PSYC 303 Exam 3 Notes
PSYC 303 Exam 3 Notes Psyc 303
Popular in Research Methods in Psychology
Popular in Psychlogy
This 10 page Study Guide was uploaded by Miranda Bostad on Wednesday March 9, 2016. The Study Guide belongs to Psyc 303 at University of North Dakota taught by Dr. Adam Derenne in Spring 2016. Since its upload, it has received 23 views. For similar materials see Research Methods in Psychology in Psychlogy at University of North Dakota.
Reviews for PSYC 303 Exam 3 Notes
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 03/09/16
Chapter 7 Lets say a psychologist is interested in how nonphysical variables affect attractiveness. What these variables are, and how they work, may vary as a function of age, gender, culture, etc. Rather than ask participants what they value, she has them evaluate a series of individuals with different characteristics. o Bill is a 40-year old… salary (good income or bad)…out-door activities…hasn’t been on a date in almost a year. (Picture) Rate how interested you are in Bill (1-5) o Participants are shown different salaries with same description Experiment 1: o Relatively high income vs. relatively low income. Experiment 2: o Relatively high dating frequency vs. Experiment 3: Hypothesis: Rating with high income > rating with low income. With a single IV: Conduct a one-way ANOVA. o If the difference is statistically significant, what have we learned? .001 (cutoff is .05) We have a statistically significant difference The difference between the groups is not likely due to chance. o What else do we need to know Mean Low Income: 2.2 High Income: 3.4 Yes hypothesis is confirmed. Hypothesis: Rating with high income > rating with average income > rating with low income o What type of analysis should be performed? o If the difference is statistically significant, what have we learned? P value = .001 Can’t tell if this is statistically significant. o What else must we do to determine whether the hypothesis is confirmed? Post-hoc Comparisons Factorial Designs Each experiment could follow a relatively simple, between-groups design… …Or as a factorial design, with two or more IVs Chapter 7 Possibility 1: High (Frequency of Low Dates) High (Income) Low Possibility 2: High Average Low Factorial Designs o Examples 2 X 2 design 2 X 4 design 4 X 4 design st o Number of levels of the 1 nd (first number) o Number of levels of the 2 IV (second number) Factorial Designs o Freedman (1975) How crowding effects behavior Crowding during a boring speech would make people feel even worse; however, feeling crowded during an enjoyable speech would heighten people’s experience. High (Pleasantness) Low High (Density) Very High Very Low Low Somewhat High Somewhat Low Factorial Designs o A two-way ANOVA is performed, which entails three comparisons. o Two-Way ANOVA Main Effect for Variable A o Interaction sets us apart more than anything else Types of Experiments Between-Subjects o Source of error is individual differences. o Solution is random assignment. Within-Subjects o Sources of error are order effects, time of measurement, subject attrition, carryover effects. o Solution is counterbalancing. True Experiments o Researcher manipulates the independent variable. o Confounds are eliminated Quasi-Experiments o Researcher has imperfect control over the I.V. o Confounds are not completely eliminated. o Example: A researcher measures the self-esteem of 200 kids enrolled in a T- ball league both before the season begins and again after it ends. The researcher hypothesizes that playing sports improves self- esteem. Time1: Self-esteem measured Between Time 1 & 2: Kids participate in the T-ball League Time 2: Self0esteem measured again Why is this not a true experiment? No control for confounds (variables)/other things that can effect self-esteem o Ex: Weather (summer months may increase self- esteem), other fun experiences (vacations) True experiments and quasi-experiments exist on a continuum No experiment involves perfect experimental control o Low control High Control o Quasi-Experiments----------------------------------------------- True Experiments Experiments with higher control are preferred over experiments with lesser control. However, “perfect” experiments are not necessary for science to advance. When studies are imperfect, scientists rely on systematic replication and convergent evidence. Quasi-Experiments Why are quasi-experimental designs used? o It may be unethical to conduct a true experiment. o It may be impractical to manipulate the variable of interest. Example: o Do elderly adults living in nursing homes show greater cognitive deficits than elderly adults living with family members? o Researchers cant randomly assign elderly adults to different living conditions o What kinds of “imperfect” comparisons can be made? Possible Measures: Measures of cognition: Verbal ability, working memory, reaction times, critical reasoning, spatial orientation Physiological measures: PET scans, MRI, lung capacity, blood pressure. Possible Comparisons: Compare older adults in different living conditions (1X) Compare trends over times in older adults in different living conditions. Compare older adults before/after moving into a nursing home. Nonequivalent Control-Group Design o Compares treatment group to a control group, but there is no random assignment (e.g., living in nursing home vs. with relatives). Before-and-After Design o Compares participants before and after an event of interest (e.g., entering a nursing home). Nonequivalent Control-Group Example: o An anti-drug program is implemented at GF RRHS. Both before and after the program s administered, students at RRHS and CHS complete anonymous surveys about their use of alcohol and other controlled substances. RRHS 1 Assessment Anti-Drug 2 ndAssessment (5%) Program (1%) st nd CHS 1 2 (5%) (5%) o If drug use falls at RRHS and not CHS, can we be sure that the program was effective? No, possible different atmospheres in schools Maybe there was an overdose at RRHS Maybe supplier at RRHS was arrested Before-and-After o Participants are compared before and after some event. Simple before-and-after (1 observation before/after) Time series design (repeated observations before/after) o A researcher is interested in the effects of exercise on the livers of older adults residing in a nursing home. The research hypothesizes that after the exercise program has been introduced, the older adults Simple Before-and-After 1.) Observation (1X) 2.) Program Introduced 3.) Observation (1X) Time Series Design 1.) Multiple observations 2.) Program Introduced 3.) Multiple observations Takes more time, but permits comparisons in trends. o Interrupted time-series design A times series design in which the event of interest is naturally occurring. Ex: A research is interested in the effects that safety measures have has on the frequency of car accident fatalities. Quasi-Experimental designs o Multiple-group before and after design Observations are made before and after an event of interest. Some participants are exposed to the event (treatment group) and some are not (control group) Experimental Designs True experiments: Researcher has complete control over the independent variable; random assignment is possible. Quasi-experiments: Research selects (rather than manipulates) the levels of the independent variable; random assignment is not possible. Quasi-Experimental Designs Natural treatment: The independent variable is an event outside the researchers’ control. Subject Variable: The independent variable is an individual characteristic. o Gender o Ethnicity o Age o Personality o Socioeconomic status Random assignment is not possible with natural treatments and subject variables; therefore, these independent variables have built-in confounds. So how is progress made? (with quasi-experimental designs) o 1.) Perform systematic replications, seek convergent evidence. If there is a real gender difference in risk aversion, it should appear under widely varying circumstances. o 2.) Minimize the faults in design Match males and females on key factors (e.g., past experiences, current earnings, perceived vulnerability, etc.) Matching Procedures Individual matching: Each participant of one gender is matched with a person of the opposite gender on key factors. o Ex: Male, low income but high expected earnings, minimal experience with insurance Female, low income but high expected earnings, minimal experience with insurance. Aggregate matching: The two genders are equivalent, on average, with respect to key factors. o 3.) Examine the interaction of subject variables with researcher- manipulated independent variables Example (For quasi-experimental) o A researcher is interested in the cognitive declines during old age. Specifically, the researcher wants to know whether the declines result from an inevitable weakening of the brain and nervous system or from disuse of certain cognitive abilities. o Age is a subject variable (younger adults vs. older adults) o Practice can be a researcher-manipulated IV (participants have no practice or practice) Developmental Research Age is the independent variable. Quasi experimental Designs o Cross-sectional design: Persons of different ages are compared at the same time. Is practical. Confounding is due to uncontrolled individual differences (no random assignment). o Longitudinal design: The same individuals are observed at different ages. Not simple, long process usually (Similar to within-subjects) Confounding is due to uncontrolled differences across time (no counterbalancing) o Cross-sequential design: Combines the two approaches. More certainty Research with Single Participants Two basic approaches: o Case study: Descriptive, non-experimental o Experimental single-subject design: Meets criteria for a true experiment Case Studies Intensive study of subject with unusual characteristics Data can come from a variety of sources Examples: o Phineas Gage Rod through head, took out his frontal lobe but he survived but personality changed o “Anna O” Psychodynamic Psychology o Victor The “wild boy” of Aveyron Itard attempted to teach him language and human emotion Victor demonstrated empathy, but never mastered language o H.M. His hippocampus was removed to treat his epilepsy Lost ability to create new declarative memories Single-Subject Designs Distinguishing characteristics: o Each participant is exposed to every level of the IV (like a within- subjects design) o Data are analyzed graphically Dependent Variable on Y-axis (measure of behavior?) Consecutive Sessions on X-axis Nomenclature: o A: First condition; a baseline condition o B: Second condition; a treatment condition o C and up: Optional additional treatment and conditions The duration of the conditions may be tailored to the individual subject. Mort than one participant is used to replicated the results. Experimental Designs Why do traditional research methods us large –n designs o A large sample size reduces the importance of individual differences and the risk of Type I and Type II error. Why would researchers ever want to use small-n (i.e., single-subject) designs? o Concern ultimately is with the individual and the individual may be “lost” in large-n designs. Comparing Methods Point: Inferential statistics determine objectively whether differences are not likely due to chance Counterpoint: Inferential statistics do not show whether differences are important Point: Large numbers of participants allow the researcher to more confidently generalize the findings. Counterpoint: Generalization is best established through systematic replication. (Also, large-n designs mask individual differences). Point: With large-n within-subjects designs, confounding can occur due to subject attrition, carryover effects and the time of measurement problem. These can be eliminated through counterbalancing. Counterpoint: One can use instead stability criteria and replication to establish internal validity. Single-Subject designs: The AB Design: a baseline condition precedes a treatment condition o Simplest possible experimental single subject design o EX: Ward and Carnes examined whether public-posting of self-set goals would improve the playing ability of college football players (linebackers) STRENGTHS: useful when it is unethical to long impose the baseline WEAKNESSES: Imperfect control over the time of measurement Multielement design: treatments repeatedly alternate o EX: Using two different types of reinforcement contingencies to treat 3-year old children suffering from chronic food refusal. The children were given 5 seconds to accept each bite of food before the contingency was imposed. STRENGTHS: Frequent transitions WEAKNESSES: should not be used if carryover effects expected The Reversal (or withdrawal design): o There is more than one baseline condition o EX: Examined inappropriate mealtime behavior of children suffering from eating disorders is the result of positive attention (adult attention; escape from aversive situation) STRENGTHS: repeated baselines WEAKNESSES: should not be used if carryover effects expected Multiple baseline designs: o The transition points between conditions occur at varying times o The procedure is repeated using different participants, behaviors, situations o EX: “stair case” like graphs STRENGTHS: Varying transition times WEAKNESSES: No major weaknesses
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'