The homepage Internet
Resources for Institutional Research includes more than 550 links in 41
topical areas, some with annotations and comments. Reviewing these many
resources, I have thought about how each of them, from U.S. News and World Report
to Dilbert, offers
critical information for higher education administration. Given competing
expectations and priorities for IR, why should you bother spending the time to
track down these links? In this paper I try to provide concrete examples of the
value of knowing how to access and use these resources.
In the process of developing the case studies, I collected survey data using
a web-based
questionnaire form in which I asked respondents to describe and evaluate
ways in which they used the Internet to do their work. Some of the more
interesting project data
are included here. I am very grateful to institutional researchers throughout
the country for their many email discussions, links, and ideas about ways to use
the Internet, particularly to Lucinda Potter at West Virginia University. Tod
Massa, Jose Cruz, Larry Nelson, and Bob Daly have also been very helpful in this
process.
My hope for this paper/homepage is that through these case studies I will be
able to share the growing excitement and enthusiasm for the World Wide Web. I
believe that the Internet is the future for institutional research and planning
in higher education. In this age of decreased funding, continued legislative
scrutiny, and the student right to know movement, the success of institutional
research lies in our ability to serve as complex information brokers using the
best tools available (the Internet).
The following case studies are presented in this paper. If you have comments
or want to contribute a case study of your own, please email me and/or send in a case study of
your own.
You know that your SHEEO has a homepage where they have been putting up SAS
output from analysis of comparison data. You go to their homepage and click
through the list of reports to find student enrollment data. You wander around
each report till you find one which breaks out in-state vs. out-of-state, but
for headcount only. For each of the past five years, you save the report to
disk. Then you parse the data from the pre-formatted ascii text used for the
HTML document into a spreadsheet. You bring all five years of data into one
large spreadsheet, cutting and pasting only the bottom line columns of total
headcount. You write a formula to calculate average annual rate of growth.
When you present the report, you say that FTE data are not readily available,
but would headcount do? If you need FTE you will have to get the data from your
SHEEO. The VP doesn't want to tip her hand, so she says not to call the SHEEO,
especially since the headcount data make the point she wants to show. After you
go back to your office, you remember that approved enrollment projections are
also up on the web. You save these from HTML into ascii and parse them into a
second spreadsheet, mentioning to your budget director that they are available
if the headcount trend data end up being useful and the discussion focuses on
the effect of declining state appropriations on projected in-state
enrollments.
You've seen the AAUP faculty salary data which Arizona State University put
up on the web. You choose the peer institutions to include and submit the Perl
script which is available at the site to select your data and calculate means.
After printing the table, you share the results. Your boss makes the point that
while the data are interesting, the salaries are not weighted to be similar to
your institution in distribution by rank, gender, tenure, discipline, or
contract length. You go back to the office looking for datasets which will help
you make your point.
While the Oklahoma data are useful, most of your peers do not participate and
institutions are anonymous. You could request a copy of their CUPA faculty
salary by discipline and IPEDS SA reports for the current year and plug
different numbers into a spreadsheet, hoping something useful would come out.
Only the IPEDS SA data are available electronically. You pull out the IPEDS
CD-Rom, but the data are old. So you go to the NSF homepage for CASPAR data,
hoping the vendor has put out a more recent version of the SA data, but it is
the same as the IPEDS CD-Rom. Next, you go to the NCES gopher site, where you
download the IPEDS SA dataset for a more recent year.
Reading the documentation, which is mostly for SAS, you think about assigning
it to one of your programmers, but you don't want to take her off of other
projects. After downloading and unzipping the file, you bring the data from
aasci into Dbase, which you can manipulate more easily. You discard unwanted
rows until you get one record for each fice code. Using the documentation, you
create dbase fields for only the variables you are interested in and begin to
look at the data.
Although you could use the combination of control and highest level of program offering which are included in the IPEDS data, you need to sort peers by the newest Carnegie classification. You copy this from a disk you bought to your C drive, wondering if you could have found it for free on the web. Then you write a query to join the file to your SA dataset using FICE code as the unique variable. It takes you a while to figure out why you lose some records. This is caused by different UNITID codes and levels of rollup to dummy FICE codes. Eventually, you are able to get reports of salary data for all of the different possibilities of peer institutions by control and Carnegie classification which you are interested in. You end up dumping the data from DBF to spreadsheet format to do more analysis, since the report generator you are used to isn't very sophisticated and you don't have time to learn a new one. Finally, you are able to present spreadsheets of different scenarious of peer comparison choices and the effect each has on determining the all ranks, total university faculty salary average, weighting the data as much as possible with gender, rank, and contract length. To share your work with others in your institution, you export the report from the spreadsheet to HTML tables using a macro you downloaded off the net. Then you include the table as a menu choice on the homepage for your IR office.
You search the Web, looking at the NSF homepages. You find not only the detailed tables (expenditures by discipline) that you have not yet received for FY 1993, but also ALL of the FY 1994 tables. You are able to complete your project with the latest available data.
You begin by checking out the web versions of the major college admissions
guides to find out what peer and regional institutions already offer this
program. Once you narrow this list, you go into the homepages for these
departments to gather any kind of information which is out there - enrollment,
course offerings, number of faculty, degree audits, anything tied to the ESL
program. While some schools have a special homepage for this program, most do
not. So you look to see if they have an electronic factbook which would provide
trend data on student headcount enrollment for this major. As you go, you both
print out pertinent homepages and save them to disk, so you can import them into
a spreadsheet for further analysis and presentation.
You give the provost a breakdown on competitors and their enrollment data for
the proposed program. You offer to contact your counterpart at each school and
inquire about the program, but are told to stop working on the project and move
on to another new priority.
After the planning session, the dean is given the green light for further study. He casually mentions that there are plenty of jobs out there for people with ESL degrees. Is there some way to document it, you wonder? You know that there are many listservs out there for different disciplines. Using one of the list of listservs, you locate an ESL discussion group. To investigate, you subscribe, read the current day's postings, and request an index of archive files at the listserv's FTP site. After a few hours, you have downloaded the past two years of discussion in a second listserv which was mentioned in the first one and is dedicated to discussion about employment. A cursory reading suggests that there is an open market for teachers with an master's in ESL, as long as they are willing to move to the southeast and southwest, but that the competition is pretty fierce in your region. You summarize your thoughts in a one-page memo to the dean, with a blind carbon copy to the provost.
You have read versions of the SRTK legislation wherever you find them, using
the Internet search engines. Several listservs mention them and you pay
attention to what others are posting. You print these messages and circulate
them. A special AIR listserv is created just for monitoring SRTK and you join,
trying to help stimulate discussion. Finally, you read the December, 1995
regulations online, print them, save them to disk, and section by section
analyze what your institution would have to do to comply with the regulations.
You cut and paste the text of the act into your word processing document for a
presentation of recommendations to major administrators.
The current version of the IPEDS Finance Report is available on the NCES
gopher site. You download it, unzip it, make it into a SAS dataset, decide what
variables you want to look at based on the print form and the documentation,
then create a basic report structure. You cut and paste the results from the
output screen of SAS for Windows into a spreadsheet. To get the student FTE
data, you have to decide which calculation of FTE you want to use - the official
NCES definition based on a combination of full and part-time headcount or a
calculation based on student credit hours. You decide on SCH, since that is what
you use internally, and download the same year's IPEDS Institutional
Characteristics data. As in the Finance report, you unzip the file, create a SAS
dataset, decide on the variables to use, and bring the data into DBase and then
a spreadsheet, calculating the student FTE data for your peers.
After calculating the ratios, you are asked if there is any way that you can
obtain more current data. Looking at your peers' homepages, you find that none
of them include official IPEDS revenue data, although some FTE data are
available. Most of the peer institutions are in the East and their SHEEO's
participate in the National Cooperative Data Share (NCDS) service run by John
Minter Associates which shares early reports of IPEDS data. Using this gopher
site, you are able to obtain electronic versions of the Finance report for half
of your peers. You end up printing the results and plugging them into your
spreadsheet because it seems easier to do. For the rest of the schools, you have
someone call each of them and request them to fax you a copy of certain pages
from the Finance and Institutional Characteristics reports ASAP. You enter the
data into another spreadsheet with a current version of the ratio results (which
are almost identical to the previous year).
After vowing that you will refrain from opening your mouth in meetings from
now on, you settle down to take a fresh look at the sequence of homepages users
find when they get to your primary URL. What does this say about the
institution? You print each page and begin noticing the repetition of certain
key words. Many of the pages were written by the same office and you recognize
the style and choice of words. Getting beyond that, you begin to see an implied
set of assumptions about what the institution has to offer. While there are no
data used to document these assumptions, and you question some of them, it soon
becomes obvious that a person reading these pages would reach certain
conclusions about your school.
Then you begin looking at other "quasi-official" homepages for departments
and programs. You find faculty resumes, course syllabi, and pictures of
students. At times, the image is very different or disjointed from that
portrayed in the university-wide homepages. Using the techniques of content
analysis, you find a way to code the qualitative look and feel of each homepage.
You use these keywords and images to show that users reading these pages for the
first time are told several types of stories/narratives about the institution.
You wait for the president and provost to move this discussion along the way
they want to, recognizing that you have now unwittingly helped set de facto
standards for how the institutional mission should be portrayed in its
officially-sanctioned, world wide web homepages.
Using the online admissions guides, you review program offerings for schools
within a reasonable commuting distance. Drawing a circle on the map, you find
five competitors with bachelor's degree for adults. You use the homepages for
each school to search for information about each program, including admissions
requirements, enrollment, and program offerings. Two of the schools have
electronic factbooks and another has a gopher site, so you are able to locate
headcount enrollment trends for the these program for the past five years. You
compare the enrollment to the data in your own factbook and do a quick
spreadsheet.
You know that census data include population statistics by age and
educational attainment rates, so you use the U.S. Gazetteer and Census Bureau Home Page to get 1990 census
data for each county in the commuting circle. This lets you estimate the number
of adults who have high school diplomas but not bachelor's degrees, by age
group. You choose additional census data to flush out some demographics about
this population.
One of the county extension agents helps you find a report prepared by the
state agency about lifelong learning in the region and you are able to obtain an
electronic copy of it from the FTP site at your state's land grant institution.
You begin tweaking assumptions about how many more students could conceivably
enroll in the program, using the program's current demographics as a guide. You
could go further with it, but are not sure whether it is useful, so you take
what you have to the next meeting of the task force.
After doing a quick literature search with the homepage for ERIC, you find
several articles to read. You access your online library catalog and find which
journals are in your library. The others you will have to read on fiche or
request full-text copies of. You are interested in knowing whether other schools
have such a policy in place, so you write up a quick survey and send it to a
group of email addresses for your peer institutions. Some of them respond by
e-mail and you take their responses and content analyze them. Those who don't
respond you contact by phone. You write up a quick report summarizing
frequencies for each response to your survey and citing the literature. After
talking about the problem with your registrar, you try and estimate the effect a
policy might have on your enrollment. It appears that a policy which limits the
worst offenders will not have much impact and that most students would not be
effected because they do not earn many W grades. You brief your boss about the
results and move on to another project.
You call up every homepage you can find for these institutions and print out
any data on admissions. For about half of the schools, you construct a
spreadsheet showing that applications have remained stable over time, but that
the number of admitted students has increased at some institutions in order to
produce increases in enrollment. Several institutions show striking changes. Are
they real? What did they do and what was the effect? You go to the homepage for
your SHEEO and find a table comparing state institutions' new student enrollment
by level. The data conflict with your spreadsheet. You end up calling your
counterpart at the peer institution, where they admit that they too opened the
floodgates to meet projected enrollment increases in order to justify capital
outlay expenditures. In a future meeting, you have a better handle on the
phenomena and its possible effect on the institution.
After reviewing the literature with the ERIC homepage, you find that the
standard sources are IPEDS degrees conferred data by discipline, NRC doctoral
recipient data, and U.S. Census data. You locate the homepages for NCES, NRC,
and Census, but are not sure where to go in them. So you use Internet Resources for
Institutional Research to find the right page. The CASPAR data intrigue you
too, so you download the main files and those for the NRC and IPEDS data. After
learning to use the data, you are dismayed that they do not break out all of the
disciplines the way you thought they would. For the NRC data, you have to locate
a current version of the doctoral recipient report. For the IPEDS data, you try
using the CD-Rom which NCES sent you in the mail. After e-mailing some of your
colleages who do affirmative action reports, you present your options for data
to the affirmative action officer. This person decides that you should use the
trend IPEDS data in CASPAR for the disciplines which match your departments and
the trend NRC data for the other disciplines. In a few cases, you need to use
the newest NRC book. A colleague e-mails you about the Oklahoma data and offers
to fax you the pages for the CIP codes you need. You run this by the AAO, who
agrees that they are better than the current year's NRC data, which would be
appropriate only for new assistant professors.
In a few more hours, you are able to present a spreadsheet which documents
the choice of data source by discipline and, for those departments where you
have multiple disciplines, reweights some disciplines based on the number of
faculty in the department. You've done your best in the time frame, though you
know it is "down and dirty" and not how you would normally like to produce these
important faculty hiring statistics. For those departments which are not in
compliance and have fewer minorities or women than the labor market figures
calculated in the eight factor analyses, you make sure to verify your
documentation of faculty availability.
You read about several SQL query software packages in PC Magazine. The online
version has the links to the software which is reviewed, so you play with the
links. You download trial versions of several software packages and read the
user manuals that come with them. Several sites have their own newsgroups to
discuss software issues and you check these out. One package, Cold Fusion by Allaire, particularly
interests you. When you review their site, you see that they also offer an
online Forums package and have an example of it in place for user support. You
read comments and questions about the software.
You look at sites which have implemented the software. Not many are in higher
education, but the commercial homepages are interesting. To use Cold Fusion, you
have to run Windows NT on a web server with at least 24 megs of RAM. You plan on
using Netscape's Commerce server, which is free for higher education
institutions and begin following the newsgroups which discuss the server. You
download the software and install it. Meanwhile, a discussion in the Allaire
Forums suggests that the Windows NT version of the Commerce server crashes with
Cold Fusion. Another web server, Purveyor, is recommended as working better with
NT and Cold Fusion. You find the homepage for this software. In a newsgroup,
someone writes that Purveyor is free to higher education institutions. This is
not mentioned anywhere on the homepage and you contact the vendor, which ends up
sending you free disks and manuals.
In a learn as you go mode, you end up installing Purveyor and Cold Fusion on
your web server. You don't really have CGI programming skills, but extend your
knowledge of HTML to work with the Cold Fusion tags. You try writing scripts to
access an extract of the Fall 1995 human resource database. When you encounter
problems, you post messages on the Forums and the vendor responds. Reading other
threads of posts, you get more of a feel for what it takes to do drill down
access. You really need to know more about SQL. While you do searches on this as
a key word, you come up pretty empty. Finally, someone in a newsgroup has
written that MS Query is bundled with Access and Excel and that you can use this
to cut and paste SQL syntax into your Cold Fusion template. It works - a little.
You buy a book for Access which includes chapters on querying and SQL. Finally,
the computer person in human resources lends you a simple SQL book which
explains why you keep having the same types of problems with your code. Soon the
code is fixed and you are demonstrating the project to other offices and
offering to write HTML forms and Cold Fusion templates to query human resource,
student, finance, and space data.
You sit down with the staffer and show them the legislative updates which
were prepared last year by the SHEEO. A copy of the last update is available on
the SHEEO's homepage. You show the person how other universities in your state
monitored the status of bills and published their own updates on gopher and web
page sites. The staffer is already subscribed to the legislature's online system
and you show her how to telnet to it from within Netscape. You show her the
homepage for the state legislature which lists the status of bills in a way
which is GUI-based and much more user-friendly than the older, online system.
You show her how to cut and paste text from the online system, the legislature
homepage, legislative updates from other universities, and the SHEEO updates
into a word processing document. She can then either send this document as an
e-mail attachment to major administrators at your institution or cut and past
the text into the body of an e-mail message. When you check back with her a week
later, you see that she has developed a system which far surpasses what you did
last year and you stop worrying about whether you should do more on the
project.
Using the calendars of upcoming conferences which are available on listservs,
the CAIR homepage, and at the Chronicle's Academe Today site, you make a list of
everything that meets in the next two months. After you make the first cut based
on travel costs or too exotic of a location, look at the association homepages
which offer the conferences. Is there an electronic version of the program? Are
there sessions you really want to go to? Are the authors of session papers
someone you really want to pay attention to?
You find a conference with three papers and a panel about a hot IR topic -
data warehousing in higher education. But you have never heard of the
presenters, so you e-mail them asking what they will cover in their sessions. If
a well-known scholar is giving a keynote, do you really think you'll learn more
at the speech and follow-up session than you will if you take the time to read
what she/he has already written. (Hint: do an ERIC search and see whether their
work interests you). Once you have narrowed down your list of choices, look for
homepages that describe travel attractions in the region. Is the conference
meeting at a single hotel? It may not be your best choice. Look for homepages
for all major hotels in the premier conference cities. Use the homepages to find
the best price. Use homepages for travel services to evaluate what you want to
do on your trip. Check the local weather service online to know how to dress.
When it comes time to make travel arrangements, check online travel agents
and/or homepages for specific airlines and rental car agencies which may offer
better pricing.
Using the online versions of PC Magazine and other computer trade journals,
you read the latest hardware reviews. If there's a specific vendor you get
interested in, say Micron, check out their homepage for the latest options and
discount values. Once you've selected the system you want, think about buying
directly from the vendor. If you want to order from one of the computer chain
stores which will provide on-site service in your region, go to their homepage
for pricing, though they may not have all of the vendors you want. Don't forget
the computer discount stores/mail order chains, all of which have their own
homepages with pricing specials and technical support for your questions.
Once you buy a PC and the vendor tells you it's been shipped, use the
shipping agent's homepage to track the progress of your package. After you get
the boxes unpacked and have trouble getting one of the driver's to work
correctly, use the vendor's homepage forum or newsgroup for support without
waiting 2 hours at lunch time to speak with a telephone representative.
First, you clarify which institutions the survey needs to go to using control
and Carnegie classification. You use the recent Carnegie classification data,
which you get with the institutional demographics included in the CASPAR
datasets. You create a dataset with institutional name, control, and Carnegie
classification. To get the presidential names and addresses, you download the
most recent version of the IPEDS Institutional Characteristics data from the
NCES gopher site. Using the documentation, you strip out everything but the name
and address fields you want. You import these data from ascii into DBase, then
create a secondary mail merge file with a DBase report. You provide the
president's secretary with the secondary mail merge file to use for the cover
letters and mailing labels. For those institutions which for some reason fell
out of the dataset or did not get matched between the CASPAR data and the IC
data, you look the president's name and address up in the HEP Directory. After a literature search in ERIC, you read the recent New Directions for Institutional
Research piece on college guidebooks. The Internet search engines lead you
to examples of standard survey responses at five institutions. You print these
homepages and compare them to the ten most important college guides your office
completes. Someone tells you that there was a report on standard survey
responses in an edition of the Electronic AIR. You get the URL for the IBM ISAAC
site where these are stored and look at the table of contents for each edition
for the topic. The Electronic AIR talks about work being done by Bob Daly and
you read his piece in New Directions. You have some questions about data on
faculty workload and SAT scores, so you e-mail him. He mentions discussion of
the topic in one of the new AIR listservs and you subscribe to it and look for
archives (there are none yet). You send an e-mail to your peer group asking whether they too are considering
this alternative. You get back responses with links to drafts of this type of
document on various office homepages. After your staff help you mock up one for
your institution, you put it up in an obscure subdirectory on your web server
and ask some of your friends and colleagues to take a look at it. One day you
are asked to talk about it to several major administrators. It is your chance to
push using the Internet. They go for the standard survey response, since you
also have the data on the web to convince them that your office is being
productive, efficient, and fulfilling its mission. The standard survey response
also holds potential to become your own "college guide."
Though you mean to locate the classic IR essays from the AIR Professional
File and New Directions about the mission and purpose of IR, you get sidetracked
and don't have time to get them. At night, you log on and slowly work your way
through the 100+ IR office homepages which are on the web, looking for mission
statements. Some offices label a mission statement as such, while others include
something like a mission statement on their primary homepage. It goes too
slowly, so you use a search engine to select pages with the criteria
"institutional research and mission." Some hits are nonsense, but many are
exactly what you want. You save each of them to disk. Remembering that different
search engines get different numbers of hits, you try the same search criteria
with SavvySearch, Metasearch, and Alta Vista. After saving the results to
disk and visiting a hundred different homepages, you have examples of thirty
mission statements. As you prepare to content analyze them, you cut and paste text from each
statement into your word processor. While you think you'll have time to look for
key words, themes, and competing expectations and priorities, you don't. You end
up taking copies of the 30 mission statements to the retreat, where your VP
passes them out to subgroups to look at and comment on. Mission becomes a hot
topic of discussion, but the groups don't have time to report on the differences
between statements. A week later, your VP has drafted a proposed mission statement. She shares it
in staff meeting, you and others suggest changes, and it goes to her boss the
next day. You wake up to realize that your job has been redefined in light of
the new mission statement and that your project workload needs a major overhaul.
Something tells you to go online and look at a synopsis of the Seven Habits book
by Steven Covey.
John H. Milam, Jr.,
Ph.D.
Designing a standard survey response
Like all IR
offices, yours is swamped with external surveys and ad hoc requests that, while
important to some constituencies, are not part of the reward system in your
mission. You want to create a standard survey response, but are unsure what to
do and how it will impact your institution's ratings with the college guides and
image with prospective students. What do you do?
Creating a mission statement for an IR office
Your
vice president takes you and your staff on a retreat to her house. She is very
involved with ideas about continuous quality improvement and downstream/upstream
work processes and wants to update your office's mission statement. She asks you
to see if there are any IR mission statements out there on the Internet. What
do you do?
To contact the author of this paper/homepage about Internet
resources for institutional research, planning, and higher education research,
send e-mail to jmilam@wpgate.gmu.edu,
or snail mail to:
Information Management and Reporting
George Mason
University
MS3D2, 205 Mason Hall
Fairfax, VA 22030-4444
(703) 993-8837
(phone)
(703) 993-8835 (fax)