Criminology Exam 3 Study Guide PDF

Essay questions Unit 9 ○ Whether the researcher should use an experimental design or quasi design ○ Ethical consideration Experimental design Involves random assignment and a comparison group All variables...

Essay questions Unit 9 ○ Whether the researcher should use an experimental design or quasi design ○ Ethical consideration Experimental design Involves random assignment and a comparison group All variables besides the manipulated variable are controlled or held constant A design is NOT an experimental design without random assignment ○ Leads to probabilistic equivalence The assumption that groups are equal at the start ○ Also includes experimental groups The group that receives the treatment or the manipulation Quasi design Is not a true experimental design, but it is similar to one in that there is a comparison group and intervention Has a comparison group but does not involve random assignment Often used in real world settings Higher in ecological validity, lower in internal validity ○ Non equivalent group design Most common type of quasi experimental design A design where a treatment group is compared with a comparison group, but there is no random assignment to groups Unit 10 ○ Whether the researcher should do a survey or interview ○ 3 justifications Survey Means of observation of or measuring of phenomena through questions to a respondent  Questionnaires Surveys completed by the respondent without direct input from the researcher conducting this survey Can be administered in a variety of ways: ○ By email ○ On paper ○ On computers (e.g airport, bathrooms) ○ In group settings Close ended questions: ○ Respondent is offered choices Open ended questions: ○ Respondent free to give responses Interviews Where the interviews actually asks questions to the respondents See unit 10 notes Unit 11 ○ What qualitative approach the research should approach and qualitative design ○ Justification Qualitative research Research that utilizes qualitative methodology Yield information about crime and offenders Less common in criminology and social science More idiographic than nomothetic Qualitative design Three qualitative approaches ○ Ethnography Most seen in the field of anthropology An approach that emphasizes broad aspects of the culture being studied in a great deal of depth with the researcher acting as a participant in the studied activities Time consuming, $$$ Not common in criminology ○ Phenomenology Rejects the basic premises of positivism by suggesting that there are many things we cannot easily observe or measure and stating “.. and thats okay.”  Emphasizes subjective experiences and interpretations and seeks to understand those rather than attempting to find some objective truth A traditional approach to studying drunk driving might be to simply ask someone if they have ever driven while intoxicated. This is a positivist approach. It assumes there is a point of intoxication after which people should not drive. ○ Field research Studying topics in the real world, the environment where the events actually take place Example: A researcher might go into courtrooms to observe behaviors rather than look at databases of outcomes of cases or trials Unit 12 ○ Research questions, which statistical test they should use Pearson’s r Independent variable: quantitative Dependent variable: quantitative Positive correlation: as x increases, so does y Negative correlation: as x increases, y decreases No correlation: no statistical relationship between variables Values vary from -1.00 to 1.00, bigger values more likely to be significant Examples: ○ There is a positive correlation between number of prior offenses and amount of months a person is sentenced to jail ○ There is no correlation between sales of adult magazines and sexual assault rates ○ There is a negative correlation between number of guns a person owns and self-reported fear of crime on a 7 point scale  T-test Independent variable: categorical and binary (two attributes) Dependent variable: quantitative Tells us if values differ across groups or attributes of the independent variable Example: ○ Men commit nearly twice as many violent crimes in their lifetimes than do women, on average ○ Fear of crime on a 7 point scale is significantly higher among those who live in urban areas than those who live in rural areas Chi-square test Independent variable: categorical Dependent variable: categorical Tells if the group proportions are different Example: ○ Women over the age of 21 are significantly more likely to report having ever been pregnant than women under the age of 21 ○ Violent offenders are significantly more likely to recidivate than are nonviolent offenders ○ Black jurors are less likely to state someone is not guilty than are white and hispanic jurors Chapter 9- Group Designs Group Designs ○ Research designs where the independent variable is some categorical variable where groups can be formed and the dependent variable is some variable that is measured Example: Independent variable ○ Police work shift, with attributes of “day shift” and “night shift” Dependent variable ○ Arrests made, stress level Treatment or intervention ○ Something that changed or that we think might influence the dependent variable Examples: ○ The independent variables is/are group designs are inherently categorical in nature ○ All group designs will have a posttest, measurement of the dependent variable after the treatment ○ Some will have a pretest, a measurement of the dependent variable before the treatment Independent variable Groups Dependent variable (intervention/treatment) Pill Tylenol, aspirin, aleve Pain ratings Police shooting Simulator, live range Officer confidence training LSAT tutoring Class, no tutoring LSAT scores Use of stop and frisk Has policy, no policy Crime rates policy Sanction for crime Probation, prison Recidivism One-Group Pretest, Posttest Design ○ One treatment group that receives some treatment or intervention ○ Allows researcher to demonstrate covariability when there are changed from the pretest to posttest ○ Demonstrating temporal precedence Two-Group Posttest Only Design  ○ Introduces a comparison group into a research study A group that we compare to the treatment group ○ Often, this is a control group A group that receives no intervention at all Know this Example ○ When we begin to add comparison groups, we might demonstrate that factors other than the treatment or intervention might be responsible for changes from the pretest to the posttest. Yes, crime rates dropped in New York before and after the stop and frisk policy was in place. However, we can add a second city (Los Angeles) or even a third, fourth, fifth, sixth, or more. Crime rates fell throughout a large portion of the world during this same time period. What is clear is that New York City’s stop and frisk policy cannot possibly explain the drop in crime rates in Los Angeles or Las Vegas or Amsterdam or in Berlin. Adding a comparison group supports other possible causes. Scholar, Jan van Dijk, argues that the drop is explained by changing technology (e.g., alarms, cameras, etc.). Random assignment ○ Placement by the researcher into research groups ○ Involves randomly placing each participant (or other element) into a given group or condition in research ○ Example: Individuals in a study take an experimental medication or a placebo Random assignment would involve something like flipping a coin for each participant to determine whether they get the drug or the placebo Three categories of research design ○ Experimental design Involves random assignment and a comparison group All variables besides the manipulated variable are controlled or held constant A design is NOT an experimental design without random assignment Leads to probabilistic equivalence ○ The assumption that groups are equal at the start Also includes experimental groups ○ The group that receives the treatment or the manipulation ○ One way factorial design Type of experimental design that is not much more complex than a two-group experimental design Has one independent and dependent variable Utilizes an independent variable with more than two groups Example: ○ In the offender race study, there could also be a group with a picture of an Asian offender and one with a picture of a Hispanic offender. Now there are four groups instead of two. The researcher might find that Black offenders and Hispanic offenders are found guilty at a significantly higher rate than White offenders who are, in turn, found guilty at a significantly higher rate than Asian offenders. That tells us more than a difference between two groups. ○ Two way factorial designs Type of experimental design that utilizes two independent variables People are still randomly assigned to a condition (group) We can examine the main effects and interactions Main effect ○ The isolated effect of a single dependent variable on the dependent variable Interaction ○ Occurs when the effect of one variable changes with the level of another variable ○ Quasi-experimental design Is not a true experimental design, but it is similar to one in that there is a comparison group and intervention Has a comparison group but does not involve random assignment Often used in real world settings Higher in ecological validity, lower in internal validity Non equivalent group design ○ Most common type of quasi experimental design A design where a treatment group is compared with a comparison group, but there is no random assignment to groups ○ Non experimental design  A design without random assignment to groups or a comparison group Design type Random assignment? Comparison group? Experimental design Yes Yes Quasi-experimental No Yes design Non-experimental design No No Field experiments ○ An experimental design that takes place in the field (the real world) rather than in a laboratory setting Topic 10 -Survey Research Survey ○ Means of observation of or measuring of phenomena through questions to a respondent Respondent ○ Person who completes a survey Survey methods ○ Process of actually surveying people Good vs Bad surveys ○ One goal is to differentiate between good surveys and survey questions and bad surveys or surveys questions Bad surveys are mostly useless Bad surveys due to researcher inexperience and/or incompetence Bad surveys intentionally misleading Questionnaires  ○ Surveys completed by the respondent without direct input from the researcher conducting this survey ○ Can be administered in a variety of ways: By email On paper On computers (e.g airport, bathrooms) In group settings ○ Close ended questions: Respondent is offered choices ○ Open ended questions: Respondent free to give responses Interviews ○ Where the interviews actually asks questions to the respondents Unit of analysis- survey method to use ○ Individuals ○ Households ○ Companies ○ Cities ○ States Choosing between questionnaire and interview ○ Considerations include: Sampling techniques and goals Need for clarification Speed and geographic scope of collection Population cooperation Cost of collection Ease of analysis Type of information desired Sensitivity of questions Question and response length Needs of the respondent Some other survey method choices: questionnaire concerns ○ Writing the questions ○ Use of filler questions Determines whether or not someone should be in a study Relevant with quota sampling ○ Question order Bias concerns  ○ Social desirability Where someone answers question in a way people want them to answer questions Appropriate questions Example: Should we defund the police ○ Interviewer effects Tone of voice Example: Are you voting for Kamala or trump ○ False respondents People who aren't supposed to be in the study Types of survey questions ○ Structured Questions that are generally close-ended Meant to limit misinterpretation Forced choice answers Quantitative Examples: Fill in the blank ○ Unstructured Open-ended Tend to allow for more liberal interpretation Particularly valuable in focus groups or group interviews Qualitative ○ Dichotomous Questions with two possible responses yes/no questions agree/disagree Vote between two candidates Matters for purposes of analysis, but also good for true forced choice Used sparingly Filter or contingency questions ○ Question asked to determine if respondent is qualified enough to answer a subsequent question Wouldn't ask how you feel about a course if you have not taken the course Double barreled question-AVOID  ○ Questions that potentially ask two (or more) things at one time ○ Difficult to interpret ○ Can fix with two questions ○ Example: How much do you enjoy the content in Howard Smith's classes and dad jokes? Interpretations of questions ○ Include necessary context to answer the question Respondents need information ○ Avoid asking questions about things respondents might not know Specific questions ○ Be specific What do you think of the last text? VS how difficult was the last test? General questions ○ Avoid being too specific How did you feel about the class last period? Avoid biased or loaded questions ○ Questions that lead the respondent in some way What is the worst part about following social distancing? Response format ○ How the respondent will respond Multiple choice Multiple select Open ended Rank ordering Semantic differential Question wording ○ Respondent need to know what you are asking Are you active on social media? Where are you from? What do you think about crime? What assumptions does the question make? ○ Sometimes people do not know something ○ Read questions carefully ○ Is there a need for a filter question? (Or more than one) Is the time frame specified?  ○ Word choice matters here Will May Could Should Might ○ Can also provide a time frame “Past 6 months” “Past year” “Ever” How personal is the wording? ○ A few words are sufficient to go from impersonal to personal “How do people like you” “How do you” Is the wording too direct? ○ Consider how questions might make the respondents feel, particularly with sensitive issues “Do you ever fantasize about perverted stuff?” Question placement ○ Mutually exclusive When all options are different from one another What is your age? ○ 18 or younger ○ 19-25 ○ 26-30 ○ 31-37 ○ 38 or older ○ Exhaustive How old are you 18-21 21-25 25-30 30 or older Every possible answer is there Opening questions ○ First impressions are important ○ First few questions set tone ○ Opening questions should Be easy to answer Might be descriptive  Not be sensitive questions Sensitive question ○ Some questions are uncomfortable or difficult for respondents ○ Develop trust or rapport ○ Warm up questions can help, but sensitive material should not come up abruptly or appear unconnected with the survey ○ Transition sentence can be helpful Topic 11- Qualitative design Qualitative research ○ Research that utilizes qualitative methodology ○ Yield information about crime and offenders ○ Less common in criminology and social science ○ More idiographic than nomothetic Types of qualitative data ○ Interview data Products of in depth interviews Direct audio recordings, written word for word transcripts ○ Direct observation Focuses on actions rather than words Includes documentation of people's behaviors rather than their responses to questions ○ Written documents or video/audio recordings Articles, social media, posts, videos Qualitative approaches ○ What is an approach? The plans and procedures for research that span the steps from broad assumptions to detailed methods of data collection The hows and whys of research Three qualitative approaches ○ Ethnography Most seen in the field of anthropology An approach that emphasizes broad aspects of the culture being studied in a great deal of depth with the researcher acting as a participant in the studied activities Time consuming, $$$ Not common in criminology ○ Phenomenology  Rejects the basic premises of positivism by suggesting that there are many things we cannot easily observe or measure and stating “.. and thats okay.” Emphasizes subjective experiences and interpretations and seeks to understand those rather than attempting to find some objective truth A traditional approach to studying drunk driving might be to simply ask someone if they have ever driven while intoxicated. This is a positivist approach. It assumes there is a point of intoxication after which people should not drive. ○ Field research Studying topics in the real world, the environment where the events actually take place Example: A researcher might go into courtrooms to observe behaviors rather than look at databases of outcomes of cases or trials Approaches vs methods ○ Approach More about the philosophy of conducting research ○ Method Has to do with the act of doing the research ○ Approaches and methods are connected Observational research (type of qualitative methods) ○ Individuals or groups are observed by researchers, quite often in their natural environments ○ Common in ethnography ○ Observational research in the field, in the real world, is known as naturalistic observation Participant observation ○ The researcher observes while also participating in the given observed activity ○ Interacts with the people and the environment being studied ○ One concern: Researcher somehow interferes with the research findings, leading to a less than natural situation True observations ○ Where the researcher has minimal interaction with those being studied Difference between naturalistic observation from controlled observation  ○ In a controlled observation, the researcher sets up an artificial environment to observe interactions in a controlled (laboratory) setting, naturalistic is outside of a laboratory ○ This is ideal for a researcher who wants to maximize internal validity Interviews ○ Can be structured or unstructured ○ Structured Predetermined set of questions asked to all interviewees from a script ○ Unstructured Start with a few basic questions then are intended to flow naturally like a conversation Content analysis ○ Unobtrusive analysis of written, verbal, or other physical content that looks for themes, word usage, or tone ○ Unobtrusive There is no interaction between the researcher and those being studied Word clouds ○ Look for common words in a set of words Case studies ○ An in depth study on a specific person, condition, organization, or social context ○ Not intended to generalize ○ Idiographic Topic 12- Statistics Statistics ○ Ways of summarizing date within datasets ○ Ways of explaining patterns within or between data points or datasets Descriptive statistics ○ Ways of summarizing (describing) the data in a given dataset in simplified form ○ Frequencies and percentages Works well with nominal data Summarizing groups (categories) ○ Measures of central tendency Measures of central tendency  ○ Ways of describing the typical or average values in a set of data, generally focusing on the middle point ○ Grade point average ○ Three most common types: Mode Most common value in data set Value that shows up most frequently Can use with all levels of measurement 5, 8, 8, 10, 9 ○ Mode is 8 Median The middle value in a set when numerical values are placed in order from smallest to largest Example: ○ 40, 60, 70, 80, 90, 100, 100 ○ median= 80 (3 higher and 3 lower levels) ○ Easiest to calculate when there are an odd number of values Can use with any numerical values Ideal measure when there are outliers in the data or data are skewed (lots of large or small values) Mean Arithmetic average of values Add up all values then divide by number of values present Example: ○ 50, 60, 70, 80, 90, 100, 100 ○ Add up values (470) ○ Divide by total values (470/7=17/14) Distribution of data ○ Frequency of each number (or a range of numbers) in a set of data ○ Shape of a graph with all values placed on a frequency chart ○ Where the values fall (in visual form) Normal distribution (The Bell Curve)  ○ Used to describe a data distribution shaped like a bell ○ The mean, median, and mode are theoretically the same or almost the same values Variability ○ Directly related to distribution of data ○ How far values diverge from the central point in the data (the spread of the data) Variability: Range ○ Calculated by subtracting the smallest value from the largest value in a set ○ Gives a rough estimate of the spread of data but does not provide the shape of the distribution ○ Example: 100, 180, 200, 3000, 1900, 1500, 150 3000-100 Variability: Skewness ○ How far the shape of a curve deviates from a bell-shaped curve- essentially this is how much it leans in one direction or the other ○ How much it “leans” in a specific direction ○ Exam: List of values whether the mode is bigger or smaller than the mead, Variability: kurtosis ○ How closely the distribution clings to the middle point ○ bell-shaped,flat, (values spread all over), steep (most values at middle point without much spread) Variability: percentiles ○ Where a particular value values on a continuum in relation to the other values in the continuum ○ How much of the distribution is as high as or smaller than the given value ○ Ex- 75th percentile scores as well as or better than 75% of those who took a test Outlier ○ A value that is far away from most numbers in the set, it is a statistical anomaly Central tendency and outliers  ○ Outliers are values that stray far from most values in the set ○ Can have a large impact on the mean value ○ Skewed data are examples, but can even be due to a single outlier Variance and standard deviation ○ Variance: A value we calculate to obtain the standard deviation ( this is an oversimplified definition) Standard deviation Solved as the square root of the variance Demonstrates how far values fall from the middle value, on average Range ○ Set of data is the largest value in the set minus the smallest value; it is the difference between the highest and lowest values ○ Can only be calculated with quantitative data Percentile ranking ○ Where a particular value falls on a continuum in contrast to other values on the continuum Inferential statistics ○ Statistics that allow us to estimate population parameters ○ Statistics that allow us to test relational or causal research hypotheses Things happen by chance ○ Winning powerball ○ Waking up from a coma after being “clinically dead” ○ Struck by lightning ○ All have a low probability Significance testing ○ Examines the probability that the findings of a statistical relationship occur due to chance ○ Estimated likelihood that a finding is due to chance alone ○ Is it a statistical fluke? Factors that influence significance ○ Effect size ○ Sample size More people in a study, more happy on result ○ Variability Nice bell curves Inferential statistics for hypothesis testing  ○ Statistical test chosen depends on many factors including: Number of variables (we will focus on bivariate tests) Type of independent variable (categorical or quantitative) Type of dependent variable (categorical or quantitative) Goals of research and research design Pearson’s r ○ Independent variable: quantitative ○ Dependent variable: quantitative ○ Positive correlation: as x increases, so does y ○ Negative correlation: as x increases, y decreases ○ No correlation: no statistical relationship between variables ○ Values vary from -1.00 to 1.00, bigger values more likely to be significant ○ Examples: There is a positive correlation between number of prior offenses and amount of months a person is sentenced to jail There is no correlation between sales of adult magazines and sexual assault rates There is a negative correlation between number of guns a person owns and self-reported fear of crime on a 7 point scale T-test ○ Independent variable: categorical and binary (two attributes) ○ Dependent variable: quantitative ○ Tells us if values differ across groups or attributes of the independent variable ○ Example: Men commit nearly twice as many violent crimes in their lifetimes than do women, on average Fear of crime on a 7 point scale is significantly higher among those who live in urban areas than those who live in rural areas Chi-square test ○ Independent variable: categorical ○ Dependent variable: categorical ○ Tells if the group proportions are different ○ Example: Women over the age of 21 are significantly more likely to report having ever been pregnant than women under the age of 21 Violent offenders are significantly more likely to recidivate than are nonviolent offenders Black jurors are less likely to state someone is not guilty than are white and hispanic jurors Independent variable Dependent variable Inferential statistic Quantitative Quantitative Pearsons’ r Categorical and binary quantitative t-test Categorical and nonbinary quantitative Analysis of variance (3 or more attributes) (ANOVA) Categorical Categorial Chi square quantitative categorical Test depends on other factors and beyond the scope of this course