-No harm to participants

-Voluntary particpation

-Anoyity and confidentiality

-Deceiving subjects

-Morality

-What is right and wrong

-Conforming to the standards of conduct

What to be aware of in criminal justice research

Why is it challenging in CJ?

-Be aware of the general agreements shared by researchers about what is proper and improper

-Challenging in CJ because research frequently deals with illegal behavior

-Psychological harm

-Embarrassment for people who're asked to reveal information such as sexual abuse

-Physical harm

-Biological studies and DV studies

-Issues of privacy

-Crime-mapping software

-Zimbardo's simulated prison experiment

-Male undergrad students

-Assumed the role of guards or prisoners in a mock prison at Stanford University

-Humphrey's Tearoom Trade

-1970

-Researchers studied homosexual behavior

-Pretended to be a 'watch queen' voyeur of this behavior in public restrooms

-Tuckegee Syphilis Study

-1932-1972

-Tuskegee, Alabama

-400 poor African Americans with syphilis

-Disease was allowed to progress without treatment

-Guatemalan STD Study

-Between 1946-1948

-Where people in Guatemala were intentionally infected with STDS

-Latane and Darley Study

-Staged crimes to find out which witnesses would intervene

-Monahan and Associatives

-1993

-3 groups at potential risk of physical harm in their research on violence

-Research subjects, researchers, as well as others

Voluntary participation

Why might participants volunteer?

-Agreed upon tenent of conducting research where it is voluntary

-Subjects may volunteer because they think theyll personally benefit

-Money

-Anonymity

-Researcher cant associate info from subjects with the subject's identity/know the identity

-Examples

-Mail survey.. No identifying information

-Web based survey.. No login

-Confidentiality

-Researcher knows person's identity but can't reveal identity and responses

-Examples

-Interview survey

-Electronic monitoring

Legal liability

Which types of study/research?

-For Field Research

-Under criminal law, you can be arrested for obstructing justice or being an accessory to crime

-Knowledge that subjects have committed illegal acts

-For Self Report studies

-Knowledge that subjects have commited illegal acts

-Possibility exists that your info will be subject to a writ by the government, or subpoena to appear in court and required to reveal information

-But Federal law protects researchers and the data they collect

-For legal action in most circumstances, you are protected

-Staff misbehavior

-Research causes crime

-Withholding of Desirable Treatment

-Is it ethical to deny subjects treatment or intervention that may help them?

-Solution.. Interrupt experiments if preliminary results indicate improvement in treatment group

-Mandatory reporting

-Of thing such as child abuse

-Both ways of promoting compliance with ethical principles

-Institutional review boards

-Codes of Professional Ethics

-Describe whats acceptable and unacceptable

Institutional Review Boards

What is it and purposes

-Government/University organizations that researcher must get approval from to conduct research with humans

-Two purposes

-Decide if risks to humans warrant benefits of the study to society

-Decide if study design has safeguards to ensure safety/confidentiality 

Evaluation Research

What is it, what is it not, and what can you use?

-Research purpose, not a specific methodology

-Purpose to evaluate impact

-Can use

-Survey

-Longitudinal

-Qualitative field research

-Experimental

-Form of applied vs. theoretical research

-What works and doesn't work

-What works for what types of people and not for others

-Whats works in what situations and what doesnt work in other situations

-What we need to change in a program or policy to make it work better

-Outcome measures

-Experimental context

-Specifying interventions

-Specifying population

-Operationalizing success over failure

-Must be able to reliably/validly measure your outcome variables. 

-If program intended to reduce or increase a behavior, you must be able to accurately measure it

-Measure aspects of context of an experiment that may affect the outcome variable

-Example

-Evaluating a program designed to trail unskileld workers and trying to find out if it increases employment levels

-Measure employment after program completion

-Would also need to measure employment levels in community overtime

-To see if unemployment levels changed before or after the program

-Accurately measure program intervention or the experimental stimulus

-If a program is your experimental stimulus

-Then you need to measure extent and quality of participation in the program

-Example

-Studying effectiveness of private prison in terms of recidivism

-If there are different companies operating the private prisons, you'll want  to compare outcome measures between each one to see if the company that accounts for differences in the effectiveness of the private prisons

-Will want to also determine how the prisons operated differently and their quality

-Define population of possible subjects the program is appropriate for

-Need to control for other variables such as sex and race

-Example

-Recidivism and and private prisons

-Lons Lonza Kaduce vs. the Knight study in terms of specifying what a public vs. a private inmate is

-Defying success over failure

-Measuring of these two is not clear cut

-Sometimes comes down to cost-benefit analysis. Do financial benefits outweigh cost of services?

-Does it have a positive return on the investment?

-Examples

-Changes in sentencing/punishment policy evaluation

-The public vs. private prison and recidivism

-Experimental Designs

-Quasi-Experimental Designs

-Time-Series Designs

-Multiple Time-Series Designs

-Qualitative Evaluations

Experimental Design type of Evaluative Research

What is it and ethical and legal problems

-Randomly assign subjects to control and experimental groups

-Measure before and after experimental effect

-Deals with Internal Validity

-Ethical/Legal programs

-Could randomly assign inmates to a prison faith program and control group with no faith program

-Measure level of prison violence before and after. 

-Examples

-Classical Experimental Model

-Pre-test Post-test Control Group

-Time Series Design

-Studies that involve measurements taken over time with intervention during time series

-Examples

-Study change in sentencing/punishment impacts crime rates

-How does the reenstatement of capital punishment impact murder rate?

-Multiple Time-Series Design

-Measurements of variable(s) over time in different locations or among different groups, ones who got experimental stimulus and ones who didn't

-Examples

-Evaluating effectiveness of a change in punishment on crime rates in Florida

-Can also measure crime rates over same before/after period where similar interventions not implemented

-Compare crime rates before and after the invention of the experimental state and control state.

-Can also be less structed and more qualitative than quantiative

-Argue the most effective Evaluation Research is one that combines qualitative and quantitative components

-Making stat comparisons is useful, so is gaining an in-depth understanding of the processes producing results.. or preventing expected results from appearing

1. Know audience

-Taylor data findings of abilities, attention spans, and knowledge of audience

2. Let data answer your question

-Dont present data/stats just to show it. Stats help answer questions

3. Make presentations of data sample

-Complex graphs and tables will confuse audience

4. Parsimony in numbers

-Fewer the numbers, better your point is understood

-Graph with adjacent columns with equal heights to the class frequencies

-Widths equal to size of class intervals

-Derived from a frequuency distribution where you group your data in equal intervals

Measures of Dispersion

What are they and types

-Refer to how values are distributed in each of your variables

-Types

-Range

-Standard Deviation

-Index of the variability in values in a variable around the mean of the distribution

-Higher the SD, more disperse data is around the mean

-Lower the SD, less disperse data is around the mean

-Smaller the SD, the more representative the mean is of the values on the variable

-Involve inferring something about characteristics of a population based on an examination of only a portion of the population

-We select a sample for convience and cost, goal is to make inferences about various pop. parameter on basis of known sample stats

-Used for gaining knowledge about a large class of things from a small class of the same things

-Random Sample

-Every other sample of same size has same probability of being selected

-N = Number of elements in population

-n =  Number of elements in sample

-Population

-Totality of observations for which a research project is concerned

-Measures of variables on the population are called population parameters

-Example: Average age of all convicted murders of the percentage of the entire population that received a death sentence

-Sample

-Small part of poluation

-Measures of variables from a sample called sample stats

-Example: Average age of convicted murders in sample of the percentage that recieved a death sentence are sample stats

-Estimating population parameteres

-Finding it based on value of sample stats

-Based on sampling theory and sampling distributions, we can say a sample stat is within a certain range of a population parameter

-Population parameter 'x' percent of the time

-Testing hypothesis

-For stat purposes, we test the Null Hypothesis, which is a statement of no difference in relationship (0)

-Example

-If we want to test how the type of punishment policy affects the likelihood of recidivism among robbers

-With cutoff point of 5% or 1% .. .05 and .01

-Means if our sample state has only 5% or 1% likelihood of being different from population parameter

-Want to generalize any sample stat to a population

-Means 5 out of 100 times will population parameter be different than samples of the same size

-Stat significance tells how often we'll be wrong in this statement

-

Measures of Association

What is it and difference between Stat Significance

-Stats that tells us strength or degree of relationship between 2 variables

-Different from level of stat significance because stat significance is dependent on

-Strength of relationship

-Size of sample

-Higher the sample size, more likely youll find stat significance, with given level of strength in relationship

-Represent level of correlation between 2 variables or degree to which 1 variable changes in another variable

-Ranges from -1 to +1 with -1

-Higher the number, strong relationship

-0 = No relationship

-Negative means as independent variable increases, dependent variable decreases

-Unlike percentages, they have absolute values so comparisons between different correlations can be made.

-Easy method of summarizing relationship

-Elaborate: We physically control for variables by selecting each category of a control variable and crosstab independent variables

-Statistical Models: Put all variables in a model that tells unique effect of each independent variable on the dependent.

-You're controlling for effects of other variables on the dependent

-Analysis of Variance .. (ADI: Analyze Dichotomous Intervals)

-Independent: Dichotomous

-Dependent: Interval

-Multiple Regression .. (MIA)

-Independent: Any

-Dependent: Interval

-Logistic Regression (LID: Logic Is Dead)

-Independent: Any

-Dependent: Dichotomous

-Logit Coefficient: Tells relative impact of each variable on dependent variable (recidivism) and tells whether effect is statistically significant

-Odds Ratio (OR): Tells percent likelihood of dependent (recidivsm) for each independent variables

-Convert OR to percent

(1-OR)*100

-Both hold all other variables constant

-Specified what a public vs. a private inmate is