interactive keys and others

[Mike Dallwitz]

Heike Vibrans wrote:

> Are there any quantitative data or publications out there on the relative
> merits of different ways of identifying plants (perhaps comparing
> conventional keys, interactive keys and flower-color-sorted pictures, as in
> many popular field guides)? In terms of % correct identifications and time
> necessary? Perhaps even distinguishing between trained botanists and
> laypeople?

Here are some references.

Fermanian, T.W., Barkworth, M. and Liu, H. 1989. Trained and untrained
individual’s ability to identify morphological characters of immature
grasses. Agronomy Journal 81, 918–922.

Morse, D.R., Tardivel, G.M. and Spicer, J.I. 1996. A comparison of the
effectiveness of a dichotomous key and a multiaccess key to woodlice.
Technical Report 14–96, Computing Laboratory, University of Kent at
Canterbury, UK. http://www.cs.ukc.ac.uk/pubs/1996/44/index.html

Stucky, J.M. 1984. Comparison of 2 methods of identifying weed seedlings.
Weed Science 32: 598–602.

Tardivel, G. M., and Morse, D. R. 1998. The role of the user in
computer-based species identification. In ‘Information technology, plant
pathology and biodiversity’, pp. 247–259. (Eds P. Bridge, P. Jeffries, D.R.
Morse, and P.R. Scott.) (CAB International: Wallingford.)

Wright, J.F., Morse, D.R., and Tardivel, G.M. 1995. An investigation into
the use of hypertext as a user-interface to taxonomic keys. Computer
Applications in the Biosciences 11: 19–27.

All of these experiments have serious defects in design - for example, using
programs that don't have a 'best characters' option, or using exactly the
same data in the interactive key as in the conventional key. So they tell us
little or nothing about the merits of properly constructed and used
interactive keys, compared with conventional keys. One unpublished
experiment, using Intkey, actually stipulated to the testers that they
shouldn't use the 'best characters' and 'error tolerance' options!

Here's an analogy. Supposing experiments were being proposed to test whether
the use of microscopes was helpful in identification. What would you think
of experiments that included one of the following methods?

(1) The microscope has no mechanism for focusing (= no 'best characters').

(2) The microscope has a focusing mechanism, but the users are not
instructed in its use and benefits (= no training in strategies for
interactive identification).

(3) The microscope has automatic focusing, but the users are instructed to
turn it off (= we're really determined to show that interactive keys are no
good).

--
Mike Dallwitz
Contact information: http://delta-intkey.com/contact/dallwitz.htm
DELTA home page: http://delta-intkey.com