Research & the Role of Statistics
Variables & Levels of Measurement
The Structure of Research & The
Role of Statistics
Begin with Broad Questions
 Most social research
originates from some
general problem or
question
 Curious/troubled
about some aspect of
society
Begin with Broad Questions
 Example: What
influences how a child
does in school?
 General question that
can’t be adequately
addressed by 1 study
Narrow Down, Focus In
 Next, we come up with a
more specific research
question
 one we can realistically
address
 Here, a review of the
scientific literature can
serve as a guide
 Tells you what other
researchers have found
 Gives “bearing” to your
research study
Narrow Down, Focus In
 Example: What is the
relationship between
family structure and
school performance?
Narrow Down, Focus In
Also can be stated as a
causal theory –
 an explanation of the
relationships b/t
phenomena
 Example: Children
with more parental
support/guidance will
tend to perform better
in school.
Theory
 Children with more parental support/guidance will
tend to perform better in school.
 Underlined terms are concepts – abstract ideas
 concepts are ambiguous
Operationalize
 operationalize – define
a concept in a way that it
can be measured
Operationalize
 Put another way: turning
concepts into…
 variables
 something measurable
 any trait that can change values from
case to case
 Some concepts easier to
operationalize than others
 Examples:
 Parental support/guidance  #
parents in home (1 or 2)
 School performance  GPA (1 to 4)
 OTHER OPERATIONALIZATIONS?
Group Exercise: “Operationalization”
 Working with the person (or 2) closest to
you, come up with variables (something
measurable) that could be used as
indicators of the following concepts:
1. Healthy lifestyle (of an individual)
2. Economic health of Duluth
3. Success of UMD grads
Operationalize
 Hypothesis:
 derived from theory
 statement about a
relationship between
variables
 therefore:
 it is more specific/exact than
a theory
 it is testable
Operationalize
 Hypothesis example:
 Students living in homes
with 2 parents/guardians
will tend to have higher
GPA’s than students from
1-parent households.
 Independent variable (x)
 cause (i.e., # of parents)
 Dependent variable (y)
 effect or outcome measure
(GPA)
 xy
Observe
 Observations allow for
hypothesis testing
 Science is a systematic
method for explaining
empirical phenomena
 Empirical means
measurable & observable
Observe
 Research methods are
the tools used at this
stage
 How are data to be
sampled & gathered?
 Lab experiment?
 Survey?
 Analysis of existing data?
 Observations produce
data
 Observation vs. Anecdote
Analyze Data & Reach Conclusions
 Our focus in this class:
 hypotheses are tested by
comparing observations to
theoretical predictions
 Statistical procedures
give the ability to tell:
 whether the data support
our hypotheses
 & by extension, whether
our theory is supported
Analyze Data & Reach Conclusions
 Two classes of
statistical techniques:
1. Descriptive – used to
summarize/organize/
describe data.
 Example: What is the avg.
# of hours per week people
spend on face book?
Analyze Data & Reach Conclusions
 Two classes of
statistical techniques:
2. Inferential – used to
generalize findings from
a sample to a population
 Example: polling just a few
hundred voters to predict
how a presidential election
will turn out.
Generalize Back to Questions
 What do the results tell
us about our original
broader question?
 Determined by:
 How theories are
operationalized
 The nature of the
observed sample
Variables 101
 VARIABLES are any trait that can
change values from case to case

Attribute – specific value on a variable


Example: sex has 2 attributes, male & female
Variables ALWAYS should:


be exhaustive – variables should consist of all possible
values/attributes
have mutually exclusive attributes; no case should be
able to have 2 attributes simultaneously
Levels of Measurement
1. Nominal – mutually exclusive & exhaustive
categories that cannot be meaningfully
ordered (e.g., sex, religion, political affiliation)
– Categories need to be relatively
homogenous
Levels of Measurement
Scales for Measuring Students’ Living Arrangements
A
*With parents
B
*With parents
D
*With parents
*With roommates *Dorms
*Dorms
*Apartment
*Dorms
*House
*House
*Apartmnt
*Other
*Other
*House
Levels of Measurement
2. Ordinal – categories can be ranked in
addition to being categorized.
 Example: “I would rather get beat with a lead
pipe than attend this class.”





1 = strongly disagree
2 = disagree
3 = neutral
4 = agree
5 = strongly agree
Levels of measurement
 What’s Wrong with This Question:

How long have you been attending UMD?
1. 1 to 11 months
2. 1 to 2 years
3. 2 to 3 years
4. 3 to 4 years
5. 5 or more years
Levels of measurement
3. Interval-Ratio – categorical units are equal
 Examples: prison sentence in months, population
of Duluth, age
 This level permits all mathematical operations
(e.g., someone who is 34 is twice as old as one
17)
 Pointy headed issue
 Interval = no meaningful zero point
 Ratio = meaningful zero point
 DOESN’T MATTER ONE BIT FOR DATA ANALYSIS
 SPSS calls both interval and ratio variables “SCALE”
Group Exercise
 Research Hypothesis = Males who
experience hair loss become more likely to
experience depression.
 What is the IV? What is the level of measurement for this
variable?
 What is the DV? Operationalize the DV so that it is measured at
the nominal, ordinal, and interval/ratio levels.