Methods

CRA gathers survey data during the fall. In the 2025 survey cycle, all eligible Taulbee Survey respondents were contacted via the Alchemer survey management platform. Most of the academic units in this report are departments, but some are colleges or schools of information or computing. Throughout this report, we use the term “unit” to refer to the department, college, or school offering the program.

Degree production and enrollment data (PhD, Master’s, and Bachelor’s) refer to the previous academic year (2024-25). Data for new students and projected degree production in all categories refer to the current academic year (2025-26). Salaries are those effective on July 10, 2025.

Survey Design & Overview

The Taulbee Survey is conducted annually by the Computing Research Association (CRA) and consists of two surveys: the Taulbee Main Survey and the Taulbee Salary Survey. Together, they provide a comprehensive view of the enrollment, production, employment, and compensation of computing faculty and students at academic units across the U.S. and Canada.

The Taulbee Main Survey (which you can find as a PDF here) collects data on:

Reporting Structure and Data Release (Section A),
Other Department Questions (Section B),
PhD Degree Titles, Awards, Enrollment, and Other PhD Questions (Sections C-F),
Master’s Preliminary Questions, Degree Titles, Awards, Enrollment, and Other Master’s Questions (Sections G1-J),
Bachelor’s Degree Titles, Awards, Enrollment, and Other Bachelor’s Questions (Sections K-N),
Faculty Positions, Tenure/Tenure-Track Faculty Numbers, Non-Tenure-Track Faculty Numbers, and Newly Appointed Faculty (Sections O-R), and
Research Expenditures and Graduate Student Support (Section S).

The Taulbee Salary Survey (which you can find as a PDF here) collects data on:

Preliminary Questions (Section A),
Full-Time Faculty Salaries (Section B),
Adjunct Rates (Section C), and
Doctoral Student Tuition Rates and Stipends (Section D).

In Section B, participating units were given the option to submit either individual-level faculty salary data or aggregate department-level data. For units submitting individual-level data, we first calculated years-in-rank based on hiring year-in-rank and role type. We then computed the average salary and headcount for each specific role and rank within that unit. These derived aggregates were then pooled with the standard aggregate submissions to calculate the final aggregated statistics.

Data Collection

Survey Distribution

This year, data collection for both the Taulbee Main Survey and the Taulbee Salary Survey was migrated to the Alchemer platform to streamline the submission process. The Salary Survey was open from October through December 2025, and the Main Survey was open from October 2025 through February 2026 following a deadline extension from the original January 24, 2026 close date.

In a departure from previous years, the 2025 cycle utilized a Pre-Registration phase, where academic units confirmed participation and designated a Data Liaison prior to the survey launch through a Pre-Registration Form. Survey access links were distributed directly to the designated Data Liaisons, replacing the previous user-login system. Data Liaisons were responsible for distributing specific survey sections to the appropriate staff members within their units using the platform’s Section Navigator, which allows the Data Liaison to forward each section to the relevant staff member with the constraint that only one staff member be assigned per section to preserve data integrity. More information on this process can be found here.

Sample & Response Rates

Of the 226 units invited to participate, 141 units submitted data to the Taulbee Main survey, yielding a total response rate of 62.3%. The responding sample includes 133 U.S. units and 8 Canadian units.

Table 1: Number of Respondents to the Taulbee Survey (2020-2025)
Year	CS	CE	CA	IN	Total
2020	150/193 (78%)	6/35 (17%)	8/29 (28%)	15/22 (68%)	179/279 (64%)
2021	142/195 (73%)	6/35 (17%)	8/29 (28%)	15/23 (65%)	171/282 (61%)
2022	146/205 (71%)	7/35 (20%)	14/34 (41%)	15/23 (65%)	182/297 (61%)
2023	144/210 (69%)	6/38 (16%)	11/36 (31%)	15/30 (50%)	176/314 (56%)
2024	129/210 (61%)	3/38 (8%)	11/36 (31%)	14/30 (47%)	157/314 (50%)
2025	120/176 (68%)	2/15 (13%)	8/19 (42%)	11/16 (69%)	141/226 (62%)

Sample Characteristics

The figures below summarize the characteristics of the responding sample for the most recent survey year. Carnegie Classification and locale percentages are computed among classified units; institution control and unit size are reported across all responding units.

Institution Control Type. A majority of the institutions that responded to the Taulbee Survey 2025 were public institutions, making up 73.8% of the sample. Private institutions made up the remaining 26.2% of the sample.

Carnegie Classification. To classify the units, we referenced the 2021 Carnegie Classification of Institutions of Higher Education dataset. The Carnegie Classifications are a system for organizing the degree-granting colleges and universities in the United States. The responding sample is concentrated among doctoral research universities (R1 and R2).

Locale. We used locale data from the 2024 Integrated Postsecondary Education Data System (IPEDS) from the National Center for Education Statistics (NCES) to classify the units that responded to Taulbee. A majority of the institutions that responded to Taulbee this year were located within cities, making up 77.4% of the sample. Units located in suburbs made up 18% of the sample, and the rest were located in towns.

Unit Size. We define unit size as the number of Assistant Professors, Associate Professors, Full Professors, Other Instructors, Teaching Professors, Researchers, and Postdoctorates. A majority of units had a unit size of 26-75 faculty members, with 53.2% of units falling into this group, followed by 1-25 faculty members (16.3%), 76-100 faculty members (15.6%), and >100 faculty members (14.9%).

Data Analysis

CRA staff downloaded the data from Alchemer for cleaning and analysis. We built a custom data cleaning pipeline in R to handle validation, transformation, and statistical analysis of the unit-level data. Tables and visualizations were generated using Quarto to provide a statistical summary of the data.

Trend Analysis

A recurring concern with multi-year survey data is that the set of responding units changes from year to year: units skip taking the survey or skip individual items within their survey response. Year-over-year comparisons on the full sample therefore can conflate true changes in the underlying quantity (e.g., enrollment, degree production) with compositional changes in who responded that year. Because Taulbee response patterns are non-random, as larger and more research-intensive units may tend to report more consistently, these compositional shifts can plausibly drive a measurable share of any apparent trend.

To address this, most longitudinal trend figures in this report are restricted to section-level cohorts, which we define as a consistently responding set of units that reported the section’s defining variables in every year of the comparison window (2020-2025). Trend statements throughout the report should therefore be read as within-unit changes. Each section ends with a “Section Cohort” block that compares cohort composition (by control, locale, unit size, and other characteristics) to the full sample for the most recent year, so readers can assess representativeness directly and adjust their interpretation of trend figures accordingly.

Section Cohorts

A unit enters a section cohort if, in every year of the window, it reported degrees awarded and enrollment at the given program level for at least one computing program type, specifically in CS, CE, or IN. We use this rule rather than a stricter “all three fields every year” rule because most units offer fewer than all three programs, and requiring all three would discard a substantial share of programs in CS-only or CS+IN units without improving the validity of within-unit comparisons in their primary field. The resulting section cohort is the intersection of these two reporting requirements applied per level, per field across every year of the window.

Section membership is a necessary condition but not always sufficient. As an example, a unit may satisfy the section cohort rule via degrees awarded and enrollment reporting in CS, but may report computer engineering enrollment only sporadically. Therefore, for trend figures that aggregate a specific variable or set of variables, we apply a second matching step per figure: the unit must also have reported that figure’s key variable(s) as non-missing in every year of the window. For multi-line figures whose lines represent independent metrics — most commonly the CS, CE, and IN totals plotted together — each line is built from its own per-variable cohort, so the unit count on each line is constant across years (though it differs between lines). Each figure’s footnote reports the exact unit count(s) contributing.

Demographic breakdowns (gender, race/ethnicity, and gender × race/ethnicity intersection plots) use a different per-figure matching rule: the major gender categories must be reported in every year, ensuring that observed shifts in the demographic composition reflect changes within a stable set of units rather than the entry or exit of units that report demographic detail.

Please note that all tables and “current year” figures (i.e., those summarizing only the most recent survey year) draw on the full responding sample rather than the cohort.