The source for these data comes from the Bureau of Labor Statistics Employment Projections (http://www.bls.gov/emp/tables.htm). The projection for new computing jobs is 597,100 from 2016-2026. Projections for all other STEM jobs combined is 433,100 over the same period. This is a 58:42 ratio of jobs in Computing versus the rest of STEM.
For STEM occupations, we use the SOC codes that the BLS defined as STEM in the “Science, Engineering, Mathematics, and Information Technology Domain” (http://www.bls.gov/soc/Attachment_A_STEM.pdf and http://www.bls.gov/soc/Attachment_B_STEM.pdf).
For computing occupations, we use all of the occupations listed under “Computer Occupations” SOC 15-1100, as well as additional individual codes in other categories that are clearly computer science occupations. Specific codes for both classifications are listed below. Note that these codes include occupations at all degree levels.
Computer Science codes:
Computer and Information Systems Managers
Computer Hardware Engineers
Computer science teachers, postsecondary
Computer and information research scientists
Computer systems analysts
Information security analysts
Software developers, applications
Software developers, systems software
Network and computer systems administrators
Computer network architects
Computer user support specialists
Computer network support specialists
Computer occupations, all other
STEM Codes (Science, Engineering, Mathematics, and Information Technology Domain)
Computer and Information Systems Managers
Architectural and Engineering Managers
Natural Sciences Managers
Mathematical Science occupations
Surveyors, Cartographers, and Photogrammetrists
17-3000 (except 17-3011)
Drafters, Engineering Technicians, and mapping Technicians
19-3000 (except 19-3093)
Social Scientists and Related Workers
Life, Physical and Social Science Technicians
Math and Computer Teachers, Postsecondary
Engineering Teachers, Postsecondary
Life Sciences Teachers, Postsecondary
Physical Sciences Teachers, Postsecondary
Social Sciences Teachers, Postsecondary
Sales Representatives, Wholesale and Manufacturing, Technical and Scientific Products
The number of STEM and Computer Science graduates comes from the National Center for Education Statistics (NCES) IPEDS Completions Survey, obtained using the National Science Foundation (NSF) WebCASPAR tool (https://ncsesdata.nsf.gov/webcaspar/index.jsp?subHeader=WebCASPARHome). Choose the table “NCES Degrees Awarded by Degree Level and Field” under “Frequently Requested Tables,” modify analysis variable for “Degrees/Awards Conferred (NCES population of institutions),” and modify classification variables for 2015 (the most recent year available), bachelor’s degrees only, and (under “institutional control (survey-specific)”), public institutions or nonprofit private institutions. The national count of CS graduates includes the U.S territories Guam, Puerto Rico, and Virgin Islands.
The classification of STEM degrees comes from two NCES tables (http://nces.ed.gov/pubs2013/2013152.pdf and http://nces.ed.gov/programs/digest/d14/tables/dt14_318.45.asp), and includes the following degrees:
Other Physical Sciences
Mathematics and Statistics
Other Life Sciences
Other Science and Engineering Technologies
Interdisciplinary or Other Sciences
According to these data, there were 580,940 bachelor’s degrees earned in STEM in 2015, and only 49,291 of those—8.48%—were in Computer Science.
The source of this data is the Gallup research study (commissioned by Google) Trends in the State of Computer Science in U.S. K-12 Schools, released in 2016 and found here: http://csedu.gallup.com/home.aspx. According to this report, 40% of principals report having at least one computer science class in which students learn programming or coding.
Prior data (1 in 4 schools) came from the report Searching for Computer Science: Access and Barriers in K-12 Education. According to this report, 3/4 of principals surveyed said that their schools do not offer courses with computer programming and coding. Of the 8811 principals who responded to the survey, 4745 said they offer computer science, 3761 said they don’t offer computer science, and 305 said they didn’t know. However, when asked about the content of these computer science courses, only about half of the 4745 principals who offer computer science said that these courses include programming—bringing the fraction of schools that offer computer science courses with programming to 1/4 (i.e., 4745 x 0.5 = 2,372 and 2,372/8,811 = 0.26). Note that we used 1/4 rather than 1/2 due to the confusion about what constitutes computer science. Programming is a critical component to include in a computer science course, and thus only those classes that have some programming should count as computer science.
The source of this data is a series of Gallup research studies (commissioned by Google). The study Trends in the State of Computer Science in U.S. K-12 Schools was released in 2016 and found here: http://csedu.gallup.com/. According to the study, 93% of parents “see CS education as a good use of resources at their child’s school” (p. 8). Previous data from the study Searching for Computer Science: Access and Barriers in K-12 Education (released in 2015) found that 91% of parents wanted their students to learn CS and 90% of parents wanted their child’s school to teach CS.
This data comes from a study by Change the Equation and C+R Research, with analysis completed by Code.org. More information about the study can be found from Change the Equation at http://changetheequation.org/students-stem-methodology. High school students were asked, for each course they have taken or plan to take, about whether they “like it a lot,” “like it a little,” “dislike it a little,” or “dislike it a lot.” When comparing computer science courses to other courses, more students like graphic arts, performing arts, and computer science courses. Code.org’s analysis is described in more detail at this link: https://docs.google.com/spreadsheets/d/15c-xozSgav1s38KcmxfLbe1CrHw4BsJdQEU3HqOFIGI/edit?usp=sharing.
The data about AP computer science exams compared to other subjects, high schools offering the exam, and participation by female, Black, Hispanic/Latino, Native American/Alaska Native, and Native Hawaiian/Other Pacific Islander students comes from the College Board National and State Summary Reports: http://research.collegeboard.org/programs/ap/data.
We consider the following exams STEM: Biology, Calculus AB, Calculus BC, Chemistry, Computer Science A, Computer Science Principles, Environmental Science, Physics 1, Physics 2, Physics C: Elec. & Magnet., Physics C: Mechanics, and Statistics.
The net present value for lifetime earnings for high school graduates, college graduates, and computer science majors is from The Hamilton Project (Brookings): http://www.hamiltonproject.org/charts/career_earnings_by_college_major/
The number of current open computing jobs comes from the sum of the per-state jobs data from The Conference Board’s Help Wanted OnLine®service (see below for more details on the data from The Conference Board).
The projected rate of this job growth comes from the Bureau of Labor Statistics Employment Projections data (http://www.bls.gov/emp/tables.htm) for 2014-2024, released in 2015. These data predict an employment change of 12.5% for computing occupations, and an employment change of 6.5% for all occupations. For computing occupations, we use SOC codes 11-3021, 15-1100, 17-2061, and 25-1021 (see more details on these codes in the table above).
This comes from the number of job openings in April 2016 as reported by The Conference Board’s Help Wanted OnLine®service (click here for more information about HWOL and their data collection methods). The number of job openings in each category were multiplied by the average salary (from the Bureau of Labor Statistics 2015 OES data). For computing occupations, we use SOC codes 11-3021, 15-1100, 17-2061, and 25-1021 (see more details on these codes in the table above).
The sources for this infographic are the College Board (2015 AP Computer Science A exams), the National Center for Education Statistics (college graduates by degree in 2015), and the Bureau of Labor Statistics Current Population Survey (people employed in computing occupations in 2017). From these sources, females represent 22% of AP CS A exam-takers, 18% of CS bachelor’s degrees, and 24% of people employed in computing occupations. Underrepresented minorities represent 13% of AP CS A exam-takers (9% Hispanic, 4% Black), 18% of CS bachelor’s degrees (9% Hispanic or Latino, 8% Black, less than 1% American Indian or Alaska Native), and 15% of people employed in computing occupations (8% Black, 7% Hispanic).
Women who try AP Computer Science in high school are ten times more likely to major in it in college comes from a 2007 research study by the College Board (http://research.collegeboard.org/sites/default/files/publications/2012/7/researchreport-2007-4-ap-students-college-analysis-five-year-academic-careers.pdf). This research study also revealed that students who take AP Computer Science in high school are 6 times more likely to major in computer science than those who do not, and Black and Hispanic students are 7 or 8 times more likely.
The original source of this data was the ACM Running on Empty report: http://www.acm.org/runningonempty/. However, thanks to the advocacy efforts by Code.org and sister organization Computing in the Core, the list of states that allow computer science to count towards graduation credit has increased monthly, and at this point Code.org is the definitive source of the data. The latest list of states is reflected at http://code.org/action
These data come from the Horizon Media’s WHY group survey summarized here: http://www.prnewswire.com/news-releases/horizon-media-study-reveals-americans-prioritize-stem-subjects-over-the-arts-science-is-cool-coding-is-new-literacy-300154137.html. The group reported that “three in four Americans agree that ‘today, science is cool in a way that it wasn’t ten years ago.’ And computer science is a major driver of this new perceived cool factor -- with 73% agreeing that ‘in the future, all the best jobs will require knowledge of computer coding languages.” And “when asked which two subjects other than reading and writing are critical to ensure the next generation is prepared for the future, 70% chose math, and 50% said computer science. More than two thirds (65%) went as far to agree that ‘most students would benefit more from learning a computer coding language than a foreign language.’ ”
The source of this data is the Gallup research study (commissioned by Google) Searching for Computer Science: Access and Barriers in K-12 Education, released in 2015 and found here: http://csedu.gallup.com/. When asked, “Do you think most students should be required to learn computer science?” 67% of parents answered yes (p. 24). Teachers were asked how much they agreed with the statement “Most students should be required to take a computer science course” on a 5-point scale from strongly disagree to strongly agree. 33% of teachers answered strongly agree and 23% of teachers answered agree, for a total of 56% of teachers. Demand is highest among parents of low-income students (p. 13), and teachers of low income students (see the related Gallup research report, Images of Computer Science: Perceptions Among Students, Parents and Educators in the U.S., p. 22, http://csedu.gallup.com/).
According to an analysis by the National Association of Colleges and Employers, graduates with computer science degrees earn the second highest starting salaries (just after mechatronic engineering graduates). Data from 2018 can be found at http://www.naceweb.org/job-market/compensation/the-top-paid-majors-for-the-class-of-2018/ and 2015 at http://www.naceweb.org/job-market/graduate-outcomes/first-destination/class-of-2015/
Another analysis by Forbes in 2013 reported that the best-paying degree in the USA was a computer science degree from Carnegie Mellon. http://www.forbes.com/pictures/mkl45kkeg/1-carnegie-mellon-school-of-computer-science/#550c690b174f
According to this IDC study in 2014, or this easier to read summary, there are 11M software professionals in the world. 19.2% are in the U.S., which means 2.1M software professionals in the US. According to NCWIT, 26% of these professionals are female, which is about 550,000.
This statistic was included in the MSFT National Talent Strategy document and taken from a Georgetown University Center for Education and the Workforce Report on STEM (October 2011) by Anthony Carnevale, Nicole Smith, and Michelle Melton - see http://cew.georgetown.edu/stem/. The relevant quote is: "Computer occupations are the most widely represented across industries. For example, 9 percent are in Information Services, 12 percent are in Financial Services, 36 percent are in Professional and Business Services, 7 percent are in Government and Public Education Services, and 12 percent are in Manufacturing". (12 + 36 + 7 + 12 = 67%)
At Code.org, we do not count unique student IDs perfectly when tracking participation in the Hour of Code. Why? Partly because we don’t want the friction of prompting to “login / register” before a student or classroom tries learning for the first time, and partly because there are many activities we cannot track online. We do take certain steps to reduce double-counting, but without a login prompt, this can’t work perfectly. At the same time, there are MANY student activities in the Hour of Code that aren’t tracked at all. For example: (1) students who use a mobile/tablet app to try the Hour of Code are typically not counted, (2) students who share a screen for pair-programming or group-programming may be counted as one, (3) students trying an unplugged classroom activity cannot be counted online, and (4) teachers who create their own Hour of Code activities aren’t tracked. As a result, there is some under-counting and some double-counting, and so we do not view the Hour of Code tracker to be an exact measure of usage. It is certainly directionally correct, and shows that many tens of millions of students have participated. And our “lines of code” counter tracks very real usage in our learning platforms.
The source data is from the US Dept of Labor, Office of Foreign Labor Certification: http://www.foreignlaborcert.doleta.gov/pdf/H-1B_Selected_Statistics_FY2014_Q1.pdf.
This is how we classified the job types:
computer systems analyst
software applications developer
computer occupations other
software developers - system software
computer information system manager
network/computer system administrator
The average salary for a computing occupation versus the average salary in the state is from the Bureau of Labor Statistics May 2016 State Occupational Employment and Wage Estimates (found at http://www.bls.gov/oes/current/oessrcst.htm). For average state salary, we use “Annual mean wage” for all occupations. For average salary in computing occupations, we calculate the weighted arithmetic mean of all computing occupations (BLS codes 11-3021, 15-1100, 17-2061, and 25-1021) using the “annual mean wage” and total employment for each occupation code. That is, instead of simply finding the mean of the four occupation codes, we multiply the average salary for a given code by the number of people employed in that occupation and divide the sum of these salaries by the total number of people employed in computing occupations.
The number of open computing jobs in each state comes from The Conference Board’s Help Wanted OnLine®service (click here for more information about HWOL and their data collection methods). It represents the number of open jobs in the previous month (seasonally adjusted) for Bureau of Labor Statistics’ (BLS) Category SOC “15-0000 Computer and Mathematical Occupations”). This is a conservative estimate of the number of computing occupations as it excludes three BLS categories that include computing occupations: Computer and Information Systems Managers 11-3021, Computer Hardware Engineers 17-2061, and Computer Science Teachers, Postsecondary 25-1021. However, the 15-0000 SOC also includes some mathematical occupations that are not considered computing occupations. (This is due to limitations with our agreement with the Conference Board.) This data is cross-sector.
The comparison to the state average demand rate is comparing the job demand (% open jobs / # of existing jobs determined in the May 2015 BLS’s OES survey) in computing occupations vs the state average.
The national jobs data is the sum of the 50 states plus D.C.
The number of CS graduates in a given state and the percent female comes from the National Center for Education Statistics (NCES) IPEDS Completions Survey, obtained using the National Science Foundation (NSF) WebCASPAR tool. The table “NCES Degrees Awarded by Degree Level and Field” was formatted for “Degrees/Awards Conferred (NCES population of institutions),” 2015, bachelor’s degrees only, and public institutions or private institutions: nonprofit.
Number of Hours of Code completed in a state comes from our Numbers of Hour of Code data (see above). IP addresses are used to determine the state.
The reported number of Code Studio accounts (by teacher and by student) includes all accounts that have been created and that have logged in at least once. IP addresses are used to determine the state.
48% of our students come from traditionally underrepresented minorities. This includes Black, Latinx, Hispanic, Native Hawaiian or Other Pacific Islander, American Indian or Alaskan Native. 49% of students are on free and reduced meal programs (FARM). To protect the privacy of our youngest students, we measure the diversity for students under age 13 for the entire classroom by surveying teachers. This means these numbers are based on teachers’ estimates of the actual student ethnicities. For older students, the students self identify their race.
Similarly, for student privacy, we do not ask individual students if they are on free or reduced meals. Instead, we have an optional survey for teachers. This means these numbers are based on their knowledge of which students have subsidized meals. To protect privacy, our surveys are optional and do not represent all Code.org teachers.
Furthermore, this survey method doesn’t reflect the ethnicity of students under age 13 who are using Code Studio at home, without a classroom teacher. Our organizational focus is on bringing computer science into K-12 schools, and that is also what we are measuring.
Lastly, our ethnicity surveys do not measure international diversity because our focus is the U.S. and ethnic questions would be different outside the U.S. Similarly, the free and reduced meal program is specific to the United States. We do not measure similar programs internationally.
Our “45% female” measure of gender diversity in CS Fundamentals courses on Code Studio is based on student accounts, and thus represents all active Code Studio students worldwide. This number is updated annually and reflects active student accounts for the previous year.
The previous version of this document (prior to early 2015) can be found here.