Exam 2 materials

Question 1

Descriptive statistics

Answer 1

• Describe the data
  
  • Measures of central tendency
  • Dispersion
• Don’t want to list every data point when describing findings
• Organize and summarize data so your audience can understand
• Can’t measure the population, so we estimate parameters (numerical characteristics of a population) by using sample statistics (using a sample)

Question 2

Measures of central tendency

Answer 2

• Mean, median, and mode (focus on mean)
• Will include normal distribution

Question 3

Dispersion

Answer 3

• Range, standard deviation, variance
  
  • Variance and SD reflect consistency of scores

Question 4

Inferential statistics

Answer 4

• Make inferences about true difference in population based on sample data
  
  • Based on probability
  • Based on sampling distibution
• Want to draw conclusions beyond just description
• Create meaning for data
• Is there a difference between my groups?
• Did manipulation do anything (have an effect?)
• Are our means different?

Question 5

Sampling distribution

Answer 5

• Theoretical probability distribution of possible values of some sample statistic which would occur if we were to draw all possible samples of a fixed size from a given population and only chance was operating

Question 6

How to test for differing means

Answer 6

• Use hypothesis testing - set up null and alternative hypotheses
• Determine likelihood of obtaining our means because of measurement error
• Because we don’t know the population mean or population standard deviation, we must calculate an inferential statistic
• Then compare that statistic to a sampling distribution of that statistic
  
  • How likely is it that we’ll get a statistic that big by chance?
  • Can use T-tests or ANOVAs

Question 7

Hypothesis testing

Answer 7

• H_o = Null hypothesis
  
  • No difference between groups
• 2. H₁ = Alternative/research hypothesis
     * Difference between experimental and control groups
• Create decision rule to determine
  
  • E.g., reject H_o if means are different/ manipulation had an effect
  • E.g., reject if p < .05
• Reject H_o if there is only a very small probability (p value) that we could get particular sample statistic when H_o is true

Question 8

How do we know probability?

Answer 8

• We get T-statistic, compare it to sampling distibution, and determine how likely it is to get a T of this size just by chance
• SPSS calculates T-statistic for us

Question 9

How do we know probability if more than one group?

Answer 9

• Calculate F-statistic (using ANOVA)
• Compare to F-distribution
  
  • Takes into account sample size
• F = between group variance (variance attributed to our manipulation) / within group variance (random noise)
• If small probability of getting an F this large, we reject the null, and conclude that the manipulation did something

Question 10

Type 1 error

Answer 10

• Deciding groups are different when they are really the same
• Rejecting the null when it is true

Question 11

Type 2 error

Answer 11

• Deciding groups are the same when they are really different
• Failing to reject the null when it is false

Question 12

What influences power?

Answer 12

• Sample size
• Variability
• Effect size
• Design issues
  
  • Sensitivity of measures
  • Strength of manipulation

Question 13

Factorial design

Answer 13

• Design including 2 or more factors, each with discrete possible values or “levels,” and whose experimental units take on all possible combinations of these levels across all such factors
• Independent variable=factor
• E.g., AMP study

Question 14

Advantages of factorial design

Answer 14

• Rather than letting other variables be extraneous or controlled–manipulate them
• Allows us to measure how 2 variables function in relation to each other
• Interaction - when relationship between one IV and DV depends on level of second IV

Question 15

Disadvantages of factorial design

Answer 15

• More time, more money
• More difficult to interpret - may want to examine variables in separate experiments

Question 16

Between-subjects factorials

Answer 16

• Different participants in each of your conditions
• Need a lot more participants
• Still need to use random assignment
• Issues of random assignment and subject #

Question 17

Within-subjects factorials

Answer 17

• All participants participate in each trial
• Now people are participating in even more trials
• Issues of order and sequence effects

Question 18

Mixed factorial design

Answer 18

• Using both a between and within subjects design
• One group of participants is only seeing images in the pleasant, neutral, and unpleasant condition
• Th other group is only seeing words in the pleasant, unpleasant, and neutral condition

Question 19

Main effect

Answer 19

• Effect of IV on DV averaging across levels of any other IV
• Is there an effect of IV1 ignoring IV2?
• Is there an effect of IV2 ignoring IV1?
• E.g., what do pleasantness ratings look like based on just prime valence alone?

Question 20

Interactions

Answer 20

• Response to one factor depends on level of other factor
• E.g., do pleasantness ratings across prime valence depend on the prime format?

Question 21

AxBxCxD complex factorial design

Answer 21

• Must look for:

1. Main effects of each IV by itself
2. 2-way interactions: all possible combinations of interactions between 2 different variables
3. 3-way interactions: all possible combinations of interactions between 3 different variables
4. 4-way interactions: interaction of all 4 variables

• Very hard to interpret

Question 22

What do we measure with surveys?

Answer 22

• Attitudes and beliefs
• Facts and demographics
• Behaviors
• Will useful info be gained from each question?

Question 23

What are some different ways of administering surveys?

Answer 23

• Questionnaires
  
  • Face-to-face
  • Mail
  • Internet
    
    • problem of multi-tasking
  • Other technologies (e.g., PDAs)
• Interviews
  
  • Face-to-face
  • Phone
  • Focus group
    
    • Nice b/c you can get responses building off others in the group

Question 24

How to sample population?

Answer 24

• Probability sampling
  
  • Simple random
  • Stratified
  • Cluster
• Nonprobability sampling (high likelihood that those people in your sample are somehow different from those you have not sampled)
  
  • Haphazard/convenience
  • Purposive
  • Quota
  • Snowball

Question 25

Simple random sampling

Answer 25

• Everybody in the population has an equal chance of being chosen to be in the sample
• Very rarely can we do this
• Try to do this with large-scale election votes

Question 26

Stratified sampling

Answer 26

• Create groups within our population and you call these groups strata
• E.g., want to make sure you have people from each ethnic group
• Sample people from each of the groups, then randomly sample from each group to create sample

Question 27

Cluster sampling

Answer 27

• Still create groups (clusters), and rather than sampling from each of those clusters, you randomly sample clusters, and survey everybody in that cluster (which becomes your sample)
• Still using probability, but random sampling now occurs at the level of the group
• Usually sample several clusters (e.g., counties of Missouri or dorms at Wash U)

Question 28

Haphazard/convenience sampling

Answer 28

• People who volunteer to do it - those available to participate in your study
• People nice enough to answer
• Can use for studies like levels of processing (same for everyone)
• Maybe couldn’t use this for learning about sexual behavior

Question 29

Purposive sampling

Answer 29

• Intentionally seeking out certain groups of people to be in your sample
• Maybe you’re interested in studying people that have 2 kids, or with a certain income level

Question 30

Quota sampling (example of purposive sampling)

Answer 30

• Trying to fulfill a quota - may be interested in having a sample that does reflect the racial makeup of the population
• Enroll people until you make a quota

Question 31

Snowball sampling (another example of purposive sampling)

Answer 31

• Tell participants to tell other people who fit characteristics to enroll in the study

Question 32

2 issues to think about when sampling a population

Answer 32

1. Sampling frame:

• the actual population of individuals you’re sampling from
• When you sample, you’re trying to sample from the pop, but often your sampling frame is not equal to the population
• E.g., if cell phone study, younger people are more likely to be over-represented

1. Response rate

• What percentage of people who are asked to complete a survey actually complete it
• If low response rate, it is much more likely that the people who are not completing it are somehow different from those that do complete it

Question 33

How to word questions?

Answer 33

Do:

• Simplify wording and structure
• Use common vocabulary
• Define terms if necessary

Don’t:

• Use negatives
• Use double-barreled questions
• Use biased/loaded questions

Question 34

How to structure responses?

Answer 34

Open-ended:

• Can provide lots of detail
• Freedom for respondent
• Difficult to code/translate data for analysis
• May produce less info - laziness or cuing

Close-ended:

• Easier, faster to answer and code
• Options may not be inclusive
  
  • Could create more response options
  • Could run a pilot/ask for feedback
  • Could give option to provide more info
  • Could use rating scales

Question 35

Graphic rating scale

Answer 35

• With verbal and nonverbal cues (e.g., smiley face and “no pain” underneath)
• Gives people a line, and the way you ask people to respond is a graphic representation of where they fall on a scale

Question 36

Non-verbal scale

Answer 36

• Wong Baker face scale
• Indicate where you sit on the scale based on smiley and frowney faces

Question 37

Likert scale

Answer 37

• Has numerical responses to it
• Anchored at disagree and agree (anchored in terms of agreement: agree on one side, disagree on the other)

Question 38

Semantic differential scale

Answer 38

• Has anchors at the 2 sides which are opposites of one another; Ss asked to respond within the range
• E.g., cold and warm, wet or dry
• “Same” on one side, “different” on the other

Question 39

Issues that influence responding

Answer 39

• Response set
• Question wording
• Question context
• Question order
• Adjacent questions
• Response format
• Response alternatives

Question 40

Response set

Answer 40

• “Faking good” or social desirability
  
  • Misrepresent themselves to appear more positive than they are
• “Faking bad”
  
  • Maybe to make symptoms sound more severe than they are
• “yea-saying” or “nay-saying”
  
  • Can reverse-code to catch these individuals

Question 41

Influence of question wording

Answer 41

• Wording of question can fundamentally change people’s interpretations/responses
• Death penalty example
• School vaccination example

Question 42

Influence of question context

Answer 42

• Norbert Schwarz gave people news story about mass murder, and asked Ss to explain why mass murder occurred
• Explanations varied drastically across surveys depending on whether it was from “institute of personality” (focused more on the murderer’s personality) or “institute for social research” (focused more on environmental aspects)

Question 43

Influence of question order

Answer 43

• Norbert Schwarz
• Asked Ss “how satisfied are you with your….life or marriage?” (depending on which comes first)
• When life came before marriage, r=.32
• When marriage came before life, r=.67
• When asked about life first, you consider many different aspects
• When you get to marriage, then ask about life, you consider marriage more heavily when you make the life decision

Question 44

Influence of adjacent questions

Answer 44

• Colin Powell example:
  
  • Question with additional information yielded different ratings of republican party

Question 45

Influence of response format

Answer 45

• Open vs. closed
• E.g., most important thing to help prepare children for life?
  
  • To think for themselves
• Close-ended: 62% of participants chose “to think for themselves”
• Open-ended: 5% of participants chose “to think for themselves”

Question 46

Influence of response alternatives

Answer 46

• E.g., “how often are you really irritated?
• Low frequency (e.g., never to more than twice a month) vs. high frequency (e.g., twice a month to several times a day)
• 39% report more than twice a month for low freq question; 62% for high frequency question

Question 47

Correlational research

Answer 47

• Used to describe the relationship between two or more naturally occuring variables
  
  • E.g., where do people tend to sit in class/how do they do on an exam
• Useful when manipulating variables is not ethical or practical
• Requires 2 sets of measurements with same individuals or paired values
  
  • Need what row person is sitting in and exam scores

Question 48

30 million word gap by age 3

Answer 48

• Hart and Risley observed 42 children for an hour each month in their homes from 7-9 months through 3 years
• Also examined SES of families
• Found that SES was correlated with language abilities
• More specifically, number of words kids heard was correlated with language abilities
• Found that children’s scores at 3 years were strongly correlated with their 3rd grade performance (vocab growth at 3 correlates with language abilities in 3rd grade)

Question 49

2 issues with correlational studies

Answer 49

1. Directionality
2. 3rd variable

Question 50

Pearson correlation coefficient (r)

Answer 50

• Quantitative expression of strength of relationship between 2 variables
• Tells you direction of relationship (positive or negative)
• Values range from +1.00 to -1.00

Question 51

Effect size

Answer 51

• Strength of association between variables
• With correlations, pearson r is measure of effect size
• Small effect size: r = .1 to .2
• Medium: r ~ .3
• Large: r ~ .4
  
  • This rule of thumb differs by area of research
• Coefficient of determination: r²

Question 52

5 issues with correlations

Answer 52

1. Outliers
2. Sample size
3. Restriction of range
4. Nonlinear relationships
5. Reliability of measure

Question 53

Issue 1 - outliers

Answer 53

• Can spuriously affect your correlation coefficient
• Falls on or near regression line = drives up r (makes it bigger than it should be)
• Falls far from regression line = drives down r (smaller than it should be)
• Regression line is best fit line for all data

Question 54

Issue 2 - sample size

Answer 54

• Smaller n = greater influence of outliers that might be in your data
• Harder to disprove null
• Stronger relationships sometimes not significant if sample size is too small

Question 55

Issue 3 - nonlinear relationships

Answer 55

• A low correlation coefficient does not necessarily mean that no relationship exists
  
  • It means that no linear relationship exists
• Important to plot data for this reason

Question 56

Issue 4 - Restriction of range

Answer 56

• Restriction of range will produce low correlation coefficients (artificially)
• Variability is good for finding patterns

Question 57

Issue 5 - reliability of measure

Answer 57

• An unreliable measure can reduce correlations
• If she gives us a measure of self-esteem and correlates with GPA, if the measure is unreliable, correlation coefficient will be lower than if accurate measure is used

Question 58

Partial correlation

Answer 58

• Addresses the “3rd variable problem”
• Statistically control for third variable
  
  • Removing the influence of one variable from other ones
• E.g., correlation between birth weight and developmental delays
• If there’s some other variable that can explain correlation better (e.g., mother’s health) you can partial out influence of mothers health on the other correlation
• This makes other correlation much smaller, which shows that low birth weight effect is really explained better by the mother’s health rather than a direct relationship

Question 59

Multiple correlation (multiple regression)

Answer 59

• Single variable – can create equation to predict behavior
• Correlation between several predictor variables and a single criterion variable
• Leads to better accuracy in prediction
• Multiple regression
• E.g., success in college (factors in high school GPA, SAT, letters of rec…)

Question 60

Cross-lagged panel correlation

Answer 60

• Uses correlation to make inferences about causation
• Correlation between 2 variables is calculated at 2 time points
• Cross-lagged: look at behavior 1 at time point 1, and behavior 2 at time point 2, as well as the opposite (behavior 2 at time point 1, and behavior 1 at time point 2)
• Allowing time to help us say something about cause and effect
  
  • Seeing what effect one variable has on the other variable across time
• E.g., agression and TV:
• Shows that if you look at preference for violent tv in 3rd grade, and its correlation for aggression in 13th grade, it’s pretty small
• More likely that direction influence is that preference for violent tv leads to aggression, rather than watching violent tv leads to preference for aggression

Question 61

Single case studies

Answer 61

• Not a case study (in which you aren’t trying to change a behavior)
• Assess effectiveness of treatment for one/few patients
• Rare condition or targeted treatment
• Can be used when ethical concerns (e.g., clinical treatment or educational intervention)
  
  • Can use +/- reinforcement

Question 62

Typical intervention study – pretest-posttest design (not single case)

Answer 62

• Treatment group
  
  • Pre-test → Treatment → Post-test
• Control group
  
  • Pre-test → No treatment → Post-test
• Random assignment with control/comparison group

Question 63

Single case design (and types)

Answer 63

• AB, ABA (reversal), ABAB, interactive, multiple baselines
• No control/comparison group
• Repeated measures (time series)
• Participant serves as own control (will see how applying intervention changes behavior)
• Design phases:
  
  • Baseline
  • Intervention

Question 64

AB design & potential problems

Answer 64

• A: measure behavior (baseline)
• Intervention/treatment
• B: Measure behavior again
• Potential problems:
  
  • History
  • Maturation

Question 65

Threats to internal validity

Answer 65

• Is effect on behavior due to manipulation?
• History effects:
  
  • Event that takes place during course of study, but influences dependent variable of interest
  • E.g., examine effectiveness of therapy on anxiety, client finds job during therapy
• Maturation effects:
  
  • Participant grows older or more experienced throughout the course of the study
  • E.g., see reduction in tantrums

Question 66

ABA (reversal) design

Answer 66

• Measure baseline, apply treatment & measure behavior, then remove treatment and measure again
• Demonstration of return to baseline of the behavior
• Unlikely that maturation is causing the return to baseline, making it more likely that what’s causing that change is removal of the treatment

Question 67

ABAB design

Answer 67

• Measure baseline, apply treatment & measure behavior and look for corresponding change, reverse treatment and look for corresponding change back to baseline, then introduce treatment once more
• Every time we show that there is a time locked change in behavior we can say it’s due to intervention / removal of intervention

Question 68

Interactive design

Answer 68

• Look at the influence of multiple interventions
• Allows us to look at the interactive effect of 2 interventions at once versus the effect of a single intervention
• A-B-BC-B-BC
  A = baseline
  B = intervention 1
  BC = intervention 1 + 2
  B = return to intervention 1 (remove intervention 2)
  BC = Intervention 1 + 2 again
• Problem: we don’t know what intervention 2 is doing by itself (no C-phase because this would take lots of time)
• E.g., anorexia study

Question 69

Issues with baseline (4)

Answer 69

• Stability of baseline
• Length of baseline and intervention
• Target behavior
• Carryover of intervention (return to baseline)
  
  • Want to use something else for this

Question 70

Issue of stability of baseline

Answer 70

• Can be:
  
  • Stable
  • Variable
  • Accelerating
  • Decelerating
• Must characterize baseline before we begin because it affects explanation of behavior

Question 71

Issue of length of baseline and intervention

Answer 71

• Must measure the behavior at least 3 times

Question 72

Issue of target behavior

Answer 72

• Sometimes so severe that you don’t want to ever withdraw the treatment

Question 73

Multiple baseline design

Answer 73

• No withdrawal of intervention
• No problem demonstrating reversal (because we never reverse)
• Effects are replicated across behaviors or subjects
• Demonstrate change occurs when, and only when, an intervention is directed at that behavior or subject
• Rather than shifting behavior with one person multiple times, we shift behavior with multiple behaviors or multiple subjects
• For each subject, vary the amount of time you extend the baseline behavior (delay until intervention)
  
  • Want to see a time-locked change in the behavior
  • Unlikely that something else (maturation,…) is causing change in behavior

Question 74

Problems with single case designs

Answer 74

• Can’t necessarily generalize to anybody but that single person
• Good if you’re the parent of the child it helps, but no good for anyone else

Question 75

Quasi-experimental designs

Answer 75

• Looking at variables of interest that can’t be manipulated
• In general, comparing 2 groups of individuals

Question 76

What quasi-experimental designs can measure

Answer 76

Observe either:

• Effects of natural treatments (e.g., differences in school curriculums across districts)
• Subject variables (e.g., gender, weight)
• IV examined without random assignment

Question 77

Advantages & disadvantages of quasi-experimental designs:

Answer 77

Advantages:

• May be the only feasible way to study variable of interest

Disadvantages:

• Limited experimental control
• Less internal validity (less able to identify cause and effect)

Question 78

Types of “Pre-experimental” designs

Answer 78

• One group posttest only design (“one-shot case study”)
• One group pretest-posttest only design
• Nonequivalent control group design (posttest only)
• Nonequivalent control group pretest posttest design is a quasi-experimental design

Question 79

One group posttest only design (“one-shot case study”)

Answer 79

• Uses one group, and you only measure behavior after something has happened
• E.g., earthquake happens, then you measure stress level of city

Problem:

• No baseline, no control or comparison group
• Therefore, you can’t make any conclusions based on this design

Question 80

One group pretest-posttest only design

Answer 80

• Measure behavior before and after design
• Still doesn’t address issue of control/comparison group

E.g., administer survey on attitudes toward smoking ban:

• Select participants with “anti” ban attitudes
• Smoking ban questionnaire –> “pro” smoking ban film –> Smoking ban questionnaire
• Conclusion: attitudes shifted toward more “pro” smoking ban…film works

Question 81

One group pretest-posttest only design: threats to internal validity

Answer 81

• Could produce effect on DV rather than IV
• History
• Maturation
• Testing/practice effects (participants learn from the experiment)
  
  • Especially if pretest and posttest are the same
• Statistical regression (to the mean)
  
  • High and low scoring participants will move toward the mean on a second test
  • Smoking ban study: this is bad because you want the effect of the intervention to move the extreme participants (anti) to shift in the “pro” direction…hard to tell if statistical regression

Question 82

Nonequivalent control group design (posttest only)

Answer 82

• Improvement on one group posttest only design because now we add a control group
• No random assignment - non-equivalent control group

E.g., group 1: earthquake –> measure stress

     group 2:   (nothing)   --\> measure stress

• Problem: differences may be due to selection effects

Question 83

Nonequivalent control group pretest-posttest design

Answer 83

• Quasi-experimental design
• Simply introduce a pretest to nonequivalent control group (posttest only)

Group 1:

Stress —- Earthquake —- stress

Group 2:

Stress —- —- stress

Problems:

• Potential differences at first observation, but can at least quantify
• E.g., if you just look at 2 time points, it may appear that your program is working, but you must look at overall trend
• One thing we can use to address this change in baselines is an interrupted-time series design

Question 84

Interrupted-time series design

Answer 84

• Measuring behavior multiple times across time, and at a single time point you interrupt with treatment and see what effect the treatment has on the behavior of interest
• Helps see delayed effect, but must measure past delay
• Problem: no control group, history/maturation effects (don’t know how much the manipulation is changing behavior)
  
  • Can use control series design for this

Question 85

Control series design

Answer 85

• Quasi-experimental design
• Introduce control group to interrupted time series design
• No random assignment - non-equivalent control group

Question 86

Using age as subject variable/ 3 developmental designs

Answer 86

• Can’t randomly assign subjects to different ages, but we are intersted in age-related changes
• Age is not manipulated, no random assignment
• Special developmental research designs:

1. Cross-sectional
2. Longitudinal
3. Cross-sequential

Question 87

Cross-sectional design

Answer 87

Study: does age affect the ability to learn a computer application?

• May conclude that the older you get, the harder it is for you to learn things, but:

Problem with cross-sectional designs:

• Cohort effect - different cohorts have had different experiences growing up which can effect results

Question 88

Longitudinal design

Answer 88

Study: does temperament change with age?

• Same cohort
  
  • Gets rid of any cohort effects because subjects are all growing up at the same time
• However, it does confound age with time
  
  • Maybe there’s some sort of cultural shift as kids grow up that effects results (not just age)
• Expensive, time-consuming, and mortality/attrition
• Testing effects
  
  • If you’re giving people the same test every time, there is a possibility that they’re learning from that test

Question 89

Cross-sequential design

Answer 89

• Measuring different cohorts longitudinally
• E.g., measuring the same group of people at 5-year intervals
• Same problems as longitudinal designs

Question 90

Descriptive research

Answer 90

• Not concerned with relationships
• Concerned with describing individual variables/phenomenon
  
  • E.g., 61% of adults in US drink alcohol
  • For college students, suicide rate is 7.5 a year per 100,000 students
• Often starting point for further research
• Qualitative (just talking about content of behavior or responses) vs. quantitative

Question 91

Qualitative descriptive research

Answer 91

• E.g., from Nathan’s (2005) My Freshman Year:
• A good question, I learned, is one that voices a concern shared by other students or that asks for clarification of upcoming work

Question 92

Quantitative descriptive research

Answer 92

• Almost 1/3 of all discussion topics reported were about boys, meeting boys, and sex
• She’s quantifying the number of responses that fit into a particular category

Question 93

Observational research

Answer 93

• Observe and systematically record natural behavior
• Can be a huge range of behaviors that one may be interested in studying
• Can be used in observational, correlational, and experimental designs
• E.g., birds of paradise

Question 94

4 ways observation may influence behavior

Answer 94

1. Issue of reactivity - must make sure setting is natural
2. Concealment - Must hide (either person or camera)
3. Habituation - Put a person/researcher or camera in an environment for so long that the people we’re observing no longer realize it’s there
4. Participant observation - Researcher goes in and becomes a participant in the group they are observing (like Rebecca Nathan)

Question 95

3 ways to increase reliability with observational research

Answer 95

1. Coding system - clear operational definitions of behavior categories
2. Trained observers
3. Multiple observers - inter-rater reliability

Question 96

4 ways to quantify observations

Answer 96

1. Frequency method
2. Duration method
3. Interval method
4. Content analysis

Question 97

Frequency method

Answer 97

• Count the number of times a person engages in a specific behavior

Question 98

Duration method

Answer 98

• Measure how long a person engages in a particular behavior

Question 99

Interval method

Answer 99

• Divide observation period into intervals, then count the number of intervals in which the child engages in the specific behavior

Question 100

Content analysis

Answer 100

• Measure behaviors/events/words in media, text, or speech
• Can use frequency or duration methods (frequency more common)
• E.g., Number of aggressive acts during saturday morning cartoons
• Can be qualitative or quantitative

Question 101

Advantages/disadvantages of observational research

Answer 101

Advantages:

• Observe actual behavior rather than relying on self-reports
• High ecological validity – can you apply your findings to the real world (b/c it’s primarily conducted in the field)
• Comprehensive description of behavior, holistic

Disadvantages:

• Simply describe behavior, no causal statements
• Data more difficult to deal with
• Can take more time (have to watch behavior for longer to get enough data to do something with)

Question 102

3 types of observational research

Answer 102

1. Naturalistic observation
2. Participant observation
3. Contrived observation

Question 103

Naturalistic observation, advantages, and disadvantages

Answer 103

• Observe behavior in natural setting
• Inconspicuous, unobtrusive observation
• E.g., Jane Goodall

Advantages

• Characterize behavior without changing it
• Observe behavior that can’t practically or ethically be manipulated

Disadvantages

• Time needed can be prohibitive
• Can be difficult not to influence behavior

Question 104

Participant observation, advantages, and disadvantages

Answer 104

• Researcher becomes participant, interacts with subjects that he/she is studying
• Inconspicuous observation not possible
  
  • E.g., cult, gang, college students

Advantages:

• Can get information not accessible to outside observation
• Observer gains unique perspective (because they become part of the group)

Disadvantages

• Time consuming (leave your life to become a part of this group)
• Can be potentially dangerous
• May alter behavior of people you’re studying
• May lose some of your objectivity by becoming part of the group (why rigorous coding system is important)

Question 105

Contrived observation, advantages, and disadvantages

Answer 105

• Set up situation likely to produce behavior
• Can be in lab or “real world”
  
  • E.g., helping behavior after violent movies, child-parent interactions in lab

Advantage

• Do not have to wait for the behavior to occur naturally – saves us time

Disadvantage

• Behavior may be less natural since environment is contrived

Question 106

Case studies

Answer 106

• Description of one individual, not group
• Rare, unusual phenomenon
  
  • Rare clinical disorders
  • Brain injury
  • Extraordinary abilities
• Detailed description of observations during diagnosis/treatment
• May interview client and/or relatives
• Observe client, give surveys/tests, gather archival data
  
  • Medical records…

Question 107

Patient H.M.

Answer 107

History:

• Epilepsy started at age 10
• Late 20s became severe
• Scoville performed bilateral temporal lobectomy (took out both of his temporal lobes)
  
  • Milner was his sidekick

What happened:

• Produced anterograde amnesia
• Remembered childhood, but could not remember events after surgery
• Other cognitive functioning intact

Question 108

Advantages & disadvantages of case studies

Answer 108

Advantages:

• Lots of detail
• Generate new hypotheses
• Can falsify theories
• Extremely powerful/convincing/compelling

Disadvantages

• Simply descriptive, does not explain – need additional tests/studies to understand why/how
• Lack internal validity
• Difficult to generalize to a population because they are only studying this one person

Question 109

Generalizing results beyond college students

Answer 109

WEIRDest people in the world

• Western, Educated, Industrialized, Rich, and Democratic
• Psych studies are almost completely with this population
• Because of this, we should limit our conclusions/indicate limitation of using WEIRD group in our paper

Question 110

Muller-Lyer illusion

Answer 110

• “low-level” processing task that shows cultural effects

Question 111

Generalizing results across age groups

Answer 111

• Circadian rhythms of oler/younger people differ (optimal time of old people is early; later for younger people)
• Therefore, should test both groups during their optimal time periods

Question 112

Generalizing results across genders

Answer 112

• Many studies have shown that females are worse at math/science
  
  • Helpern said it’s a cultural (rather than biological) phenomenon

Question 113

Generalizing results across geographical locations

Answer 113

• Apparently people in the midwest are more extraverted
• Midwestern people may be more willing to sign up for studies than eastern people

Question 114

Generalizing results across cultures

Answer 114

• How do individuals in USA/Russia self-reflect on emotional events (do they ruminate, or let it go)
• More people from Russia reported self-reflecting
• Russians reflect on past experiences from 3rd person (which makes event less emotional-not bad to ruminate in this case); Americans from 1st person perspective (bad to ruminate-more emotional)

Question 115

Generalizing results beyond the lab

Answer 115

• Influence of violence/helping behavior
• Woman dropping crutches outside of nonviolent/violent movie
• Found that difference seen is really being driven by the violent movie

Question 116

2 kinds of replication (and descriptions)

Answer 116

1. Exact replication - use the exact same methods/operational definitions (re-run a study that someone else did)
2. Conceptual replication - use different procedures but same underlying conceptual variables
   
   1. E.g., maybe want to know if there are any long-lasting effects of violent movies

• Rarely see exact replications (often add something new)