Research Methods pt 2 Flashcards
Chapter 5
Self-report Measure
Reporting a person’s answer to a question you measured
How much caffeine have you consumed? How many caffeinated beverages have you had in the last week?
Ask parents or teachers for child self-report
Risk of biases; exaggerated results; self-fulfilling prophecy
Observational Measure
Behavioral measure
Operationalizing a behavior by recording observable behaviors
Classroom engagement
Conceptual variable
How many times did you see a child raising their hand?
Or observing a child in a school (natural setting)
How often does someone sneeze?
BIAS: people may act different when being observed
Physiological measure
Operationalizing a variable by recording biological data
How much cortisol is released in a saliva?
Categorical Variable
Nominal variable
Levels are qualitatively distinct categories
Assign numbers to variables; have no meaning; just to categories
Quantitative Variables
Coated with meaningful numbers
Height weight scale of someone’s wellbeing
Ordinal Scale
Ranking something in order
Only but so much
ex: socio economic status (“low income”,”middle income”,”high income”), education level (“high school”,”BS”,”MS”,”
Interval scale
Equal distances between numbers; between levels
0 does not have a value, not true 0, TEMP, pH, SAT score, credit score
Ratio Scale
Equal intervals and true zero: miles per hour, unemployment rate, time to repurchase , numbers of errors
Test-retest reliability
Consistent scores everytime you measure something
IQ test beginning of a semester, at end
Score should be relatively consistent
Interrater reliability
Having consistent scores no matter who is taking the measurement
Everybody that observes how often the child is smiling should have the same amount
Internal reliability
Going to provide consistent pattern of responses
Ask how lonely how you feel, asking in different ways,
Should be the same level of loneliness
Chapter4
Why do we need ethics?
Researchers are not always objective
Just because something advances the field doesn’t mean it doesn’t also harm the participants
Checks and balances: Not just one person holding the power and deeming the study as good/bad
Nazi human experimentation (1942-1945): Medical experiments in concentration camps.
Twins: control and experimental
Vaccines
Human limits
Freezing
Mustard gas: Finding effective treatment
Chambers of high altitude simulation
Poisoning
Blood coagulation
Tuskegee Study (1932-1972): Effects of untreated syphilis
Only used Black men and weren’t told that they had syphilis.
Told that they had bad blood and weren’t medically treated for syphilis
Wanted to understand the damage it caused and if it actually needed to be treated
Study continued after penicillin (treatment for syphilis) was created
Justification was that they did nothing to worsen the disease, they just didn’t treat it.
Henrietta Lacks (1951): Cell harvesting without consent
Mother of 5 had vaginal bleeding and had cervical cancer
Used radiation treatment
Took her cells and tested them without her knowing
Her cells doubled every 20-24 hours, unlike the others which would die quickly
Refrigerator Study (1958): Can children escape from refrigerators?
201 children; age 2-5
Locked kids in the fridge to see which release devices worked best
Milgram Obedience Study (1961)
Post APA guidelines
Learner was a confederate. Gave mainly wrong answers and wasn’t actually shocked
Teacher (research participant): shock learner each time they answer wrong
Increase shock voltage after each wrong answer
Reminded them what they signed up for is someone said they wouldn’t shock them
Psychological harm
Zimbardo Prison Study (1973): How strong does a role influence a person’s behavior?
Prisoner or guard
Chose similar people as participants
Guards went power crazy; prisoners became very stressed
No checks and balances
Nuremberg code: Process of informed consent.
Participation in research is voluntary and participants must be given information about the risks involved in the research
Research must contribute to scientific knowledge and be conducted by qualified researchers
Researchers need to avoid unnecessary harm, take precautions against risk, ensure the benefits outweigh the risks of the study, and terminate the study if unforeseen harm comes to the participants
Participants have the right to discontinue their participation in the study for any reason
APA Ethics Code:
1953 APA developed their own code for psychology
Researchers must reduce harm due to deception
Ensure confidentiality of participant data
Leave the responsibility of overseeing research studies to the researchers
Belmont Report
Beneficence:
Treated in an ethical manner not only by protecting their decisions and protecting them from harm to make efforts to secure their well-being.
This is often understood to cover acts of kindness or charity that go beyond strict obligation
Two general rules have been formulated as complementary expressions of beneficent actions in this sense: (1) do not harm and (2) maximize possible benefits and minimize possible harms
Belmont Report Justice:
Injustice occurs when some benefit to which a person is entitled is denied without food reason or when some burden is imposed unduly. Another way of receiving the principle of justice is that equals ought to be treated equally
Belmont Report
Respect for persons:
Respect for persons incorporates at least two ethical convictions: Autonomous agent. Even individuals with diminished autonomy are entitled to protection
The principle of respect for persons thus divides into two separate moral requirements: the requirement to acknowledge autonomy and the requirement to protect those with diminished autonomy
Institutional Review Board (IRB)
Performs risk-benefit analysis to ensure all studies meet community standards of ethical behavior. Checks and balances.
Institutional Review Board (IRB) Composition
University instructors and researchers
University staff with research protocol expertise
Community members (minister, business person, etc.)
IRB Proposal
Non-technical summary of the practical and theoretical research goals
Detailed description of research procedures
Key ethics issues addressed in IRB:
Informed consent
Reduce harm
Confidentiality
Lack of coercion
Lack of deception
Debriefing
Ethics in reporting results
Minor giving consent is called assent
Chapter 6
Construct Validity Of Surveys and Polls
Survey/poll
Choosing question formats
Writing well-worded questions
Open-ended questions
What do you think of this class?
Con: process code and analyze that data, time consuming, may restrict the answers
Forced choice format
Do you like this class so far? Yes or no
Likert Scale
Ex: Asked to reflect their degree of agreement
Anchored by the terms strongly agree, agree, neither agree neither disagree, disagree, strongly disagree
If there are more or less answers than stated above, it is a likert-type scale
How strongly they agree and disagree; not using likert scale
Semantic Differential Format
Rate a target object using a numeric scale.
Anchored by adjectives
RATE MY PROFESSORS 1-5, hardest class I have ever done, taken this class before
How difficult is this class?
Easy 1 2 3 4 5 Hard
Double-barreled question
Do you enjoy swimming and wearing sunscreen?
How well that variable was captured or measured
Clear and straightforward
How questions are asked
Negatively Worded Question
More cognitively difficult a question is to answer the more difficult it will be
People who do not drive with a suspended license should never be punished
Disagree 1 2 3 4 5 Agree
Not that clear, can reduce the construct validity by adding a cognitive load, get in the way of someone’s opinion
QUESTION order
Earlier items can influence the response of later items
Primed by first item
Reduce this problem by creating different versions of the survey with questions in different orders to reduce order effect
Response sets
Weaken construct validity, aren;t saying what they think
Answering a number of questions in the same way
Acquiescence (yea-saying)
Type of response set by just saying yes
Just answering positively to all of the questions ]
Fence Sitting
Not answering one extreme or the other, choosing neutral for every answer. Not getting everyone’s true opinion
Use force choice or remove neutral response questions to avoid it
Reversing the wording of item to see who is
ROSENBERG SELF ESTEEM
Trying to look Good
Social desirable responding/faking good
Do something different than what we are
Fake bad as well \trying to look bad
Make it anonymous, making it more comfortable for true response
Self report
Asking people more than they even know issue
Self reporting memories of events
Might not get data on past events
Where were you when you got acceptance letter?
Rating consumer products
Observational research
Watches an individual and systematically recording their actions
Some people believe that it is better
Frequency claims are based on observational data
Observing how much people talk
Observer Bias
when observers see what they expect to see
WAtch a videotape of a man; half was told he was a criminal half was a graduate
Bias comes in, when you are watching something with those terms
Observer effects
when participants confirm an observer’s expectations
Children in school, teacher expectations,
If one teacher expect high performance, they get high performance
Clever Hans
Masked design,
Clear rating scales or clear code books
Observers won’t know what positions your participants are assigned to or what the study is about
Reactivity
when participants react to being watched: changing how they are act because someone is watching
Solutions for Reactivity
Solution 1: Blend in
Unobtrusive Observations
Try to make yourself less obvious by sitting in the back of the classroom or being in a double mirror room.
Solution 2: Wait it out
Goal is to get your participant used to your presence
Solution 3: measure the behavior’s results
Exam Review
Open-ended
too much data too much time to sort through
Physiological
changes in brain waves, heart rate, respiration
Two research assistants watching some behaviors
their scores are not very correlated, internal reliability, need to be retrained, or that the code book is clear
Categorical way to operationalize some variable
People are anxious
Not really anxious
Very anxious
CATEGORIES
Double-barrel question
two questions in one
Issue: hard to measure, some people might interpret in different ways, have people answering 1 question, but you don’t know what question the person is answering
Ordinal Scale 2
Ordinal scale is using numbers to rank individuals 1 2 3 4
Test-retest reliability 2
measure things at different time points, we should get similar test results
Ratio scale and interval scale
ratio-has a true zero
Interval: no real zero
Good internal reliability
have a measure of happiness ask it in different ways; all ways are highly correlated with each other for each person
TUSKEGEEE
Belmont report 2
Belmont report: beneficence, justice, respect for persons REMEMBER WHAT THESE ARE AND WHAT WE MEANS
How to be coercion in results
give an incentive, money, so hard for them to say no to
ZImbardo
researchers were given authority about theri own studies, he didn’t have to go through IRB oversight over the study
Purpose of the Institutional Review Board
to make sure studies met standard met ethical behavior and protect human subjects
Studies about UNC: like or don’t like UNC, participants only like unc, nothing bad, what might this suggest?
Yay- saying, individuals agree with everything, short way to get through studies
Construct validity
How was depression measured-want to make sure everyone is very clear when you ask them how they experience or define depression
External Validity
Generalizability
If there are particular study and you focus on one demographic of college students in a specific region that you need to consider
Are we able to apply the conclusions of the study outside of the context of that study?
Context of which a study can be generalized to the larger population
Statistical Validity
Statistical
The extent to whether or not the conclusion of the study can be reliable and accurate
Talking about the stats
We may been looking at means and other statistical variables which we have thresholds for-those thresholds tell us how statistically valid those thresholds are. Do we see high mean
Internal Validity
Try to understand if we have considered all types of variable that may contribute to the results we get. Even with all of that we have to consider if the variable we manipulated caused the outcome we got. Are the results we obtained solely affected by the changes that we manipulated as researchers. (not from the some confounding variable