2 messages, 95 lines:

(1)---------------------------
Date:         Mon, 16 Feb 1998 14:32:53 -0500
From:         Susan Zappen <szappen@SCOTT.SKIDMORE.EDU>
Subject:      Re: Journal need survey

Joanna,
If you want staff to identify titles for possible cancelation, be sure to
indicate the importance of ranking titles. I've experienced a similar review
where the faculty simply said every title was essential. That kind of
response gives you no guidance. If you can promise some new subscriptions to
those who carefully review and identify unneeded titles, you can create a
win-win situation. Good luck!

          ********************
Susan H. Zappen
Head of Technical Services
Lucy Scribner Library
Skidmore College
Saratoga Springs NY 12866-1632

Phone: 518-580-5521
Fax:      518-580-5540
szappen@skidmore.edu
          ********************

(2)---------------------------
Date:         Mon, 16 Feb 1998 17:12:17 -0500
From:         Ward Saylor <ward.saylor@jcu.edu.au>
Subject:      Re: Journal need survey

We've tried various ranking methods, including the 3-pointer. Its major
drawback was a skewed distribution leaving few cancellation candidates.

The method we found best suited to our needs was to allow every respondant
to list as many journals as they liked in rank order. To make it easier,
five titles are allowed in each rank. So a respondant's top five titles
are scored "A", the next five as "B" and so on as long as they wish.

These ranks are then turned into numerical scores as the reciprocal of the
letter place in the alphabet. So an "A" scores 1, a "B" 0.5, a "C" 0.333,
a Z 1/26 (whatever that is). This formula was arrived at after a number of
trials and seems to give the best results.

This is done on a spreadsheet and the scores summed.

Staff/departments/schools can then be given a listing in rank order of all
titles in which they expressed an interest with a line drawn at the
required cancellation savings point. They can then juggle between the
"safe" and "at risk" categories.

We refine this by applying other formulae which take into account cost
etc. For example, you might accept a title with an everage cost as 'safe'
if it scores at least one A but not if it scores less. A title costing
twice the average might need two "A"s while one costing half the average
might need only a "B" - the actual relationships are determined by your
cost/scoring/savings profile.

Ward Saylor
Associate Librarian (Resources)
James Cook University
Townsville, Queensland, Australia
ward.saylor@jcu.edu.au

---
Joanna Tousley-Escalante wrote:

>Our library will undertake a badly needed review of journal use/need
>this year.  As we consider instruments to weigh the "need" value, I
>would like to hear from those of you who have undertaken this exercise
>in the recent past, or who have a regular review set up for use
>periodically.
>
>We are considering how to ask them to rank their need and are
>considering offering them a scale of 1-3 [3 being need very much; 1
>being only some need].
>
>What other ranking or grading options do you consider that allow staff
>to indicate value to them while giving us the option to cut if budget
>requires this?
>
>Also, has anyone found software to assist in such an endeavor?
>
>Our thanks in advance for your thoughts on this question.
>
>Regards,
>Joanna Tousley-Escalante
>
>*       Head Technical Services Unit
>*       VIC Library - Vienna, Austria
>*       International Atomic Energy Agency
>*       j.tousley@iaea.org