false

Class Notes
(834,756)

Canada
(508,705)

University of Toronto St. George
(43,993)

Philosophy
(1,521)

PHL201H1
(40)

Dan Dolderman
(1)

Lecture

Unlock Document

Philosophy

PHL201H1

Dan Dolderman

Summer

Description

Measurement as a random variable:
Let’s agree that
• a measurement is an activity, which depends on certain, known or unknown
conditions
• the outcome of each measurement is understandably varying due to the
changing parameters
• measurement is at the heart of a scientific experiment
• An essential characteristic of a scientific experiment is its reproducibility.
Due to the variability of the outcome of each measurement, of the same phenomenon
under the same conditions, we consider a measurement as a ‘random variable’. A
random variable is a process, the outcome of which is a random selection of a
number from a collection of numbers (known as the range of the random variable.)
Of course, a random variable is selecting numbers based on absolutely no algorithm,
and this is what characterizes the random variable. As soon as any element of
uncertainty involves in a decision making, random variables are used as
mathematical models. We usually build a non-deterministic model for our knowledge
of the reality, and at the center of this model there resides a random variable. But
often we face a certain decision making that is only partially random. The existing
information about the nature of the selection/experiment usually influences the
outcome of the selection; as such this randomness is no longer absolutely random,
yet it is random as it does not obey an algorithm for its decision making. Thus, in
various investigations we use one of the many types of random variables that may
better suit the nature of our selection.
The outcome of a random variable is a number, but what is it that this number is
supposed to be representing? Is there an absolute measurement that this random
variable is trying to capture, or are these random values collectively presenting us
with a kind of information about what we are trying to measure? Or perhaps these
random values are collectively trying to invent the measurement for us! In other
words, is there one actual number that we do not know (but we believe it exists) that
we are trying to achieve, or is it that there is no actual number and we are just
1
constructing it by trying to measure it ? Here is an example to facilitate this thought:
1
This is an example of the mystic gaze, ancient among the mystics, and wathextensively
presented in a book of verses called the ‘conference of the birds’ by the 12 century mystic
Attar of Nishapur. In this epic, the birds are searching their leader, the legendary Phoenix. In
this search many of them perish, but only 30 of them have enough endurance and enough
faith in the existence of their leader (the absolute truth), and only this group finishes the using the projector we project the picture of a dart board on the wall and ask people
to try to throw dart at the imaginary/projected dart board. Of course many try it and
some come very close to hitting the center and some even manage to hit the center.
Of course at this point we are sure that only few of the darts landed on the center,
and in this sense the rest of the darts landed off the target, and as such they are
‘incorrect’. Then we turn the projector light off, and we look at the dart holes on the
wall. At this point it is only the holes that lead us to believe that there was a center
and that the dart holes were many attempts to hit the center. But it is not clear which
dart holes belong to the ones that hit the target: which ones are the ‘correct’ ones
and which ones are the ‘incorrect’ ones. But still collectively the dart holes will tell us
a story as to where the center could be; and to be sure we really need all of the dart
holes to complete the picture. In a sense the truth belongs to the entire population of
the dart holes: together they tell a story; it is together that the population talks about
being close to the center or being far from it. Each dart hole, even the furthest away
from the center is having a share in indicating where the center could be, and as such
each dart hole carries a portion of the truth. Now without any knowledge of the center
we try to use the dart holes to actually recreate the center. In practice we do not
insist on the knowledge of the center but we use the probabilities of the dart holes
being close to it or far from it: this is the language of the modern sciences; closer to
political elections when they report the results a poll they generally mention a
number with an error together with a probability (for example the conservatives
are gaining 40% of the votes, with the error of 2%, 18 time out of 20.) This is how a
measurement is to be discussed in the modern sciences. Let’s go back to the
example of the imaginary dart board and that the furthest point carries some truth: if
we eliminate the furthest point then the chances of other points being more true is
increasing, and as such the distribution will become distorted (and away from the true
reality of itself.) here is somewhat related paradox: three prisoners A,B, and C, are in
a cell when the guard announces that two of the prisoners will be released and the
third one will be sentenced harshly, but the guard is not willing to reveal the identity
of the prisoner who will be harshly sentenced. As such each prisoner will have 1/3
chances of being sentenced harshly and the 2/3 chances of going free. At this time
prisoner A approaches the guard and asks the guard: “ we both know that surely one
of the two prisoners B or C will be set free; will you be kind to tell me which one of the
prisoners B or C will be set free?” The guard suggests that if he declares one of the
prisoners B or C as one who will surely go home, then the chances of A being harshly
sentenced will increase from 1/3 to ½ and that is not fair to A. As you can see the
elimination of one of the options (who was surely going to go home) will alter the
nature of the distribution.
Any experiment seems to suggest one type of randomness inherent in it. For example
the experiment of flipping a coin can have only two outcomes: {Head,Tail} or if you
journey to the mount Olympus, where they hear a disembodied voice which welcomes them
and calls them the various attributes of the Phoenix: they had become the truth (which they
searched so faithfully!) wish to represent them numerically, the domain of the associated random variable is
{0,1}. Here is a classification of various types of random variables suitable for
various experiments:
•Bernoulli random variable has range of values {0,1}; otherwise known as
Success and Failure, or True and False.
•Uniform random variable has a range of values {1,2,3,…,n} for some number
n. For example throwing a fair die presents us with 6 outcomes each of which
is equally likely. This can help modeling the experiments that have one of
the n different outcomes all are equally likely because we have absolutely no
specific knowledge about the details of the experiment.
•Binomial random variable has a range of values {0,1,2,3, …, n} for some
number n, and this is used to model the number of successes in n trials of a
Bernoulli experiment. For example we flip a coin n times and count the
number of the Heads; this number can be modeled by a binomial random
variable.
•Geometric random variable has infinite range {1,2, 3, … } of values and can
be most naturally used to model an experiment that involves an attempt to
get one success only. We try the Bernoul

More
Less
Related notes for PHL201H1

Join OneClass

Access over 10 million pages of study

documents for 1.3 million courses.

Sign up

Join to view

Continue

Continue
OR

By registering, I agree to the
Terms
and
Privacy Policies

Already have an account?
Log in

Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.