 134–141. https://doi.org/10.1016/j.acalib.2014.02.004 Tran, C.Y., Miller, C.-A., & Aveni, D. (2018). Baseline assessment: Understanding WISE freshman students’ in- formation literacy skills in a one-shot library session. Science & Technology Libraries, 37(3), 302–321. https:// doi.org/10.1080/0194262X.2018.1460651 Tomaszewski, R. (2021). A STEM e-class in action: A case study for asynchronous one-shot library instruction. Journal of Academic Librarianship, 47(5), 102414. https://doi.org/10.1016/j.acalib.2021.102414 *Walker, K.W., & Pearce, M. (2014). Student engagement in one-shot library instruction. Journal of Academic Li- brarianship, 40(3): 281–290. https://doi.org/10.1016/j.acalib.2014.04.004 Wang, R. (2016). Assessment for one-shot library instruction: A conceptual approach. portal: Libraries and the Academy, 16(3): 619–648. https://doi.org/10.1353/pla.2016.0042 Wiggins, G. (1989). A true test: Toward more authentic and equitable assessment. Phi Delta Kappan, 70(9), 703–713. *Wilhite, J.M. (2004). Internet versus live: Assessment of government documents bibliographic instruction. Journal of Government Information, 30(5–6): 561–574. https://doi.org/10.1016/j.jgi.2004.10.002 * = included in the meta-analysis https://doi.org/10.1016/j.acalib.2013.08.013 https://doi.org/10.1016/j.acalib.2013.08.013 https://doi.org/10.1016/j.acalib.2014.02.004 https://doi.org/10.1080/0194262X.2018.1460651 https://doi.org/10.1080/0194262X.2018.1460651 https://doi.org/10.1016/j.acalib.2021.102414 https://doi.org/10.1016/j.acalib.2014.04.004 https://doi.org/10.1353/pla.2016.0042 https://doi.org/10.1016/j.jgi.2004.10.002