ERIC ED502709: Latino Achievement in the Sciences, Technology, Engineering, and Mathematics. ETS Policy Notes. Volume 14, Number 2, Summer 2006

2006·0.97 MB·English

by  ERIC

POLICY NOTES
News from the ETS Policy Information Center
Volume 14, Number 2 Policy Evaluation & Research Center Summer 2006
The Problematic Pipeline This issue — HigHligHts
ets s
from tHe ymposium
Providing a context for the theme of the conference, Indiana
l a
on atino cHievement in
University’s Jorge Chapa, professor and Director of the Latino
s , t ,
tHe ciences ecHnology
Studies program, spoke of “The Problematic Pipeline: Demographic
e ,
ngineering and
Trends and Latino Participation in Graduate Science, Technology,
m .
atHematics
Engineering, and Mathematics Programs.” Chapa and Belinda De
La Rosa, research analyst, coauthored the research, which posed For more than two decades,
the question: will Latino population growth translate into increased ETS, the College Board, and the
Hispanic Research Center at
participation in higher education, and particularly in graduate science,
Arizona State University have
technology, engineering, and mathematics (STEM) programs?
worked together to promote the
Chapa described a leaky pipeline, where graduate degree recipients success of Latino students. To
from the nation’s colleges and universities do not reflect the racial that end, the organizations joined
and ethnic diversity of the population. Demographic and educational forces on November 20-22, 2005
trends confirm that the percentage of Latinos decreases at each in Princeton, NJ, to convene an
higher education stage in the pipeline. invitational conference devoted
to Latino Achievement in the
Figure 1: Latino Demographic and Education Trends
Sciences, Technology, Engineering,
and Mathematics (STEM). The
20
conference had three major goals:
17.5 1980 1990 2000
• explore the challenges, practices,
os15 14.2
n policies, research, and advocacy
Lati 12.5 11.9 at the national, state, and
on of 10 9 10.8 9.9 institutional levels required for:
porti 6.4 6.8 6.5 8.1 6.6 6.3 (1) improving preparation of
Pro 5 4.5 4.35.3 4.2 3.8 Ltoa tcinolole sgteu;d (e2n)t es nfhora naccicnegs sth e
3.2
2.4 2.5
1.6 achievement, retention, and
0 persistence of Latino college
Total Population High School Two Associate’s Four Bachelor’s Ph.D.s
Population 18 to 24 Graduates Years Degrees Years Degrees students; (3) improving
College College
preparation of Latino students for
Source: U.S. Bureau of the Census and National Center for Education Statistics graduate and professional school
data as presented by Chapa and De La Rosa.
Listening. STEM programs; and
Note: Data for two years of college and four years of college are not available for 1980.
Learning.
Leading. —continued on page 
Dr. Alfredo G. de los Santos Jr., Research Professor at the Hispanic Research Center of Arizona State University, was a member of the Conference Planning
Committee. He served as conference convener and master of ceremonies.
Latinos are a young population, Chapa noted. group in several states and many cities. They
Many are still in their childbearing years and are moving to all regions of the United States,
many come from backgrounds with factors increasing their numbers in areas that previously
generally thought to limit educational success. had relatively few Latinos. The Latino populations
Population counts from the 2000 Census indicate of North Carolina, Arkansas, and Georgia, for
that the Latino population grew by more than example, all increased by more than 300 percent
57 percent between 1990 and 2000, compared to between 1990 and 2000. Despite this trend toward
a 13 percent increase for the total population. It geographic dispersion, the major portion of the
has continued this rapid growth since the 2000 Latino population is concentrated in just a few
Census: While the rate of growth for the U.S. states: California and Texas are home to half of
population as a whole was 2.5 percent, the Latino the national Latino population.
population grew 9.8 percent. As has been true
Turning his attention from overall population
since 1980, about half of Latino growth was due
trends, Chapa examined the educational pipeline,
to international migration, and the other half due
starting from K-12 and moving through doctoral
to natural increase. By all projections, the Latino
study. He stated that only a relatively small
population will continue to grow at a much
proportion of all Latino high school graduates
faster rate than the U.S. population for many
are undocumented, and that legal status in itself
more decades.
is not the most widespread reason that Latino
As the largest minority group in the United States, high school graduates, in general, have difficulties
Latinos are fast becoming a majority population entering higher education. Latinos are relatively

well represented at the community college level, education (1,157 vs. 1.5 million) suggests that the
where in 2000 they constituted 14.2 percent of all number of future Latino Ph.D.s could increase
community college students and earned close to greatly. However, this will depend more on the
10 percent of the Associate of Arts (A.A.) degrees future we create than on any discernible trend. By
granted. Accessing higher education through addressing the pervasive achievement gap and high
enrollment in community colleges is a path drop out rates for Latino students, we can make sure
often chosen by first-generation college students, that they persist in schooling at the postsecondary
although transfer to four-year programs remains level and graduate to join an educated labor force
problematic for Latino students. Chapa noted that able to compete in a global economy.
opportunities to continue beyond the A.A. degree
are often missing because of miscommunication
and misunderstandings regarding prerequisites
for majors, degrees, and transfers. He suggested
that efforts to increase Latino Ph.D. production
Strengthening the Pipeline
could start with focusing attention and resources
on community college students. Patricia Gándara, Professor of Education at
the University of California, Davis, leads the
Although the increases in Latino postsecondary
Education Policy Studies program at that
degree attainment have been most pronounced
campus. Her research focuses on access to higher
at the associate degree level, there has also been
education for low-income and minority students,
growth at the higher levels. Chapa focused further
and on the education of English learners. Her
on science and engineering (S&E), from 1994 to
paper, “Strengthening the Academic Pipeline
2001. He noted that a small group of institutions
Leading to Careers in Math, Science, and
granted the most S&E bachelor’s degrees to
Technology for Latino Students,” took a closer
Latinos. He also commented that if only a few
look at the achievement gap issues within the
other institutions were to become similarly
pipeline – their origins and their persistence – and
productive, the number of Latino S&E bachelor’s
offered some bold proposals to address
degrees would increase sharply.
the problems.
A 31.6 percent increase in the attainment of
To strengthen the pipeline, she said, we must
doctoral degrees by Latinos between 1994 and
begin at the very beginning: focusing on broad
2001 could be viewed as a cause for celebration,
academic skills, attending to language issues, and
until we see that the number of S&E doctoral
realizing that school reform efforts alone will not
degrees awarded to Latinos is still only in the
be sufficient to close the gaps. Latino children
hundreds. Chapa stated that this pipeline might
enter school significantly behind their White and
be better termed a “pipette,” since its yield is
Asian peers, in large part because of nonschool
so small. Added to this is the fact that Latino
factors that affect achievement. Latino students
males are severely underrepresented in higher
are poor – 28 percent of them live in poverty – and
education, and all the more so at the doctoral
Latino parents have the lowest overall education
level. Chapa cautioned, too, that among Latinos,
levels of any ethnic group in the United States.
42 percent of S&E Ph.D.s granted in 2001 went
A strong sense of family, as well as linguistic and
to temporary visa holders, rather than to Latinos
cultural traditions, combine to make many Latino
educated in the United States.
parents reluctant to send their young children
Chapa concluded that the extremely low numbers
to preschool programs. As a result, they enroll
of Ph.D.s granted to Latinos compared to the
their children in kindergarten at younger ages,
very large number of Latinos enrolled in higher
placing them at a further disadvantage compared

—continued from page 1 to children who have the benefit of preschool and more time to
() increasing the supply of Latino develop language and learning skills.
prospects for STEM tenure-track faculty
Gándara reviewed early intervention efforts, citing research
• create a vision for comprehensive conclusions that such programs can have a long-term effect
programs of research and policy on cognitive functioning and achievement levels, as well as
initiatives to increase the number of intellectual, social, and developmental outcomes for children.
Latino students pursuing STEM careers Without these interventions, achievement gaps open early and
remain or grow worse. Figure 2 shows data from the Early
• strengthen the partnership of the College
Board, ETS, and the Hispanic Research Childhood Longitudinal Survey that illustrate the achievement
Center of Arizona State University in gap in reading and math at the beginning of kindergarten. Latino
focusing on expanding the success of children are underrepresented among the top scorers. This gap
Latino students persists into high school, where SAT data show that the average
verbal and mathematics scores of the top Latino quintile are
Representing a wide network of scholars
598 and 646, respectively, compared to 663 and 729 for the top-
and policy-makers who focus on
scoring White quintile.
Latino achievement issues, conference
participants came from 18 states, 21
K-12 interventions, according to Gándara, are almost always
colleges and universities, 17 national
directed at whole-school improvement. They focus on reforming
and regional organizations, and 5 local
schools that enroll large numbers of poor and ethnic minority
and state governments. The conference
students, rather than on raising the performance of particular
was built around six commissioned
groups of students. Whether or not these interventions are
papers by prominent scholars, including
effective is difficult to say, however, since most studies of school
Jorge Chapa (Indiana University); Patricia
reform do not study schools for a long enough time or employ
Gándara (University of California – Davis);
adequate comparisons to be able to draw solid conclusions.
David A. Payne (ETS); Michael T. Nettles
and Catherine M. Millett (ETS); Jamie P. As a contrast to the school reform model, Gándara reviewed
Merisotis (Institute for Higher Education three intervention programs targeted at Latino students: High
Policy); and Gary D. Keller and Antonio School Puente (a college-access program); ALAS (a middle
A. Garcia (Arizona State University). For
each paper, two or three policy-makers or
Figure 2: Percentage of Kindergartners in the Highest Reading and
practitioners, knowledgeable about Latino
Mathematics Quartiles, by Racial/Ethnic Group, Fall 1998
educational issues, participated as reactors
who offered their opinions and engaged the
39
Asian American
audience in full discussion of the ideas set
38
forth in the research. See the box on page 5
for a list of the reactors.
30
White
32
The Conference Planning Committee
was drawn from the three sponsoring
organizations: the College Board, ETS, and 15
Latino
the Hispanic Research Center at Arizona State 14
University. Five of the research papers have
appeared in the Journal of Hispanic Higher Black 15 Reading Mathematics
10
Education, Volume 5, Number 3 (July 2006).
This issue of ETS Policy Notes presents an
0 10 20 30 40 50
overview of the conference and summaries of
Percentage
its six papers for a broader audience.
Source: U.S. Department of Education, National Center for Education Statistics,
America’s Kindergartners, 2000.

The reactors included the Honorable Ruben Hinojosa, U.S. Congressman (TX-15); Brian L. Foster, Provost, University of
Missouri – Columbia; Raymund A. Paredes, Commissioner of Higher Education, Texas Higher Education Coordinating
Board; Amaury Nora, Professor and Director, National Center for Hispanic Student Success, University of Houston; Richard
P. Duran, Professor, University of California, Santa Barbara; Antonio R. Flores, President, Hispanic Association of Colleges
and Universities; Jeanett Castellanos, lecturer, Department of Social Sciences and Chicano/Latino Studies and Director,
Social Science Academic Resource Center, School of Social Sciences, University of California – Irvine; Richard O. Hope,
Vice President, The Woodrow Wilson National Fellowship Foundation; Raul Yzaguirre, Presidential Professor in Practice
in Community Development and Civil Rights, Arizona State University; Richard Kazis, Vice President, Jobs for the Future;
Roosevelt Y. Johnson, Director, Division of Human Resource Development, National Science Foundation; and Richard A. Tapia,
University Professor, Department of Computational and Applied Mathematics, Rice University.
school intervention aimed at reducing drop out the pipeline to college and STEM careers, it will
and truancy in a school serving almost all Latino not be enough to close the considerable gaps in
students); and AVID (a college access program opportunity and outcomes for these students.
serving a variety of students, but where Latino
According to Gándara, one finding that recurs in
effects could be disaggregated). Each of the
many studies of specific interventions is that the
programs had been able to prove empirically
longer a student is exposed to the “treatment,”
that outcomes for participants were significantly
whether it is school reform or individualized
greater than for nonparticipants. Certain features
enrichment, the better the outcomes appear to
were common among the three programs: (1)
be. She suggested that rather than segmenting
intensive monitoring of program participants
intervention efforts by level of schooling, such
by an adult; (2) an articulated program that
programs should reach across school sectors
reached across grade levels and extended over
to provide continuous support for students,
the secondary years; (3) consistent messages to
serving cohorts of students from 7th or 8th grade
raise aspirations; (4) building of group cohesion
through high school graduation. Additionally, they
among peers and a sense of membership in the
should provide services that promote academic
school and the program; and (5) in the case of the
performance in core areas including math,
two college-access programs, access to a rigorous
information about the necessary costs to attend
curriculum and scaffolded support to succeed in
college, and professional development for teachers
that curriculum.
and community members in the program.
Access to a rigorous curriculum continues to
Strengthening the pipeline to college and beyond
be a critical issue for Latino and other minority
for Latino students will require far more than
students. Gándara cited evidence that Latino
bringing in the cavalry during high school and
and other underrepresented students are more
offering special college-access programs to
likely to be assigned to low curriculum tracks
inspire, prepare, and guide them into college.
independent of their test scores than are White
And, moving them into the math/science pipeline
students. This is especially true in the case of
will require a broad strategy that extends beyond
mathematics courses, where algebra continues to
those disciplines to preliteracy skills. Some
be the major gatekeeper for entry into the college-
researchers argue that neural patterns established
preparatory track. Certainly Latino students will
in the highly sensitive early years of life may be
not go on to STEM careers if they do not have
impossible to modify with any of the interventions
rigorous preparation in math; but without such
we have devised to date. Others are more
preparation they are not likely to go on to college
optimistic about the plasticity of the human brain
at all. Thus, while school reform aimed at offering
and argue that intensive intervention can change
greater academic opportunities to Latino students
the course of development. No one, however,
must form part of the solution to the leaks in

argues seriously that such a thing can student population in the STEM fields at both
be accomplished without significant investment the undergraduate and graduate levels. Minority
of resources. students, in particular, are not enrolling in
numbers proportional to their presence in the
What would it cost to really make a difference?
overall population. From the business community
Given the data we have on the costs of
and from the federal government there are
interventions, it is reasonable to assume that we
calls for increasing the number of STEM
would have to invest twice the national average
graduates, and the academic community is
of per-pupil expenditures the nation now invests
calling for additional STEM faculty, especially
in K-12 education, at least for the early formative
minority faculty.
years (from birth to age 3) where intensive
intervention is necessary, and half of that (or 1.5 The GRE General Test is one of several measures
times our current K-12 investment) from age 3 used by graduate admissions committees to
to high school graduation. Such an investment determine which students are most likely to
would provide a seamless web of support and be successful in their programs. It contains a
enrichment for those students whose families Verbal Reasoning (VR) section and a Quantitative
cannot provide the resources needed to even the Reasoning (QR) section. While the VR does
playing field for them. a respectable job in differentiating among
applicants, the current QR measure, based
on high school-level mathematics, does not
differentiate among applicants that cluster at
the top of the scale. Currently, 50 percent of
international examinees score between 760 and
Developing an Assessment
800 on the GRE Quantitative measure. For all
of and for Learning
applicants (i.e., domestic and international),
David G. Payne, Executive Director of the the median score for the QR is well over 700 for
Graduate Record Examinations® Program students in engineering and the physical sciences.
at ETS, discussed the paper he coauthored This means that for many STEM programs, the
with Jacqueline Briel, John Hawthorn, and GRE Quantitative measure does not provide
Karen Riedeburg: “A GRE® Test for the STEM useful information to admissions committees.
Disciplines: Developing an Assessment OF and Because of this, there is a growing concern raised
FOR Learning.” To set the context, Payne reviewed in the STEM disciplines about the relevancy of the
the challenges facing the STEM fields on the GRE for admissions purposes. Other issues center
institutional, national, and international levels, as around the achievement gaps recorded on the QR
well as the challenges for the GRE program. by underrepresented minorities, women, and first-
generation college students. These pose a major
Among the international challenges, he cited
challenge for members of those groups who
increased competition for international students,
wish to enter and complete graduate degree
greater opportunities for STEM students and
programs in the STEM disciplines. Thus, there
faculty in universities worldwide, and increased
are concerns both with the measurement at the
competition for STEM graduates in industry
top of the scale and the achievement gap for
and the academy. He noted the growth in the
underrepresented minorities.
United States of foreign-born science and
engineering (S&E) doctorates, foreign and In response to these concerns, ETS is considering
foreign-born tenured and tenure-track faculty, the development of a new GRE® STEM test.
and the slow increase of the overall domestic This effort would support the goal of broadening

participation in the STEM fields while, at the • Developing the test and learning scaffolding
same time, providing admissions information to
• Effectively utilizing the evidence-centered-
graduate programs that will promote improved
design and scale-score anchoring procedures
student success. If ETS develops the GRE STEM
to improve student learning
test, it will be important to keep in mind that
the new test may reveal a large achievement gap Evidence-centered-design (ECD) seeks to clarify
between minority and majority students, between what is being measured by a test, support
first-generation students and those whose parents inferences made on the basis of evidence derived
have college educations, and between male and from the test, and develop claims about what
female students. But ETS is proposing to link the test taker knows and is able to do. It allows
the assessment of learning with assessment for test developers to make explicit the claims about
learning with this test. The combination of a new students’ knowledge, skills, and abilities that
GRE STEM test and a well-articulated learning would be useful in evaluating applicants across
scaffolding that supports student progress all of the STEM disciplines. The score scale
toward mastering critical skills will represent an anchoring process is used to confirm the validity
of an operational assessment and strengthen its
important contribution to graduate programs
links to curriculum. The process involves selecting
and to the faculty and staff working with
questions that anchor at given locations on the
undergraduate students.
score scale.
Payne outlined earlier GRE efforts to develop
It will be challenging to link the assessment
a more demanding mathematical reasoning
specifically to undergraduate curriculum in
assessment, and the lessons learned from those
the various disciplines, but this is something
previous attempts. He described current thinking
ETS has done frequently with other tests.
and alternatives under consideration that would
Much more difficult, however, will be the next
increase the level of presupposed mathematical
step: feeding this information about what it
knowledge and increase the complexity of the
takes to succeed in graduate school back to the
reasoning involved in answering the questions.
undergraduate curricula to strengthen STEM-
Since the current Quantitative measure on the
discipline programs nationwide. To do this,
GRE test already presents a daunting challenge
ETS will partner with the higher education and
to candidates with a weak mathematical
graduate communities, working with faculty in
background, the remedy almost certainly requires
STEM disciplines who are involved in graduate
offering a separate measure for STEM candidates
education and with faculty and staff involved in
and continuing to offer the current measure
teaching, tutoring, and mentoring undergraduates
to others. Payne described the options under
in the STEM fields, particularly those who work
consideration for a new assessment and discussed
with underrepresented minorities, first-generation
the pros and cons of each alternative.
students, and women. The idea is to see where the
According to Payne, there are three major steps
skills measured in the GRE STEM test are covered
involved in developing a GRE STEM assessment
in the undergraduate curriculum and in tutoring
that will aid admissions committees and help
programs. As experience and data from the STEM
undergraduate students as they prepare for
test are gathered, ETS staff will have a better
careers in the STEM fields:
understanding of what areas need strengthening,
• Identifying the specific skills that need to and these statements of skill areas can be used
be assessed for admissions and student aid as the basic learning scaffolding to effect change
decisions in the STEM disciplines in undergraduate programs and campus tutoring

services. The goal here is to feed back information Nettles and Millett identified four critical aspects
on performance on the GRE STEM to help with of the doctoral student experience and then
improving learning at the undergraduate level. examined the Hispanic STEM experience through
that prism. The four aspects are:
• Research assistantships
• Faculty mentors
• Research productivity and publication, and
Expanding and Cultivating the
Hispanic Doctoral Workforce • Doctoral degree completion
From their extensive research on doctoral student Research productivity and degree completion
experiences, Catherine M. Millett, ETS Research can be seen as outputs produced from the
Scientist and Michael T. Nettles, ETS Senior resources (funding and mentoring) invested in
Vice President and Edmund W. Gordon Chair acquiring a doctoral education. Because funding
for Policy Evaluation and Research, focused and mentoring (inputs) influence research
on funding, mentoring, publishing, and degree productivity and degree completion (outcomes),
completion for Hispanic doctoral students. They they are examined to determine what level of
also described some of the social and academic influence they each have on these outcomes.
challenges that confront Hispanic students along In other words, they function as both inputs
the way to a doctoral degree. and outcomes.
Offering statistics supported by several of the By focusing on these four critical experiences,
conference speakers, Millett and Nettles noted the researchers hope to contribute to the national
that Hispanic individuals may have the potential dialogue on how best to prepare entering doctoral
to have the greatest effect upon changing students for their doctoral program experience, as
the demographic makeup of the U.S. STEM well as to consider what elements are critical for
workforce over the next 50 years. The estimated all doctoral students.
41.3 million Hispanics in the United States now
Equitable access to these critical experiences
constitute one-seventh of the total population,
is central to the process, so that all students
and the Hispanic population is growing at a rate
— regardless of race/ethnicity, gender, and
of 3.6 percent per year, compared to the overall
personal background — have the opportunity
population growth rate of 1 percent. The national
to participate. Nettles and Millett’s recent book,
imperative to diversify the doctoral-trained
Three Magic Letters: Getting to Ph.D., is based on
workforce in the STEM fields is also a policy
a survey of more than 9,000 doctoral students in
priority for the U.S. Department of Education, as
five major fields of study (education, engineering,
reflected in its designated areas of national need.
humanities, science, mathematics, and the social
Doctoral degrees often lead to teaching positions, sciences) from 21 doctoral granting institutions.
but the numbers for Hispanics are small. In the Two questions guided the analyses of the
STEM fields, Hispanic faculty represented from Hispanic STEM doctoral experience: (1) What
2 to 4 percent of all faculty. In 2003, Hispanic are the differences between Hispanic and White
citizens earned just 1,270, or 3 percent, of the doctoral students in having research assistantship
total doctoral degrees awarded in the United opportunities, in faculty mentoring, in research
States. Within the STEM fields, the numbers are productivity, and in degree completion? (2)
even smaller. In mathematics, for example, a What contributes to these differences? They
total of 516 doctoral degrees were awarded, with also examined issues that various constituents
Hispanics earning just 16. of the graduate enterprise can address either

individually or as a community. While the four student had done at least one of four particular
experiences listed above pertain to individual measures: presented a paper, published an article
students, they can be influenced and shaped by in a refereed journal, published a chapter in an
the larger graduate community. edited book, or published a book. Half of the
students had research productivity. More than
Research Assistantships. For many students,
a third had presented a paper at a conference,
an enduring concern is financial. The prevalence
and slightly less than a third had published a
of research assistantships varies by field of study,
refereed article. Publishing articles was more
with students in engineering (82 percent) and
prevalent in engineering (47 percent) and
science and mathematics (69 percent) having
sciences and mathematics (44 percent). Overall,
a greater opportunity for assistantships than
Hispanic students reported lower publishing rates
students in the humanities (33 percent) or
compared to White students (23 percent vs. 30
education (28 percent). When the researchers
percent). Within field, however, Hispanic students
compared Hispanic students to White students
had similar experiences of presenting a paper
across all fields, they found that Hispanic students
(either sole- or joint-authored) compared to their
were disadvantaged with respect to being research
White peers.
assistants (44 percent vs. 51 percent). However,
in engineering and science and mathematics, Degree Completion. The sample included
Hispanic students are not disadvantaged students who completed at least their first year
compared to White students in receiving of doctoral study, the assumption being that
research assistantships over the course of their some of the initial disharmony between students’
doctoral programs. expectations and the reality of being a doctoral
student would have been resolved. For the sample
Having a Faculty Mentor. Overall, 70 percent
of doctoral students beyond the first year, 62
of doctoral students report having a mentor. In
percent earned their doctoral degrees by 2001.
the aggregate, when examining differences by
Engineering and sciences and mathematics
race/ethnicity, Hispanic and White students were
have the highest completion rates among the
comparable (67 percent vs. 71 percent), and this
five fields of study (75 percent and 72 percent).
pattern holds when comparing Hispanic with
Looking at the sample as a whole, Hispanic
White students in the five fields of study.
students have lower degree-completion rates
Research Productivity and Publishing. than White students (50 percent vs. 61 percent),
Considering that one of the goals of graduate and by field, Hispanics were less successful than
school is to train future researchers in industry, their White peers in engineering (56 percent
government, or the academy, Nettles and Millett vs. 79 percent). In further analyses by field to
examined the acquisition of skills and human examine the predictors of degree completion,
capital during the training process. Their goal engineering was the only field in which race/
was to learn about students’ exposure to the ethnicity was associated with degree completion.
research enterprise and the opportunity they had In engineering, Hispanic, African American, and
to participate. Research productivity was broken Asian American students were each less likely
into its component parts, and students were asked to complete their degrees compared to their
about their participation in more than 20 different White peers (6.7 times, 2.5 times, and 1.8 times,
types of scholarly activities. They received credit respectively).
for sole-authored or joint-authored work. An
In summary, the good news is that there appear
aggregate measure of research productivity was
to be relatively few differences in the experiences
also developed, taking into account whether a
of Hispanic and White students in engineering

and sciences and mathematics. The one troubling Overview
exception, even when all appears to be equal, The collaboration was forged in 1979, a time
is that Hispanic students have a lower degree- when the merits of standardized testing
completion rate in the field of engineering. But were being challenged. The Hispanic Higher
even if the experiences of Hispanic and White Education Coalition (HHEC) and other groups
STEM doctoral students are similar, there still representing Latinos were vocal on the efficacy
remains much to do to increase the number of of standardized tests and their appropriateness
Hispanic doctoral degree recipients in the STEM for use with linguistically diverse communities.
fields. Working with faculty and administration on The HHEC made several recommendations to
doctoral campuses, education policy-makers can the testing industry that led to the development
help reduce the obstacles that stand in the way of of standards and practices for fairness in test
the growing population of Hispanic youth. development, administration, and scoring. For
example, ETS led the industry in responding to
the fairness issue by creating a procedure called
differential item functioning (DIF). In 1982, the
College Board responded positively to the HHEC
recommendation that the distribution of scores
It’s Not Rocket Science!
by race and ethnicity be published. The HHEC
Providing details and personal insights into the concluded with an expression of support for the
25-year collaboration between the College Board, testing industry and for objective tests, including
ETS, and the Hispanic Research Center at Arizona standardized tests, although at the same time
State University (ASU), Gary Keller offered a it fully supported the movement toward more
three-part account of the collaboration. His paper accountability in testing.
was titled, “It’s Not Rocket Science: 25 Years of
The responsiveness of both organizations to
Successful Latino Student Advancement … and
HHEC’s recommendations resulted in positive
the Prospects for 25 Years More.” Keller, Regents’
changes in policy and disclosure, as well as a
Professor and Director of the Center, authored
wide-ranging collaboration among the College
the paper with his colleague, Antonio G. García,
Board, ETS, and the Hispanic professional
Associate Dean in ASU’s College of Engineering
community, the latter headquartered since 1986
and Associate Director of the Center.
at the Hispanic Research Center of Arizona State
Keller shaped his remarks to cover a historical University (ASU).
overview, development of a national academic
The interactions were informal from 1979 to
intervention project to recruit Latino students
1984, when the HHEC, College Board, and ETS
into STEM fields and retain them from
agreed to establish a formal partnership. The
secondary school through the Ph.D., and a call
HHEC polled its organizational members on
to action. According to Keller, the collaboration
the priority issues, and this helped establish an
demonstrates that educational aptitude and
agenda for the three organizations for practical,
achievement tests can be and have been used
interventionist projects and for research studies.
productively for the advancement of Latino
At the same time, the College Board and ETS
students. This has been especially true for the
each set aside an annual budget for work with the
battery of Advanced Placement® Program (AP)
Latino community. The three organizations have
examinations and for the Graduate Record
met periodically and completed projects since 1984.
Examinations (GRE) program — all integral parts
Projects
of the partnership programs.
From the beginning, the three organizations
worked as operational partners, with joint
10