Understanding bullying among nurses and its implications for staff turnover. For a Ph.D. dissertation,
a nurse-researcher had surveyed 500 nurses to learn about ways in which they might have been mistreated
by fellow nurses; how these harmful behaviors might have related to one another; and how such mistreatment,
combined with demographic factors, might explain some nurses' intention to leave the profession.
Item distributions were heavily skewed, and so we searched for effective power transformations to
make the variables more amenable to modeling. With this accomplished, I guided her through a series
of factor analyses that helped sort two dozen behaviors into four underlying factors (roughly: punishment,
belittling, exclusion, and undermining). We used these factors in multiple regression and analysis
of covariance models to explain the outcome of intention to leave the profession, after generating
and thoroughly checking diagnostic statistics and graphs to see whether the strength or direction
of any relationships hinged on gender, ethnicity, experience, or nursing specialty. I advised and
supported her as she wrote up her findings, interpreted their larger significance, and underwent
a successful Ph.D. defense.
The research led to publication of three articles in peer-reviewed journals, the
exploring the most valid and efficient ways of determining the extent to which bullying affects an organization. A new instrument
comprised of just four survey items was found to be superior to a popular one that required 22. In
response to this paper we have fielded many requests by other researchers asking for permission to
use our instrument.
Explaining the incidence of surgical errors. A researcher working with a national nursing organization
forged the opportunity to collect a wealth of data on the occurrence of wrong-site surgeries (e.g.,
removal of the wrong kidney) as well as on staffing, staff education, and climate in several hundred
US hospitals. She sought to assemble these many types of objective and subjective data into a model
that could help predict incidence of these catastrophic errors. Because such errors are rare given
the vast number of surgeries performed, standard linear regression would have been ineffectual. I
helped her devise more general linear models -- using logistic and poisson regression -- that would
be more sensitive to distinctions between the rare and the extremely rare. I also advised on scale
development (combining survey responses into scale scores for greatest validity and reliability)
and on model specification (maximizing the model's stability and sensitivity by choosing the most
promising variables and the best number to include).
Predictive modeling on university enrollment. A large Mid-Atlantic university needed analytical
assistance to uncover factors affecting accepted students' decisions as to whether to enroll. The
university's admitted pool was unusually "price-insensitive," meaning that the size of financial
aid awards had surprisingly little connection to matriculation decisions. I have used logistic regression
interaction analyses, among others, to search out "pockets of price sensitivity"--subgroups for whom
the award mattered more and for whom the university would thus have more leverage in using aid to
attract students. The university has contracted with me on multiple occasions to explore issues affecting
undergraduate and graduate populations, students at risk of academic probation, and international
students. The work has influenced a $45-million awarding process and has filled an important gap
in this enrollment management operation.
Investigating faculty and staff qualifications. In 2005 a consortium of over 100 private schools
needed to inventory the qualifications of several thousand employees. Results would be used to support
a statewide staffing initiative. I worked with the organization's leadership to devise a systematic,
highly defensible strategy for coding key information from over 2,000 program documents completed
each year. These documents were not standardized; they were completed by several dozen school administrators
using different reporting formats and guidelines. Extracting the relevant information required a
nontrivial amount of subjective decisionmaking, which undercut the precision desired for this high-stakes
evaluation. I compensated by creating a Monte Carlo simulation to model the ways in which 10 types
of error might have biased the estimates obtained and to show how these estimates could be adjusted.
As a result of this analysis the organization was able to present stakeholders, including the state
legislature, with better-supported, more authoritative statements on a vital topic. The study helped
enable many member schools to successfully negotiate with the state for better staff compensation.
Additional iterations of study. In the years since 2005 I have created annual reports both on
qualifications and on retention/turnover. My analyses have included data graphics illustrating job
movement patterns typifying employees with particular characteristics and circumstances. I have also
expanded the use of simulations so as to investigate longitudinal changes in qualifications and retention,
since conventional statistical tests would not apply. Results have been used by the central organization
as well as by many of its member schools. For this association I have also designed studies and/or
conducted analyses on demographic trends likely to affect enrollment and on outcomes for students
in private as opposed to public special education programs.
Comparing student achievement in charter schools and conventional public schools.
A Ph.D. researcher in education sought to evaluate the proliferation of charter schools by
analyzing the state assessment scores of all Massachusetts charter and conventional schools. An important
goal was to control for socioeconomic factors -- some at the school level, which is commonly done
and sometimes sufficient, but also some at the individual student level. This task was highly complicated
by the nature of charter school enrollments, with a given school accepting children from as many
as 20 different towns or regional districts. It was also complicated by having different sets of
students and student characteristics involved in the analysis depending on grade level and subject.
We devised a complex process by which students' income levels would be estimated based on their home district, with each
charter school and each conventional public district assigned SES covariates that were weighted composites
of students' backgrounds. (We later found that our method was essentially the same as one developed
independently by a research unit of the Massachusetts Department of Education.) In addition to helping
with design and analysis for this study, I advised on manuscript editing and other aspects of the
process, helping bring about a successful doctoral defense.
Surveying the US market for staff development trainings.
A firm offering staff trainings and instructional materials wanted to make the most of an online survey to be sent to 20,000
professionals. I helped their marketing director formulate questions so as to obtain the clearest
and most useful information, and I set up the survey using an online tool. There were dozens of analytic
topics of interest to the firm; I designed a set of multivariate statistical procedures that each
allowed investigation into several topics simultaneously. These linear and logistic regressions saved
time and money compared to one-by-one investigation, and they had the added advantage of providing
control for potentially confounding variables.
Devising a sampling plan to reduce corporate auditing expenses.
An international manufacturer of agricultural equipment had hired a financial services firm for advice on how to save money
on a tax audit. Of many thousands of transactions possibly eligible for tax deductions, the firm
needed to identify samples that could be audited, since checking each transaction for possible deductions
would have been completely cost-prohibitive. I used IRS guidelines, empirical characteristics of
the body of transactions, and inferred characteristics (from a Monte Carlo simulation I designed)
to help the firm make informed recommendations to their client on how to create the most advantageous
Helping attorneys prepare rebuttal of expert testimony.
In a felony trial, the state's expert witness had given extensive testimony based on a set of research findings in experimental
psychology. I identified a series of weaknesses in this testimony, helping defense attorneys to bring
about an acquittal.
Conducting analysis for a book on developmental psychology.
I helped a researcher analyze extensive data collected on sibling and family interactions.
While a great deal of attention had been paid to validity of measurements, the sample size presented
some problems in light of the research questions being addressed. To respond to these challenges,
I found ways to improve upon the factor analysis previously conducted, and I devised ways to strategically
use cluster analysis, multinomial logistic regression, and 3-D scatter plots to shed greater light
on the researcher's questions. The following year the book was published by a major university press.
Data analysis in archaeology. An Ivy League archaeologist was studying artifacts found on a
Greek island and dating to 2,000 to 3,000 years ago. He needed to better understand the relationships
between objects, their historical periods, and the areas where they were found within the archaeological
site. Initially, complex correlation matrices contributed to his understanding of these relationships.
Next, several iterations of cluster analysis, conducted collaboratively, shed light on the types
of objects that tended to appear together and on the profiles that might be attached to certain locations,
based on the abundance of their artifacts and on the wealth and variety of trade that they indicated.
All in all, the findings helped explain both the values held by these ancient peoples and the dynamics
of their economies.
A committed undergraduate researcher was studying the progression of societies in ancient Mexico. Her thesis depended on
her ability to distinguish between greater and lesser specialization exhibited by members of these
societies, as reflected in artifacts such as potshards. I helped her evaluate several methods of
comparing variabilities of different artifact measurements and to incorporate the most applicable
variance tests in her dissertation.
~ Copyright 2007-2018 by Roland B. Stark. ~