The ORIGINAL Balanced Affective Word List Project
No longer in development - kept for historical purposes
By Greg Siegle
San Diego State University / University of California, San Diego
This project was originated in 1994 to establish a list of words with normed
emotional valences for use in psychology experiments. A number of researchers
throughout the world had gone to great pains to establish the "normative"
emotional valence or valences of words (e.g., "positive" or "negative").
By collecting these lists of normed words in a single place, I hoped to
provide a corpus which will be of use to future psychology researchers.
Since that time, the The
Center for the Study of Emotion and Attention (CSEA) has created
such a large normed corpus of over 2000 words. It is described briefly
at their Media
Core website. Those interested can obtain the "Affective Norms for
English Words" (ANEW) list from them directly. I have written a program to create
lists of words balanced for valence, frequency, and word length from
the ANEW list. If you are hoping to do a project using affective
words I recommend you follow that link.
THIS PAGE REFLECTS THE ORIGINAL WORDLIST I created, before the ANEW
list was around, in 1994. It does not reflect
more recent work including information about the ANEW wordlist and a program
that uses that list to create balanced wordlists. Rather, it reflects
the original older project, and a program to use that list to generate
affective word lists from it balanced for valence, word length and frequency.
Words for the original list were taken from:
-
a corpus of words collected by Mark Shibley and I consisting of words normed
on 50 undergraduates on a likert scale with ratings for positivity and
negativity between 1 and 7. Only words with means of 1-2,5,6-7 were used
and no words with standard devations > 1 were taken.
-
a list Gerald Matthews sent me which was also normed for affective valence
-
a list published by Carolyn John: John, C. (1988) Emotionality ratings
and free association norms of 240 emotional and nonemotional words. Cognition
and Emotion (2):49-70.
Different contributors seem to have calculated word frequencies differently.
That is, some used the frequency of any form of the word while others used
just the frequency of the given form of the word, and still others used
just the frequency of the word as a given part of speech. I haven't had
time to standardize it all yet.
Getting the word list
Currently, the word list is in a very basic, and somewhat impovrished format.
It consists of an ASCII file with 3 tab-delimited columns including: word_valence,
length, and word frequency (according to the Francis and Kucera (1982)
corpus).
Word valences are coded: 1=positive 2=negative 3=anxious 4=neutral
Click to obtain
the word list
Contributing words
If you are interested in contributing a list of normed affectively valenced
words to the list please mail or e-mail me the following:
-
a description of your norming procedure
-
an ASCII (text) file containing the words and normed ratings.
Ideally, each line in the file would contain a word and its associated
ratings (e.g., for emotionality, or for various valences) separated by
a delimiter such as a comma. The first line in the file would be a key
to the fields on forthcoming lines. Other formats are fine too though –
I’m most concerned about getting the words and ratings themselves.
-
(optional) it is useful to have not only an index of the central tendency
of ratings for words but the spread also, so, for example, a user can look
for words that are rated very negatively by almost everyone.
-
(optional) references to any publications using the word list (optional)
an address or e-mail address where you can be reached
In return you get:
-
mentioned as a contributor on this web page
-
the knowledge that future researchers won’t have to go through what you
did to create the word list
-
my undying gratitude
You can e-mail me at:
gsiegle+@pitt.edu
A program which uses the original word list
This program generates word lists balanced for affective tone (positive,
negative, & neutral, with optional inclusion of anxiety words too),
as well as word frequency (according to the Francis and Kucera (1982) corpus)
and word length.
To use the program just type "wordlist" followed by the total number
of words you want. You can specify the word file the program should use
by typing it after the number. If you want anxiety words too, type the
word ANX after the word file. For example, the command:
wordlist 16 words.prn ANX
would generate a balanced list of 16 words including positive, negative,
and neutral words. The resulting printout is of the form word valence length
frequency.
Click to obtain
the DOS executable program
Click to obtain
C++ source code.
Note: To use the program you MUST obtain the word list, and store
it in the same directory with the program. If you have not yet obtained
the word list, do it now: obtain
the word list
Caveats
I have put in under 2 hours writing this program. It's not a professional
thing. It has some minor glitches (e.g., returning too many words for very
low numbers of words in a list, not handling numbers of words that are
not multiples of 3 or 4 gracefully, etc.). That will all be changed in
future versions when I have some time. Feel free to edit it, distribute
it, etc. but if you do, please include ALL the files in this directory.
You can reach me with any comments or questions by e-mail at gsiegle+@pitt.edu
or by surface mail at: Greg Siegle, Biometrics Research 151R, 7180 Highland
Dr., VA Pittsburgh, Pittsburgh PA, 15206.
Copyright
Finally, the following copyright notice is begrudgingly included for the
sake of beurocratic completeness. The copyright notice applies only to
the wordlist program and not to the words.prn file which is entirely public
domain. Copyright (c) 1994 by Greg Siegle and the San Diego State University/University
of California, San Diego Joint Doctoral Program in Clinical Psychology
Permission to use this software is granted subject to the following restrictions
and understandings: 1) There is no warantee or statement that the operation
of this software will be error free. Greg Siegle and the SDSU/UCSD JDP
in Clinical Psychology are under no obligation to provide any services
by way of maintainence, update or otherwise. 2) Any user of such software
agrees to indemnify and hold harmless Greg Siegle and the SDSU/UCSD JDP
in Clinical Psychology from all claims arising out of the use of this software
or arising out of any accident, injury or damage and from all costs, counsel
fees, and liabilities incurred in or about any such claim, action, or proceeding
brought thereon 3) Users are requested but not required to inform Greg
Siegle of noteworthy uses of this software
Comments or problems with this page can
be mailed to Greg Siegle:
gsiegle+@pitt.edu