Study Guides (248,000)
Canada (121,218)
Criminology (359)
CRIM 220 (19)

Final review.docx

9 Pages
Unlock Document

CRIM 220
William Glackman

Chapter 8: Unobtrusive & Archival Methods Unobtrusive Measures: Measures that are less influenced by the intrusion of researchers Non-reactive  The ‘data’ are typically produced without thought that the ‘evidence’ might someday be scrutinized by social scientists  The researcher arrives on the scene after, and sometimes long after, the participants have left Two categories 1. Physical trace measures: • Erosion: Wearing away or removal of products or materials because of our physical presence or activity Ex. Using the rate of tile replacement in a museum as an index of exhibit popularity • Accretion: addition to or building up of products or materials because of our physical presence or activity Ex. Using the setting of buttons on car radios as an indicator of listening preferences • Advantage: inconspicuous and anonymity • Disadvantage: invasion of privacy, issues of confidentiality, issues of consent & unknown representativeness of the information that survives i. Selective Deposit: Of all the individuals and societies that have ever existed on this planet, some individuals and some societies are more likely than others to have placed their beliefs and experiences into the historical record Ex. During Renaissance, we are likely to look primarily at the experience of the wealthy ii. Selective Survival: Among those things that are initially put into the historical record, some have a better chance of surviving the ravages of time than others Ex. Wealthy and powerful have advantage th iii. When studying life expectancies in 14 c. France, the marble tombstones of the wealthy would have survived longer than the wooden crosses of the poor 2. Archival measures: Any information that is contained in ‘hard copy’ records or documents • Advantages: o Encourage thinking in social processes o Allow longitudinal analysis to be done now o Sources exist (others can go back to a given document) o Costs less • Disadvantages: o Subject to selective deposit & selective survival  Government documents are more likely to be available than those of smaller groups or individuals o Possible shifts in how particular data series are designed and in the procedures for recording or saving material Memory Fade: The tendency for events to be forgotten Telescoping: The propensity of respondents to bring events that were outside the sample period into it • Ex. Respondents including events that happened more than a year ago in victimization surveys within the last year • Bounding: instead of saying ‘in the last six months’, you might cite a more memorable event such as ‘since Christmas’ Chapter 9: Manipulative control and the logic of experimentation Manipulative Control: the active and intentional manipulation of the setting by the researcher in order to maximize clarity of inference by the controlling rival plausible explanations. Manipulative control is epitomized by the laboratory experiment, where the experimenter exerts control over every aspect of the setting Analytic Control (Contrast to Manipulative control): One of the two general approaches to research that attempts to make inferences about ‘causes’. Researcher takes an existing situation and, through their analytic powers alone, tries to make sense of the causal patterns that exist INDEPENDENT & DEPENDENT VARIABLES Independent (treatment) variable: A variable or set of variables whose effect we wish to assess Dependent (outcome) variable: Some way to measure the effects of the independent variable INTERNAL VALIDITY The extent to which differences observed in the study can be unambiguously attributed to the experimental treatment itself, rather than to other factors • Can we be confidant that the differences observed are caused by the independent variable? • Rival Plausible Explanation: alternative factors that might also have accounted for the results you observe. Sound research design aim to minimize these Threats to internal validity • History: pretest/posttest designs to the specific events occurring between the first and second measurement in addition to the independent variable o What happened between the pretest & posttest that might also account for the results? • Maturation: participants themselves that change as a function of time per se (such as growing older, more tired, more hungry, etc.) • Testing: the effects of taking a test on scores in the second testing o Pretest sensitization: You may become sensitive to the issue involved in a way that you wouldn’t have been otherwise o Practice effects: practice administered during pretest may effect your posttest. Abilities may have improved because of the practice the pretest gave you • Statistical Regression (Regression toward the mean): the propensity of extreme scorers on the first testing to score closer to the mean of the group on the second testing o If a group is picked because of the extremity of their scores • Selection Bias: violates pretest equivalence • Instrumentation Changes: changes in the way data are collected or organized during the course of a study Controlling Rival Plausible Explanation • Control Group: Treated the same way as experimental group but does not receive independent variable o Comparison groups that help us isolate and assess the impact of a particular variable o Don’t always have to be a ‘do nothing’ group o Experimentalist Logic: If you have two groups who are equal to begin with and if you treat them equally in all respects except one, any subsequent differences between the groups must be attributable to the one variable on which the group differed o Assumption of pretest equivalence: If we can assume out experimental and control groups were equal to begin with and if we use appropriate controls, high internal validity will result • Selection Bias: Makes the locus of the onserved change more ambiguous. Can be solved by: o Random Assignment: letting ‘chance’ be the sole determination of which group any given person is a member of  Allows for equal distribution across all the groups in the study  Posttest-only control group design: pretest is considered redundant in this design, since one can assume that the groups are equal to begin with thus, this design circumvents the problems associated with the pretest sensitization o Matching: intentionally R ­­­­­­­­­­­­­­­­­­­­­­­­­­­­­­­­­­­Experimental Group O1 creating pretest Random assignment of participa­­­­­­Control Group O2 equivalence Posttest only, given   Identify to both groups pairs of individuals who are matched on some variable and then randomly assign one person from each paarit to the experimental group and the other to the control group.  Can get out of hand if you try to match participants on more than one variable EXTERNAL VALIDITY The generalizability of results beyond the specifics of the study • Does not necessarily have to do with how representative ones sample is • Ex. Would we obtain the same results if we did it with other students? In other courses? In another semester? ECOLOGICAL VALIDITY Representativeness of the treatments and measures you use in relation to the particular environment to which you wish to generalize • The study must approximate the real-world that is being examined STATISTICAL CONCLUSION VALIDITY • Many statistical tests are designed to tell us the exact probability of obtaining the results we observe if chance alone is operating • Statistical significance does not imply that what you found is important or internally valid Chapter 10 - From Manipulation to Analytic Control: Quasi- Experimentation and Case Study Analysis Quasi-Experimentation •Donald T. Campbell supplied a vocabulary and set of dimensions on which research might be evaluated, and argued that researchers should not confuse the trappings of experimentation with its underlying logic o E.g. "internal validity", "external validity" •If manipulation of the independent variable and random assignment to groups become difficult to accomplish in field settings, we should not resign ourselves to the feeling that causal inference is impossible o Should look to underlying logic of experimentation for alternative procedures that will fulfill these same experimental objectives •Campbell states that the temporal precedence criterion of causality is relatively easily dealt with in many field situations by the acquisition of time-series data The Logic of Quasi-Experimentation •Was there a change in the dependent variable? •Was the independent variable really manipulated? o 2 of Mill's criteria must be established: • The presumed "cause" was temporally precedent to the alleged "effect" • The cause and effect did covary in time •Eliminating Rival Plausible Explanations o Make a cognitive shift in our approach to experimentation o Threats to internal validity • Can address it by gathering time-series data that show the maturational trend of the data o History • What events might have occurred between the pretest and the post-test other than the independent variable? o Statistical regression/regression to the mean o Emphasis is on informed critique as a source for generating rival plausible explanations • Onus is on the researcher both to recognize potential rival explanations that are actually plausible and to identify the sorts of data that will address those possibility o Resulting analysis will leave you with either: • No rival plausible explanations remaining • Some (or many) rival plausible explanations remaining New Threats to Internal Validity • Manipulative control was not the be all end all of experimentalist research o The ability to create situations at whim and randomly assign people to this or that group •Analytic control could accomplish the same means o Using our heads to anticipate and address all rival plausible explanations •Diffusion or Imitation of Treatments o In the field, it's not unusual to be faced with intact groups that can't be isolated from one another • May find that the boundaries between the groups are or become somewhat blurred o Any communication between groups will make each group a little more like the other o Might leave you concluding erroneously that the treatment variable is ineffective • Compensatory Equalization of Treatments o When the treatment is being evaluated involves goods and/or services deemed desirable, and where there is a large disparity between groups • Administrators, those charged with implementing the treatment, or some of the recipients may reroute some goods or provide access to services among some or all members of the disadvantaged group in an effort to alleviate disparity o May act to obscure differences between groups, leading to an erroneous finding of "no difference" •Compensatory Rivalry o Knowing that one is in the disadvantaged group may spur a competitive spirit to overcome adversity and perform well • New programs and legislative changes are often much more public than experiments o Also called the John Henry effect o Obscures "real" differences between groups •Resentful Demoralization o Individuals in the disadvantaged will perceive the result as adverse and inevitable o May not even try to compete, and may even intentionally reduce their performance o Researcher may overestimate the act
More Less

Related notes for CRIM 220

Log In


Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.