Textbook Notes (362,880)
Canada (158,081)
Psychology (9,549)
PSYB01H3 (585)
Anna Nagy (283)
Chapter 9-12

PSYB01 Notes Chap. 9-12.doc

21 Pages
Unlock Document

University of Toronto Scarborough
Anna Nagy

Chapter 9 - Method used to choose subjects for experiment  profound effect on external validity (study generalized to other populations/settings) - Important to accurately describe the population = PROBABILITY SAMPLING (scientific polls) - Setting the Stage: o Supply participants with information necessary for them to give their informed consent to participate  Underlying rationale of the study (does not have to be completely truthful to AVOID reactivity) • Must tell truth in debriefing o Experimental setting must seem plausible to participant o No clear-cut rules to setting the stage - To manipulate the IV = construct operational definition (Chapter 4) o Operational Definition  turn a conceptual variable into a set of operations: specific stimuli, event, instructions constructed that will be presented to the research participants  Set of procedures used to measure or manipulated variables (:. Studied empirically) e.g. pain questionnaire  Ensures abstract concepts are discussed with concrete terms  Help researchers communicate their ideas to others (understanding of definitions and terms) o E.g. Researcher changing the conditions to which participants are exposed to o How variable is manipulated depends on:  Variable  Practicality  Cost  Ethics - Manipulating the IV (introducing different levels if the IV) o Straightforward Manipulation  Low realism  Usually involve stimuli in the form of photos, video, text, etc.  More commonly done  quick  Participants don’t require to do too much  Most experiments use straightforward manipulations o Staged Manipulation (High Impact Experiment)  Intended to involve the participant in what is going on • Put participants in a psychological state  frustration, anger, lower self esteem • Tp simulate situations that may happen in the real world  High level of realism  Participant experience something • Milgrim’s study + electrical shocks + obedience • Solomon and Ash + study of conformity + lines of different lengths +% of time the participant agreed to the wrong answer?  Behaviours participants engage in when they don’t know what’s going on  Use of confederate(s)  used for conformity studies, field experiments, lab research etc. • Actors  Problems: • Ensuring that confederates act exactly the same for every participant (extraneous variable) o Subtle interpersonal communication difficult to REPLICATE THE EXPERIMENT • Complex staged manipulations are DIFFICULT TO INTERPRET o Many things happen in the experiment, what ONE thing was responsible for the result? • Time consuming - Strength of Manipulations o Simplest experimental designs have two levels of the IV  Must choose the levels (as a researcher) o Strong manipulation = maximizes differences between the 2 groups + increases the chance that IV  DV is statistically significant  To ensure that a relationship does exist  E.g similarity in attitude vs. measure of liking  level 1 = not much similar, level 10 = VERY similar • 2 Considerations: 1. external validity of a study - Strongest manipulation must entail a situation that rarely, if ever, occurs in real world  is this manipulation reflective?? 2. ethics - Strong as possible within the BOUNDARIES of ethics 3. costly - Afford equipment + confederates? - Measuring the DV - DV in most experiments = 3 types  Self Reports, Behavioural, Physiological o Self-report measures (written or verbal  on thoughts and behaviour)  Rating scales with descriptive anchors (example)  Advantages: • Convenient, easy to construct and administer • Allows greater precision (instrument is highly developed) questionnaires • Private behaviour, opinions, thoughts, emotions • Exposes thoughts, attitudes that people may not acted on  Disadvantages: • Rely on honesty, memory • Social desirability responding say what they think they want the researcher to hear (socially aware)  e.g doctor asks how many drinks you drink  you lie o Behavioural Measures  Measure how people behave in a situation  Observable Behaviours  Can include facial expressions, proximity, large movements , micro-movements, reaction time  Must be • (1) concrete & (2) codeable • E.g. People walking in a train station in Tokyo o Want to see how many are dressed in business/casual attire, alone/companion  Must be easy to code WHAT is considered “having a companion”/ “business” (boundaries = coding)  Advantages: • Visible, external indicators of inner states (hard to fake inner state  hard to disguise) • More spontaneous, less filtered than verbal measures • No verbal skills required (children that are preverbal, people who speak diff language, animals)  Facial Expression as a Behavioural Measure - muscles of face expand and contract in different configurations depending on the emotion - particular patterns of expansion & contraction of muscles create certain expressions (even if someone is trying to conceal expression, muscles still subtly move) - -see if people are lying  Implicit and physiological measures • Implicit measures access automatic reactions and evaluations. o IAT (implicit association test (google) look at reaction)  To measure attitudes  very little control on reaction time/influence test • Pairing stereotypes association with female/male &career/family • Bogus pipeline o Makes a machine that looks like it can see if one is lying or not, participant is more likely to speak the truth o Physiological measures (data must be applicable to what you’re studying) o galvanic skin response (GSR) general emotional arousal and anxiety (electrical conductance of skin) sweat, heart rate, respiration o electromyogram (EMG)  measures muscle tension to frequently measure stress and tension o electroencephalogram (EEG) measure electrical activity of brain  arousal of different parts depending on situation o MRI + fMRI • Autonomic arousal  measure body state of readiness to fight or flight  Stress • EMG (data related to the functioning of the heart), cortisol level (stress horomone)  Testosterone level • Under threat = more testosterone released  Neural activity • Depends on level of technology equipment EEG, MRI, fMRI data (shows structures +patterns of activation in brain) o Considerations: o Usually experiments have MORE THAN ONE MEASURE OF THE DV  variable can be measured in more than one concrete way  if IV has the same measure for several measurements of DV = confidence in results increase  ORDER EFFECT? • Counterbalance it • Present the most important measures first, less important at the end (if there is no indication that order is a serious problem in the research) o Sensitivity of DV  DV should be sensitive enough to detect differences between groups  Sensitivity IMPORTANT when measuring human performance  memory = reaction time, recall, recognition  Ceiling Effect: the IV seems to have no effect on DV only because the participants quickly reach the maximum performance level  Floor Effect: task is so difficult, hardly anyone can perform well o Some measures more costly than others  self reports = inexpensive, behavioural + physiological measures = $ - Additional controls needed to ensure more internal validity/address alternative explanations ( besides just experimental + control group) o Controlling for Participant Expectations  Demand Characteristics: any feature of the experiment that might inform the participants of the purpose of the study  p’s reactivity changes  want to prove the hypothesis as correct • Deception  o To make p’s believe the study to be about something else  disguise through cover stories + filler items in questionnaires • Assess Demand Characteristic o Asking the p’s what they perceive as the purpose of the study  Demand characteristics eliminated when people are unaware that an experiment is going on (field settings/observational research)  Placebo Effect: administer fake pills + real pills and compare (Placebo group added to take placebo pill). If there is the same effect of IV  DV, then the result is caused by the placebo effect (just administering a pill or an injection may be sufficient to cause an observed improvement in behaviour • Balanced placebo design  no alc = no alc, no alc = alc, alc = no alc, alc = alc o Belief that one consumed alc is more important in determining behaviour than alcohol itself • Beneficence: control group will receive treatment as soon as they finish study (ethical) o Controlling for Experimenter Expectations  Expectations of experimenter can turn into biases experimenter bias or expectancy effects  Knowing condition = • Treat participants differently when conducting experiment • Record of behaviours  experimenters may interpret behaviour slightly differently  Teacher expectancy affects pupil performance  Expectancies communicated in VERBAL and NONVERBAL ways  Solution: • Experimenters practice behaving consistently with all participants • Run all conditions simultaneously so experimenter’s behaviour is the same (feasible for some situations  experimenter must give instructions to all p’s) • Procedures are automated • Single-blind experiment: participants unaware of experiment purpose • Double-blind experiment: participants + experimenter unaware of hypothesis/type of group (hire experimenters) - Research proposal (present tense): literature review that provides background on the study (similar to intro+method of journal article) o Why the research is done (what answers will the research give) o Procedure to measure results o Plans of analysis - Pilot Study: trial run of experiment with a small # of p’s o Reveal if participants understand the instructions o Total experimental setting is plausible? o KINKS  P’s think aloud protocol + ask p’s about the experience • Experimenters figure out kinks, become more comfortable in that role - Manipulation Checks: attempt to directly measure whether the IV manipulation has a intended effect on the DV o Provide construct validity of manipulation  Used in pilot study + ACTUAL study (towards the end of the test to avoid distraction/revealing of purpose of study)  Advantage: • Can see if manipulation is successful in pilot study (change accordingly) • Reveal non-significant results  Two ways to check manipulation  explicit measure of the independent variable  self esteem measure, measure mood in a mood inducing experiment  questionnaire for participants after the experiment whether they perceived and how they interpreted the manipulation - Debriefing: interaction between experimenter and participant AFTER the experiment o Discuss ethical and educational implications of study o View the perception the participant has on the study  Can be written or oral • Must have: o Purpose of experiment, some background, logic of experiment o Contact info for key people/investigator in the research o Option to withdraw data • Can also include (optional): o Options for counselling or other assistance o Request to maintain study details confidential (request p’s to not tell people what happened in the study) - Statistical Analysis  to reveal if there is an actual relationship between the IV and DV - Communicating Research to others: o Professional Meetings o Journal Articles (past tense)  peer review (2+ reviewers)  90% of papers submitted to prestigious journals are rejected - Participants: o How many?  A bigger sample is better UP to a point • 1500 similar to 15000  30 per conditioned considered minimum (no reason why)  E.g. if it’s 3x2 design  6 conditions x 30 =180 participants o Who?  Convenience  undergraduates from university  Snowball  you ask your friends to ask their friends etc., sending through email word of mouth  Internet-based  studies online (less control on the participants who will be doing it o Probability or non-probability sample?  Need something done fairly quickly  non-probability sample  Sometimes probability samples aren’t necessary Chapter 10 - More levels of IV can = more information on the exact relationship between the IV and DV - 2 IV levels = linear - Multiple IV levels = monotonic - Curvilinear = at least 3 levels of IV (inverted-U = like a triangle) o Exist in psychology Factorial Designs • More than one i.v. • Simplest form  2x2 factorial design (2 IV with each having 2 levels) = 4 conditions o 2x3  2 IV, one having 2 levels, the other having 3 levels  6 conditions o 2x3x4  3 IV, one with 2 levels, 1 with 3 levels, 1 with 4 levels  24 conditions Experimental Design Names • 2 levels of the first iv, 3 levels of the second o “2x3 design” o Ex. Coffee drinking x time of day DV: how long it takes for you to sleep that night  Coffee drinking 2 levels: coffee or water  Time of Day 3 levels: morning, noon, night • 3 levels of the first iv, 2 levels of the second,4 levels of the third o “3x2x4 design” o Must have a lot of participants unless you’re doing repeated measures o Ex. Coffee drinking x time of day x exam duration  Coffee drink has 3 levels: 1, 2, 3 cups  Time of day 2 levels: night and day  Exam duration has 4 levels: 30, 60, 90, 120min o 1 cup, night, 30 1 cup, night, 60  1 cup, night, 90 ETC. • Types of Information: o Main Effect: effect of the IV (2 IV = 2 main effects) by itself  Overall relationship between the IV and DV o Interaction: interaction between two IV  effect of one IV depends on the particular level of another IV • Need to consider independent effect of each i.v, and joint effects. (i.e. does effect of the first IV depend on level of 2nd? ) o Ex. Personality of person (introvert/extravert) and level of noise in room (both are IV) and how it affect the memory (DV) o Ex. If we were looking at GENDER and TIME OF EXAM, these would be 2 IVs. GENDER : 2 levels, TIME OF EXAM: multiple levels  morning, noon, evening  DV = performance on exam • IV x PV Designs (IV by Participant V): factorial design that includes experimental and nonexperimental variables o Allow researchers to investigate the different types of individuals respond to the same manipulated variable  PV  gender, age, ethnic groups, personality, clinical diagnosis category etc  Simplest form  1 manipulated IV with 2 levels x 1 participant variable with 2 levels • E.g. Distraction x Participant Personality o Distraction levels: TV on vs. Silence o Participant Personality levels: extravert vs. introvert • Interaction discussed through moderator variable (influences the relationship between the 2 variables) o E.g. distraction = poor reading comprehension  moderator variable is introvert (moderates relationships between the other variables)  Distracted resulting in poor reading comprehension ONLY when the participant is an introvert • Several possibilities of results for 2x2 Factorial Design: o May or may not be significant main effect of variable A o May or may not be significant main effect of variable B o May or may not be a significant interaction between the 2 IV (A x B) • Analysis of variance: to assess the STATISTICAL SIGNIFICANCE of the main effects and interaction o Significant interaction  look at simple main effects (examines mean differences at each level of the IV) • Assignment Participants to Conditions (2 basic ways) o Independent groups design: different participants assigned to each of the conditions  2x2 design. 4 conditions. 10 participants per condition = 40 participants in total o Repeated measures design: same participants in each of the conditions  2x2 design. 4 conditions. 10 participants needed for all 4 conditions o Mixed factorial design: mixture of both designs  2x2 design. 4 conditions. 10 participants do repeated measure of one variable A in 2 levels (extroverted participants experience distracted and silences), 10 other participants do repeated measure of one variable B in 2 levels (introverted experience distracted and silences) • Graphs used when horizontal variable is QUANTITATIVE Chapter 11 - 3 types of special research situations: - Single-Case Experimental Design (NOT THE SAME AS CASE STUDY): o Came from research on operant conditioning by Skinner o Often seen as clinical, counselling, educational, applied settings etc. o Determine whether an experimental manipulation had an effect ON A SINGLE PARTICIPANT/GROUP o TIME SERIES DESIGN  Subject behaviour measured overtime during a baseline control period  Manipulation introduced in the treatment period • Effectiveness of manipulation = behavioural changes from baseline and treatment period Advantage of Small-N Design - Great for Participants from hard to find populations - Results easy to interpret (often no statistics) - Can focus on helping one (few) participant(s)   Problem: • There may be many other explanations for behavioural changes (alternative explanations) besides experimental treatment o Single-case designs addresses problem o Reversal Design: demonstrate reversibility of the manipulation  A(baseline)  B(treatment)  A(baseline)  ABA design • A single reversal is not extremely powerful evidence for effectiveness of treatment :. o Random fluctuation of behaviour?  ABAB design  ABABAB design • Ethical reason points out that it doesn’t seem right to end the experiment without treatment o Multiple Baseline Designs: • Some reversals to baseline may be IMPOSSIBLE or UNETHICAL • Treatments may be irreversible  Effectiveness of the treatment is demonstrated when the behaviour changes only AFTER the treatment is administered (such change must be observed under MULTIPLE circumstances (to rule out extraneous variables)  Multiple baseline across: • Subjects  behaviours of several subjects measured over time (each subject, manipulation introduced at different points in times) = rule out explanations based on chance, history, events, etc. o Same behaviour for 2 or more individuals (2+ individuals) • Behaviours  several behaviours of the same subject measured over time (different times) o 2 or more behaviours for same individual • Situations  same behaviour measured in different settings (different times) o Same behaviour for 2 or more individuals (2+ individuals)  Individuals asserting own control o Replication = generalizability of results o Single-case Design  present results from each subject individually rather than a group mean (group means can sometimes be misleading to representing individual responses to manipulations)  Valuable for someone who is applying change technique in a natural environment (teacher new style) - Program Evaluation: research on programs that are implemented to achieve some positive effect on group individuals o Implemented in schools, work settings, entire communities (E.g. DARE in schools reduce drug use) o 5 types of evaluation:  Needs Assessment  is there, in fact, problems that need to be addressed in a target population?  Program Theory Assessment  program designed to address problem (must have valid assumptions for the causes of the problems and the rationale of the proposed program • Involve collaboration of researchers to determine if proposed program does address problem of the target population in an APPROPRIATE WAY  Process Evaluation  program researcher monitors program to ensure that the target population is being addressed  implementation of the program okay?  Outcome Evaluation  intended outcomes of the program being realized?  must measure outcome (DV) and see the relationship between IV and DV  Efficiency Assessment  is the program worth the resources it consumed (after being proven that program IS effective), resources put in better use when implemented, cost worth it? - Quasi-Experimental Designs o Study the effect of an IV in settings which the control features of the true experimental designs cannot be achieved o Examine impact of IV to DV  causal inference more difficult because lacking true experimental designs (such as random assignment to conditions) o Group is usually NATURALLY OCCURING (same nature as an experiment, but no tight control over experiment) - One Group Posttest ONLY Design: o One-shot case study, lacks a control group/comparison group o Sitting on the bus with a stranger (closely), seeing how long it takes for them to leave (9.6 sec)  DV due to IV??  stranger would have sat that long anyways? Sat longer ‘cause he liked you? o Lacks internal validity (without the comparison data)  No causal inferences - One Group Prettest-Posttest Design: o Measure participants before the manipulation, then once after  E.g. Participant  measure smoking  relaxation program  measure smoking o Problem: Threats to internal validity • History o Anything that happens between the pretest and the post-test o Event affects study outcome  E.g. Amanda Todd who committed suicide after bullying (made viral video) • If this happened during your study on bullying, the results may be affected for the post-test (less instances of bullying) • Maturation o Change in subjects over time th th  E.g. Doing a study of bullying on students from 8 grade to 12 grade, the participant may mature (change due to experiences they learn, education) o Systematic changes over time • Testing / Repeated testing o Repeated testing of IV in different level  bullied, having bullied (questionnaire will allow students to think more/analyze/understandg bullying) which ultimately affects the frequency of bullying o Testing itself in pretest affects the participant’s behaviour • Instrument Decay o Basic characteristics of the measuring instrument changes over time  E.g human observers used to measure behaviour  gaining skill, boredom, change standard of observations • Mortality o Subjects dropping out  If the subjects are a specific type of people =problem • E.g. smoking sensation study and hypnosis o A specific TYPE of person (one that is low in suggestibility – thus less impact from hypnosis treatments) will more likely drop out • Regression to the mean o High or low measurements followed by measurements closer to group mean  E.g. intervention study for bullying • You look for a school that has a high score of bullying and you apply the invention o You see that the invention has reduced bullying to the average amount  Problem: the high score of bullying may be due to day-specific reasons (perhaps, that day, the instances of bullying is oddly higher…but normally, it is at the mean) therefore, may give you results of more success than there actually is  Occurs when you gather a set of extreme scores taken at one time, and comparing it with another point in time  reliability of measure?, measurement error - Nonequivalent Control Group Design: o Employs a separate control group, but the participants in the two conditions are NOT equivalent o Differences = confounding variable o Selection Bias: participants from the two groups in the experiment are chosen from NATURAL groups  Researcher has no control which participants are in each group • Participants are not equivalent PRIOR to the program - Nonequivalent Control Group Pretest-Posttest Design: o Assignment to groups is NOT random o One of the most USEFUL quasi-experimental design o Advantage: knowing pretest scores  Even if groups are NOT equivalent, researchers can look at the CHANGES in scores from pretest to posttest - Solving Nonequivalent Groups o Matching pairs between treatment and control group (propensity score matching) o matching instead of randomization • match participants based on variables key to your study (in both control and experimental group) o time series analysis (interrupted time series design)  improve the interrupted time series design with a control series design  find a similar control group o unit of analysis not people • Can be an organization, program, etc. - Developmental Research Designs o Cross-sectional Method  Persons of different ages are studied at ONE point in time (independent group study)  Less expensive  Cohort  group of people born at about the same time, exposed to the same events in society, influences by the same demographic trends such as divorce and family size  Differences between ages = to cohorts effect confounds age with cohorts effect o Longitudinal Method  Same group of person are measured at different points in time (as they grow older) (repeated measures)  Attrition Mortality! o Sequential Method  First phase = cross-sectional method  Second phase = longitudinal method (participants tested AT LEAST one more time after) Choosing a Research Method - Every research method has positives and negatives: The choice is affected by: • Resources like time and money • Ethical concerns • E.g. studying child abuse  can’t have a control and experimental group… • The research question • Under idea condition, your choice should be based on research question How does the question guide the choice? : Description: asking what or how many, getting ideas about why, developing theory • Case stud
More Less

Related notes for PSYB01H3

Log In


Don't have an account?

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.