By Any Measure

Introduction

“What gets measured gets managed.”^[1]

The things that get measured by an organization are a tell for what matters to the organization. Librarianship has been grappling with an underrepresentation of Black, Indigenous, and People of Color (BIPOC) librarians for decades,^[2] and while libraries have developed a plethora of recruitment initiatives, the numbers appear to have not changed meaningfully, remaining static over the decades. A contributing factor may be that demographic information about the profession is not routinely measured and reported.

What is collected is dispersed across multiple data sources and research initiatives, covering different time periods and using a variety of methodologies. In this issue brief, we take a closer look at how diversity in librarianship has been measured and the shortcomings and complexities of the data. We then use four different starting points from the data to construct four models to project out the diversity of the profession ten years from now, in 2033, as well as what would be necessary to push the needle on demographic diversity even further.

The status quo by numbers

In 2006, The American Library Association (ALA) launched an initiative to begin to measure the demographics of the profession: Diversity Counts.^[3] This website reports racial demographic data, and since its release, Diversity Counts has been used by many researchers, program managers, and advocates as a “tool to demonstrate the need for increased diversity recruitment within the profession.”^[4] Diversity Counts was last updated with data from 2009-2010 (see Table 1 below).

From the beginning, ALA recognized that providing a snapshot of the profession was not sufficient: “instead of attaching expectations to specific numbers within the data, it may be more accurate to explore trends from year to year, to see if there are consistencies in the growth or change in the population.”^[5] Yet, to date ALA’s Diversity Counts initiative is limited to that single year. Kevin Strowder, director of ALA’s Office for Diversity, Literacy and Outreach Services is hopeful that will change, noting “we are in the process of bringing on a contractor to handle the quantitative and qualitative aspects of the Diversity Counts initiative.”^[6]

Table 1: Number of Credentialed Librarians by Characteristic, 2009-2010^[7]

While not explicitly as focused on racial and ethnic diversity, an ALA member demographic survey launched in 2005 does track race, among other factors. The latest publication from this effort, a report released in 2017, presents data from that same year and compares it to the data from 2014 (see table 2 below).

Table 2. ALA Members by Race or Family Origin (% of members self-identifying)^[8]

The demographic percentages are consistent across the two surveys, with changes of less than one percentage point across all categories. In methodological terms, ALA’s public report notes that “as of January 9, 2017, 37,666 current members have participated,”^[9] accounting for 74 percent total members at the time. While the survey may have reflected the demographics of ALA members, most working librarians are not members of ALA. ALA membership is also not exclusive to librarians, as ALA members include trustees, publishers, and others. The 2017 demographic report also does not specify who respondents are, in contrast to Diversity Counts which differentiates between credentialed librarians (defined as “librarians who report that they have completed a master’s degree or another degree higher than a master’s) and non-credentialed librarians (“librarians are all other persons whose occupation is librarian but have not completed a master’s degree”).^[10] Moreover, Given the membership dues and the tight correlation between wealth and race, these data may not be representative of the field as a whole. In spite of the limitations, the actual reported figures tightly align with the data from other sources and different timeframes.

Turning to data sources outside of the profession, we investigated governmental agency data. Using the Current Population Survey (CPS), the Bureau of Labor Statistics (BLS) issues monthly demographic profiles of the current state of the workforce, including statistics for “librarians and media specialists,” discussed here. It is worth noting that BLS-CPS also reports data on three other categories: “library technicians,” “library assistants (clerical),” and “other educational instruction and library workers,” but we do not include their counts here. The percentage of white librarians has remained about the same since 2017 when the public reports first included data broken out by race and ethnicity (see table 3 below).

Table 3: Percentage of employed librarians by detailed occupation, sex, race, and Hispanic or Latino ethnicity (From BLS-CPS Table 11, 2017-2021).^[11]

As we can see from the table above, while the percentage of white librarians is fairly stable, the counts of Black and Hispanic librarians fluctuate substantially from year to year. This has to do with several methodological details having to do with how the BLS-CPS is constructed, which have implications for our understanding of the diversity of the field. First, the 158,000 count of librarians and media specialists that the Bureau of Labor Statistics reports in the Current Population Survey for 2021, for example, represents the average of 12 monthly counts for that year. Yet, that average, without any insight into the full range (i.e. the minimum and maximum counts) in that given year, does not provide a full picture. Second, since race and Hispanic origin are not mutually exclusive categories, librarians classified as Hispanic may also be white, Black, Native American, or Asian. As a result, the percentages reported in BLS-CPS add up to over 100 percent since individuals may be counted under more than one race/ethnicity group. Third, the counts and percentages reported are estimates for the full population based on very small samples, depending on the professional group.

Given the limitations of the BLS-CPS, particularly the issue of double counting individuals, we turned to the Census’s microdata files in order to extract the unique demographic counts, including race and Hispanic-origin combinations. Microdata are sample files that are publicly and freely made available by the Census Bureau. Based on the American Community Survey, they allow individuals to explore the data in more detail than the pre-tabulated data the Census produces. We extracted the data from the Institute for Social Research and Data Innovation, a University of Minnesota division that oversees IPUMS (originally an acronym for Integrated Public Use Microdata Series). Specifically, we used IPUMS-CPS,^[12] which harmonizes microdata from the Current Population Survey dating as far back as 1962.^[13]

Given our interest in long-term employment and demographic trends in the field, we used the Annual Social and Economic Supplement (ASEC) of CPS. ASEC is a survey conducted in March each year and is primarily used by researchers to track longitudinal or year-over-year changes. Furthermore, CPS is a monthly survey of roughly 60,000 households, while ASEC features a larger sample size of 75,000 households. It is also worth noting that the Bureau of Labor Statistics reports data from CPS, while the Census Bureau reports data from ASEC for its year-over-year reports.

It is worth highlighting that reproducing the exact counts reported by the Bureau of Labor Statistics from the ASEC data is not possible due to two factors. First, the Census Bureau routinely introduces seasonal adjustments to their data and weights.^[14] Second, the small count of individuals in the profession surveyed (around 100 individuals) leads to sampling variation in the estimated full population,^[15] both year-over-year, as well as within any given year’s monthly estimates. Across each year in our study, the total estimated count of library and media collection specialists ranges from 168,000 to 227,000.^[16]

For our analysis, we extracted annual employment and demographic data for librarians and media collection specialists for each year between 2010 and 2022, inclusive. In order to capture non-duplicated counts of race and ethnicity, we captured each combination of two variables—race (all 26 categories captured by the Census) and Hispanic origin (a binary yes/no category). Thus, we were able to capture individuals of each race, as well as Hispanic origin status. Given the small percentage of individuals in some race and Hispanic origin combinations, we rolled them up in two ways: first, all Hispanic-origin individuals, regardless of race were grouped under a “Hispanic” group, and second, any two or more non-Hispanic races were grouped into a “Two or More Races” category.^[17] The extracted data is presented in table 4 below.

Table 4: Percentage of employed librarians by race-ethnicity, 2010-2022 (from IPUMS-CPS).

What does the future of the profession look like?

The demographic data that we have about the profession, as we have shown, are not completely up to date and may not provide the most accurate accounting. However, the data do confirm that the vast majority of librarians are white and that the racial and ethnic makeup in the field has changed little over the past decade. As we can see in table 4 above, the percentages fluctuate year to year, and in the case of Black librarians, there has been a steady decline since 2018. As with BLS-CPS data, the IPUMS-CPS data are based on small samples on which estimates to the full population are constructed. In the particular case of librarians, some of the samples include only 100 or so individuals, which are then used to estimate the full population of librarians. With all their complexities and built-in assumptions, these datasets offer the most comprehensive snapshot we have of librarianship—and they all point to a consistent status quo when it comes to the predominantly white racial makeup of the profession—particularly when we focus on credentialed librarians. Looking ahead, are there signs of change?

While considering multiple variables and making a host of assumptions, it is possible to project out and model the racial demographics of the profession in the early 2030s. One such assumption has to do with educational attainment as proxy for a profession. While not all librarians hold an MLS degree—after all, the IPUMS-CPS sample is not exclusive to MLS degree-holders—we can make a reasonable assumption that most individuals pursuing an MLS degree plan to pursue a career in librarianship. According to ALA’s demographic report from 2017, 61 percent of ALA members who responded hold an MLS degree.^[18] Seventy-one percent of library positions advertised in 2022 required or preferred an MLS degree, with 55 percent of employers making it a requirement (up from 49 percent in 2021).^[19] Given this, looking at the demographics of current MLS students can help us predict change on the horizon.

Every year, the Association of Library and Information Science Educators (ALISE) reports the number of graduates of LIS programs along with racial demographic characteristics.^[20] Looking at the 2021 data, released in 2022, we calculated that there are 5,332 MLS graduates per year, excluding international students and students whose race was unknown.^[21] The race and ethnicity of these graduates in 2021 is presented in table 5 below. Because the percentage of white MLS graduates (65.7 percent) is lower than the Current Population Survey percentage of white librarians (85.7 percent), it suggests that over time the gap between white librarians and BIPOC librarians will close. The question then is, at what rate is this occurring?

Table 5: MLS Graduates by Race/Ethnicity, 2021 (percentages, excludes international students and those who preferred not to disclose race/ethnicity)^[22]

The Bureau of Labor Statistics Occupational Outlook Handbook predicts growth in librarian and media specialist positions to increase at a rate of 800 per year through 2031.^[23] We held constant the Occupational Outlook Handbook growth rate of 800 librarians per year through 2032 and held constant the total number of MLS graduates per year. We used as a variable the percentage of white MLS grads and BIPOC MLS grads each year. We make the assumption that the difference between the annual number of MLS graduates and the predicted growth rate is equal to people retiring or leaving the field for other reasons. Using the 2021 BLS-CPS data as a starting point, we project librarian racial demographics out to the year 2033. Over this period, the total number of librarians is expected to increase from 158,231 to an estimated 168,631, and the percentage of white librarians to decrease from 85.7 percent to 82.5 percent. We then ran other scenarios, to project the data using the ALA Diversity Counts and IPUMS-CPS initial data as starting points, as well as an initial starting point based off of the growth rate or the years in the sample we extracted through IPUMS-CPS (2010-2022).^[24] We also did sensitivity analysis to see under what circumstances the percentage of while librarians would drop further,^[25] for example to 75 percent by 2033, given that at the national level, the US is predicted to be 55.8 non-Hispanic white in 2030.^[26]

Within these parameters, an additional 500 to 1,100 librarians of color will need to graduate with the MLS degree every year for the next 10 years for the profession to become 75 percent white. Currently around 1,250 new BIPOC librarians graduate every year. To reach 75 percent white as a profession would mean an increase from the currently estimated number of annual BIPOC MLS graduates to between 1,750 – 2,350. Under the constraints established by our model, there would need to be an offsetting reduction of 500 – 1,100 white librarians graduating every year.

Table 6: Demographic Projections for MLS-bearing librarian through 2033^[27]

Conclusion

The above-referenced projections highlight the complexities of trying to benchmark and project the demographic trends of a profession. To do so, we have had to work with multiple datasets, all with their own assumptions, sampling variation, adjustments, and different operationalizations of race and ethnicity categories. We had to construct our own assumptions about who enters and who leaves the profession, as well as sensitivity analysis in absence of readily available data that speaks to the growth and attrition rate of the librarian workforce. Fundamentally, we had to grapple with who is defined as a librarian in all these datasets. Thus, in absence of consistently measured and reported data, it becomes very difficult to offer a current benchmark of the profession—let alone project its future with confidence.

Given that faculty diversity correlates with higher levels of student success, particularly on metrics such as student retention, persistence, and completion rates, as well as the positive impact on students’ sense of belonging,^[28] it is worth raising if librarians’ diversity would also correlate with positive outcomes when it comes to student success metrics. To study this hypothesis, on one hand we would need more sustained, systematic data collection in order to confidently track progress, while on the other hand, we would need to take stock of some of the effectiveness of initiatives to be able to fund and make the profession more attractive to BIPOC students given the differential experiences of BIPOC students in the library, something Curtis Kendrick explores further in a companion issue brief.^[29]

Appendix A: Methodology

We extracted the data using IPUMS-CPS’s data extraction page. We used the following variables and samples to create our extract. For the samples, we used ASEC 2010 through ASEC 2022. To capture librarians and media collection specialists, we filtered the dataset to include variable OCC=2435 (2430 for years prior to 2020). We also used RACE and HISPAN to capture all 26 races the Census operates with, as well as Hispanic origin status. We also restricted the dataset to include EMPSTAT= (10,12), to capture only actively employed librarians. For weights, we used the standard ASECWT for all years except 2019-2021. For these years we used ASECWTCVD, a weight designed to account for pandemic-related non-response rates.

For percent change, we used the following formula:

And for growth rate, we used:

Appendix B: Detailed 75 percent scenario projection

We began our projection with the demographic breakdown of MLS degrees awarded in 2021, by race and ethnicity.^[30] We then calculated what percentage each non-white group represents out of the total of non-white graduates (columns 3 and 4 in table 7 below). We then conducted sensitivity analysis with different values of potential count of MLS graduates to get to 75 percent white graduates, which for this scenario is an extra 1,000 MLS graduates. We then multiplied the percentages in column 4 below by 1,000 (column 5) to calculate how many additional MLS graduates of each race-ethnicity would be necessary to approach the 75 percent white threshold. In columns 6 and 7, we depict the revised total count and percentages of graduates (column 2 + column 5) necessary to result in the profession being only 75 percent white by 2033.

Table 7: Projected Demographic Trends for Librarians

We then referenced the anticipated growth—8,000 new librarian jobs between 2021-2031—from the US Bureau of Labor Statistics’ Occupational Outlook Handbook.^[31] Thus, we use an anticipated annual growth rate of 800 new positions (8,000/10 years). Thus, the number of MLS-degree holders who leave the profession comes down to 5,332-800=4,532 per year.

In order to project the number of MLS-degree holders by race-ethnicity who then leave the field we multiplied 4,352 by the percentage each race-ethnicity represents from the initial 2022 counts, as extracted from IPUMS-CPS (raw counts and associated percentages are in columns 2 and 3, table 8 below), giving us the projected MLS-holders who leave the field in column 4 below. Finally, we then subtracted these new numbers from column 4 from the new adjusted total of MLS graduates from column 5, table 7 above, giving us the net adjusted change in MLS graduates, depicted in column 5, table 8 below.

Table 8: Projected Annualized Growth

Then, we took these net adjusted counts from the fifth column of table 8 and added them to the 2022 initial count of librarians, as extracted from IPUMS-CPS (column 2, table 9 below). The new counts represent the starting counts for the following year, 2023. We then repeated the steps covered by table 9 for these new counts/new initial percent of librarian to then calculate the 2024 beginning counts, and so forth through 2033. We see that in 2033, the count of white librarians becomes 75.12 percent (161,712/215,260 x 100).

Table 9. 2022-233 Full Projection, Revised CPS Counts, End of Year Counts

Endnotes

1. The quote is from V. F. Ridgway in “Dysfunctional Consequences of Performance Measurements,” Administrative Science Quarterly 1, no. 2 (September 1956), 240-247, https://www.jstor.org/stable/i341217.  Note: Curtis Kendrick is a member of ITHAKA’s board of trustees.↑

2. We use BIPOC librarians or librarians of color interchangeably throughout this essay. When reporting government or ALA data we use the nomenclature of the source documents.

   ↑

3. American Library Association, Diversity Counts 2009-2010 Update, last accessed 10 April 2023, https://www.ala.org/aboutala/offices/diversity/diversitycounts/divcounts.

   ↑

4. Ibid.

   ↑

5. Ibid.

   ↑

6.  Personal email correspondence between Kevin Strowder and Curtis Kendrick, January 12, 2023.

   ↑

7. According to Table A-2, there is a higher level of racial diversity among library assistants, where a little over 26 percent are BIPOC, American Library Association, Diversity Counts 2009-2010 Update, last accessed 10 April 2023, https://www.ala.org/aboutala/offices/diversity/diversitycounts/divcounts,

   ↑

8. American Library Association Office for Research and Statistics, 2017 ALA Demographic Survey, last accessed 10 November 2022, https://www.ala.org/tools/sites/ala.org.tools/files/content/Draft%20of%20Member%20Demographics%20Survey%2001-11-2017.pdf.

   ↑

9. Ibid.

   ↑

10. Ibid, p. 5.

    ↑

11. US Bureau of Labor Statistics, Labor Force Statistics from the Current Population Survey, 7 November 2022, https://www.bls.gov/cps/tables.htm#annual,.

    ↑

12. Sarah Flood, Miriam King, Renae Rodgers, Steven Ruggles, J. Robert Warren and Michael Westberry, Integrated Public Use Microdata Series, Current Population Survey: Version 10.0 [dataset], Minneapolis, MN: IPUMS, 2022, https://doi.org/10.18128/D030.V10.0.

    ↑

13. US Census Bureau, Current Population Survey Microdata, 2023, https://data.census.gov/mdat/#/.

    ↑

14. US Census Bureau, Time Series and Seasonal Adjustment, 2023, https://www.census.gov/topics/research/seasonal-adjustment.html#:~:text=Seasonal%20adjustment%20is%20the%20estimation,removal%20from%20the%20time%20series.

    ↑

15. Sampling variation refers to how each estimate or statistic can vary due to how different each sample is from the others.

    ↑

16. We are grateful to Ranae Rodgers and Matt Bombyk from the Minnesota Population Center for their assistance with IPUMS data extracts.

    ↑

17. For full variable selection and data, see Appendix A.

    ↑

18. American Library Association Office for Research and Statistics, 2017 ALA Demographic Survey, last accessed 10 November 2022, https://www.ala.org/tools/sites/ala.org.tools/files/content/Draft%20of%20Member%20Demographics%20Survey%2001-11-2017.pdf.

    ↑

19. San Jose State University School of Information, “MLIS Skills at Work: A Snapshot of Job Postings Spring 2022,” https://ischool.sjsu.edu/sites/main/files/file-attachments/career_trends.pdf?1655307854.

    ↑

20. We used the 2021 count of MLS graduates per year, as reported by ALISE, as a proxy for the racial demographic profile and number of graduates. We base projections for growth in librarian positions on the US Bureau of Statistics Occupational Outlook Handbook. We assume the difference between the number of graduates per year and the increase in positions anticipated in the handbook projections is equivalent to the number leaving the profession per year, either through retirement or to pursue other opportunities.

    ↑

21.  Association of Library and Information Science Education, 2022 Statistical Report, last accessed 26 November 2023, https://online.fliphtml5.com/gjade/xyno/#p=10.

    ↑

22. Ibid.

    ↑

23. US Bureau of Labor Statistics, Occupational Outlook Handbook, last accessed 10 November 2022, https://www.bls.gov/ooh/education-training-and-library/librarians.htm.

    ↑

24. See Appendix A for more details.

    ↑

25. Sensitivity analysis is a statistical technique used to determine how different values of an independent variable–here the number of BIPOC MLS graduates–would impact the values of a dependent variables—here percent of white librarians in 2033.

    ↑

26. Jonathan Vespa, Lauren Medina, and David M. Armstrong, Demographic Turning Points for the United States: Population Projections for 2020 to 2060, March 2018 (Revised 2020), https://www.census.gov/content/dam/Census/library/publications/2020/demo/p25-1144.pdf.

    ↑

27. For detailed counts on projected scenarios, see appendix B.

    ↑

28. Jasmín D. Llamas, Khoa Nguyen, and Alisia G.T.T. Tran, “The Case for Greater Faculty Diversity: Examining the Educational Impacts of Student-Faculty Racial/Ethnic Match,” Race Ethnicity and Education 24, no. 3 (October 2019): 375-391, https://doi.org/10.1080/13613324.2019.1679759; see also Jinann Bittar, Gabriel Montague and Lauren Ilano, “Faculty Diversity and Student Success Go Hand in Hand, So Why Are University Faculties so White?” Education Trust, 1 December 2022, https://edtrust.org/resource/faculty-diversity-and-student-success-go-hand-in-hand-so-why-are-university-faculties-so-white/. ↑

29. Curtis L. Kendrick, “Changing the Racial Demographics of Librarians,” Ithaka S+R, 18 April 2023, https://sr.ithaka.org/publications/changing-the-racial-demographics-of-librarians.

    ↑

30. ALISE, 2022 Statistical Report. https://online.fliphtml5.com/gjade/xyno/#p=10, consulted on January 26th 2023.

    ↑

31. US Bureau of Labor Statistics, Occupational Outlook Handbook, last accessed 10 November 2022, https://www.bls.gov/ooh/education-training-and-library/librarians.htm.

    ↑

32. Sarah Flood, Miriam King, Renae Rodgers, Steven Ruggles, J. Robert Warren and Michael Westberry, Integrated Public Use Microdata Series, Current Population Survey: Version 10.0 [dataset], Minneapolis, MN: IPUMS, 2022, https://doi.org/10.18128/D030.V10.0.

    ↑