PSYCOLOQUY V2 #8 (2.8.5 Queries/Replies: 296 lines)
Stevan Harnad
2.8.5.1 Query: Cognition and Forecasting
2.8.5.2 Reply : IQ inquiry
2.8.5.3 Query: Educational/Psychological Association email addresses
2.8.5.4 Query: Help Needed for Croatian Refugees
2.8.5.5 Query: Soviet Infant Cognition Collaboration
2.8.5.6 Query: Social Work Directory
----------------------------------------------------------------------
From: Mary Ann Metzger <MET...@UMBC.BITNET>
Subject: Query: Cognition and Forecasting
Evolution of cognition: factors affecting forecasting
Mary Ann Metzger
Department of Psychology
University of Maryland UMBC
Baltimore, Maryland 21228
This is a query, but a few paragraphs are needed to set the context.
Consider the consequences, for the evolution of cognition, of
limitations on the possibility of long-term forecasting of natural
phenomena (Gleick 1987; Grebogi, Ott, & Yorke 1987). Examples of
biologically important natural phenomena are weather, water, terrain,
and animal behavior. These are governed by nonlinear dynamical
processes which may be comprised of several different states each of
which may have both deterministic and chance components. If accurate
forecasts about natural processes are possible, then it would be
beneficial for humans to have developed good intuitions about factors
that affect the success of such forecasts. If forecasts are not
possible, then the absence of good intuitions would be no mystery.
Consider also one aspect of the formal methods that have been used to
generate forecasts: There is an expectation that success in forecasting
is greatly influenced by the precision of estimates of quantities to be
used to make the forecasts. The quantities are usually estimated on the
assumption that a particular set of equations (a model) approximately
describes the process. Examples of quantities for which precision is
important are the coefficients of the model and the average values of
variables. Theoretically, when these quantities are being estimated,
precision is greatly affected by the number of observations used in the
calculation.
To illustrate: When random samples of observations have been used to
calculate estimates of the mean of a variable (e.g., height) for a
given population of almost any type, mean values that are slightly
erroneous occur several times more often when they have been calculated
from samples of 10 observations than from 1000 observations, while
means that are highly erroneous occur frequently with samples of size
10 but almost never with samples of size 1000. The high degree of
precision associated with random samples of large size, then, connects
sample size to success in forecasting.
If it is valuable (in the evolutionary sense) to forecast the future
and if the size of random samples determines the success of
forecasting, we might expect people to have developed good intuitions
about the importance of sample size, but people have not done so. When
asked to solve certain types of problems, subjects generally fail to
appreciate the influence of very large sample size (Tversky & Kahneman
1971). Those who have not had specific training in statistics are very
unlikely to have good intuitions about it and even experts fail to
apply their knowledge about sample size in all cases where it is
appropriate. Instead, subjects judge that highly erroneous means
calculated from random samples occur about as frequently with samples
of size 10 as they do with samples of size 1000.
To illustrate with a typical problem, a subject might be asked to
imagine a situation in which, in one year of driving every day through
both a small town (10 traffic lights) and a large city (100 traffic
lights), a driver has recorded the percentage of encountering green
rather than red lights in each. The subject would then be asked whether
the driver will have recorded more days of encountering 90% green
lights in the town or more such days in the city. Assuming that the
traffic lights operate independently and are equally likely to be green
or red at any given time (for a true percentage of 50%), the figure of
90% is very discrepant from the true percentage and therefore much more
likely to occur in the small town. Therefore, the correct answer is
that more such days will have been recorded for the town, but subjects
generally respond that the number of such days should be the same for
both town and city. This failure to appreciate the influence of sample
size needs an explanation.
One explanation is that humans have limited information processing
capacity and so there has been an evolutionary trade-off: good
intuitive understanding of the behavior of large random samples was
traded off to conserve capacity for something else and a less true but
often useful intuition was adopted instead. The substituted intuition
is that means of samples of all sizes (except means of very small
samples, say around 7 or fewer observations) are equally likely to be
close to the population mean (Kahneman & Tversky 1971). If
understanding the properties of means of large samples is truly
important to forecasting the future, then this would seem to be a
mighty trade-off.
A second explanation (one I favor) is that stable estimates of means
calculated from large samples are not particularly useful for
forecasting future events in natural processes. It is, after all, only
a theory that the precision of estimates is significantly related to
success in forecasting and a theory can be wrong. To take a different
approach to forecasting the future, note that for nonlinear dynamical
systems, that is, for many important natural processes, the far future
is inherently unpredictable (Gleick 1987). Similarly, the far past is
not inherently useful for predicting the near future. It is the
properties of natural processes, not the limitations on information
processing, that confine us to forecasting the near future from the
near past. For the purpose of forecasting the near future, having very
precise estimates of quantities of approximate models may not be as
useful as having good models with rough estimates. As an example, in
sailing it would be useful to know the average depth of the water to
several decimal places (an approximate model with precise estimates),
but it would be more useful to know the approximate depth and
approximate location of reefs and channels (a better model with rough
estimates).
Finally, my question: I am looking for counterexamples to the
theory that I favor. That is, I am looking for examples of
important natural phenomena for which it would be useful (in an
evolutionary sense) to have a very precise estimate of some
quantity calculated from a random sample. The example need not
involve forecasting, but it must involve natural processes, not
human record-keeping. The demonstrations of the failure of
subjects to appreciate the properties of large samples have
generally required subjects to reason about records assembled and
maintained by humans. Such records are static (or are presumed
static for the purpose of the demonstration) and therefore do not
have many of the interesting properties of natural processes. If
I have not defined the problem away by imposing this restriction,
I would like to have examples involving only natural processes in
which the precision associated with large random samples is
important for drawing conclusions or making judgments.
References
Gleick, J. (1987) Chaos: Making a new science. New York NY: Viking.
Grebogi, C., Ott, E., & Yorke, J. A. (1987) Chaos, strange attractors,
and fractal basin boundaries in nonlinear dynamics. Science, 38,
632-638.
Tversky, A., & Kahneman, D. (1971) Belief in the law of
small numbers. Psychological Bulletin, 76, 105-110. reprinted in
Judgement under uncertainty: Heuristics and biases (1982), D.
Kahneman, P. Slovic, and A. Tversky (eds.), Cambridge University
Press.
------------------------------
From: GREE...@TWSUVM.BITNET
Subject: 2.8.5.2 Reply : IQ inquiry
So, why measure IQ at all unless it is to be able to give someone a
label of smart or not, with all the socio-political baggage that carries?
Read Gould (1981) and Lewontin, Rose and Kamin (1984) about this aspect
of the controversy about measuring intelligence.
Gould, S. J. (1981). The Mismeasure of Man. New york: Norton
Lewontin, R. C., Rose, S., & Kamin, L. J. (1984). Not in Our Genes.NY: Pantheon
Gary Greenberg
------------------------------
From: HI...@EGFRCUVX.BITNET
Subject: 2.8.5.3 Query: Educational/Psychological Association email addresses
I am looking for the e-mail addresses of the headquarters of the
following associations:
* American Psychological Association
* American Educational Research Association
* National Council of Measurement in Education
* Psychological Corporation
* Educational Testing Services
* Consulting Psychologists Press
I wonder if you can help. By the way, I am not sure if some or all of
them have e-mail addresses. Thank you in advance.
Ahmed K. Higab
The National center for Educational
Research and Development
Cairo, Egypt
Bitnet: HIGAB@EGFRCUVX
------------------------------
From: SEIZMOLOSKI_ZAVOD <seizmolo...@uni-lj.ac.mail.YU>
Subject: 2.8.5.4 Query: Help Needed for Croatian Refugees
Dear Collegue,
I am writing to you in a very hard time for all of us in
Croatia. You must have heard about the terrible situation in our
country. The war is something we never expected and it found us
unprepared as professionals, too.
At the moment we have about 20,000 children that had to leave
their homes. A lot of them are separated from their families and
some have even lost one of the parents. We are now trying to begin
the school year with them and they really need a psychological help.
The literature about working with refugees and PTSD is very
scarce here. Maybe you or one of your collegues have some experience
with these kind of problems. We would appreciate your sending us any
kind of advice, research results or therapeutic suggestions for
working with that kind of traumatized population.
If there is any paper or book that you find helpful for us,
please send to us on the following address.
With many thanks
Ivanka Zivcic
Department of Psychology
School of Education
University of Rijeka
Brusica 1
51000 Rijeka
Croatia, Yugoslavia
tel. +38 51 446692
fax +38 51 446611
E-mail: seizmolo...@uni-lj.ac.mail.yu
------------------------------
From: mc...@weber.ucsd.edu (Mike Cole)
Subject: 2.8.5.5 Query: Soviet Infant Cognition Collaboration
The research group from Institute of Psychology of USSR, whose
interests are in the field of Infant Cognition is interested in
communicatng with other groups of individuals and persons working at
this area. Our previous research is connected with recovery from
temporal visual deprivation (innate cataracts) in infancy. Now we are
exploring the problem of audio-visual coordination using innate
cataracts as a model of the influence of environmental factors and
twins as a model of genetic and experience interaction. In our research
we use physiological indices (EEG, EOG, GSR) and behavioral estimates.
If you are interested in this problems, please send us information
about your research and your computer address and inform us whether you
would like to receive our papers. We are interested in joint research.
We have ideas and projects but at present we have some difficulties
with equipment and we need sponsors for our investigation.
Dr. Elena Sergienko
write to psy...@comlab.vega.msk.su <for Sergienko>
------------------------------
From: Natalie Strangelove <441495%UOT...@acadvm1.uottawa.ca>
Subject: 2.8.5.6 Query: Social Work Directory
I am presently compiling the _Directory of Networked Resources for
Social Work Studies and am seeking information pertaining to any
networked (ie - available via Internet/Bitnet and affiliated networks)
resources that could potentially facilitate Social Work studies. Such
resources would include Listserv Lists (Special Interest Groups),
online archives, electronic journals and newsletters, and any relevant
organizations that maintain an Internet accessible e-mail address.
If you are aware of any such resources that should be included in
this Directory, please contact:
Natalie Strangelove
School of Social Work
Carleton University
<441495@UOTTAWA>
<441...@ACADVM1.UOTTAWA.CA>
------------------------------
PSYCOLOQUY
is sponsored by
the Science Directorate of
the American Psychological Association
(202) 955-7653
Co-Editors:
(scientific discussion) (professional/clinical discussion)
Stevan Harnad Perry London, Dean, Cary Cherniss (Assoc Ed.)
Psychology Department Graduate School of Applied Graduate School of Applied
Princeton University and Professional Psychology and Professional Psychology
Rutgers University Rutgers University
Assistant Editor:
Malcolm Bauer
Psychology Department
Princeton University
End of PSYCOLOQUY Digest
******************************