Competency L
Demonstrate understanding of quantitative and qualitative research methods, the ability to design a research project, and the ability to evaluate and synthesize research literature.
Understanding the Competency
Competency L makes three demands, and it is worth separating them before doing anything else. The first is that I understand both quantitative and qualitative research methods well enough to explain what each is for. The second is that I can design a research project, meaning I can move from a question to a plan capable of answering it. The third is that I can evaluate and synthesize research literature, which are two different acts rather than one.
Research, in the sense this competency uses, is systematic inquiry undertaken to answer a question with evidence rather than with intuition or precedent. The word doing the work here is systematic. What separates research from an informed opinion is not the sophistication of the conclusion but the fact that the procedure by which the conclusion was reached is visible, repeatable, and open to challenge. Connaway and Radford (2021) organize the whole field around this sequence. First a question is formulated, then a design is chosen to suit it, then a population is defined and sampled, followed by data being collected through an instrument, and finally the results are analyzed and reported in a way that lets another person judge whether the conclusion follows as expected.
The two families of method differ in what they are built to find out. Quantitative research measures and counts. It produces numeric data, analyzes it statistically, and answers questions of magnitude and relationship (how many, how often, how much, and whether two things vary together). Its ambition is usually generalization from a sample to a larger population, which is why sampling is the hinge on which its credibility turns. Qualitative research investigates meaning, process, and context. It produces textual data, most often from interviews, observation, focus groups, or documents. Qualitative research analyzes the data interpretively, answering questions of how and why and what. Its ambition is depth and fidelity to participants’ own accounts rather than generalization.
A persistent misunderstanding, which Wildemuth (2016) is careful to dispel, treats qualitative work as the softer or less rigorous of the two. It is not. Qualitative research has its own standards of rigor (i.e., systematic coding, attention to disconfirming evidence, transparency about the researcher’s position, etc.). A poorly designed survey is no more trustworthy than a poorly conducted interview study. What the competency asks is not that I prefer one, but that I can say which question calls for which method, and why.
The phrase applicability to a specific environment is the part practitioners most often skip. Methods are not universally good or bad. They are appropriate or inappropriate to a setting. A design that works in a university with a research office and an institutional review board may be unworkable in a small public library with one program coordinator and forty participants. Choosing a method means choosing among real constraints, not selecting the most impressive option.
During my MLIS program at San José State University, I developed this competency through two courses in particular. INFO 285-Research Methods: Evaluating Programs and Services, required me to build an evaluation framework for a real grant-funded program, then to analyze real assessment data collected from its participants. INFO 287-Evidence Based Research, required me to locate, appraise, and synthesize a body of published research for a defined professional audience. Before either, my undergraduate major in social science with a minor in psychology gave me my first grounding in research design and statistics. Additionally, my professional work in digital asset management involves reporting on usage metrics that were, in effect, descriptive statistics presented to people who would make decisions on them.
Quantitative and Qualitative Research Methods
Quantitative work in library and information science most often takes the form of surveys and of analysis performed on data an institution already holds. Its foundation is descriptive statistics, meaning the summary measures that characterize a data set. This includes things like measures of central tendency such as the mean, median, and mode, and measures of dispersion such as the range and standard deviation (which describe how widely the value scatters around the center). Frequency distributions, which report how many cases fall into each category or interval, are the workhorse of practitioner research. Cumulative frequency, which accumulates those counts across excessive intervals, lets researchers say what proportion of a group falls at or below a given point. Conway and Radford (2021) treat these as the baseline competence, since most library research questions are descriptive rather than causal. Inferential techniques go further, testing whether an observed pattern is likely to reflect something real in the population rather than chance. The chi-square tests, for instance, examine whether two categorical variables are associated more strongly than would be expected at random.
Wildemuth (2016) is useful precisely because it resists organizing methods by tradition and organizes them instead by the question they answer. This moves from research questions and study design through sampling, and then through the individual techniques of data collection and analysis. The implication running through it is that method selection is a downstream decision. A researcher who chooses a survey because surveys are familiar, and only then decides what to ask, has inverted the process.
On the qualitative side, the most practically valuable of my sources is Hsieh and Shannon (2005), who distinguished three approaches to qualitative content analysis (the systematic interpretation of textual data through coding). In conventional content analysis, the codes emerge from the data itself. The researcher reads without a pre-existing scheme, and allows categories to form from what respondents actually said. In directed content analysis the researcher begins from an existing theory or prior research, and uses it to generate an initial coding scheme. They test and extend it against new data. Summative content analysis has the researcher counting occurrences of particular words or content, and then interpreting the underlying meaning of those patterns. The distinction matters, because the three answer different questions and carry different risks. Conventional analysis can surface categories no existing framework anticipated, but offers no theoretical scaffolding. Directed analysis is efficient and cumulative, but the researcher may see only what the framework predisposed them to see.
Mixed methods research combines the two families. Hayman and Smith (2020) examine how it is actually practiced within library information science. Their central point is one that is widely misunderstooMixed methods are not simply doing some of each (qualitative and quantitative), but what makes a study genuinely mixed is integration. Meaning that the quantitative and qualitative strands inform one another rather than sitting side-by-side in the same report. A survey that establishes the extent of something, followed by interviews that explains why it occurs, is integrated. A survey in some interviews reported in separate sections, and never brought into contact, are two studies bound together.
Evaluating and synthesizing literature is a methodological problem in its own right. Grant and Booth (2009) demonstrate why. Their typology identifies 14 distinct review types and characterizes each among four dimensions–search, appraisal, synthesis, and analysis. Showing that a systematic review, a scoping review, a rapid review, a critical review, and a narrative review are different instruments with different claims to comprehensiveness. The practical consequence is that “literature review” names a family rather than a method, and that a review which searches selectively cannot claim the authority of one that searched exhaustively. Price (2022), writing in the same evidence-based practice tradition, takes up the proliferation of synthesis approaches, and the difficulty this creates for practitioners trying to judge which body of synthesized evidence they are actually looking at. Read together, the two established that evaluating literature means asking how our view was built before accepting what it concluded.
Finally, in public service institutions, research most often appears under the name of evaluation. Haley Goldman's Evaluation Guide for Public Libraries, published by the Urban Libraries Council, addresses this audience directly. Goldman offers library staff without formal research training a structured way to decide what a program is trying to change, what evidence would demonstrate that change, and which methods are feasible given the resources at hand. Evaluation is research constrained by practice. The question is set by a funder or program rather than by the researcher's curiosity, the timeline is fixed, and the sample is whoever participated. That does not make it less rigorous, but it does mean the design work happens under conditions that textbooks treat as exceptional.
Why It Matters to the Profession
Libraries, archives, and cultural institutions make continuous decisions that rest on claims about people (i.e., that a population needs a service, that a program produced a benefit, that a collection is underused, etc). Every one of those is an empirical claim. Without method, they are assertions backed by whoever argues most confidently, and the institution has no way to discover that it is wrong.
The stakes are highest where money is involved. Grant-funded programs are required to demonstrate outcomes, not merely activity, and the distinction is exact. An output is something the program produced, such as workshops held or devices distributed. While an outcome is a change in the people it serves. An institution that reports only outputs has described its own effort, rather than its effect. Producing outcome evidence requires deciding in advance what change is expected, and how it would be detected–which is research design performed before any data exists.
Most information professionals will also consume far more research than they conduct, and that is a skill in itself. Reading a study well means asking who was sampled and who was therefore left out, whether the instrument measured what it claimed, and whether the population studied resembles one’s own well enough for the findings to transfer. A professional who cannot make that judgement risks importing recommendations developed for users nothing like theirs. This has an equity dimension that is easy to miss. Sampling decisions determine whose experience becomes evidence, and populations that are hard to reach are quietly excluded from the research base that shapes services intended for them.
Applying Research Methods in a Particular Information Environment
My particular environment is the library, archive, and cultural heritage institutions, and the evaluation of publicly funded programs and services. Research in this setting is shaped by constraints that are not defects to be apologized for, but conditions to be designed around.
Participation is voluntary, so samples are self-selected, and skew toward the already engaged. Populations are small, so a program with two dozen participants will never support inferential statistics, and should not pretend to. There is rarely a control group, because withholding a service from half of an eligible population in order to establish a comparison is usually neither practical nor ethical. Attrition between a pre-assessment and a post-assessment is routine, since participants move, withdraw, or simply do not respond a year late. Self-reported measures, which are frequently the only instrument available, carry known weaknesses. People rate their own ability against a standard that shifts as they learn more, so a participant who becomes genuinely more competent may rate themselves lower than they did at the start.
The realistic response to these constraints is a mixed design of the kind Hayman and Smith (2020) describe. Closed-response items, analyzed descriptively, establish the shape of a group and the direction of change. Open-ended items, analyzed through content analysis in one of the forms Hsieh and Shannon (2005) distinguish, explain what the numbers cannot (i.e., why a participant did not meet a goal, what obstacle intervened, what the program meant to them in terms they chose themselves). In a small program, the qualitative strand is often the more informative of the two, because with 24 participants a three-point shift in an average is within noise, while 24 explanations of what happened are 24 pieces of genuine information.
Designing for this environment therefore begins before data collection, with a statement of what the program is expected to change and what evidence would show it. It continues through instrument design, where the questions asked determine what can later be analyzed. It ends in an analysis that is honest about its own limit, reporting the number of responses actually received rather than the number of participants enrolled, and declining to generalize beyond the people who answered.
Evidence 1
The first piece of evidence demonstrating my mastery of Competency L is my report The Role of the LIbrary in the Digital Age: Makerspaces, Libraries of Things, and Community Hubs, completed in INFO 287: Evidence Based Research and dated August 23, 2026. It was solo work.
Link: https://docs.google.com/document/d/1on3gt3z-pnVNISP-pAqrNmT2X3l5dHM7TIiLNkxSYwo/edit?usp=sharing
Description of the Artifact
The report is a synthesis of research on makerspaces, Libraries of Things, and the library’s role as a community gathering place–three related expansions of public library service. It opens by establishing the audience explicitly, addressing LIS students and early-career professionals who are likely to be asked to plan, pitch, or evaluate exactly these kinds of initiatives. The stated scope is ten research studies and ten web resources. A glossary of key terms follows, defining eight concepts and attributing each to the source it came from. Four of the eight are research-methods terms rather than subject terms (cross-sectional study, ethnography, phenomenography, and systematic review).
The body is an annotated bibliography of ten peer-reviewed studies, each annotated in two parts. The Summary describes what the study set out to do and how it was conducted. It names the design, the sampling strategy, the number of participants, the instrument, and the analytics technique. The Application reports what was found, states who in the profession would find it useful and for what purpose, identifies the study’s limitations, and notes the authors’ recommendations for further research.
The ten studies were chosen to span the full methodological range. One is a multi-site ethnographic dissertation, built from participant observation, document analysis, and interviews at three libraries. Then it is analyzed using grounded theory. Another is a phenomenographic interview study of 22 tool-lending patrons. A third analyzed 23 interviews through reflexive inductive thematic analysis, coded in MAXQDA. A fourth used qualitative content analysis, starting from 800 initial codes, and refining them through iterative comparison until the data reached saturation. A fifth is a systematic review guided by the PRISMA protocol, which screened 838 records down to 43 studies. A sixth is a sequential explanatory mixed-method study. It paired a 605-respondent survey, analyzed with descriptive statistics and chi-square tests, against interviews and focus groups with 12 participants. A seventh surveyed 225 remote workers recruited through the Prolific platform, analyzing the closed questions in SPSS and the open responses thematically. The remaining studies are case studies. The most heavily triangulated of these combined 20 in-depth interviews, 14 staff-run focus groups, a 902-response community survey, planning observations, and internal planning documents.
The report closes with ten annotated web resources. These come from the Pew Research Center, the Brookings Institution, the Urban Libraries Council, IFLA, PLA, YALSA, ALA, WebJunction, Shareable, and Local Tools. Each carries a note on what the resource offers, and who would use it.
Justification and Connection to the Competency
This artifact demonstrates two parts of the competency at once. It shows most directly that I can evaluate and synthesize research literature. It shows nearly as strongly that I understand both qualitative and quantitative methods, and can tell when each is the right choice.
It demonstrates evaluation rather than summary, because the annotations judge the studies rather than restate them. Nearly every Application closes by naming what the study cannot support. I note that one case study’s interview sample skewed toward existing, favorable inclined library users, and that the authors themselves recommend future work centering non-users. I note that a major survey’s respondents skewed toward the already-engaged, white, and university-educated. I note that another study’s participants had uneven hands-on experience with the technology in question, and that data collection was constrained by reduced in-person contact during the pandemic. Identifying the population a study cannot speak for is precisely the judgment that separates evaluating research from believing it, and it is the judgment a practitioner needs before importing a finding into a different institution.
It demonstrates understanding of both methodological traditions, because I describe how each study was conducted rather than reporting only what it concluded. Doing this across ten studies required distinguishing ethnography from phenomenography, grounded theory from reflexive thematic analysis, and a cross-sectional survey from a systematic review. Then I went further to explain each in language a reader outside the field could follow. The glossary makes that translation explicit.
The report demonstrates synthesis in two ways. The first is the selection itself. Choosing ten studies that together span ethnographic immersion at one end and SPSS-analyzed survey data at the other is an argument about what the evidence base for this topic consists of. That argument is made by the shape of the bibliography rather than stated in a sentence. The second is the inclusion of Kim et al. (2022), a systematic review that coded 43 makerspace studies for their methodology, theoretical framework, setting, and participant type, and reported that qualitative single-data-source designs dominate the field, that staff and facilitators are studied far more often than patrons, that only 8 of the 43 used any explicit theoretical framework, and that research on rural and K-12 school makerspaces is nearly absent. Engaging with a study of this kind required me to reason not about individual findings, but about the condition of an evidence-base as a whole, including what it systematically fails to examine. That is the difference between reading the literature, and assessing it.
Finally, the report demonstrates that I distinguish kinds of sources by what they can be asked to do. The ten peer-reviewed studies and the ten web resources are annotated differently, and for different purposes. With the latter presented as practitioner tools and data sources rather than as evidence of the same standing. Grant and Booth (2009) make the underlying point at the level of reviews, which is that the authority of a synthesis depends on how it was assembled. The same discipline applied at the level of individual sources is what keeps a toolkit from being cited as though it were a finding.
Evidence 2
The second piece of evidence demonstrating my mastery of Competency L is my logic model for “Braille Literacy for the Next Century,” completed in INFO 285: Research Methods–Evaluating Programs and Services (Fall 2024). It was solo work.
Description of the Artifact
The artifact is a completed logic model for a real grant-funded program at the Braille Institute, budgeted at $38,175 across $25,355 in Library Services and Technology Act funds, and $12,820 in cash match and in-kind contributions. A logic model is a planning document that lays out a program’s theory of how it will work, mapping what goes in, what the program does, what it produces, and what changes as a result.
The model’s four columns run across resources, activities, outputs, and outcomes. Resources itemize funding, staffing, 42 Orbit Reader 20 devices, the Braille Institute’s technology centers, the digital platforms participants would use, and a recruitment partnership with Los Angeles City College. Activities are grouped into instruction, content acquisition, planning and evaluation, and procurement, each with a month range establishing sequence. Outputs are quantified by devices distributed, participants trained, workshops held across centers, marketing materials circulated to community organizations, and assessments conducted. Outcomes are divided into short-term and long-term. The short-term set is stated as changes in participants’ access to digital braille materials, technology literacy, knowledge of adaptive platforms, braille reading rates, and engagement. The long-term set extends to independence, institutional scalability, and statewide policy goals.
Below the main table, the model sets out ten assumptions on which the program’s success depends and a six-category analysis of external factors–covering technological, economic, social, political and legal, environmental, and institutional threats.
A reflection essay follows, in which I describe my prior work on visual impairment issues as VP of Foundation for my chapter of Delta Gamma, assess the model’s strengths, and identify two weaknesses. The first weakness is that the training curriculum treats participants as uniform in technological proficiency when they arrive, while in reality they could have widely varying comfort levels. The second weakness is that the sustainability plan relies on post-grant fundraising where a formal long-term agreement with the National Library Service or state libraries would be more durable.
Justification and Connection to the Competency
I selected this artifact because it demonstrates the design half of the competency, which the other two artifacts cannot show. A logic model’s real function is to convert a program’s implicit reasoning into explicit, checkable statements. Building one required me to say what this program assumed, what it would produce, and what would count as evidence that it had worked. That is research design in its practitioner form, and it is exactly the function Haley Goldman’s Evaluation Guide for Public LIbraries assigns to evaluation planning (i.e., deciding what change is expected before deciding how to look for it).
The clearest demonstration is the separation of outputs from outcomes. Outputs in this model are counts of what the program did–devices distributed, workshops held, materials circulated. Outcomes are statements about change in participants. Keeping these apart is the central discipline of outcome-based evaluation, because a program that reports only outputs has documented its own activity and said nothing about its effect. Writing the short-term outcomes in terms such as braille reading rates and engagement with library materials, rather than in unmeasurable terms such as awareness, was a deliberate choice to state them in a form an instrument could actually detect.
The assumptions list is where the model most clearly becomes a research instrument rather than a planning one. Each assumption identifies a condition the program’s logic depends on, and each therefore identifies something an evaluation would need to test. One of them states that the pre- and post-program assessments will effectively measure changes in participants’ skills, satisfaction, and quality of life. That is a claim about instrument validity, and articulating it as an assumption rather than treating it as given is what allows it to be examined. Others name dependencies that lie outside the program’s control, including participant willingness to engage, device reliability, and staff capacity to deliver training. The six-category external factors analysis extends this into the language of threats. These threats are technological obsolescence, supply chain disruption, funding cuts, language barriers, policy change, and staff turnover. These confounding conditions could produce a null result for reasons having nothing to do with whether the intervention works.
The reflections section demonstrates that I can evaluate a design rather than only produce one. Identifying that the training component assumes uniform technological proficiency is a critique of construct validity, since a program delivering identical instruction to participants at very different starting points will produce change that varies for reasons the model does not account for. Identifying the sustainability plan’s dependence on post-grant fundraising is a critique of the long-term outcomes’ plausibility, since those outcomes assume a continuation the resources column does not secure. Turning that same scrutiny on my own work, and documenting it rather than defending the design, is the same evaluative habit this competency asks me to bring to published research. Here I brought it to a design instead.
Evidence 3
The third piece of evidence demonstrating my mastery of Competency L is my analysis of pre- and post-assessment data from the Braille Institute’s Brailliant 14 pilot program, completed as Assignment 4 in INFO 285: Research Methods–Evaluating Programs and Services (Fall 2024). It was solo work. The dataset was supplied by my instructor from a real program. The analysis, the coding scheme, and every analytic decision in the workbook are my own.
Link: https://docs.google.com/spreadsheets/d/1u-m8qm8ltnEVqNp-rog5eY-LG8vdJzOaH5yOPbkNw8I/edit?usp=sharing
Description of the Artifact
The workbook analyzes paired pre- and post-assessment data from 24 participants who received refreshable braille displays–devices that render digital text as physical braille characters that reset line by line–through a pilot program operating across five Braille Institute centers in California.
The first sheet holds the survey instrument itself, showing the paired pre- and post-assessments, and the item types they combine. These are nominal items on how participants currently access braille and what device they would pair with the display, ratio items on books read per month and hours spent reading per day, 10-point and 5-point rating scales on confidence and satisfaction, and open-ended items on goals at the outset and whether those goals were met a year later. The second sheet holds the data set with direct identifiers removed and participants pseudonymized.
The analysis sheets that follow each take one variable. Two handle multiple response items where participants could name several access methods or several devices. For these I decomposed the combined responses into individual mentions, and recalculated the denominator to the number of mentions rather than the number of participants–producing bases of 43 and 40 against a sample of 24
One sheet handles the paired confidence ratings for using a refreshable braille device. It reports full descriptive statistics (i.e., N, mean, median, mode, minimum, maximum, range, and standard deviation) for both waves. Then builds grouped frequency distributions for each with frequency, percent, cumulative frequency, and cumulative percent across five intervals charted as paired bar graphs.
The final two sheets handle the open-ended items. For the pre-assessment goals, I read all 24 responses and coded them inductively into thematic categories including braille literacy, technology skills, access to resources, education, teaching others, independence, and employment. Then I converted the coded results into a frequency distribution and chart. For the post-assessment question on whether goals were met I built a two-part scheme in which a numeral records the outcome (i.e., met, partially met, or not met), and a letter records the reason, with responses carrying more than one reason coded accordingly.
Justification and Connection to the Competency
This artifact demonstrates two parts of the competency together. It shows that I can understand both qualitative and quantitative methods, and it shows that I can apply appropriate ones inside a real work setting. It does both by bringing the two traditions to bear on a single data set, and then integrating them.
The quantitative work is straightforward descriptive analysis executed correctly, and two decisions in particular demonstrate understanding rather than mechanical application. The first concerns the multiple response items when a respondent names three access methods, the number of responses exceeds the number of respondents, and percentages calculated against the sample size will not add up meaningfully. Recalculating the base to the count of mentions is the correct handling, which is easy to get wrong in a way that inflates every figure on the table. The second concerns attrition. 24 participants completed the pre-assessment and 19 completed the post-assessment. The workbook reports each wave against its own N rather than carrying the large figure through. Refusing to conceal non-response is a small decision that determines whether the analysis can be trusted at all, and it reflects exactly the condition I described earlier as ordinary in this environment. In voluntary programs, people leave, and an analysis that hides it is reporting a sample that does not exist.
The qualitative work is where the artifact demonstrates the most. Coding 24 open-ended goal statements into thematic categories is conventional content analysis in the sense Hsieh and Shannon (2005) define. This is evidenced by how the categories were derived from what participants actually wrote rather than imposed from an existing framework. That choice was appropriate here, because no prior scheme existed for what adults learning a new assistive technology hoped to gain. Imposing one would have risked seeing only the goals the researchers expected.
The coding scheme I built for the post-assessment question goes further. A single question asked whether participants met their goals, and why or why not. That means every response contained two distinct pieces of information–an outcome and an explanation. Rather than flatten these into one code, I built a scheme in which numerals capture the outcome and the letters capture the reason. That way our response can carry a partial outcome alongside more than one cause. For example, a participant who partially meets their goals for two separate reasons receives a code recording all three facts. Constructing this required recognizing that forcing responses into a single category would have destroyed the information that made the response worth collecting. Which is a judgment about the relationship between a coding scheme and the data it is meant to represent, rather than a procedural step.
Having coded the qualitative responses, I then quantified them into frequency distributions and charts. This is the move Hsieh and Shannon (2005) identify as summative, in which counted occurrences become the basis for interpreting patterns in the underlying content. It is also what makes this analysis genuinely mixed in the sense Hayman and Smith (2020) require. The strands are not merely adjacent. The rating-scale data established the direction and spread of change and participants' confidence, while the coded explanations account for it. This shows that a substantial share of participants attributed unmet goals to a device recall, to training that had not yet occurred, or to time restraints rather than to any failure of the technology itself. Neither strand supports that conclusion alone. The numbers alone would show a program that underperformed. The explanations alone would be 24 anecdotes. However, together they identify a specific, actionable cause, which is what integration is for.
Conclusion
Taken together, these three artifacts trace a complete research process rather than three unconnected skills. The INFO 287 report establishes what is already known, and how reliably it is known. It appraises 10 studies across the full methodological range, and assesses the condition of the evidence base as a whole. The INFO 285 logic model performs the design work that proceeds data, converting a program's implicit reasoning to explicit assumptions, measurable outcomes, and named threats. Then subjects that design to critique. The analysis executes on data that has been collected by applying descriptive statistics and inductive coding to the same participants, and integrating the two so that each explains what the other cannot. What is evident across all three, is the discipline of stating how a conclusion was reached, and of being explicit about what the evidence does not support.
In my future career I expect to apply this competency in archival, library, and cultural heritage settings. Nearly every significant decision in those environments rests on a claim about users that could, in principle, be tested. Grant-funded work in particular will require me to specify intended outcomes before our program begins, to build instruments capable of detecting them, and to report results honestly when they fall short. Just as often, I will be the consumer rather than the producer of research, reading a study and having to judge whether its sample resembles the community I serve. That judgment is not a specialist skill reserved for researchers, but the difference between adopting a practice because the evidence supports it or adopting it because someone published it.
To remain current I will follow Evidence Based Library and Information Practice, the open-access peer-reviewed journal published by the University of Alberta. That journal is the principal form for methodological work in this area, and the source of two of the reading sightings here. I will read Library & Information Science Research, which publishes empirical LIS studies along methodological reviews, and is where several of the studies in my own synthesis appeared. I will consult the American Library Association's Office for Research and Statistics, which produces the profession's national data and survey instruments, and offers practitioners guidance on collecting their own. Finally, I will follow the research and evaluation work of the expectations that govern LSTA-funded programs of exactly the kind analyzed in two of these artifacts. Continued engagement with these sources will keep both my practice and my judgment aligned with current standards as methods and expectations continue to develop.
Note on the Use of Generative Artificial Intelligence
Artificial intelligence (Anthropic, 2026) was used in preparing this essay to help identify relevant scholarly literature, to suggest the essay’s organizational structure, and to check grammar and APA formatting. All writing, evidence selection, and analysis are my own.
References
Anthropic. (2026). Claude (Opus 5) [Large language model]. https://claude.ai
Connaway, L. S., & Radford, M. L. (2018). Research methods in library and information science (7th ed.). Libraries Unlimited. https://doi-org.libaccess.sjlibrary.org/10.1086/697711
Grant, M. J., & Booth, A. (2009). A typology of reviews: an analysis of 14 review types and associated methodologies. Health Information and Libraries Journal, 26(2), 91–108. https://doi.org/10.1111/j.1471-1842.2009.00848.x
Haley Goldman, K. (n.d.). Evaluation guide for public libraries. Urban Libraries Council. https://www.urbanlibraries.org/files/KHG-Evaluation-Guide.pdf
Hayman, R. & Smith, E. (2020). Mixed Methods Research in Library and Information Science: A Methodological Review. Evidence Based Library and Information Practice, 15(1), 106–125. https://doi.org/10.18438/eblip29648
Hsieh, H.-F., & Shannon, S. E. (2005). Three Approaches to Qualitative Content Analysis. Qualitative Health Research, 15(9), 1277–1288. https://doi.org/10.1177/1049732305276687
Kim, S. H., Jung, Y. J., & Choi, G. W. (2022). A systematic review of library makerspaces research. Library & Information Science Research, 44(4), Article 101202. https://doi.org/10.1016/j.lisr.2022.101202
Price, C. (2022). [Rev. of Syntheses Synthesized: A Look Back at Grant and Booth’s Review Typology]. Evidence Based Library and Information Practice, 17(2), 132–138. https://doi.org/10.18438/eblip30093
Barbara M. Wildemuth. (2016). Applications of Social Research Methods to Questions in Information and Library Science (Barbara M. Wildemuth, Ed.). Bloomsbury USA (Minor Textbooks).
