Class Notes (1,000,000)
CA (620,000)
UTM (20,000)
SOC (4,000)
SOC221H5 (200)
Lecture 9

SOC221H5 Lecture Notes - Lecture 9: University Of Western Ontario


Department
Sociology
Course Code
SOC221H5
Professor
Jayne Baker
Lecture
9

This preview shows pages 1-3. to view the full 15 pages of the document.
LEC 9
Slide: Experimental Research
second quantitative method that we studied in the course, last week we looked
at surveys
this type of research is what many of us think about when we imagine a visual
of science, especially natural science
we would picture a laboratory environment
although the experiment is very common in the natural sciences, it looks a little
bit different in the social sciences
is used in social sciences like sociology and criminology, a little bit less
commonly though
the logic of the experiment is the same, whether in the natural science, hard
sciences or social sciences
the idea is that you’re measuring/observing something, introducing
something new/manipulating a condition, then measuring/observing
again
if at that second point of measuring/observing you notice something
different, then that means most likely you’ve identified a variable
that has an effect/ a causal relationship
the beauty about this model is that you can establish causality
you have this isolated laboratory type setting
in other approaches, causality may be implied, but in experimental
research, causality can be observed/measured
Slide: Why experiment?
To demonstrate causation- it means that x causes a change in y. cause and
effect. x is the presumed cause, and y is the presumed effect
Easiest method for meeting the three conditions of causality
1. Temporal order
2. Association
3. No alternative explanations
Slide: 3 Types of Experiments
Laboratory experiment
sometimes also called a kind of classic experiment
high degree of control over the setting, artificial setting (lower
external validity)
high internal validity- the conditions of the experiment are being
controlled
external validity on the other hand is lower- because classic experiments
tend to happen in artificial settings

Only pages 1-3 are available for preview. Some parts have been intentionally blurred.

because its such a controlled environment, its unlikely that you
would be able to replicate these results outside of a
laboratory
not as commonly used in sociology as the field experiment and the
natural experiment
pros:
high internal validity, so you can be really conclusive about any
kind of causality that you find
cons
but there’s a question about generalizability, would you actually
be able to repeat that study and arrive at the same finding
that you did in a diff environment?
doesn’t work well in social world- so not as common in sociology
as you would find a field experiment and a natural
experiment
Field experiment
natural setting; less control (higher external validity)
takes place in the natural setting- happens out in the field
e.g experiment that you conduct in the street
not artificial lab setting
p.210- example of field experiment on the accuracy of eyewitness
identifications
the researcher arrange for a target to approach men and women
in public places and ask for directions/help, and these
people wore dark sunglasses and a baseball cap
what the researchers did was once that interaction has taken
place, half the time they went up to the people that had been
approached and asked them to give a recollection of the
person who had asked them for directions
and then for the other half of the participants, they were
approached 4 hours later and asked to give the same
example 2- video
Dalton Coloney
Devah Pager-. Milwaukee hired young men to apply for jobs,
posed as high school students with limited job experience.
created profiles, e.g convicted,
black man without a criminal record feared no better than a white
man just getting out of jail
replicated study in New York- found similar results
because it takes place in a natural setting, higher external validity,
because you can try it in different contexts
Natural experiment
natural change in society takes place, can measure before and

Only pages 1-3 are available for preview. Some parts have been intentionally blurred.

after; less control
its when social scientists can measure something before and after,
there’s natural change in society
so they are not the ones actively introducing something/ manipulating
things, but rather that change it already taking place in society, and
so they can measure before and after that change takes place and
look for the differences
e.g when 5$ a day daycare was introduced in Quebec- discussed in
textbook
some of the key ways that they differ are in terms of their measures of external
validity and internal validity p.119 READ IT, even though not in chapter on
experiment
internal validity: the ability of experimenters to strength a causal explanation’s
logical rigour by eliminating potential alternative explanations for an
association between the treatment and the dependent variable [textbook]
basically what all that means is: do you have a solid experiment? do
you have a really good design when it comes to your experiment?
so that you know that if you observe a change here as a result of
this introduction/manipulation, so that you know that you’ve
measured causality - you know that you’re measuring what you
intended to measure
internal validity is about the design of the experiment, is it solid?
External validity: the ability to generalize from experimental research to settings
or people that differ from the specific conditions of the study
its about a kind of generalizability
if you had an experiment in a particular setting, would those results hold
up in a different environment?
This plays not why we have these 3 different types of experiments
Slide: Research questions
only appropriate for certain kinds of research questions
given that this method allows you to establish causation, it is of course most
appropriate for questions that are trying to establish causation- that are
attempting to establish causation - does x cause a change in y
Emphasis is on causal relations
causal relationship needs to be testable
however, that kind of causal relationship has to be able to be tested
within ethical boundaries
sometimes you might have a question that is suited for an experiment,
but you’re crossing ethical boundaries by using the experimental
method
e.g Facebook tweaked peoples news feed, without consent
some people received more positive emotional content, and other
You're Reading a Preview

Unlock to view full version