Review Article

A Rapid Review of the Reporting and Characteristics of Instruments Measuring Satisfaction with Reference Service in Academic Libraries

Heidi Senior

Reference/Instruction Librarian

Clark Library

University of Portland

Portland, Oregon, United States of America

Email: [email protected]

Tori Ward 

MLIS Graduate 

Syracuse University 

Syracuse, New York, United States of America 

Email: [email protected]

Received: 13 Feb. 2019 Accepted: 6 Nov. 2019

2019 Senior and Ward. This is an Open Access article distributed under the terms of the Creative Commons‐Attribution‐Noncommercial‐Share Alike License 4.0 International (http://creativecommons.org/licenses/by-nc-sa/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly attributed, not used for commercial purposes, and, if transformed, the resulting work is redistributed under the same or similar license to this one.

DOI: 10.18438/eblip29556

Abstract

Objective – The objective of this review was to examine research instrument characteristics, and to examine the validity and reliability of research instruments developed by practicing librarians, which measure the construct of patron satisfaction with academic library reference services. The authors were also interested in the extent to which instruments could be reused.

Methods – Authors searched three major library and information science databases: Library and Information Science Technology Abstracts (LISTA); Library Science Database (LD); and Library Literature & Information Science Index. Other databases searched were Current Nursing and Allied Health Literature (CINAHL); Education Resources Information Center (ERIC); Google Scholar; PubMed; and Web of Science. The authors identified studies of patron satisfaction with academic library reference services in which the researcher(s) developed an instrument to study the satisfaction construct. In this rapid-review study, the studies were from 2015 and 2016 only. All retrieved studies were examined for evidence of validity and reliability as primary indicators of instrument quality, and data was extracted for country of study, research design, mode of reference service, data collection method, types of questions, number of items related to satisfaction, and content of items representing the satisfaction construct. Instrument reusability was also determined.

Results – At the end of the screening stage of the review, a total of 29 instruments were examined. Nearly all studies were quantitative or mixed quantitative/qualitative in design. Twenty-six (90%) of the studies employed surveys alone to gather data. Twelve publications (41%) included a discussion of any type of validity; five (17%) included discussion of any type of reliability. Three articles (10%) demonstrated more than one type of validity evidence. Nine articles (31%) included the instrument in full in an appendix, and eight instruments (28%) were not appended but were described adequately so as to be reusable.

Conclusions – This review identified a range of quality in librarians’ research instruments for evaluating satisfaction with reference services. We encourage librarians to perform similar reviews to locate the highest-quality instrument on which to model their own, thereby increasing the rigor of Library and Information Science (LIS) research in general. This study shows that even a two-year rapid review is sufficient to locate a large quantity of research instruments to assist librarians in developing instruments.

Introduction

Reference services are a primary function of nearly every library. Library staff make themselves available to patrons through multiple communication modes such as in-person, chat, phone, and email in order to “recommend, interpret, evaluate, and/or use information resources to help others to meet particular information needs” (Reference and User Services Association, 2008). They might gather statistics relating to the number and type of questions patrons ask, and perhaps the difficulty of answering those questions according to the READ Scale (Gerlich & Berard, 2007), but these statistics do not express whether patrons were satisfied with the answer. To determine if their library’s patrons are satisfied with the provided service, librarians need to obtain patrons’ opinions directly through data gathering methods such as surveys or

interviews, known collectively as research tools or instruments. They might then publish the results of their study to help fellow librarians develop their own patron-satisfaction tools. One study found that reference topics represented 9.5% of all library and information sciences research (Koufogiannakis, Slater, & Crumley, 2004).

Systematic instrument review has been a common practice in health science research and has developed to the extent that standards exist for specific topic areas, such as the Consensus-based Standards for the Selection of Health Measurement Instruments (COSMIN) Initiative (2018). This type of study uses systematic review methodology to identify and analyze the psychometric characteristics of research instruments. While anthologies of research instruments produced by librarians and measuring satisfaction with reference services exist, such as those found in The Reference Assessment Manual (American Library Association, Evaluation of Reference and Adult Services Committee, 1995, pp. 255-345), to date it seems that no one has published a systematic instrument review that would obtain an overall image of the state of instrument development in this area. We therefore decided to conduct a review to gain an understanding of the quality of instruments produced by academic librarians studying patron satisfaction with reference service.

Literature Review

While at the time of our study no reviews of instruments fully using the systematic review methodology had appeared in LIS literature, we found that researchers had mentioned instruments and evaluated them to varying extents in articles on faculty attitudes toward open access publication (Otto, 2016); information literacy (Beile, 2008; Schilling & Applegate, 2012); information seeking behavior (McKechnie, Chabot, Dalmer, Julien, & Mabbott, 2016); satisfaction with chat reference (Lasda Bergman & Holden, 2010); and assessment of individual research consultations (Fournier & Sikora, 2015).

Of these researchers, only Lasda Bergman & Holden (2010) followed a systematic review protocol in their research criteria and search methods, retaining after their final appraisal stage 12 studies regarding user satisfaction with electronic reference. However, because they did not present details of each instrument in an evidence table, we were unable to reproduce their data extraction process. Schilling and Applegate (2012) identified 27 tools in their survey of academic library literature on student learning assessment from 2007 to 2012 but did not take a systematic approach and emphasized each instrument’s content rather than construction and measurement concerns. Similarly, Fournier and Sikora’s 2015 scoping review located 20 studies using various methods to assess individual research consultations but did not review instrument characteristics beyond the type of assessment method. Beile’s (2008) report covered widely-known information literacy assessment tools that would provide data “considered acceptable evidence for program reviews” (p. 1) such as Standardized Assessment of Information Literacy Skills (SAILS) and Educational Testing Service’s iSkills, but did not describe a process for identifying the seven tests and four rubrics included in the paper. McKechnie et al.’s approach to evaluating research rigor involved using a checklist that asked whether authors attached or included their instrument – an element we included in our study – and whether the instrument had undergone pre-testing, an important component in demonstrating an instrument’s validity (2016). While conducting a literature review prior to studying effective faculty outreach messages regarding open access publication, Otto (2016) realized that the studies reviewed did not accurately reflect faculty understanding due to flaws in their underlying instruments such as failing to define terms, adapting previous surveys without updating questions, and inserting inadvertent bias into survey questions and response options. Although Otto did not report evidence of the instruments’ validity and reliability specifically, several of the issues Otto identified might have been resolved had the instruments’ developers paid closer attention to determining their validity.

Shortly before completing our manuscript, we learned of the publication late in 2017 of a systematic review of 22 self-efficacy scales assessing students’ information literacy skills (Mahmood) from 45 studies published between 1994 and 2015. Because Mahmood’s review was limited to studies in which authors reported the use of any validity and also any reliability indicators, it differs from our relatively unrestricted approach. Mahmood’s study likely omits scales and does not provide a full picture of the state of instrument development in this area.

We identified two instrument reviews from the field of education (Gotch & French, 2014; Siddiq, Hatlevik, Olsen, Throndsen, & Scherer, 2016), the second of which served as a preliminary model for the data extraction stage of our pilot study. Gotch & French (2014) reviewed 36 measures published between 1991 and 2012 of classroom teachers’ assessment literacy, using “evaluation of the content of assessment literacy measures beyond literature review and solicitation of feedback,” “internal consistency reliability,” and “internal structure” to demonstrate validity, and “score stability” (p. 15) to demonstrate reliability of each instrument. We decided not to use Gotch and French as a model because the authors did not rigorously follow a systematic approach in database searching or in presenting their results in evidence tables. The second systematic instrument review (Siddiq, et al., 2016) covered 38 information and communication technology instruments aimed at primary and secondary school students, and was a useful framework to emulate because, like our study, it was descriptive rather than evaluative in design. Furthermore, the authors carefully documented their search strategy and findings in a way that adhered closely to systematic review methodology. Like our study, the authors appeared to be concerned to represent the state of the field and included instruments whose developers did not address evidence of their validity or reliability.

Our review of librarians’ studies examining instruments determined that the instrument review methodology is under-used in librarianship, and that our pilot study identifies a new area of research. By drawing on similar reviews in education, we demonstrate the usefulness to librarian-researchers of breaking out of disciplinary compartmentalization for assistance with promising methodologies.

Research Questions

We began this study with a basic question:

What is the quality of research instruments produced by librarians? We developed the following more specific questions using patron satisfaction with reference services in academic libraries as a focus. We defined reference service as librarians helping others to meet particular information needs in-person at a desk, roaming, or via consultations; through virtual methods such as chat and email; or over a telephone.

Q1: How did LIS researchers gather data on patron satisfaction with academic library reference services in the years 2015-2016?

Q2: To what extent did the instrument developers document the validity and reliability of their instruments?

Q3: To what extent are the instruments provided in an appendix or described in the publication, to assist in reuse?

Method

Selection of Review Type

The systematic review is considered the most rigorous methodology for gathering and synthesizing information based on predetermined inclusion/exclusion criteria, clear and reproducible search methods, and quality assessment, with results presented in an evidence table (Xu, Kang, & Song, 2015; see also McKibbon, 2006; Phelps & Campbell, 2012). Traditional systematic reviews, however, aim to be comprehensive in coverage. Because the lead author wanted to accomplish as much work as possible during a sabbatical, we elected to perform a rapid review, which follows the systematic review methodology (predetermined inclusion/exclusion criteria, clear and reproducible search methods, and quality assessment, with results presented in an evidence table) but is limited in time (Grant & Booth, 2009, p. 100). See Table 1 for distinctions between systematic and rapid reviews.

Table 1

Differences Between Systematic and Rapid Review Types

Review Type	Description	Search	Appraisal	Synthesis	Analysis
Rapid Review	Assessment of what is already known about a policy or practice issue, by using systematic review methods to search and critically appraise existing research	Completeness of searching determined by time constraints	Time-limited formal quality assessment	Typically, narrative and tabular	Quantities of literature and overall quality/direction of effect of literature
Systematic Review	Seeks to systematically search for, appraise and synthesis research evidence, often adhering to guidelines on the conduct of a review	Aims for exhaustive, comprehensive searching	Quality assessment may determine inclusion/exclusion	Typically, narrative with tabular accompaniment	What is known; recommendations for practice. What remains unknown; uncertainty around findings, recommendations for future research

Grant & Booth, 2009, p. 95.

Inclusion Criteria

We assembled and agreed upon the following criteria:

• Quantitative, qualitative, or mixed-method research studies measuring satisfaction with reference service carried out in any type of academic library including health science libraries, addressing any type of patron. We included studies measuring satisfaction with several library services, as long as one question asked about reference service.
• Studies published in 2015 or 2016
• Instruments developed by front-line librarians, including adaptations of a standardized instrument such as SERVQUAL or SERVPERF
• English, French, or Spanish language

Search Strategy

EBSCO’s Library and Information Science Technology Abstracts (LISTA) and ProQuest’s Library Science Database (LD) were the primary sources of studies; we also searched Library Literature & Information Science Index, CINAHL, ERIC, Google Scholar, PubMed, and Web of Science. In addition to these databases, we searched the American Library Association, the Association of College and Research Libraries, and assessment conference programs that were published online for the years of interest.

When developing our search strategies, we kept in mind the caveat raised by LIS authors that database thesauri might be incomplete or that subject headings might not be applied uniformly. VanScoy and Fontana noted in their 2016 study of reference and information service (RIS) research that “This method relies on the RIS research articles being correctly assigned the relevant descriptor in the databases” (p. 96). This warning echoes that of McKechnie, Baker, Greenwood, & Julien in 2002 who said “Both [EBSCO and ProQuest] indexes used terms … that were too general to be useful” (p. 123) and found that indexing terms were incorrectly applied in 28-34% of the articles they examined, as well as Greifeneder who warned in 2014 that one of the studies in her literature review might have a biased retrieval set because it included articles indexed under only two subject terms rather than searching more widely (Background, para. 8). We therefore decided to run both subject and keyword searches.

After a careful examination of subject terms used in either LISTA or LD, and heeding past research on effective search strategy, we performed the following searches:

LISTA: (academic AND librar* AND (reference OR "user satisfaction")) AND (SU (research or surveys or questionnaires) OR AB (study or survey* or interview* or research*))  

LD: all ((academic* AND librar* AND (reference OR "user satisfaction" OR "customer satisfaction" OR "customer services"))) AND su(research or surveys or questionnaires) AND ab(study or survey* or interview* or research*). Note that the LD search is identical to the LISTA search except for the inclusion of “customer satisfaction” and “customer services,” which are subject terms not used in the LISTA database.

Given the inconsistent application of subject terms in library literature databases, we note that articles given the subject term “academic libraries” might not describe undergraduate or community college libraries. However, we found no additional articles when we re-ran searches with the subject terms “community college libraries” and “undergraduate libraries.”

We then examined abstracts and developed a free-text keyword search that we adapted for use in all of the databases, in an attempt to find all articles that might not have had correct subject-term labels: (reference or "research consultation") AND (satisf* or evaluat* or assess* or improve*) AND (experiment* or survey* or qualitative or servqual or instrument or investigat* or analysis or questionnaire*) AND "academic librar*". We also ran a broad search for librar* AND reference AND satisfaction, being mindful that the “academic libraries” label might not be uniformly applied and that some articles might use “college” or “university” instead, or that various labels might be used to represent different categories of library patron, or different types of data-gathering instruments. We removed search terms related to research methodologies to have broad retrieval.

Conforming to our inclusion criteria, we limited results in each database to journal articles from the years 2015 and 2016 and checked each database for conference papers as a separate source type. We did not apply language or geographic location limiters and were prepared to examine articles in French or Spanish in addition to English, but our searches retrieved only English-language publications. In preparation for retrieving a large amount of results, such as within Google Scholar, we determined that we would review the first 300 items only. Within those 300 results, we ceased reviewing when we began encountering irrelevant items.

When searching PubMed, we applied a search filter provided by COSMIN in order to better identify all studies containing measurement properties. In Google Scholar, we utilized the Advanced Search feature to narrow our results. In ERIC and CINAHL, we utilized the database thesauri to identify subject terms. We also hand searched 10 online journals (Journal of Academic Librarianship; College & Research Libraries; Library & Information Science Research; portal; Journal of the Medical Library Association; Journal of Librarianship and Information Science; Reference Services Review; Medical Reference Services Quarterly; Reference Librarian; and College and Undergraduate Libraries), adhering to our year restriction of 2015-2016. To standardize our searches, we created a table in which both authors’ search strings were input to compare and ensure that we were staying consistent with our searches and results.

All search strategies are provided in Appendix A.

Reviewing Process and Study Evaluation

We compiled citations in a RefWorks database and removed duplicates using the RefWorks tool. We examined bibliographic information from the databases, such as title and abstract, to screen for relevant articles. To add a peer reviewing element to our searches, we kept track of our subsequent searches on a separate workbook so that each author could observe and be able to discuss the quality of each search with the other. In those workbooks, we documented the search conducted, the database in which the search was conducted, the limiters set in each search, the results of each search, the citations found from each search (if applicable), and any notes.

Data Extraction

As stated earlier, we used Siddiq et al.’s (2016) extraction sheet as a model because we aimed to be descriptive rather than evaluative in scope. Following their model, we extracted the following data: country of study; stated purpose of study; mode of reference service; age/level of students (if students were part of the targeted population); size of targeted population; usable responses; sampling strategy; research design; data collection method; types of questions, other than demographic (i.e., Likert scale format; presence of open-ended questions); demographics gathered; technical aspects, such as distribution, availability of translations, duration of survey period; time allotted to complete survey; validity indicators; reliability indicators (see below for definitions of validity and reliability); instrument availability in appendix; reusability of instrument, if not appended; number of items related to satisfaction; content of items representing the satisfaction construct.

Definitions

Generally speaking, an instrument is said to provide valid results when it measures what the instrument’s developer intended it to measure within a study’s setting and population, and reliable results when the instrument provides the same score if repeatedly implemented among the same population. Researchers have further specified various elements that assist in demonstrating the validity of information obtained via an instrument. We used these definitions when examining the instruments we gathered. Tables 2 and 3 include the codes we assigned to each element, to make our reporting table more compact.

Table 2

Definitions of Validity

Title	Code	Definition	Evidence
Face Validity	V1	“The instrument appears to measure what it claims to measure” (Gay, Mills, & Airasian, 2006 as quoted in Connaway & Radford, 2017, p. 82)	Demonstrated through pre-testing, ideally with subjects similar to the target population, and with instrument development specialists
Content Validity: Item	V2a	“…the items of the instrument or test …represent measurement in the intended content area” (Connaway & Radford, 2017, p. 81).	Demonstrated through item analysis during pre-testing
Content Validity: Sampling	V2b	“…how well the instrument samples the total content area” (Connaway & Radford, 2017, pp. 81-82)	Demonstrated through discussion of included constructs
Construct Validity	V3	“…instrument measures the construct in question and no other.” (Connaway & Radford, 2017, p. 83)	Demonstrated through factor analysis, other tests of dimensionality, to retain convergent (contributing) items and remove divergent (non-contributing) ones
Intercoder Reliability	V4	Degree to which scorers or raters agree on evaluating or observing a variable (Connaway & Radford, 2017, p. 316).	Demonstrated through percentage agreement among raters

We used these definitions of reliability, and assigned these codes:

Table 3

Definitions of Reliability

Title	Code	Definition	Evidence
Internal Consistency	R1	How well items on a test relate to each other. (Connaway & Radford, 2017, p. 84)	Demonstrated through Cronbach’s alpha, Kuder-Richardson 20 tests (Catalano, 2016, p. 8).
Measurement Reliability	R2	“The degree to which an instrument accurately and consistently measures whatever it measures” (Connaway & Radford, 2017, p. 83).	Demonstrated through test-retest correlation, meaning repeated administration to same group of the whole instrument (Catalano, 2016, p. 8) or split-half method, meaning correlation of scores obtained from each half of a tested population or from each half of an instrument that measures a single construct (Catalano, p. 9; Connaway & Radford, pp. 83-84).

We described an instrument as “resusable” only when we could answer three questions about the instrument: Are the number of questions reported? Is the full text of each question provided, and associated items? Is the format of each question described: scale values, anchor labels such as “Very Satisfied” and “Very Unsatisfied”? If we felt that we had to guess as to whether the author fully described an instrument, we labeled it not replicable. We automatically coded appended instruments as replicable. The most frequent reason for describing an instrument as “not replicable” was that authors did not supply the number of questions and items, so that we could not be sure if they had described the entire instrument.

As explained earlier, we restricted the definitions of validity and reliability to those used by Connaway and Radford (2017), with occasional details borrowed from Catalano (2016). We decided not to use the more expansive definitions that Siddiq et al. (2016) employed in which for example an instrument’s having a basis in theory could be perceived as evidence that it produced valid results. We developed our own evidence extraction sheet to avoid obscuring the definition of each of these concepts: we wanted to focus on precise definitions of validity and reliability, and Siddiq et al.’s criteria extended beyond those definitions.

After completing the process of acquiring and screening studies, we jointly read and reviewed six studies in duplicate and compared our extracted data, to ensure that we agreed. We then separately reviewed the remaining 23 studies and consulted with each other on any confusing elements. After our subsequent searches we divided responsibility similarly to review the nine additional articles. When we disagreed, we located more information on the issue to arrive at a consensus. For example, a disagreement about validity types might require refreshing our understanding of the definitions. We did not require a third party to resolve disagreements. We recorded our data in a shared spreadsheet.

Figure 1

Flow chart of review process

Results

Through our initial searches in seven databases, we found 2,189 articles that appeared to be relevant to our study. After removing duplicates, we were left with 2,108 relevant articles. We further reviewed the article titles and abstracts and found that 1,770 were not truly relevant to our study. We assessed the remaining 338 articles for eligibility and rejected 309 articles because they described instruments measuring satisfaction with only the library as a whole, or instruments measuring usage of or familiarity with reference services, or instruments measuring satisfaction with services other than reference. We extracted data from the final set of 29 studies. Figure 1 shows a PRISMA flow chart of our process, a standard component of systematic review articles (Moher, Liberati, Tetzlaff, & Altman, 2009).

We concluded during the extraction process that certain criteria were more relevant to our focus on the instrument development process and therefore decided not to report irrelevant criteria such as “purpose of study,” “age of respondents,” “mode of service,” and “response rate.” We excluded “time needed to complete the instrument” from this paper because many authors did not report it. Additional criteria included article title, journal, age/level of student, type of institution, sampling strategy, technical aspects (e.g., distribution and survey period), and demographics gathered. Although we gathered data in these categories, we found that these criteria did not further our understanding of how librarian researchers report on instrument development and implementation. Our evidence tables focus on the following criteria: study; country; research design; data collection method; types of questions; validity evidence; reliability indicators; whether the instrument is appended; whether the instrument is mentioned in the abstract as appended; and replicability of the instrument if not appended. The list of extraction elements is in Appendix B, and the full data extraction spreadsheet is available online (https://docs.google.com/spreadsheets/d/1M7MYhNcKbscrak9CqGpBg0X6z-d6V5qs-IwIE4I92SQ/edit#gid=0).

Outside of the data relating to our research questions, the data on country of study might be of interest to researchers. Ten studies took place in the United States of America; the second most common country was India with six studies. The remaining studies took place in China (2), Ghana (1), Jamaica (1), Malaysia (2), Nigeria (4), Philippines (1), and Taiwan (2).

These are the results of our data extraction as they relate to our research questions:

Q1: How did LIS researchers gather data on patron satisfaction with reference services in the years 2015-2016?

Of the 29 studies we gathered, 18 (62%) were solely quantitative in design and one (3%) solely qualitative. We labeled ten (34%) studies as combining both quantitative and qualitative designs, but this was usually because we defined “mixed methods” broadly to allow open-ended questions to be called qualitative; only two studies (7%) (Jacoby, Ward, Avery, & Marcyk, 2016; Verma & Parang, 2015) were truly mixed using the more conservative approach as defined by Fidel (2008) in which qualitative and quantitative methods were used to answer the same research question. Askew (2015) and Yap and Cajes (2016) employed quantitative methods to ask students about their satisfaction with roaming reference service and qualitative methods to ask librarians about their experience with providing the service; these studies are therefore labeled “quantitative” for the purposes of this review.

Twenty-six (90%) of the studies employed surveys alone to gather data. Construction of the surveys varied; the number of items related to satisfaction ranged from 1 to 16. Eight (29%) of the studies asked only about overall satisfaction with reference service, while another 10 (34%) included an item about overall service

satisfaction as well as other attributes contributing to satisfaction. Respondents were asked to consider aspects of librarian behaviour such as approachability and responsiveness, helpfulness, respect for confidentiality, and offering referrals; and aspects of librarian performance such as ability, accuracy, knowledge, and inspiring confidence. Five instruments (Blake et al., 2016; Butler & Byrd, 2016; Huang, Pu, Chen, & Chiu, 2015; Jacoby et al., 2016; Luo & Buer, 2015) asked students to gauge their likeliness to use, re-use, or recommend the service. Masrek and Gaskin reported presenting respondents with 10 items about Service quality, usefulness, and satisfaction (p. 42) but unfortunately did not provide the full text of the items within their article, making it impossible to determine how they conceptualized these elements of satisfaction.

Researchers commonly used 5-point Likert scales to quantify respondents’ agreement or disagreement with statements; this type of scale occurred in 13 (45%) of the 29 survey instruments. Three (10%) of these scales did not have the traditional neutral midpoint. Two (7%) of the scales offered three positive scores versus two negative (Sivagnanam & Esmail, 2015; Xie & Sun, 2015); the third scale was recoded by its developers to have three negative scores and two positive (Yan et al., 2015). Similarly, researchers employing 3-point scales did not always include a midpoint; three studies (Butler & Byrd, 2016; Ekere, Omekwu, & Nwoha, 2016; Yap & Cajes, 2016) offered two positive options and one negative. Two 4-point scales (Khobragade & Lihitkar, 2016; Yap & Cahes, 2016) were likewise unbalanced, with three positive and one negative choices. Duan (2016) used 4-point scales to measure satisfaction with different modes of reference service and 6-point scales to measure satisfaction with reference librarians’ behavior. The remaining scales ranged in size from two scale points to nine.

Most authors used typical labels for scale points, e.g., variations on “Very Satisfied,” “Somewhat Satisfied,” “Satisfied,” and “Very Dissatisfied,” “Somewhat Dissatisfied,” and “Dissatisfied,” or related labels such as “Useful,” and “Adequate,” but some authors labeled scale points differently from these norms. Duan (2016) provided explanatory text for each scale point, e.g., “Unsatisfied, because they solved few of my problems but were not willing to help me again” (p. 164). Sivagnanam and Esmail (2015) labeled their scale points “Not Satisfied,” “Not Much Satisfied,” “Particularly Satisfied,” “Fairly Satisfied,” “Absolutely Satisfied.” Yan, et al. (2015) were not clear, as it seemed they gave their scale two midpoint labels, “Neutral” and “Not Familiar.” Most 5-point scales had a midpoint labeled “Neutral” (Askew, 2015; Blevins, et al., 2016; Boyce, 2015; Huang et al, 2015; Mohindra & Kumar, 2015) or “Neither Agree nor Disagree” (Jacoby et al., 2016; Masrek & Gaskin, 2016). Three authors did not report the label used for their midpoint (Chen, 2016; Ganaie, 2016; Swoger & Hoffman, 2015).

A list of studies with their associated research design, data collection method, and Likert-scale type is available in Table 4.

Table 4

Studies Included in This Review

Study	Research Design	Data Collection Method	Types of Questions
Akor & Alhassan, 2015	Quant.	Survey	4-point scale
Askew, 2015	Quant.	Survey	5-point scale
Blake et al., 2016	Mixed	Survey	4-point scale, open-ended
Blevins, DeBerg, & Kiscaden, 2016	Mixed	Survey	5-point scale, open-ended
Boyce, 2015	Mixed	Survey	Choose from list, 5-point scale, yes/no, open-ended
Butler & Byrd, 2016	Mixed	Survey	3-point scale, open-ended
Chen, 2016	Quant.	Survey (based on SERVQUAL)	5-point scale
Dahan, Taib, Zainudin, & Ismail, 2016	Quant.	Survey (based on LIBQUAL)	9-point scale
Duan, 2016	Quant.	Survey	6-point and 4-point scales
Ekere, Omekwu, & Nwoha, 2016	Quant.	Survey	3-point scale
Ganaie, 2016	Quant.	Survey	5-point scale
Huang, Pu, Chen, & Chiu, 2015	Quant.	Survey	5-point scale
Ikolo, 2015	Quant.	Survey	2-point scale
Jacoby, Ward, Avery, & Marcyk, 2016	Mixed	Survey, Focus Groups, Interviews	5-point scale; open-ended
Khobragade & Lihitkar, 2016	Quant.	Survey	4-point scale
Kloda & Moore, 2016	Mixed	Survey	3-point scale, open-ended
Luo & Buer, 2015	Mixed	Survey	5-point scale, open-ended
Masrek & Gaskin, 2016	Quant.	Survey	5-point scale
Mohindra & Kumar, 2015	Quant.	Survey	5-point scale
Nicholas et al., 2015	Mixed	Survey	Choose from list, open-ended
Sivagnanam & Esmail, 2015	Quant.	Survey	5-point scale
Swoger & Hoffman, 2015	Mixed	Survey	5-point scale, open-ended
Tiemo & Ateboh, 2016	Quant.	Survey	4-point scale
Verma & Laltlanmawii, 2016	Quant.	Survey	3-point scale
Verma & Parang, 2015	Mixed	Surveys, Interviews	3-point scales
Watts & Mahfood, 2015	Qual.	Focus Groups	Open-ended
Xie & Sun, 2015	Quant.	Survey	5-point scale
Yan, Hu, & Hu, 2015	Quant.	Survey	5-point scale
Yap & Cajes, 2016	Quant.	Survey	3-point and 4-point scales; another scale not specified

Q2: To what extent are the instruments documented or included in the body of a publication?

Nine articles (31%) included the instrument in full in an appendix, and of the remaining studies we found that eight instruments (28% of the total) were replicable according to our criteria as described earlier. Detailed information is provided in Table 5.

We noticed that two of the instruments (Duan, 2016; Xie & Sun, 2015) were translated into Chinese as well as English; both versions were available to respondents, but the author described the English-language instrument within the publication. We were unable to determine if any differences might exist between the two versions.

Q3: To what extent are the instruments’ reliability and validity documented?

Twelve publications (41%) included a discussion of any type of validity; five (17%) included discussion of any type of reliability. Three articles (10%) demonstrated more than one type of validity evidence. See Table 6 for a complete list.

Validity Evidence

Face Validity

Face validity was the most common type of validity represented, as nine authors (31%) had pre-tested their instruments; however, we found that in two cases (7%) (Akor & Alhassan, 2015; Blevins, DeBerg, & Kiscaden, 2016) only librarian colleagues participated rather than members of the target population or instrument development specialists. The pre-testing process with potential respondents varied; Blake et al. (2016) held campus interview sessions, while Butler and Byrd (2016) informally polled library student employees. Kloda and Moore (2016) and three sets of researchers (Askew, 2015; Huang et al., 2015; Masrek & Gaskin, 2016) presented instrument drafts to members of their respondent population. Only three studies (10%) specifically reported pre-testing with a population contrasted with librarians and therefore presumably instrument development specialists: Chen (2016) met with “academic experts” (p. 319); Blake et al. worked with “experts from the university’s Educational Innovation Institute” (p. 227), and Masrek and Gaskin (2016) pre-tested their instrument with “experts in the faculty” (p. 42).

Table 5

Reusability of Instruments Within Studies

Study	Instrument Appended	Reusability of Instrument
Akor & Alhassan, 2015	No	No
Askew, 2015	No	Yes
Blake et al., 2016	Yes (Online)	Yes
Blevins, DeBerg, & Kiscaden, 2016	Yes	Yes
Boyce, 2015	Yes	Yes
Butler & Byrd, 2016	Yes (Online)	Yes
Chen, 2016	No	Yes
Dahan, Taib, Zainudin, & Ismail, 2016	No	Yes
Duan, 2016	No	No
Ekere, Omekwu, & Nwoha, 2016	No	Yes
Ganaie, 2016	No	No
Huang, Pu, Chen, & Chiu, 2015	Yes	Yes
Ikolo, 2015	No	Yes
Jacoby, Ward, Avery, & Marcyk, 2016	Yes	Yes
Khobragade & Lihitkar, 2016	No	No
Kloda & Moore, 2016	No	Yes
Luo & Buer, 2015	No	Yes
Masrek & Gaskin, 2016	No	No
Mohindra & Kumar, 2015	No	No
Nicholas et al., 2015	No	No
Sivagnanam & Esmail, 2015	No	No
Swoger & Hoffman, 2015	Yes	Yes
Tiemo & Ateboh, 2016	No	Yes
Verma & Laltlanmawii, 2016	No	No
Verma & Parang, 2015	No	No
Watts & Mahfood, 2015	Yes	Yes
Xie & Sun, 2015	Yes	Yes
Yan, Hu, & Hu, 2015	No	No
Yap & Cajes, 2016	No	No

Table 6

Validity^a and Reliability^b

Study	Validity Evidence	Reliability Indicators
Akor & Alhassan, 2015	V1	Not stated
Askew, 2015	V1	Not stated
Blake et al., 2016	V1, V2a, V3	Not stated
Blevins, DeBerg, & Kiscaden, 2016	V1	Not stated
Boyce, 2015	Not stated	Not stated
Butler & Byrd, 2016	V1	Not stated
Chen, 2016	V1	R1
Dahan, Taib, Zainudin, & Ismail, 2016	V3	R1
Duan, 2016	Not stated	Not stated
Ekere, Omekwu, & Nwoha, 2016	Not stated	Not stated
Ganaie, 2016	Not stated	Not stated
Huang, Pu, Chen, & Chiu, 2015	V1, V3	R1
Ikolo, 2015	Not stated	Not stated
Jacoby, Ward, Avery, & Marcyk, 2016	V4	Not stated
Khobragade & Lihitkar, 2016	Not stated	Not stated
Kloda & Moore, 2016	V1	Not stated
Luo & Buer, 2015	V2b	Not stated
Masrek & Gaskin, 2016	V1, V3	R1
Mohindra & Kumar, 2015	Not stated	Not stated
Nicholas et al., 2015	Not stated	Not stated
Sivagnanam & Esmail, 2015	V1	Not stated
Swoger & Hoffman, 2015	^c	Not stated
Tiemo & Ateboh, 2016	Not stated	Not stated
Verma & Laltlanmawii, 2016	Not stated	Not stated
Verma & Parang, 2015	Not stated	Not stated
Watts & Mahfood, 2015	Not stated	Not stated
Xie & Sun, 2015	Not stated	Not stated
Yan, Hu, & Hu, 2015	V3	R1
Yap & Cajes, 2016	Not stated	Not stated

^aV1 = Face validity; V2a = Content validity (item); V2b = Content validity (sampling); V3 = Construct validity; V4 = Intercoder reliability. See Table 2 for full definitions.

^bR1 = Internal consistency. See Table 3 for full definitions.

^cIntercoder reliability coefficients not reported.

Content Validity: Item

Blake et al. (2016) were the sole authors to refer to item validity as part of their instrument development process, borrowing the definition from another paper by calling it “internal structure” (Downing, 2003, as cited in Blake et al., p. 227).

Content Validity: Sampling

Luo and Buer (2015) were the sole researchers to address sampling validity; their instrument measured variables drawn from the five areas outlined in the Reference and User Services Association’s (RUSA) Guidelines for Behavioral Performance of Reference and Information Service Providers, as well as from past research on evaluation of reference service.

Construct Validity

Five publications (17%) addressed construct validity as demonstrated by factor analysis and other tests of dimensionality; three of these are described in the “multiple examples” section below because they tested construct validity along with other forms of validity. Two studies (7%) addressed construct validity alone.

Yan et al. (2015) determined the value of average variance extracted (AVE) to demonstrate convergent validity of the constructs in their instrument, and reported that “all of the AVE values range from 0.6727 to 0.8019” (p. 562), and considered these values satisfactory citing Fornell and Larcker’s 1981 publication in which 0.5 is the threshold value for AVE. While not specifically using the term “divergent validity,” Yan et al. demonstrated that they identified divergent variables, stating that “Six variables ... are dropped due to their relatively low factor loadings for its construct” (p. 562).

Dahan et al. (2016) used exploratory factor analysis, assessed using Bartlett’s test for sphericity and the Kaiser-Meyer-Olkin test, and determined that the analysis was significant (p. 41). They then conducted Varimax testing with the Kaiser Normalization Rotation method and found “that all Varimax values are greater than 0.4 and therefore reflect the valid construct of all items” (p. 41).

Intercoder Reliability

Of the ten mixed methods studies (34%), two authors (7%) presented validity evidence in their reports in the form of inter-rater agreement on thematic analysis. Two sets of researchers (Jacoby, Ward, Avery, & Marcyk, 2016; Swoger & Hoffman, 2015) showed evidence of intercoder reliability, as they both discussed and reviewed their coding process; however, they did not report reliability coefficients.

Multiple Examples of Validity Evidence

Blake et al. (2016) provided evidence of face validity, item validity, and construct validity within their study; Huang et al. (2015) demonstrated testing for face validity and construct validity; and Masrek and Gaskin (2016) also showed evidence of a combination of face validity and construct validity. Blake et al. changed their survey “to reflect the responses received from librarian reviews and campus interview sessions,” and consulted instrument development experts who helped them address content (item) and internal structure (construct) validity components (p. 227). Huang et al. invited 15 members of the college faculty to participate in their pre-test, changing the wording of some items based on the faculty’s suggestions (p. 1181), and tested for convergent and divergent validity using composite validity and average variance extracted; they determined that convergent validity was “good” and discriminant validity was “strong” (p. 1185). Masrek and Gaskin showed evidence of a combination of face validity and construct validity as they pre-tested their instrument with students who were part of the target population, as well as with experts in the faculty, and as they analyzed the scales within their instrument for convergent and discriminant validity (p. 44). We did not find evidence that any of these researchers looked for convergent and discriminant validity with similar or different instruments.

Reliability Indicators

Most (83%) of the studies did not state if they had tested their instruments for reliability. Five articles (17%) (Chen, 2016; Dahan, et al., 2016; Huang et al., 2015; Masrek & Gaskin, 2016; Yan, et al., 2015) reported measurement of internal consistency for each component of the satisfaction construct when the component was measured by multiple scale items. All of the researchers used Cronbach's ɑ (alpha) test of internal consistency, in which a value of 0.70 is commonly believed to be a basic threshold of acceptable level (Nunnally, 1978). Chen (2016) reported ɑ values ranging from 0.7305 to 0.8020, which represented a “satisfactory level of reliability” (p. 322). Dahan et al. (2016) reported alpha values ranging from 0.813 to 0.942 (p. 41). Values in Huang et al.’s study (2015) ranged from .809 to .919 (p. 1184). Masrek & Gaskin recorded ɑ “well above 0.7” (p. 42), with values ranging from 0.707 to 0.812. Yan, et al. (2015) did not report separate values for each factor, stating that “Cronbach's alphas of all factors exceed 0.8” (p. 562).

Discussion

This rapid review demonstrates that a less comprehensive and time-consuming type of systematic review of measurement properties can be a useful approach to gaining an overview of research by practicing librarians, as well as pointing to areas for improvement. Our review confirms some aspects of research studies that other librarian researchers have attended to and identifies opportunities for further research. This discussion will place our results within a broader context, followed by recommendations for improvements in practicing librarians’ instrument design.

While solely quantitative study designs continue to be most common in studies of satisfaction with reference services, approximately one-third of the studies we located also gathered patron feedback via open-ended questions. For comparison, VanScoy and Fontana determined in their study of research approaches to reference and information service that quantitative studies ranged from 56.65% in 2005 to 83.33% in 2009 (2016, p. 96). In our review, researchers used surveys alone to gather data 86% of the time, which is higher than the usage of surveys by 50.5% of practitioner researchers according to Hildreth and Aytac (2007) or the 62.3% of researchers writing about library instruction, as determined by Crawford and Feldt (2007, p. 84). However, our results are not surprising given the quantitative design and measurement goal of the studies we identified. We found an improvement over McKechnie et al.’s study in which 17.6% of articles included an appended instrument (2016), with 31% providing this service.

Mahmood’s systematic review found that Likert scales “or Likert-type scoring methods” were used in 15 of 22 scales, and that the “points for scoring options ranged from 2 to 11” but did not report further detail about the design of each Likert scale (p. 1044). It could be useful therefore to compare our results regarding Likert-scale design with studies outside of LIS. Our partial model for this study, Siddiq et al.’s systematic review of Information and Communication Technology-literacy assessment instruments, did not include this information, but Roth, Ogrin, and Schmitz (2016) reported in their systematic instrument review that of seven instruments containing Likert scales, three employed 4-point Likert scales, three contained 5-point scales, and two had 7-point scales. These findings indicate that little agreement exists as to best practices in scale formation. Research on Likert scale questions suggests that 4-point response scales with a “no opinion” option avoid the 5-point scale’s potential for central tendency bias (respondent desire to appear moderate rather than extreme) or social desirability bias (respondent desire to avoid controversial topics). This research implies that if a 5-point scale is offered, the midpoint should be clearly labeled, as otherwise respondents might assign various meanings to the midpoint such as “don’t know,” “neutral,” or “unsure” (Nadler, Weston, & Voyles, 2015).

Librarian researchers might not adequately define “satisfaction,” as only four researchers (14%) developed question items addressing more than one aspect of this construct. Lasda Bergman and Holden (2010) identified four components of the satisfaction construct: willingness to return, positivity of experience, staff quality, and willingness to recommend a service to a colleague. Luo and Buer’s instrument (2015) included 10 components to express satisfaction but did not address another potential component: ethical issues as identified by Kloda and Moore’s (2016) question item, “The consult reflected a respect for my confidentiality as a library user.” On their survey measuring satisfaction with digital library service, including virtual reference, Masrek and Gaskin (2016) included 24 items representing 6 component factors of satisfaction, in addition to three items related to overall satisfaction (Masrek & Gaskin, personal communication, April 20, 2017). Instrument developers might consider that responses to a single question about satisfaction are likely to be positive because “providing tailored individual help … will always be appreciated, which skews user satisfaction in survey results” (Fournier & Sikora, 2015, p. 255). When measuring multiple aspects of the satisfaction construct, a researcher can determine which aspect most likely detracts or adds to patron satisfaction, and initiate training, other services, and environmental improvements to address any issues.

We were surprised to find that four of the studies we examined (Akor & Alhassan, 2015; Dahan et al., 2016; Duan, 2016; Xie & Sun, 2015) contained “double-barreled questions” (Olson, 2008, p. 210) or “multi-concept” to use Glynn’s phrase (2006, p. 394), which asked respondents to agree with statements containing two themes combined with “and” such as “librarians are competent and helpful,” or to rate librarians’ “help and answers.” Because the researcher doesn’t know which aspect of librarian service respondents are rating – competence or helpfulness? help or answers? – these items cannot contribute meaningfully to statistical analysis. Moreover, respondents will likely take more time to consider each concept, potentially leading to survey fatigue. Bassili and Scott (1996) found that “questions took significantly longer to answer when they contained two themes than when either of their themes was presented alone” (p. 394). If researchers might design a survey instrument addressing the various components that make up the satisfaction construct, and thus listing several items to cover these components, it is important to make the items as simple to answer as possible, to encourage respondents to complete the survey. Researchers will usually catch multi-concept questions during a careful pre-testing process.  

Half of the studies we located contained evidence of instrument validity, while more than three-quarters did not report data on instrument reliability, which is comparable to results from Mahmood’s (2017) systematic review of instruments, and results from similar reviews in other disciplines. While explaining that the study excluded articles without validity or reliability evidence, Mahmood (2017) stated that “A large number of studies reported surveys on assessing students’ self-efficacy in IL skills but without mentioning any reliability and validity of scales” and that “the present study’s results are consistent with systematic reviews in other areas,” reporting that between 25% and 50% of studies in three systematic reviews outside of librarianship included information on validity and reliability of instruments (p. 1045):

For example, the reliability and validity were reported in only one-third of studies about evaluation methods of continuing medical education. . . . A study of 11 urbanicity scales found that psychometric characteristics were not reported for eight instruments. . . . A recent systematic review in the area of assessing students’ communication skills found that less than half of studies reported information on reliability and validity . . . (p. 1045).

Our model instrument review researchers Siddiq et al. found that 12 of 30 test developers (40%) reported validation of the test in at least one publication, and that 24 of the 30 (80%) reported reliability evidence according to the authors’ criteria (p. 75). The reporting of validity and reliability evidence can help the reader determine which instrument to use in replicating a study and could aid in future development of an instrument that might combine constructs and items identified through a similar review.

Recommendations

Obtain Training and Refer to Research-Evaluation Checklists

Compared to classroom faculty, librarians are frequently at a disadvantage in designing research projects because they lack coursework in research methods. Initiatives such as Loyola Marymount University’s Institute for Research Design in Librarianship, the Medical Library Association’s Research Training Institute for Health Sciences Librarians, and occasional professional development opportunities, assist librarians to build their research knowledge but can’t reach every librarian. For these reasons, we recommend that librarians become more familiar with existing checklists of research evaluation (e.g., those provided by Glynn, 2006; and McKechnie et al, 2016) that can ensure a basic level of structure and rigor, and further recommend that researchers expand upon these lists as the need for research guidance becomes apparent. Based on our study, we believe that checklists for librarians need to include more guidance in instrument design and in communicating instrument details, e.g., by making sure the target construct is adequately measured; by addressing validity and reliability; by designing questions and response items carefully; including the full instrument; and citing prior instruments.

Completely Measure the Construct

When designing a research instrument, a researcher needs to determine which construct to measure, and which items will best represent that construct, whether it be satisfaction or any other construct. The researcher should keep in mind that more specific items avoid the problem of confounding variables which influence the respondent’s answer, or of misinterpretation in which the respondent’s definition of a construct differs from the researcher’s intended definition. In the realm of “satisfaction” with a service, many factors could influence respondents’ opinion of the service being measured, such as librarian behaviour or performance. It is therefore important to offer several items, rather than a single question about satisfaction.

Address Validity and Reliability

After drafting questions and items, researchers will want to ensure their instrument has face validity by pre-testing it, with non-librarian subjects similar to the target respondent population and with experts in instrument design. These pilot testers should look for bias, for example avoiding questions such as “How much has this service improved your life?” which assume a positive response; for clarity and avoiding the use of jargon, defining terms as Otto (2016) recommended; and for evidence that the question or item addresses what it is intended to address. If researchers try to address all variables encompassing the “satisfaction” construct and report this effort in their paper, that will show evidence of sampling validity. An instrument with many items could be refined by performing analyses to determine convergent and divergent items, thus demonstrating construct validity. If the instrument has been translated into or from a language other than English, developers should report separate validity and reliability information for each version of the instrument.

Design Questions and Response Items Carefully

We repeat Glynn’s (2006) recommendation that not only questions but also their “response possibilities” should be “posed clearly enough to be able to elicit precise answers” (p. 389). Glynn cautions that the Likert scale (i.e., strongly agree, agree, no opinion, disagree, strongly disagree) “[lends itself] to subjectivity and therefore the accuracy of the response is questionable” (p. 394). Regardless of the scale researchers select, we recommend employing a 4-, 5-, or 7-point Likert scale. Avoiding 2-point scales allows for variance in opinion, and avoiding 9-point scales or higher avoids dilution of opinion. We further recommend that researchers label the scale points in a uniform fashion but minimally, e.g., “Strongly/Somewhat/Agree” and “Strongly/Somewhat/Disagree,” rather than offer lengthy definitions of each point as seen in Duan(2016).

Include the Full Instrument

Several authorities on research (Connaway & Radford, 2017; Glynn, 2006; McKechnie et al., 2016) also agree that, to quote Glynn, “the data collection method must be described in such detail that it can easily be replicated” (2006, p. 393). Ideally, these authors further agree, researchers would include their instrument within the body or as an appendix of their publication, or as an online appendix. We recommend expanding existing checklists for the evaluation of research in librarianship, e.g., Glynn’s Critical Appraisal Tool for Library and Information Research (2006) and McKechnie et al.’s Research Rigour Tactics (2016), to remind authors that when they include an appendix containing the instrument, they should also note its inclusion in their abstract, to increase the likelihood that future researchers will locate it. If this inclusion is not possible, then a detailed description of the instrument should be reported in the body of the paper:

• The number of questions and items
• The full text of each question and associated item
• Question format: scale range and endpoint labels, e.g., “Agree” and “Disagree”
• Where relevant, the average time needed to complete the instrument

With this information in hand, researchers can readily reproduce the instrument and use it in their own research.

Cite Prior Instruments

We recommend also that authors cite sources if they base their instrument on previous efforts, demonstrating connections with prior research and further helping to identify useful instruments. Blevins et al. (2016) wrote that “three librarians reviewed the existing literature for similar surveys and developed a set of questions to assess customer service quality” (p. 287) but did not cite the similar surveys. By citing contributing studies, librarians uphold the professional value of encouraging their colleagues’ professional development as stated in the ALA Code of Ethics (2008).

Further Research

As more librarians implement the instrument review methodology, opportunities for future research will abound. Reviews are needed in other research areas, for example to evaluate instruments gathering librarian attitudes toward teaching, collection development, or collaborating with faculty. While we have presented one model for this methodology, there is ample room for improvement and refinement of the method; we foresee that specific standards for instrument review could be developed for librarianship. As described above, another opportunity for future research would be to examine concerns of sampling validity, i.e., which items best demonstrate the patron satisfaction construct.

Limitations

As a rapid review examining two years of librarian research, this study’s results are not necessarily representative of the body of work on student satisfaction with academic library reference services. Although we ran keyword as well as subject searches, it is possible that we did not gather all possible studies presenting librarian-developed instruments due to inconsistent indexing. It is possible that we have missed relevant articles due to not manually searching all LIS journals related to our research topic.  

Our descriptive model does not extend to evaluation of the instrument’s appropriateness in different scenarios such as in-house research versus research intended for publication. While we generally recommend designing an instrument that offers questions with several items measuring the satisfaction construct, it could be appropriate to include a single question and item addressing satisfaction when service quality assurance is the goal. For example, Swoger and Hoffman (2015) incorporated a single question about the usefulness of a specific

type of reference service; in their context the single question was primarily used for local service evaluation and might have been appropriate.

Conclusion

The quality of a research project depends on valid and reliable data collection methods. In preparation for a study, librarians should search broadly and attempt to locate the best instrument exemplars on which to model their own data-gathering method. If researchers do not have time for a comprehensive systematic review, the present study demonstrates that a rapid review can reveal a range of research instruments and guide the development of future instruments. It further demonstrates that the characteristics of librarian-produced research instruments vary widely, and that the quality of reporting varies as well. If librarians do not aim to produce high-quality data collection methods, we need to question our collective findings. By following the recommendations presented here, future researchers can build more robust LIS literature.

Acknowledgements

The authors would like to thank the University of Portland for granting the lead author a sabbatical, during which much of the work of this study was accomplished.

References

Akor, P. U., & Alhassan, J. A. (2015). Evaluation of reference services in academic libraries: A comparative analysis of three universities in Nigeria. Journal of Balkan Libraries Union, 3(1), 24-29.

American Library Association. Evaluation of Reference and Adult Services Committee. (1995). The reference assessment manual. Ann Arbor, MI: Pierian Press.

Askew, C. (2015). A mixed methods approach to assessing roaming reference services. Evidence Based Library and Information Practice, 10(2), 21-33. https://doi.org/10.18438/B8F60V

Bassili, J. N., & Scott, B. S. (1996). Response latency as a signal to question problems in survey research. Public Opinion Quarterly, 60(3), 390–399. https://doi.org/10.1086/297760

Beile, P. (2008, March). Information literacy assessment: A review of objective and interpretive methods and their uses. Paper presented at the Society for Information Technology & Teacher Education International Conference (SITE), Las Vegas, NV. Retrieved from http://eprints.rclis.org/15861/1/Info%20Lit%20Assessment.pdf

Blake, L., Ballance, D., Davies, K., Gaines, J. K., Mears, K., Shipman, P., . . . Burchfield, V. (2016). Patron perception and utilization of an embedded librarian program. Journal of the Medical Library Association: JMLA, 104(3), 226-230. https://doi.org/10.3163/1536-5050.104.3.008

Blevins, A. E., DeBerg, J., & Kiscaden, E. (2016). Assessment of service desk quality at an academic health sciences library. Medical Reference Services Quarterly, 35(3), 285-293. https://doi.org/10.1080/02763869.2016.1189782

Boyce, C. M. (2015). Secret shopping as user experience assessment tool. Public Services Quarterly, 11, 237-253. https://doi.org/10.1080/15228959.2015.1084903

Butler, K., & Byrd, J. (2016). Research consultation assessment: Perceptions of students and librarians. Journal of Academic Librarianship, 42(1), 83-86. https://doi.org/10.1016/j.acalib.2015.10.011

Catalano, A. J. (2016). Streamlining LIS research: A compendium of tried and true tests, measurements, and other instruments. Santa Barbara, CA: Libraries Unlimited.

Chen, Y. (2016). Applying the DEMATEL approach to identify the focus of library service quality. The Electronic Library, 34(2), 315-331. https://doi.org/10.1108/EL-08-2014-0134

Connaway, L. S., & Radford, M. L. (2017). Research methods in library and information science (6th. ed.). Santa Barbara, CA: Libraries Unlimited.

Crawford, G. A., & Feldt, J. (2007). An analysis of the literature on instruction in academic libraries. Reference & User Services Quarterly, 46(3), 77-87.

Dahan, S. M., Taib, M. Y., Zainudin, N. M., & Ismail, F. (2016). Surveying users' perception of academic library services quality: A case study in Universiti Malaysia Pahang (UMP) library. Journal of Academic Librarianship, 42(1), 38-43. https://doi.org/10.1016/j.acalib.2015.10.006

Duan, X. (2016). How they search, how they feel, and how to serve them? Information needs and seeking behaviors of Chinese students using academic libraries. International Information & Library Review, 48(3), 157-168. https://doi.org/10.1080/10572317.2016.1204179

Ekere, J. N., Omekwu, C. O., & Nwoha, C. M. (2016). Users’ perception of the facilities, resources and services of the MTN Digital Library at the University of Nigeria, Nsukka. Library Philosophy & Practice, 3815, 1-23.

Fidel, R. (2008). Are we there yet? Mixed methods research in library and information science. Library & Information Science Research, 30(4), 265-272. https://doi.org/10.1016/j.lisr.2008.04.001

Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39-50.

Fournier, K., & Sikora, L. (2015). Individualized research consultations in academic libraries: A scoping review of practice and evaluation methods. Evidence Based Library and Information Practice, 10(4), 247-267. https://doi.org/10.18438/B8ZC7W

Ganaie, S. A. (2016). Satisfaction of library and information science students with the services provided by Allama Iqbal Library of Kashmir University. International Research: Journal of Library and Information Science, 6(3), 504-512.

Gerlich, B. K., & Berard, G. L. (2007). Introducing the READ Scale: Qualitative statistics for academic reference services. Georgia Library Quarterly, 43(4), 7-13.

Glynn, L. (2006). A critical appraisal tool for library and information research. Library Hi Tech, 24(3), 387-399.

Gotch, C. M., & French, B. F. (2014). A systematic review of assessment literacy measures. Educational Measurement: Issues and Practice, 33(2), 14-18.

Grant, M., & Booth, A. (2009). A typology of reviews: An analysis of 14 review types and associated methodologies. Health Information and Libraries Journal, 26, 91-108.

https://doi.org/10.1111/j.1471-1842.2009.00848.x

Greifeneder (2014). Trends in information behaviour research. Paper presented at the Proceedings of ISIC: The Information Behaviour Conference, Leeds, England. Retrieved from http://www.informationr.net/ir/19-4/isic/isic13.html

Hildreth, C. R., & Aytac, S. (2007). Recent library practitioner research: A methodological analysis and critique. Journal of Education for Library and Information Science, 48(3), 236-258.

Huang, Y., Pu, Y., Chen, T., & Chiu, P. (2015). Development and evaluation of the mobile library service system success model. The Electronic Library, 33(6), 1174-1192.

Ikolo, V. E. (2015). Users satisfaction with library services: A case study of Delta State University library. International Journal of Information and Communication Technology Education, 11(2), 80-89.

Jacoby, J., Ward, D., Avery, S., & Marcyk, E. (2016). The value of chat reference services: A pilot study. portal: Libraries and the Academy, 16(1), 109-129.

Jordan, J. L. (2015). Additional search strategies may not be necessary for a rapid systematic review. Evidence Based Library & Information Practice, 10(2), 150-152. https://doi.org/10.18438/B8FC77

Khobragade, A. D., & Lihitkar, S. R. (2016). Evaluation of virtual reference service provided by IIT libraries: A survey. DESIDOC Journal of Library & Information Technology, 36(1), 23-28.

Kloda, L.A., & Moore, A. J. (2016). Evaluating reference consults in the academic library. In S. Baughman, S.Hiller, K. Monroe, & A. Pappalardo (Eds.), Proceedings of the 2016 Library Assessment Conference Building Effective, Sustainable, Practical Assessment (pp. 626-633). Washington, DC: Association of Research Libraries.

Koufogiannakis, D. (2012). The state of systematic reviews in library and information studies. Evidence Based Library & Information Practice, 7(2), 91-95.

Koufogiannakis, D., Slater, L., & Crumley, E. (2004). A content analysis of librarianship research. Journal of Information Science, 30(3), 227-239.

Lasda Bergman, E. M., & Holden, I. I. (2010). User satisfaction with electronic reference: A systematic review. Reference Services Review, 38(3), 493-509. https://doi.org/10.1108/00907321011084789

Luo, L., & Buer, V. B. (2015). Reference service evaluation at an African academic library: The user perspective. Library Review, 64(8/9), 552-566.

Mahmood, K. (2017). Reliability and validity of self-efficacy scales assessing students’ information literacy skills. Electronic Library, 35(5), 1035-1051. https://doi.org/10.1108/EL-03-2016-0056

Masrek, M. N., & Gaskin, J. E. (2016). Assessing users satisfaction with web digital library: The case of Universiti Teknologi MARA. The International Journal of Information and Learning Technology, 33(1), 36-56.

McKechnie, L. E., Baker, L., Greenwood, M., & Julien, H. (2002). Research method trends in human information literature. New Review of Information Behaviour Research, 3, 113-125.

McKechnie, L., Chabot, R., Dalmer, N., Julien, H., & Mabbott, C. (2016, September). Writing and reading the results: The reporting of research rigour tactics in information behaviour research as evident in the published proceedings of the biennial ISIC conferences, 1996 – 2014. Paper presented at the ISIC: The Information Behaviour Conference, Zadar, Croatia. Retrieved from http://www.informationr.net/ir/21-4/isic/isic1604.html

McKibbon, K. A. (2006). Systematic reviews and librarians. Library Trends, 55(1), 202-215. https://doi.org/10.1353/lib.2006.0049

Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. PLoS Medicine, 6(7), e1000097. https://doi.org/10.1371/journal.pmed.1000097

Mohindra, R., & Kumar, A. (2015). User satisfaction regarding quality of library services of A.C. Joshi Library, Panjab University, Chandigarh. DESIDOC Journal of Library & Information Technology, 35(1), 54-60.

Nadler, J. T., Weston, R., & Voyles, E. C. (2015). Stuck in the middle: The use and interpretation of mid-points in items on questionnaires. Journal of General Psychology, 142(2), 71-89. https://doi.org/10.1080/00221309.2014.994590

Nicholas, P., Sterling, J., Davis, R., Lewis, J. C., Mckoy-Johnson, F., Nelson, K., Tugwell, Y., & Tyrell, K. (2015). “Bringing the library to you!”: The halls of residence librarian program at the University of the West Indies, Mona Library. New Library World, 116(5/6), 316-335.

Nunnally, J. C. (1978). Psychometric theory. New York, NY: McGraw-Hill.

Olson, K. (2008). Double-barreled question. In P. J. Lavrakas (Ed.), Encyclopedia of survey research methods (p. 210). Thousand Oaks, CA: SAGE Publications Ltd. https://doi.org/10.4135/9781412963947.n145

Otto, J. J. (2016). A resonant message: Aligning scholar values and open access objectives in OA policy outreach to faculty and graduate students. Journal of Librarianship and Scholarly Communication, 4, eP2152. https://doi.org/10.7710/2162-3309.2152

Phelps, S. F., & Campbell, N. (2012). Systematic reviews in theory and practice for library and information studies. Library & Information Research, 36(112), 6-15.

Reference and User Services Association. (2008). Definitions of reference. Retrieved from http://www.ala.org/rusa/guidelines/definitionsreference

Roth, A., Ogrin, S., & Schmitz, B. (2016). Assessing self-regulated learning in higher education: A systematic literature review of self-report instruments. Educational Assessment, Evaluation and Accountability, 28(3), 225-250.

Schilling, K., & Applegate, R. (2012). Best methods for evaluating educational impact: A comparison of the efficacy of commonly used measures of library instruction. Journal of the Medical Library Association: JMLA, 100(4), 258-269. https://doi.org/10.3163/1536-5050.100.4.007

Siddiq, F., Hatlevik, O. E., Olsen, R. V., Throndsen, I., & Scherer, R. (2016). Taking a future perspective by learning from the past – A systematic review of assessment instruments that aim to measure primary and secondary school students' ICT literacy. Educational Research Review, 19, 58-84. https://doi.org/10.1016/j.edurev.2016.05.002

Sivagnanam, R., & Esmail, S. M. (2015). User satisfaction and performance of arts and science college libraries situated in Cuddalore District: A study. International Journal of Library and Information Studies, 5(5), 145-155.

Swoger, B. J. M., & Hoffman, K. D. (2015). Taking notes at the reference desk: Assessing and improving student learning. Reference Services Review, 43(2), 199-214.

https://doi.org/10.1108/RSR-11-2014-0054

Tiemo, P. A., & Ateboh, B. A. (2016). Users’ satisfaction with library information resources and services: A case study College of Health Sciences Library Niger Delta University, Amassoma, Nigeria. Journal of Education and Practice, 7(16): 54-59.

VanScoy, A., & Fontana, C. (2016). How reference and information service is studied: Research approaches and methods. Library & Information Science Research, 38(2), 94-100. https://doi.org/10.1016/j.lisr.2016.04.002

Verma, M. K., & Laltlanmawii, R. (2016). Use and user’s satisfaction on library resources and services by UG students of Government Hrangbana College, Aizawl: A study. Journal of Advances in Library and Information Science, 5(1), 18-23.

Verma, M. K., & Parang, B. (2015). Use and user’s satisfaction on library resources and services by students of School of Physical Sciences, Mizoram University, Aizawl: A study. Information Studies, 21(2), 193-204.

Watts, J., & Mahfood, S. (2015). Collaborating with faculty to assess research consultations for graduate students. Behavioral & Social Sciences Librarian, 34(2), 70-87. https://doi.org/10.1080/01639269.2015.1042819

Xie, J., & Sun, L. (2015). Exploring Chinese students' perspective on reference services at Chinese academic libraries: A case study approach. Journal of Academic Librarianship, 41(3), 228-235. https://doi.org/10.1016/j.acalib.2015.04.002

Xu, J., Kang, Q., & Song, Z. (2015). The current state of systematic reviews in library and information studies. Library & Information Science Research, 37(4), 296-310. https://doi.org/10.1016/j.lisr.2015.11.003

Yan, W., Hu, Y., & Hu, C. (2015). Using SEM and TCM to improve services to libraries: A comparative study on Chinese academic libraries. Journal of Academic Librarianship, 41(5), 558-566. https://doi.org/10.1016/j.acalib.2015.07.006

Yap, J. M., & Cajes, M. L. (2016). Where are the librarians? The roving reference experience of DLSU libraries. Information World/Bilgi Dünyası, 17(2), 232-246. Retrieved from http://bd.org.tr/index.php/bd/article/view/506/575

Appendix A

Database Searches

CINAHL

Search ID	Search Terms
S5	librar* AND (reference or consultation or roaming or chat) AND (satisfaction or attitudes) AND (survey* or instrument* or questionnaire* or interview* or focus group*)
S4	(S1 AND S2 AND S3)
S3	(MH "Research Instruments+")
S2	(MH "Consumer Satisfaction+")
S1	(MH "Library Reference Services")

ERIC

Search ID	Search Terms
S9	((surveys or questionnaires or instruments or measures or interviews)) AND (S2 AND S4 AND S7 AND S8)
S8	(surveys or questionnaires or instruments or measures or interviews)
S7	DE "Libraries" OR DE "Research Libraries" OR DE "Medical Libraries" OR DE "College Libraries" OR DE "Academic Libraries" OR DE "Two Year Colleges"
S6	((surveys or questionnaires or instruments or measures or interviews) AND (DE "Surveys" OR DE "Attitude Measures" OR DE "Interviews" OR DE "Measures (Individuals)" OR DE "Questionnaires" OR DE "Research" OR DE "Semi Structured Interviews" OR DE "Structured Interviews")) AND (S1 AND S2 AND S4 AND S5)
S5	(surveys or questionnaires or instruments or measures or interviews) AND (DE "Surveys" OR DE "Attitude Measures" OR DE "Interviews" OR DE "Measures (Individuals)" OR DE "Questionnaires" OR DE "Research" OR DE "Semi Structured Interviews" OR DE "Structured Interviews")
S4	DE "Attitudes" OR DE "Satisfaction" OR DE "Job Satisfaction" OR DE "Life Satisfaction" OR DE "Marital Satisfaction" OR DE "Participant Satisfaction" OR DE "Student Satisfaction" OR DE "User Satisfaction (Information)" OR DE "School Attitudes" OR DE "Student Attitudes"
S2	DE "Reference Services" OR DE "Library Services"
S1	(DE "Academic Libraries" OR DE "College Libraries" OR DE "Research Libraries") AND (DE "Reference Services" OR DE "Library Services")

PubMed –With Applied COSMIN Filter

Search ID	Search Terms
S1	academic library AND reference or consultation or roaming or chat) AND (satisfaction or attitudes) AND (survey* or instrument* or questionnaire* or interview* or focus group*)
S2	academic librar* AND reference or consultation or roaming or chat) AND (satisfaction or attitudes) AND (survey* or instrument* or questionnaire* or interview* or focus group*)
S3	(academic librar*) AND (reference or consultation or roaming or chat) AND (satisfaction or attitudes) AND (survey* or instrument* or questionnaire* or interview* or focus group*)
S4	((academic library) AND consumer satisfaction) AND (survey OR instrument OR questionnaire OR interview OR focus group)
S5	(((library reference services) AND consumer satisfaction) AND (((((survey) OR instrument) OR questionnaire) OR interview) OR focus group))

ProQuest Library Science Database (LS) – Initial Search

Search ID	Search Terms
S1	all((academic* AND librar* AND (reference OR "customer satisfaction" OR "user satisfaction" OR "customer services") )) AND su(research or surveys or questionnaires) AND ab(study or survey* or interview* or research*)

ProQuest Library Science Database (LS)- Subsequent search

Search ID	Search Terms
S1	all((reference or "research consultation") AND (satisf* or evaluat* or assess* or improve*) AND (experiment* or survey* or qualitative or servqual or instrument or investigat* or analysis or questionnaire*) AND "academic librar*")

Google Scholar

Search ID	Search Terms
S1	"library reference services" AND "consumer satisfaction" AND research instruments"
S2	academic library AND (reference OR consultation OR roaming OR chat ) AND (satisfaction or attitudes ) AND ( survey* OR instrument* OR questionnaire* OR interview* OR focus group*)
S3	Librar* AND reference AND satisfaction
S4	librar* AND (reference or consultation or roaming or chat) AND (satisfaction or attitudes) AND (survey* or instrument* or questionnaire* or interview* or focus group*)
S5	librar* AND (reference or consultation or roaming or chat) AND (satisfaction or attitudes) AND (survey* or instrument* or questionnaire* or interview* or focus group*)
S6	library AND reference AND satisfaction
S7	(academic librar*) AND (reference or consultation or roaming or chat) AND (satisfaction or attitudes) AND (survey* or instrument* or questionnaire* or interview* or focus group*)

Library, Information Science & Technology Abstracts (LISTA) – Initial Search

Search ID	Search Terms
S1	(reference or "research consultation" ) AND ( satisf* or evaluat* or assess* or improve*) AND ( experiment* or survey* or qualitative or servqual or instrument or investigat* or analysis or questionnaire*) AND "academic librar*"
S2	(academic AND librar* AND (reference OR "user satisfaction")) AND ( SU ( research or surveys or questionnaires ) OR AB (study or survey* or interview* or research*) )

LISTA – Subsequent Search

Search ID	Search Terms
S1	librar* AND reference AND satisfaction

Web of Science

Search ID	Search Terms
S1	(academic librar*) AND (survey* OR instrument* OR questionnaire* OR interview* OR focus group*) AND (satisfaction OR attitudes) AND (reference OR consultation OR roaming OR chat)
S2	(library reference services) AND (survey* OR instrument* OR questionnaire* OR interview* OR focus group*) AND (satisfaction OR attitudes)
S3	(consumer satisfaction) AND (library reference services)
S4	(consumer satisfaction) AND (library reference services) AND (survey* OR instrument* OR questionnaire* OR interview* OR focus group*)
S5	(librar*) AND (consumer satisfaction) AND (survey* OR instrument* OR questionnaire* OR interview* OR focus group*) AND (reference OR consultation OR roaming OR chat)
S6	(librar*) AND (consumer satisfaction) AND (survey* OR instrument* OR questionnaire* OR interview* OR focus group*) AND (reference OR consultation OR roaming OR chat)

Appendix B

Data Extraction Elements

APA citation

Article title

Journal

Country

Stated Purpose of Study

Area of Reference Service: General, Desk, Embedded Librarianship, External, Research Consultation, Virtual Reference (Chat or Email)

Age/Level of Students -- if students were part of the targeted population

Targeted Population

Usable Responses

Sampling Strategy

Research Design

Data Collection Method

Types of Questions, other than demographic

Demographics Gathered

Technical Aspects, such as distribution, availability of translations, duration of survey period

Time Allotted to Complete Survey

Validity Indicators

Reliability Indicators

Instrument Availability

Replicability of Instrument

Comments