Class Notes
(807,339)

Canada
(492,708)

University of Toronto Mississauga
(23,314)

Sociology
(3,969)

SOC350H5
(9)

Lecture 1

Unlock Document

University of Toronto Mississauga

Sociology

SOC350H5

David Pettinicchio

Winter

Description

Lecture 1
Pick any research question – answer the question with data we have
CGSS data
SOC222 with objective of final project – think about what research project
you want to do
Labs – help think about what you’re interested in, expose you to data and do
stuff with data (step by step)
Extend knowledge of soc222 in a different/application way
Address problems with data – learn
How to do your own project
Data in analytical way
Data for lab and data for project different
3 data set for project under lab
Topics in tutorial follow lecture
Exploring Quantitative Research in Sociology
What is Quantitative Data
Eg of what is the kind of idea behind what we will be doing
Do same logic as published research and what they do with regression
In quant method – use stat data
Answer a research question
Important differences
We are interested in measuring data – operationalized
Study crime – what do you mean by that? What kind? What does it mean?
Various ways you can measure it
Think about crime and equality – variables that can measure that
The real world is not same as textbook – messy
Settle for measures that can best measure what you want to do – not
everything is that neat in real world
Real world of doing quant analysis
Generally qualitative people do their own research and generate their own
info
Don’t rely on producer of data – question them
Why do we want to quantify?
Make inferences to broad array of population – qual cant do that
We can use inferential stats to make predictions
Describe data
Not suppose to be data mining
Question shouldn’t be driven by nature of data – why are you interested in
this
Why does x lead to y Not because its in data – interesting because its things people think about
Correlation – doesn’t mean anything
Have to have a theoretical reason for your question
How do I match my research question to what I have available
Research question
Are you seeking to make general claims, uncover broad patterns, satisfy
evidence
Should be driven by what you are trying to accomplish
Population and crime
Show graphically relationships
Refresher on regression
Bivariate regression
Y= a + bx 1
B – slope/ regression coefficient
A – y intercept
X – independent
Y – dependent
^ on top of y = predicted value of y
This model allows us to plug in value of x that is not on data to predict for
potential y
We always regress for dependent variable (outcome) – not the independent
You can also figure out if they are statistically sig – generalizable
Distance between case and prediction line (data point and line) – Error
Least sq reg reduces total amount on error (distance between case and line)
Lambda and gamma – based on least sq
Regression – prediction model
Say a lot about big questions with relatively small samples in comparisons
(large patterns)
Life expectancy and wealth question
Graph shows line of fit – line for ordinarily sq reg
R squared – measure of the strength of relationship
Our independent variable is explaining 40% of change
That single variable (wealth) explains 40 percent of variation in life
expectancy – good

More
Less
Related notes for SOC350H5