Study Guides (248,246)
Canada (121,429)
York University (10,192)
Psychology (1,203)
PSYC 2030 (60)
Final

Research Methods Final Exam Notes.doc

13 Pages
647 Views
Unlock Document

Department
Psychology
Course
PSYC 2030
Professor
Krista Phillips
Semester
Fall

Description
Research Methods - Final Exam Notes Chapter 7 Randomized experiments • Done by random assignment (also called randomization) • Each person in a population of subjects has an equal probability of being chosen at every draw • Randomized experiments considered gold standard in biomedical research • Three reasons for using random assignment 1. Provides a safeguard against the possibility of researchers subconsciously letting their opinions or preferences influence which sampling units will receive any given treatment (sampling units = participants) 2. Random assignment distributes the characteristics of the sampling units over the experimental and control conditions in a way that will not bias the outcome of the experiment 3. Random assignment permits the computation of statistics that require certain characteristics of the data. Provides a mechanism to derive probabilistic properties (p values) of estimates based on the data by controlling for extraneous variables How Is Random Assignment Accomplished? • First subject gets one treatment, second subject automatically gets other. Provides equal number for treatment and control • Idea of drawing names, whether literally or computer based Box 7.1 – Imperfect Randomized Trials • Randomized trials can fluctuate with regard to their potential value • Perfect control is hard to achieve because patients who suspect that they are in a placebo control group may attempt to obtain experimental treatment • Patients who experience negative affects to treatment may alter their dosage • Assigning terminal patients to placebo group could have legal ramifications, being denied access to a potentially lifesaving drug • Knowing that randomization is being used may make some patients wary of volunteering and could jeopardize the generalizability of the results Between-Subjects and Within-Subjects Designs • Between-subjects design – subjects are exposed to one condition each o X # of subjects receive condition A, X # of subjects receive condition B • Also called nested design, subjects are nested within their own groups or conditions • If all subjects receive both condition A and condition B, within-subjects design • T-test usual procedure for analyzing such data • Also called repeated-measures design, because the subjects reactions are measured after each condition • Also called crossed design, subjects “crossed” by conditions • Not limited to two groups Factorial Designs and Latin Square Designs • Conditions as arranged along a continuum or single dimension, it is described as a one-factor or one-way design. Factor = variable of interest • Suppose women and men are randomly assigned to a drug or a placebo group, now have a two-factor design with two levels of the variable of gender and two levels of the variable of treatment • Factorial design • Two levels of two factors, described as a 2x2 factorial design, or a 2 factorial design • When one factor is between subjects, and other is within subjects, this design is called a mixed factorial design • Order of treatments can cause issues • Address this problem by system of counterbalancing, rotating the sequences • Some get condition A first followed by condition B, and some get opposite • Latin square design has counterbalancing built in (subject on Y axis, order on X axis, and which treatment in body) Causality “Shrouded in Mystery” • Aristotle identified four kinds of causality (curve ball example, batter swings and misses, what caused the ball to break that way?) • Material causality – substance or substances necessary for the movement of something (the physics of the baseball caused it to break) • Formal causality – refers to the plan or development that gives meaning to the event (the idea of throwing a curve ball, the thought) • Final causality – refers to the objective or purpose of the event (have the batter strikeout) • Efficient causality – refers to the activating force that was responsible for the event (the act of throwing the ball) • What about in human development? • Cellular structure is the material cause, DNA or genetics is the formal cause, physiological maturation is the final cause, parenting as an environmental variable is the efficient (activating or instigating) cause • Efficient causality generally in mind when talking about manipulated treatment On What Grounds Do Scientists Infer Causality? • Hume’s three circumstance of contiguity, priority, and constant conjunction • It is possible to rule out plausible rival causal explanations • Covariation – a fusion of what Hume called contiguity and constant conjunction. Cause and effect is not necessarily constant, but likely or probable • Temporal precedence – what Hume called priority, the assumption that the cause always precedes the effect • Internal validity – attempt to rule out rival explanations 1. Scientist looks for evidence that the independent variable and dependent variable are mutually related (covary) can be shown through correlation - causation implies covariation, covariation does not imply causation 2. Y did not occur until after X occurred. A later event can not be the cause of an early one (temporal precedence) 3. Third criterion is to figure out ways to anticipate and rule out threats to internal validity. Cannot anticipate all rival explanations Proceeding even within the more limited framework of the three criteria of covariation, temporal precedence, and internal validity, researchers find they must settle for the best evidence available even if the evidence is inconclusive, thus causal inference is always subject to some degree of uncertainty Box 7.2 Hume’s “Rules” • Eight rules to judge causes and effects What is the Formative Logic of Experimental Control? Mill’s Methods • Agreement and difference – together provide the formative logical basis of experimental control • When two independent groups are comparable in all respects except for some intervention that is operating in one group but not in the other, the intervention or manipulated variable is implicated as the probable cause responsible for observed differences • Method of agreement states “If X, then Y”, X symbolizing the presumed cause and Y the presumed effect. Two or more instances where Y occurs, and only X is present, X is a sufficient condition of Y • Method of difference states “If not-X, then not-Y” If Y does not occur when the presume X is absent, then X is a necessary condition of Y, meaning it is indispensable • Created idea of control group. Box 7.3 “I Shall Please” • Placebo in Latin means I shall please • Placebos truly have an effect • Receive real treatment, but believe it to be a placebo, less of an effect • Mind over matter What Are Pre-experimental Designs? • A research design which does not fit the standards of an authentic experiment • So deficient in control that they are especially vulnerable to causal misinterpretations • One-shot case study o Symbolized as X-O where X = exposure to an event or experimental variable, and O = an observation or measurement o No comparison is made, no control group • A small improvement on this would be to measure the subjects before and after exposure to the treatment. Still can do this with a small sample • One-group pre-post design. This pre experimental design symbolized as O-X-O • Still cannot rule out uncontrolled events between X and O Circumstance that Jeopardize Internal Validity Before-after design • Two randomized groups, test both pre treatment, treat one group while one is control, look at post treatment • Tells us how much improvement occurred, can calculate difference between scores Solomon design • Identified pre testing may sensitize participants to the treatment, distorting the outcome • Control for this by using the four-group composite by 4 groups, and doing the before-after design on two and not doing it on two. Posttest all groups, but only pretest two of the groups (one control and one treatment) Four threats to internal validity • History – implies a plausible source of error attributable to an uncontrolled event that occurs between the pretest and the posttest o Random events can throw off measurements • Maturation – refers to certain intrinsic chances in the research participants, such as their growing older, wiser, stronger, or more experience between pretest and posttest • Instrumentation – refers to the intrinsic changes in the measuring instruments, such as deterioration o An effect may be due to unsuspected changes in the instruments over time o Instrument could be judges, who change over time • Selection – the selection of participants, when there are unsuspected differences between participants How Can I Control for Demand Characteristics and Expectancy Effects? • Artifact – a finding that results from conditions other than those intended by the experimenter • The experimenter’s observations having an effect on participants actions • Demand characteristics – M. T. Orne’s term for the mixture of hints and cues that govern the participants perception of (a) his or her role as research subject and (b) the experimenter’s hypothesis o Volunteers acting as a good participant, one who is sensitive to demand characteristics • Quasi-control subjects be used to counteract this o Research subject who are asked to step out of their traditional roles and to serve as coinvestigators, rather than objects of study o Asked to reflect on the context in which the experiment is being conducted • Experimenter-related artifacts, such as sources of bias or systematic error • Experimenter expectancy effect • Counteract by using blind experimenters o Unaware of which subjects are to receive which treatment o Double blind procedures, neither the subjects nor the experimenters know who is in the experimental and control groups • Also can use a factorial design that assesses of expectancy effect vs. the phenomenon • Expectancy control design Chapter 8 -looking at different nonrandomized procedures How is Causal Reasoning Attempted in the Absence of Randomization? • Prospective data – collect data by following your reaction forward in time • Individuals are observed and measured repeatedly through time • Retrospective data – data collected back in time • Based on circumstantial evidence Box 8.1 Quasi-Experimental Research • Campbell and Stanley (1963) • Identify studied that do not randomly assign individual units to treatment conditions What Is the “Third-Variable” Problem? • Causality implies correlation, but finding relationship does not tell why they are related • Possibility of a third variable that is correlated with both X and Y is the reason X and Y covary: This is the third-variable problem • Big feet in children correlate with higher grades. Is it their big feet that make them smarter? No it is their age • In nonrandomized research, an uncontrolled or unmeasured variable that is correlated with X and Y may account for the association between X and Y, third variable is the actual determinant How Can Causal Effects Be Studied in Nonequivalent Groups? • Non-equivalent-group designs are between-subjects designs in which the sampling units are allocated to experimental and control groups by means other than randomization and are observed or tested before and after the experimental treatment School A NR O X O School B NR O O • Two ways to improve non-equivalent groups design: a) large samples b) relevant subgroups that are well stocked with sampling units • Subclassifcation on propensity scores reduces all of the variables on which the treated and untreated subjects differ into a single composite variable – a propensity score o A summary statistic of all the differences on all variables on which the treated and untreated subjects differ Box 8.2 – Wait-List Control Groups • Cant use a random assignment procedure due to concerns about depriving the control group of the experimental treatment (not giving a treatment that could really help) can use wait-list control group Group 1 R O X O O Group 2 R O O X O What Are Time-Series Designs and “Found Experiments”? • In time-series designs, the defining characteristic is the study of variation across some dimension over time • When the effects of some intervention or treatment are inferred from a comparison of the outcome measures obtained at different time intervals before and after the intervention, the data structure is called an interrupted time series design • Time series means there is a data point for each point in time, and an interrupted time series means there is a dividing line at the beginning of the intervention. Marks the start of treatment • Phillips, referred to his studies as found experiments because they are essentially found in naturally occurring situations • Studied clustering of suicides after a series of televised new stories and televised movies about suicide What Within-Subjects Designs Are Used in Single-Case Experiments? • Single-case experimental research, conceptualized as a subcategory of interrupted time-series designs • Only one sampling unit is studied, or only a few units, and repeated measurements are taken of the unit, random assignment is rarely used • Behaviour baseline – similar to a pretest, analyzing behaviour before any type of treatment • Instead of Xs and Os, single case researchers use A-B-A design, which evolved out of the A-B design • A phase – no treatment, B phase – treatment • First A in ABA is the baseline • Researcher observes steady continuous behaviour, the treatment (B) is introduced • Treatment eventually removed • A-B-BC-B design, B and C are two different interventions. Measures B without C and in combination with C • A-B-A-B design Box 8.3 – Superstition in the Pigeon and the Financial Market • Single case, birds developed superstitious movements as feeder was pushed in to cages • People infer causal connection between two occurrences when
More Less

Related notes for PSYC 2030

Log In


OR

Join OneClass

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

Sign up

Join to view


OR

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.


Submit