|

Measuring
paradoxical reality
Presenting polymetric analysis
Josep
Burcet © 1995
revised ©
2002 |

contact |
Introduction
Classical techniques for sample analysis in social sciences are built on the assumption according to which subjects under observation have attributes instead of paradoxes. So, measurement is addressed to assess if such attributes are present. Conventional analysis does not observe to what extent paradoxes are present and how they act.
Nevertheless, paradoxes occur
very often in complex reality, like human individuals, groups
and other sort of cultural artifacts. For instance, attitudes
towards computers, new communications and, in general, fast
emerging technology are paradoxical in the sense that the same
person may feel attraction and repulsion, hope and fear, anger
and appeasement. People may feel that way sequentially and,
even, simultaneously.
This means that each subject
in a sample may be characterized by the presence of
antagonistic attributes and, what is more important, he/she
will behave under the influence of such paradoxes.
Terminology
conventions
In the framework of this paper,
I will call monometric perception, all the forms
of insight which take into account the presence of an
attribute and ignores the presence of their antagonistic one
(i.e. she is very introverted, he is not happy, they are very
hostile).
Conversely, I will call polymetric
perception, all forms of awareness which recognize
cohabitation of antagonistic attributes. (i.e. he-she is
lovely and repulsive).
Polymetric
perception in social research
Independent
observation of antagonistic attributes allows us to asses the
degree to which each of them is present. Suppose we want to
measure feelings aroused by the Internet. We may ask to what
extent a subject feels afraid of Internet. And afterwards, we
may ask him/her to what extent he/she feels courageous in front
of it. Independent assessment of these elements will allow us
to discriminate differences which could not be assessed
if common
measurement scales were used. If we use independent scales
ranging from 1 to 9 for each antagonistic attribute,
observations may take the following form:
| John is not specially afraid of Internet. He does not
expect unwanted consequences from it. And he feels neither
courageous in front of it |
John
is 1 afraid |
>< |
1 courageous |
| Mary is specially afraid of Internet. It appears to
her as a menacing source of troubles. She does not feel
courageous. She thinks that she will never be able to
understand and handle it. |
Mary is 7 afraid |
>< |
1 courageous |
| Peter is afraid of Internet. For him it is something
that will produce a lot of changes and he feels insecure
about the future of his job. But, at the same time, he
thinks that he could be able to acquire all the required
skills and expects that perhaps he could take advantage
of it. |
Peter is 8 afraid |
>< |
8 courageous |
Criticism
to conventional measurement scales
Suppose that John,
whose reaction to Internet is 1 afraid > < 1 courageous,
is asked to place himself in an ordinal scale ranging from
afraid to courageous. His answer will be somewhere in between
the two extremes. Peter (8 afraid> < 8 courageous) should
be forced to give a similar answer. So, under monometric
scrutiny, John and Peter cases look the same when, in
fact, their emotional responses are very different.

Between John and Peter there are
differences on the precursors of their respective reactions
differences on the reactions in themselves, and,
differences on the behavior which will follow from those reactions
As far as social research is aimed at i) detecting precursors of observed
behavior ii)
understanding characteristics of observed behavior, and iii) forecasting future
behavior following from current behavior and its precursors, it is obvious that conventional
scales do not always provide serious ground for consistent analysis.
General considerations about data recollection
Let us consider the case of Mary, who was 7 afraid > <
2 courageous. Her condition fits conventional analysis. If all the members of a sample have
the same absence of paradoxical condition, then data may be treated in conventional way.
We could then say that we are dealing with monometric reality. That is, an aspect of
reality which is characterized by the fact that when one attribute is present, its
antagonistic is absent.
So, at the stage of questionnaire design we should first
realize whether
our questions address to monometric or to polymetric issues. If we are going to deal with
a non-paradoxical issue, we may use conventional scales. But if we expect paradoxical
conditions, we should prepare our questions in order to collect data of both antagonistic
attributes.
On the nature
of polymetric reality
Attitudes, values, beliefs and emotions in general
are precursors of behavior. Generally speaking, there is a relationship between the
intensity to which such precursors are present and the energy involved in the behavior
that follows. This principle applies both to monometric and polymetric conditions. But when two or more antagonistic attributes cohabit, then the level of
paradox supplies an additional load of energy. So, energy that results from polymetric
condition depends on:
- the intensity of each antagonistic attributes, and
- the existing level of paradox between them
According to this, a 5a > <3w condition will arise more energy than a 7a
> <1w condition, even if the overall intensity of
attributes involved are the same.
Strategies
for the analysis of paradoxical condition
Such considerations apply to the analysis strategy. If we collect polymetric data, we
may:
1/ look for the precursors of each antagonistic factors, which are
not contaminated by their respective antagonistic
In the Johns case, for instance, we could undertake a multivariate analysis for
explaining his score in the afraid dimension, which was 1 in a 9 point scale. And
we could do the same for Peter, whose score was 8. Note that in conventional scales we
have been forced to look for independent variables for a say 5 score in the dimension afraid/courageous
in both cases.
2/ describe the intensity of energy arousal produced by paradoxical
condition, if found
The function e=a *w +a +w gives an evaluation of the amount
of energy arousal, where e is an index of energy summoned; a
and w represent the intensity of each antagonistic attributes
3/ look for concomitant variables of energy arousal
Using a 1 to 9 scale, the e=a *w +a +w function produces an index ranging from
1 to 99. Such an index is
susceptible of multivariate analysis (see annex 1)
Designing questions aimed at collecting paradoxical data
Questions aimed at collecting paradoxical data may arise defensive reactions against
insinuated paradox. For instance:
Usually people are culturally induced to avoid incoherence and think monometrically. As
a consequence, paradoxical related questions may be difficult both to be formulated and to
be understood.
Under the pressure of the value that postulates coherence, people may be prone to hide
inner paradoxes and give congruous answers ( if one admits that he-she dislikes computers,
he-she may be prone to hide that he-she likes them as well to some extent).
As far as awareness of ones own paradoxes may induce anxiety, some people may try
to cancel distress by refusing to understand a question on paradox issue or by giving fake
answers.
--------
If you read Spanish, you may click
here to see the polymetric paradigm applied to the novelty
intake process.
---------
Annex
| Values
resulting from the e=α*ω+α+ω
function |
Energy summoned index
| |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
| 1 |
3 |
5 |
7 |
9 |
11 |
13 |
15 |
17 |
19 |
| 2 |
5 |
8 |
11 |
14 |
17 |
20 |
23 |
28 |
29 |
| 3 |
7 |
11 |
15 |
19 |
23 |
27 |
31 |
35 |
39 |
| 4 |
9 |
14 |
19 |
24 |
29 |
34 |
39 |
44 |
49 |
| 5 |
11 |
17 |
23 |
29 |
35 |
41 |
47 |
53 |
59 |
| 6 |
13 |
20 |
27 |
34 |
41 |
48 |
55 |
62 |
69 |
| 7 |
15 |
23 |
31 |
39 |
47 |
55 |
63 |
71 |
79 |
| 8 |
17 |
28 |
35 |
44 |
53 |
62 |
71 |
80 |
89 |
| 9 |
19 |
29 |
39 |
49 |
59 |
69 |
79 |
89 |
99 |
The
energy summoned
depends on:
a) the level to which every attribute is present, and
b) the amount of existing paradox
So, the α6><ω6
condition (6+6=12) summons more energy than α9><ω3,
where the overall presence of attributes is 12 as well.
α6><ω6
gives 48
α9><ω3
gives 39
Other examples:
|
α dimension |
ω dimension |
sum
of
attributes |
level
of paradox |
energy
summoned |
| case 1 |
9 |
1 |
10 |
very
low |
11 |
| case 2 |
8 |
2 |
10 |
low |
28 |
| case 3 |
5 |
5 |
10 |
medium |
35 |
| case 4 |
9 |
9 |
18 |
very
high |
99 |
|