INTRODUCTION TO RESEARCH DESIGN STUDY GUIDE FOR EXAM 2

   What factors impact research designs?

- 1) Research Question (Hypothesis)

Involves the relationship that we are attempting to examine.

Cannot be vague

Needs a clear idea of what to study and what you expect to happen

-2) Desired Specificity of answer or info gathered

Ex: Super specific= less generalizable

Neither is incorrect but how you use it and the research question

Ex: Do we explain a single situation exhaustively? Do we want a broader explanation that can fit multi similar situations?

-3) Rules of Causality (being sure the relationship is actually causal) (Cause Þ Effect)

- Explanatory Research

- Why something has happened

- What causes certain outcomes

( Indep variable Þ Dep variable) we are looking at variables, typically social in nature that contribute to our preferred outcome variable (crime!)

Are basically quality control steps

Ex: Lombroso’s Adavist (those who engage in criminal activity were atavists, less involved humans)  

   Idiographic explanations ( Idio= individual)

- Focuses on a single event, instance, etc. )

- Basically, a case study

- Can be useful

Ex: why was Jeffrey Dahmer a serial killer?

Demanding/ fussy mother

Parents had an unstable relationship

Neglected by parents

Interested in dismembering animals

Highschool outcast and alcohol abuse

Began to develop dominating sexual fantasies

( this research will not be generalizable because it is too specific)

  Nomothetic explanations ( Nomo= no more) Broader patterns

-  Focuses on a class of situations or events

- Explaining casual patterns

- Seek to explain more broadly and more efficiently

- But technically a partial explanation

Ex: Those with low self-control are more likely to engage in crime

Ex: Those with physically abusive parents are more likely to engage in violent behavior in adulthood. (cannot go into a deep dive with this)

  What are the criteria for causality?

o   Appropriate Time Order

- The variation in the independent variable must come BEFORE the variation in the dependent variable.

Ex: Tattoos= Crime

o   Empirical Association

- ( you can tell they are related because they change together)

-the two variables must be correlated (variation in one variable is related to variation in another variable)

Ex: Strain Theory

Ex: Positive association ( ÝÝ ) Good parenting quality= increases self-control

Ex: Negative association (ßß ) Bad parenting quality= decrease self-control

o   Nonspuriousness

- “Correlation does not prove Causation”

-Relationship appears to exist but something else is at play or unknown 3^rd variable causes change in both.

Ex: The Twinkie event: Ý ice cream sales = Ý crime = (3^rd var.) Temperature

Ex: Ý Shoe Size = Ý IQ = ( 3^rd var.) Age

  What are the common elements of an experimental design?

-   Groups: Experimental and Control

Exp: Exposed into the Indep. variable

Control: similar but exposed to nothing resemble Exp. group

Example: Are drug courts effective?

We randomly sample offenders charged with drug crimes that shows signs of addiction.

Pre-Test: severity of addition for both groups

For the experimental group: Drug courts

For the control group: go through the normal process

Post-Test:

For the experimental: Nurse Family partnership (helps family with high risk of criminal activities) Brings pregnant women healthy foods, single-mothers and then check for criminal behavior of child).

For control group: stayed the same, no change because there was no partnership given to this group.

  Sources of potential errors in research designs.

o   History: Society wide, large-scale events

o   Maturation: Humans are not stagnant beings (naturally change).

o   Testing: continuous measure (not more than once)

o   Mortality: Attribution (quit) Ex: Childhood physical abuse contributes to violence in adulthood or can be unstable household (poverty, dangerous neighborhoods and bullying).

Why are Criminology experiments more likely to be quasi-experimental?

- Because there are lacking one (or more) of the elements of a truly experimental design

·      Levels of Measurement

o    Ratio: (+/-, multi/division , ranked)

The most precise level of measurement

There is an absolute zero

Ratios can be formed betw. Numbers

o   Interval: (+/-, rank ordering)

Allows for degree of difference between items

Æ = no absolute zero (zero does not indicate lack of attribute and allows for negative)

Ex: Temperature in F or C

70°F, 62°F, and 0°F

 

o   Ordinal: (a bit of math)

Fist truly quantative level of measurement

Values given greater than / less than distinctions (rank ordering)

No sub/ or addition

Ex: How much do you agree with the following statement

Strongly agree, somewhat agree, strongly disagree, or somewhat disagree

(here you can say that one person strongly agrees and the other strongly disagrees)

o   Nominal: (no math)

Lowest level of measurement, least precise

Vary in kind or quality but not in amount

No mathematical interpretation

Important characteristics without numbers

Ex: Sex

Male, female, other (no mathematical interpretation)

Ex: Race/ Ethnicity

White/ non-Hispanic, Hispanic/ Latino, African American, Asian