What are the authoritative, electronic sources of data on computer
science courses in tertiary institutions in the United States? Online
and free are better but neither are essential, having data which can
be grokked by OOoCalc/Gnumeric/Excel etc is. I want to be able to
answer questions relating to "what are the components of a US CS
degree today?"
I have looked at http://nces.ed.gov/ (too general). I found the
accreditation body http://www.abet.org and its subsidiary,
http://www.csab.org/ . CSAB has information about course requirements
but I did not find this in any aggregate data set. I rang CSAB who
explained that they do not have visibility of course content, only
requirements, and subsequently emailed the current chair of CSAB
asking for pointers. |
Request for Question Clarification by
pafalafa-ga
on
14 Mar 2005 16:55 PST
It's not entirely clear to me what you're after.
Are you looking for a typical list of courses that would go into a CS
degree? Or are you looking for a more detailed break-out...and if so,
to what level of detail?
Also, is this undergrad or grad work you're interested in?
Thanks.
pafalafa-ga
|
Clarification of Question by
parasseux-ga
on
15 Mar 2005 10:03 PST
Good points.
I am looking at undergrads. Here is a typical use: I pick a course
topic such as "Networking" and ask the question "How many undergrad
students do a networking course in the second or third year of their
degree?" Then I ask the same question for "Operating Systems" and
"Computer Architecture".
It would be extremely nice if the data was good enough to also give me
some idea that the term "Operating Systems Course" means more or less
the same thing across institutions, and I am hoping there is a tie-in
with CSAB's descriptions that will allow me to do this.
Does this make it clearer?
|
Clarification of Question by
parasseux-ga
on
15 Mar 2005 10:49 PST
Actually I just slipped something else in there by mistake, instead of
"how many students" I am primarily interested in "how many
institutions have CS degree requirements which include course XYZ in
the 2nd or subsequent year". I would like to go a little deeper and
find out "which institutions", as well.
As to my "how many students" error, that's another question and I want
to answer that too. But right now and for this question I only want to
understand the shape of undergrad CS degree requirements in the US.
Here is a set of completely fake and made-up numbers that look like
the results I am trying to achieve. Note that I am not asking the
answerer of this question to deliver these numbers, just the data from
which I can extract these numbers myself.
WARNING! ALL THESE NUMBERS ARE FALSE!
This data refers to the 2003/2004 academic year.
There are 3524 colleges and other institutions in the US offering an
undergrad CS degree requiring three years or more study which meets the
requirements of certification body XYZ.
Certification body XYZ measures course units by ABC, ie a particular degree
may have 4 x ABC of Programming but only 1 x ABC of Compiler Construction.
3842 of these certified institutions require at least two ABC units of
Java Programming.
1504 of these certified institutions offer at least one ABC unit of
Compiler Construction as an option.
WARNING! ALL THESE NUMBERS ARE FALSE!
I repeat, I am not expecting the Answerer to give me these number,
just the data from which they can be extracted.
Hopefully I have finally encapsulated the problem I want solved!
|
Request for Question Clarification by
pafalafa-ga
on
15 Mar 2005 15:35 PST
I haven't yet found the type of statistics you're seeking, but I did
come across something that may be of interest if you have not already
come across it yourself.
The IEEE developed a model undergraduate CS curricula a few years ago,
and it is explained in great detail in a report available on the web.
I would be surprised if there were many CS programs in the US that
deviated in large measure from the models they propose.
The report is at the link below, along with a few excerpts that I
thought might be of particular interest. Let me know what you think.
paf
==========
http://www.computer.org/education/cc2001/cc2001.pdf
Computing Curricula 2001
Computer Science
The Joint Task Force on Computing Curricula
IEEE Computer Society
Figure 5-1. Computer science body of knowledge with core topics underlined
[Here's what some of the details from the Figure look like]
PF. Programming Fundamentals (38 core hours)
PF1. Fundamental programming constructs (9)
PF2. Algorithms and problem-solving (6)
PF3. Fundamental data structures (14)
PF4. Recursion (5)
PF5. Event-driven programming (4)
-----
Figure 6-2. Coverage of core units
Imperative-first introduction
Traditional topic-based approach
CS111
I
. Intro to Programming
CS112
I. Data Abstraction
CS115. Discrete Structures
CS210
T. Algorithm Analysis
CS220
T. Computer Architecture
CS225
T. Operating Systems
CS230
T. Net-centric Computing
CS260
T. Artificial Intelligence
CS270
T. Databases
CS280
T. Social and Prof Issues
CS290
T. Software Development
CS490. Capstone Project
-----
The heart of their recommendations are in Chapters 7-9:
Chapter 7
Introductory Courses
Chapter 8
Intermediate Courses
Chapter 9
Completing the Curriculum
Hope that helps.
paf
|
Request for Question Clarification by
pafalafa-ga
on
15 Mar 2005 16:34 PST
Ooooh! I came *this* close!
At the link here, you'll find a lot of the sorts of stats you were
looking for, except that (1) it's a bit dated, and (2) it combines
undergrad and higher degrees.
Still, it's a start...
http://stono.cs.cofc.edu/~mccauley/survey/report2001/CompRep2001.pdf
Comprehensive Report on the 2001 Survey of Departments
Offering CAC -Accredited Degree Programs
May 2002
|
Clarification of Question by
parasseux-ga
on
15 Mar 2005 18:36 PST
You did indeed come || this close, nice find :-) I have little hope of
there being a later version of this particular report but I have
emailed the authors just in case, it does say bi-annual which
indicates 2004. As it stands I regard this as a reasonable
approximation of the type of data I'm looking for but with lower
quality. I also think it is an excellent report and have told the
authors so!
This is a survey of 45 institutions, and I have no way of knowing how
representative these are from my reading. Did they pick 15 from the
top, middle and bottom of US schools? Just one school from each state
would exhaust the sample set size of this survey, so are there no
significant differences between the states or even sides of the
country? Besides, with a sample set this size just a single school
accounts for 2%.
The reason I'm thinking in terms of hundreds and hundreds of schools
is that it seems likely to be possible to get a global data set. I'm
familiar with other national systems that have the concept of global
statistics keeping either from the government that wants to make sure
education is a reasonable quality (especially where there is partial
government funding) or from independent watchdogs or even specialist
companies who just sell information about education.
I hope I'm not aiming too high here.
|
Request for Question Clarification by
pafalafa-ga
on
15 Mar 2005 18:53 PST
Thanks for your note.
I'm still looking. Not getting closer, but I do keep coming across
interesting items.
First off, did you see the link I mentioned earlier to the IEEE
document...? Worth a look, I think.
Second, I came across this overview table of numbers of students in 2002:
[nine numbers in each row: Undergrad Total/Men/Women, then ditto for
Grad level and PhD]
Computer and information sciences, total
............................................ 47,299 34,248 13,051
16,113 10,753 5,360 750 579 171
Computer and information sciences, general
.................................. 26,652 19,956 6,696 9,512 6,606
2,906 551 440 111
Computer programming
....................................................................
582 421 161 160 104 56 3 3 0
Data processing technology/technician
............................................ 256 177 79 3 1 2 0 0 0
Information science and systems
..................................................... 9,920 6,356
3,564 3,180 1,920 1,260 60 31 29
Computer systems analysis
............................................................. 253 180
73 184 131 53 2 2 0
Computer and information sciences, other
...................................... 9,636 7,158 2,478 3,074 1,991
1,083 134 103 31
Just FYI.
paf
|
Clarification of Question by
parasseux-ga
on
15 Mar 2005 19:38 PST
Ok, I've gone through the IEEE report, some areas in detail and
skimmed others. It's a very thorough piece of interesting work (I like
the quote from 1992 describing compiler and OS courses as artifacts
rather than ideas, here here!) but I'm not sure what you are
specifically drawing attention. I can see most CS courses would fit
one of their models, but unless these institutions or someone who
collects statistics on these institutions also uses the nomenclature
of the IEEE report I'm not sure how I can use the information.
Potentially it is a universal (within the US) taxonomy for courses and
levels of courses.
Have you found bodies of stats that use the IEEE descriptive tags?
If not, how do you suggest I interpret this report to help in my search?
|
Clarification of Question by
parasseux-ga
on
15 Mar 2005 19:53 PST
The table of numbers is also relevant, but I can't interpret these
category names without definitions such as we have been discussion.
For example take the very common habit of different schools even
within the same prestigious institution to offer non-CS courses with
names like "Information Sciences", "Information Systems" etc which on
the face of it could be buried in any of two or three of the
categories given. These numbers may have some value as a sanity check
on some better stats we get later on, but apart from that I can't see
much relevance. Do you?
Perhaps whoever put the table together is worth contacting? That might
be interesting.
If you can suggest a way of improving my question or assinging some
other dollar value to a part of the question I'm listening. This is my
first time using GA. I think you know what I'm looking for now and I'm
willing to pay for it.
Suggestion: you say "I keep coming across interesting things"
tangential to this question. Don't throw them away! Maybe put them on
wikipedia, even in a fairly raw state at least someone might tidy it
up.
|
Clarification of Question by
parasseux-ga
on
17 Mar 2005 11:39 PST
Renee McCauley, author of the "Survey of Departments" paper referred
to has given me permission to quote the following description of her
work:
We sent surveys to all departments offering degree programs accredited
by the national accrediting agency (which is now the ABET - the
accreditation board for engineering and technology). Initially, CS programs
were accredited by the Computer Science Accrediting Commission (CSAC) which
absorbed into ABET about 5 years ago. For more information on ABET see
http://www.abet.org/home.html
As I recall, there were about 150 accredited programs the last time we
did the survey, so 45 was not a great response, but it may have been a fair
sample since the programs were in all types of institutions - some offering
PhDs, some with masters, and some offering only baccalaureate degrees in CS.
They came from many different states.
This seems to be the end of the threads discussed so far. Its probably
a hard problem, so I'm going to increase the price.
|
Clarification of Question by
parasseux-ga
on
18 Mar 2005 13:54 PST
I have found "What Are the Sources of Data on IT Workers?" at
http://www.cra.org/reports/wits/chapter_9.html . This points to the
Computing Research Association which has some data online at
http://www.cra.org/main/cra.info.html . I rang them and they do have
some undergrad data, but only for institutions that grand PhDs and not
to the level of breakdown I'm looking for. The National Science
Foundation (NSF) keeps some stats but again the only detailed
information is on postgrads.
Next stop the "Higher Education Research Institute (HERI) at the
University of California, Los Angeles"
|
Clarification of Question by
parasseux-ga
on
18 Mar 2005 16:39 PST
Getting somewhere. See the end of this clarification for details.
Looks like using Google Answers as a sort of public whiteboard
documenting my progress is quite effective, even if it isn't the way
it was intended to be used. Still, it reduces the problem space for
anyone planning to answer this!
Negative on HERI, although I've left a description (and a pointer to
this page) with the person in charge of the data part of their
activities.
The ACM has a special interest group on Computer Science Education
over at http://www.sigcse.org/ . There are a few questions looking for
data such as I am but no answers.
I've also had some return phonecalls from various people peripherally
pointed to by resources listed in this question. Negative.
There is an emergent field called "Computer Science Education" big
enough to generate international conferences on the topic -- but
nobody seems to have thought of gathering empirical data on what's
going on in the subject area :-) I found this by the search string
"Research Computer Science Education".
A Positive hit! Educause, http://www.educause.edu/ , see under
Resources, Additional Publications then "Pocket Guide to Higher Ed". A
PDF available online. This is in conjunction with the Campus Computing
Project, http://www.campuscomputing.net/ run by Kenneth C. Green . The
Pocket Guide has some top-level data which will be most useful, and
the data behind the guide looks like being the first hit even close to
answering the first question about numbers of schools in the US
offering CS undergrad courses.
Part of the challenge is working through the fuzzy classifications of
courses (between 5 and 10 different ways of classifying a course
depending who you talk to) and correlating this with data about course
accreditation, attrition rates, pass rates, etc. I have heard (ie this
is not empirical evidence!) the following points which both make my
quest tricky: many of the best institutions refuse to undergo any
accreditation and most of the top ten CS schools in the US. In other
words it is going to take someone with years of training to interpret
all the relevant data sets, assuming they can be found in the first
place.
I'm not giving up, stay tuned (and if any GA researchers still feel
like a challenge, at least I've explored the problem space quite a lot
for you!)
|
Request for Question Clarification by
pafalafa-ga
on
23 Mar 2005 19:47 PST
Hello parasseux-ga,
Hello again.
It's an interesting process you've started up here...the "whiteboard"
as you call it.
I just wanted to let you know that I've seen your posts, have given
them some thought, but really haven't yet made progress on this beyond
what I've provided already. Perhaps I'll have a sudden flash of
inspiration! Or perhaps another researcher will come along and pick
up the challenge.
Just wanted to let you know I haven't forgotten about you...
Best of luck,
paf
|
Clarification of Question by
parasseux-ga
on
25 Mar 2005 21:19 PST
I think I have probably got the best answer I'm likely to get. Casey
Green and his Campus Computing project really do appear to be the only
ones in the business. He takes the government data from National
Office of Education Statistics, correlates it with other data such as
better geographic references, does some data cleaning, and then either
cuts up the data in various ways or ships it as-is. For a fee of
course. The real value Campus Computing has to offer is in
interpreting these numbers. It takes an expert and on the evidence I
have to date, the chances are high that if you're reading this article
and you're not from Campus Computing then you're not an expert. I'd
live to be proved wrong :-)
|
Clarification of Question by
parasseux-ga
on
29 Mar 2005 12:11 PST
The GeoGraphic information in the education from NOES / Campus
Computing data can be correlated by zipcodes to lat/long using
http://civicspacelabs.org/zipcodedb , This in turn can be turned into
dots on a map using http://tiger.census.gov/instruct.html .
|
