Class Notes
(807,463)

Canada
(492,648)

University of Toronto St. George
(42,752)

Political Science
(3,355)

POL101Y1
(1,114)

Harald Bathelt
(8)

Lecture

# POL 322 Lecture 2.docx

Unlock Document

University of Toronto St. George

Political Science

POL101Y1

Harald Bathelt

Winter

Description

1/16/2013 4:33:00 PM
Lecture 2- POL322H1
Knowledge
Observation
Collecting evidence/facts- assumes empirical approach vs normative,
nothing is self evident, knowledge acquired through experience
However, suffers from non-justification
This process of knowledge building is based on faith that it is the best
approach. No means to prove it is correct or the absolute best
Churchill: it has been said that democracy is the worst form of government
except all the others that have been tried
Confirming, challenging exercise: not about proving but failing to disconfirm,
thus assumes rational interference- no claims about universal truths
Explain and predict: nature is orderly, events have causes (determinism)
Paradigm shifts –advance theories: example- voting behavior, partisanship
nd
big idea on how you vote. 2 . Short and long term approach of
understanding voting. They also recongnised that campaings and issues that
matter. The short term factors can explain shifts in election outcomes
Issues and policies really matter
Important aspect is that values change over time, eg partisanship is now
dead as now one participates in traditional politics
What we utilize is just as important as the knowledge development. We have
to utilize instruments and sound procedures:
-systematic (vs intuition, guessing)
-rigorous
-transparent: -evaluate methods: means to which we know, -continuous
pursuit, -self-correcting, -replication
-research, hypothesis and theory
-research question, problem puzzle: why relevant?, theory development,
hypothesis-null hypothesis and falsification, alternative possibilities
-identity: concepts and concrete measures
-approximations- some error involved, example is environmental concern types of observation: individual, countries public policies, examples of
variables: age, gender party affiliation
examples of attributes grouped in that variable
selecting appropriate research techniques- how do we gather data?
-utilize specific research methods
-appropriate technique based on research question
-several types: survey, content, experimental, focus groups, interviews
determine best means of evaluating data
quantitative- look at different statistical tests
--some better than others based on situation
---also based on measurement of variables
evaluate findings
-what are the basic findings?
-do they conform to hypothesis theory?
-can we make broad generalizations?
----probabilistic findings, no absolute truths
-evaluate competing theories
publication and transparency
-put findings out for critique
-include all aspects of study
----allows replication
----furthers knowledge
----allows for debate of methods and techniques
final lecture – important for essay
- y – a+b1x1 + b2 x2
y= a+ 0.5x1t + 0.8x2 +
beauty of multiple regression
1 independent variable is often gonna explain behavior
multiple regression: how substantial a role do IVs have?
Two methods of evaluation : 1.R squared (R2) 2. beta
r square- precise estimate of effect of an IV, or set of IV’s, on DV captured
by unstandardized B values
yet doesn’t explain how much of the variation in the DV is explained by IV’s
so, how good a job does the IV’s do at explaining the DV?
This is the role of R2
How much better can we understand ‘interest in international news’ by
knowing ‘interest in federal news’
Mean score is the best guess to determine the average or prediction of
scores- r2 takes all ivs goes through a formula and explains how much
better am I to predict the values than just getting the mean
R2- runs from a scale of 0-1, question of strength
If u get 0 then your ivs are insignificant and rethink why u have them in the
model in the first place and IV provides no contribution to understanding DV
Best guess- mean, so how much better than out best guess does our IV
add?
Rsquare of 1 would be a perfect prediction, and Iv completely explains DV
Compares error from DV mean with error from regression line
Multivariate – completely 3d dimensionalizes the results/data
Can be interpreted as the proportion of the variation in the Dv explained by
the IV’s
Or the perce

More
Less
Related notes for POL101Y1