Home Biography Letters to the Community Projects & Reports Committees & Programs Office History Reporting List Staff Directory MIT Organization Chart

• The Report:

• Executive Summary

• Letter from Susan Hockfield
• Letter from L. Rafael Reif
• Report Table of Contents
• Executive Report
• Research Report

• PDF of Entire Report
  

Report of the Initiative for Faculty Race and Diversity

B. Research Design

This research project seeks to understand the complexity of faculty behavior and experience by using a multi-method design - simultaneously employing qualitative and quantitative methods - to discover whether there are local or institutional aspects native to MIT's culture, procedures or environment that affect the Institute's ability to recruit and retain underrepresented minority faculty,³ and to show how these factors may have influenced or shaped this group of faculty and their opportunities and experiences at MIT. The study uses information from four different data sets: a faculty quality of life survey, a cohort study of faculty hired between 1991 and 2009, nine-month salary data and 93 one-on-one faculty interviews. The faculty survey was administered in January 2008. Data for the cohort study were compiled throughout the study period. The salary analysis is based on salaries effective January 2009. Interviews with 47 current and 11 former URM faculty, and a comparison group of 25 White and 10 Asian faculty members, were conducted between June 2008 and June 2009.

This multi-method design ensures that this study, by triangulating the various research approaches, can successfully address the overarching research question: "How do racial/ethnic minority faculty members experience recruitment, promotion and retention at MIT?" It allows us to bring together data on statistical trends, attitudes and perceptions, as well as individual meanings and experiences. Figure B.1 illustrates this multi-method approach.

Figure B.1

Qualitative Data

One-on-one interviews were used to discover how URM faculty members experience their lives at MIT and the role of race in that experience. Interviews with White and Asian faculty allowed us to compare the URM experience to that in the non-URM groups. There are three sets of interviews that provide the data for this qualitative analysis. The primary set is the interviews with 47 current MIT URM faculty members. Then there is a comparison group of interviews with 25 White and 10 Asian faculty. Finally, there are 11 interviews with URM faculty who have left the Institute.

Current URM Faculty. The population for the interviews with current URM faculty was identified as underrepresented minority (URM) faculty members - faculty members who self-identified as African-American/Black, American Indian/Alaska Native or Hispanic/Latina(o). At the time that the population for this study was identified (January 2008), MIT employed a total of 1,009 faculty members (i.e. assistant, associate, and/or full professors). Only 59 of those faculty members identified as URM. All members of this group were targeted to be interviewed and 80% participated: three faculty members declined to be interviewed and nine faculty members did not respond to repeated invitations. The final sample consists of 26 Black faculty members, 20 Hispanic, and one Native American. Further details on this sample are available in Appendix 2.

Comparison Sample. The comparison sample of non-URMs was chosen randomly from faculty lists, constrained only by the distributions of the URM sample on sex, tenure/non-tenure and field. Since the URM sample's distribution on these factors is different from that of the majority group, the comparison sample is useful only for comparison and is not representative of the White or Asian faculty. Two such random samples were initially chosen in order to have substitutions available for faculty who did not respond or refused an interview. If it was necessary to go beyond these two, then another random choice (subject to the same constraints) was made. As Table B.1 shows, the distributions along the selected variables are quite similar.

Former MIT URM Faculty. There are 11 interviews of URM faculty members who have left MIT. They comprise those who were willing to be interviewed from a collectively arrived at list of possibly interesting people who have left MIT, a list compiled in consultation with
current MIT URM faculty members. Ten of these former MIT URM faculty members identified as either Black and/or African American and one identified as Hispanic. Four of these were affiliated with the School of Engineering, four with the School of Science, two with SHASS and one with MIT Sloan. Two left after tenure (one retired), four left as associate professors without tenure (AWOTs), and five left as assistant professors. Their time at MIT spans a long period, from the 1970s through the 1990s and into the early 2000s. The modal path was to arrive in the 1990s and leave in the early 2000s. Typically, untenured faculty left after four to seven years (within the probationary period), though a few left much earlier, and a few stayed for nine to ten years before tenure.

Table B.1
Comparison of URM and non-URM samples

		URM		non-URM
		(n=47)		White (n=25)	Asian (n=10)
Sex:
	Male	35 (74%)		19 (76%)	8 (80%)
	Female	12 (26%)		6 (24%)	2 (20%)
Tenure:
	Tenured	21 (45%)		11 (44%)	5 (50%)
	Not tenured	26 (55%)		14 (56%)	5 (50%)
Field:
	Ec/EFA	4 (8.5%)		2 (8%)	1 (10%)
	Engineering	20 (43%)		10 (40%)	4 (40%)
	Science	5 (10.5%)		3 (12%)	1 (10%)
(other)	SHASS	9 (19%)		5 (20%)	2 (20%)
	SAP	4 (8.5%)		2 (8%)	1 (10%)
(other)	Sloan	5 (10.5%)		3 (12%)	1 (10%)

Interview Protocol and Analysis. The interview protocol for this study was developed with guidance from the research literature on the experiences of URM faculty in academia, and is shown in Appendix 1. Basically, the same protocol was used in all groups, with only minor necessary modification to meet the circumstances of the comparison group and those who have left MIT. Background information was sought to create a basic profile for each interview participant; this background information included age, field of study, year of hire, tenure year, awards won, education and family commitments. In addition, the final protocol
focused on four domains: 1) coming to MIT and experience over time, 2) racial experiences, 3) climate of inclusion, and 4) family, community and social life.

In accordance with recommended research practice, interviewers for the current URM sample, as well as for those minority faculty who have left MIT, were either African American or Latina. Two White interviewers conducted the comparison interviews with the White faculty, while the Asian faculty were interviewed by an Asian interviewer.

Interviews were digitally recorded, if the participant granted consent, and each interview was transcribed verbatim. Transcriptions were supplemented by field notes and interview memos. All interviews took place at a location of mutual accord between the research participant and the researcher. Most faculty members chose to meet on campus in their offices, with the exception of two who asked to be interviewed in a conference room. On average, each interview lasted two hours, although there were some interviews that took up to three or four hours.

All interview transcripts and notes were managed through ATLAS.ti, a qualitative software package that lends itself to an interpretative and iterative analysis. The coding system to organize data consisted of a preliminary review of all data for re-occurring topics and patterns in order to generate overall coding categories. This was followed by an inductive, line-by-line coding of each interview. Through ATLAS.ti it was then possible to create coding families - i.e. summaries of all the statements made on a particular topic. Analysis of these data was a step-wise process. It consisted, first, of abstracting general themes from the coding families including the quotes that provided the basis of the theme. Based on this listing of themes and quotes related to the topic in question it was then possible to prepare a memo on the relevant findings for that topic. The final analysis, reported here, was based on these memos.

Quantitative Data

Quantitative methods were used to compare the attitudes, perceptions and career trajectories of minority and majority faculty. This analysis was based on institutional population data as well as on the sample that answered a faculty quality of life survey.

Faculty Survey. In early 2008, MIT faculty and other instructional staff were invited to
respond to a survey about faculty work life. The survey, commissioned by the Office of the
Provost and administered by the Office of Institutional Research, examined a number of issues concerning quality of life at MIT, including workload and work-related stressors, climate, mentoring, the tenure and promotion process, and the integration of work and personal/family life. Several items, particularly in reference to climate, were added by the Faculty Race Initiative. The survey was based on the core survey developed by schools in the Association of American Universities Data Exchange (AAUDE). The survey invitation came from the provost and included reminders from the chair of the faculty and individual school deans and department heads. Survey responses were treated as confidential.

The overall response rate for tenured and tenure-track faculty was 69% (708 total responses). Ethnicity is defined as three discrete groups: underrepresented minority, which includes faculty identified as Black, Hispanic/Latino and American Indian/Alaskan Native; Asian, which includes the Far East, Southeast Asia, the Indian subcontinent and Pacific Islands; and White. Within the URM group, when sufficient numbers of respondents exist, distinctions are also made between Blacks and Hispanics.

Table B.2
Response rates by rank, school, gender, and race/ethnicity

	# of respondents	Response rate	% of respondents	% of population
Overall	708	69%	100%	100%
Rank
Tenured	519	67%	73%	76%
Non-tenured	189	77%	27%	24%
School
SAP	60	73%	8%	8%
Engineering	264	70%	37%	37%
SHASS	125	80%	18%	15%
Science	185	64%	26%	28%
Sloan School of Management	63	61%	9%	10%
Whitaker	8	57%	1%	1%
Other	3	60%	0%	0%
Gender
Female	153	79%	22%	19%
Male	555	67%	78%	81%
Race/Ethnicity
URM*	39	72%	5.5%	5.3%
Black or African American	24	80%	3.4%	2.9%
Hispanic or Latino	14	61%	2.2%	2.0%
White/Not Hispanic	572	69%	80.8%	81.1%
Asian or Pacifc Islander	79	66%	11.2%	11.6%
Other**	18	86%	2.5%	2.0%
Overall	708	69%

*Note: Includes Black, Hispanic, and American Indian/Alaskan Native (n=1) faculty.

**Note: "Other" includes: Chooses not to self identify, Not available, and Other.

Table B.2 shows the individual response rates by gender, school, rank and race/ethnicity. As the rightmost two columns show, the final sample was close to the distribution of the larger population, hence the responses were not weighted in subsequent analysis. Furthermore, the response rate for URM faculty was similar to that of White faculty (72% vs. 69%), though Blacks had a higher and Hispanics a lower response rate than the average.

Of those who responded, 78% of White faculty are tenured compared to 62% of Asian and 44% of URM faculty. Hence we controlled for tenure status in most analyses. Regarding rank, 15% of White faculty are at the assistant professor level compared to 29% of Asian faculty, 46% of URM faculty, 46% of Black and 43% of Hispanic faculty.

Appendix 3 provides further details on this sample.

Cohort Analysis. The cohort data set lists every person who was hired into a faculty position from 1991 to 2009. The data were obtained through the Office of Institutional Research and provide information on hiring, promotion, and retention as well as demographic characteristics like race/ethnicity, gender, country of origin, hiring department (including dual appointments), the rank at which people were hired, and subsequent promotions, including the dates of promotion and tenure, or date of leaving MIT. This was a difficult data set to compile and some variables of interest, such as doctoral degree institution, are not yet included; nor was the "productivity" measure used in the salary analysis available historically. One of the recommendations of this report is not only to update this data set continuously, but to expand the information in it. Nonetheless, it does allow one to compare the basic career trajectories of the URM faculty to the non-URM groups.

Eighty percent of those who entered during these years came in as assistant professors, and much of our analysis will deal with their career trajectories.

Further details on this group are available in Appendix 4.

Salary Analysis. Salary data from 9-month salaries as of January 2009 were analyzed by regressing log salary on the following variables: entry cohort, time at MIT, age, gender, race/ethnicity, country of origin, current rank, department, initial rank, and if they had ever held an administrative position. Further, for the schools of Science and Engineering, a subsequent analysis also included research volume as a possible, though not entirely satisfying, proxy for productivity.

Previous:
A. Introduction

Next:
C. Hiring and Career Trajectories