PSY 2970 - Final exam
Understanding correlation
-1, +1, perfect correlation, straight line (slope) Close to zero, weaker relationship Zero, no reliable relationship
2 basic categories of responsibility
Responsibility to the individuals (human and non-human) who participate in research studies Responsibility to the discipline of science and to the public
Seven criteria reviewed by the IRB
1. Minimize risks to participants 2. Determine reasonable risk in relation to benefits of research 3. Equitable participation selection 4. Informed consent 5. Documentation of informed consent 6. Data monitoring 7. Privacy and confidentiality
Timeline of key developments in research ethics
1947 - The Nuremberg Code Produces a medical ethics document First real written down ethics guide to research 1953 - APA Ethics Code Research and Psychology guide 1963 - Milgram obedience studies Public outcry 1971 - Zimbardo Stanford Prison Experiment Public outcry 1932 - 1972 - Tuskegee Syphilis Experiment Public outcry 1974 - The National Research Act defines IRBs Oversight 1979 - The Belmont Report
Standard deviation
A computed measure of how much scores vary around the mean score
Operational definition
A definition in terms of how one measures the things (e.g. constructs) they want to study A description of an operation/process by which we're going to measure or assess a variable
Histogram
A graph of vertical bars representing the frequency distribution of a set of data.
Box plot
A graph that displays the highest and lowest quarters of data as whiskers, the middle two quarters of the data as a box, and the median
Stem and leaf plot
A method of graphing a collection of numbers by placing the "stem" digits (or initial digits) in one column and the "leaf" digits (or remaining digits) out to the right.
Secondary source
A second-hand report of someone else's observations. Literature reviews & Meta analyses Textbooks The introduction of many research reports Most television, newspaper and magazine reports
Accessible population
A subset of the target population, consisting of those individuals who are accessible to be recruited as participants in the study
Test-retest reliability
Administer a test, then retest at a later point Should remain consistent, unless external factors have great influence
Cognitive biases
Affect the validity of our interpretations of our own experiences
Ratio
All of the properties of interval, plus the existence of a meaningful zero point Zero represents the total absence of a variable being measured
Population
All the individuals of interest
Extraneous variables
All variables that are not the focus of study Eg. other media content, child's personality
Factors that affect power
Alpha (α) - alpha goes up, beta goes down, power goes up - Can't really raise alpha though, not accepted in the research community Effect size - Larger effect size = high power - Small effect size, easy to miss = low power - Uncontrollable Sample size - More participants = more power
Empirical question
Answerable with data Terms can be precisely defined
Factorial design
Any experimental design with more than one independent variable
Behavioral
Any observable behavior for which it can be effectively argued to be a representation of the construct or variable of interest
Construct validity
Are you measuring the full extent of the underlying construct? How adequate/appropriate are your operational definitions? Many types of validity can serve as indicators of good construct validity
Inter-rater reliability (observation-based scoring)
Assume that researchers are attempting to measure affection in children Decide to observe each child and count the number of displays of affection each child provides during a 50 min group activity Researchers are scoring separately, no inconsistency from participants, any inconsistency in scoring would come from researchers
Concurrent validity
Attempting to validate a new measure of intelligence Utilizing an already validated, well-established test to compare to the new test with the same construct and measure the correlation Relies on there being an already validated, well-established test Administering both tests to participants to examine consistency and correlation
Convergent validity
Attempting to validate a new measure of something without using/having an existing, well-established and validated measure of it (if there was one, we could use concurrent validity) Using 2 new tests that are measuring the same construct - consistent results that measure the same construct indicate that both tests are relatively valid and reliable Still could both be measuring different things with consistent errors, however, it is likely measuring what is supposed to measure
Divergent validity
Attempting to validate a new measure of something, in this case by showing that it is not actually a measure of a related construct Helpful to show that the new test that measures anxiety is not accidentally measuring fear Somewhat a process of elimination
Methods of acquiring knowledge/ways of knowing
Authority: faith in the experts Intuition: following your gut Tenacity: "Everybody knows that" Rational method: logical reasoning Empiricism: direct observation or experience
Central tendency
Average/mean Mode Median
Basic vs. Applied research
Basic = research intended to describe, predict, explain Applied research = applying a specific, experimental solution to a specific problem
Physiological
Brain activity - FMRI, PET, Neurotransmitters, peptides, hormones Galvanic Skin Response - Fear - Can be hard to pinpoint just one cause of the response - An inference is required - researchers can make an incorrect inference - Rather objective, people can't really control innate, physiological responses - Expensive - Necessary equipment can be restrictive
Primary source
First-hand report, usually a research report, in which the authors describe their experiment and observations Usually published in scientific journals
Ordinal
Categorical groupings that are organized sequentially Ranking, listing in order of preferred options
Nominal
Categorical groupings that represent qualitative differences in the variable measured No meaningful order; just because an option is assigned a number doesn't mean that it has any value
Converging techniques
Check to see if you can get your answer with more than one technique
If the null is true, and the researchers accept it...
Confidence (didn't find anything because there was nothing to be found)
The Nuremberg Code
Consists of 10 points 1. Voluntary informed consent of human subjects - What does "voluntary" mean? - What does "informed" mean? - Consent vs. Assent 6. Assessment of Risk vs. Benefit 9. Right of voluntary termination of the experiment 10. Responsibility of the researcher to terminate an experiment
Constructs
Constructs are hypothetical attributes or mechanisms that help explain and predict behavior in a theory
Hypothesis
Describes or explains a relationship between variables
5 categories of research strategies
Descriptive Correlational Nonexperimental Quasi-experimental Experimental
Validity
Did we measure what we intended to measure?
Replication study
Duplicates some or all of the procedures of some prior study Exact or direct replication Conceptual replication Partial replication
Quasi-experimental strategy
Goal: to obtain evidence in support of a cause-and-effect relationship, although any causal attribution is ambiguous Uses a non-manipulated variable to define groups or conditions If the researcher cannot manipulate the variable, we cannot be sure that that IV is the cause of the DV Ex. gender identity in assigning participants to groups; relying on self-report
Internal consistency reliability
Groups of questions surrounding a similar idea to test for consistency Instead of running an average test for all questions together, group them by underlying theme The average for the first group of questions should be somewhat reliable in predicting the results for the second group of questions - if the average scores are significantly different, we can assume there is error and this test is tapping into different parts of a construct instead of the overarching idea it was intended to tap in to
Directional hypothesis set example
H1: Infants will, on average, look longer at the impossible than possible event H0: Infants will look for equal amounts of time at impossible and possible events, or will look longer at possible
Nondirectional hypothesis set example
Ho example: Infants will look, on average, for equal amounts of time at impossible and possible events HA example: Infants will look, on average, for different amounts of time at the possible and impossible events
Kurtosis
How things are dispersed around the mean/median Leptokurtic - some range, a spiked look, lots of people scoring a couple scores, and then drop-offs on either side of the spike Platykurtic - very wide range, mostly equal frequency, lots of people are scoring very different scores
Statistical significance
In a research study, a result or treatment effect that is large enough to be extremely unlikely to have occurred simply by chance
APA Ethics Code
Includes the core ideas of the Nuremberg Code and additionally outlines ethical practices for: Confidentiality and anonymity Deception - Active - Passive Debriefing - Dehoaxing - Desensitizing
Laboratory vs. Field research
Lab environments give researchers the most control, but it can be artificial, participants reactions may not be natural Field research may produce more natural reactions, but researchers do not have as much control
Looking time paradigm
Looking time experiments are based on the assumption that animals direct eye gaze toward objects or scenes based on their degree of interest, and use looking behavior to infer perceptual or cognitive characteristics of subjects
Components of experimental research design
Manipulate one variable to create 2 different treatment conditions Measure a second variable to obtain a set of scores in each treatment condition Compare the scores in treatment A with the scores in treatment B
Controlling for confounding variables
Match groups - Make sure that, on average, those in one condition are matched with those in the other condition. I.e., screen for attention to ensure that the need for attention of the control group is the same as the experimental group Random assignment - A way to ensure that each participant has an equal chance of being assigned to one condition or another Restrict range/hold constant - Limit the sample to those who don't score as an extreme on the "need for attention" assessment - Problem is that our study is no longer as generalizable as it only applies to children with a mild need for attention
Measurement error
Measured score = "true" score + error Ex. IQ Score = Intelligence + error [psychological disorders, social pressure, emotional state, presence of drugs, environmental factors, sleep/alertness] Sources of error: participant, observer, environment
Self-report
Most common type: Likert-type - Rate your agreement with the following statement on a 1 to 7 scale with 1 being strongly disagree and 7 being strongly agree Constructing a good rating scale - Need both positive and negative versions of questions for same constructs - Avoid too much leading - Questions to detect dishonesty and apathy
Mundane vs. Experimental realism
Mundane = physical - does the lab look like complete normally Experimental = psychological - have we created the psychological experience necessary to get the participant into the necessary mental space
Criteria of a good hypothesis
Must be logical Must be testable Must be refutable/falsifiable Must be "positive"
Error variance
Nonsystematic variability due to: Individual differences Random, unpredictable effects within a study Conclusions in psychology are always based upon a probability Inferential statistics will allow us to investigate the likelihood that conclusions are correct
Skew
Positive skew/right skew - curve is on the left, most of the data is represented on the right Negative skew/left skew - curve is on the right, most of the data is represented on the left Bi-modal - clustering around 2 different points, even though technically, there is still only one value turning up most frequently
If the null is false and the researchers reject it...
Power (hit, researchers successfully demonstrated or showed findings that did/do exist)
The Belmont Report
Respect for persons: Acknowledging the autonomy of all persons, and the requirement to protect those with diminished autonomy Beneficence: Maximizing benefits for the research project while minimizing risks to the participants Justice: Who ought to receive the benefits of research and bear its burdens? Ensuring reasonable, non-exploitative, and well-considered procedures are administered for participant selection
Measures of variability/dispersion
Quartiles Min, max Standard deviation
Habituation
Repeated exposure to a condition until the same condition is no longer a novel stimulus
The purpose of literature research
Research doesn't exist in a vacuum To know the current state of knowledge in a particular field To identify gaps in knowledge that your research could potentially fill
Inferential statistics
Research is almost never conducted with a population, but rather conducted with a sample From the sample, we usually wish to make inferences about the population
Rational method and how it's used within the scientific method
Researchers use logical reasoning at various stages in research, which include deriving tentative conclusions or explanations from observations or results, and determining testable predictions which would logically follow from theories or hypotheses
Extension study
Resembles a prior study, focused on the same broad question, but adds at least one new feature
Institutional Review Boards (IRBs)
Responsible for reviewing human-participant research Consists of individuals not affiliated with the research project, both scientists and non-scientists Researchers are usually required to submit a written proposal for review
Sample
Selected from the population Results from the sample are generalized to the population
Shape/nature of distribution
Skew Kurtosis
Variance
Standard deviation squared
Predictive validity
Studying the number of siblings one has to the frequency of displayed aggression Need a good, reliable, valid definition of aggression If there is already literature which has found that aggression is related to disciplinary citations in school, we can use that
Ethical conduct with animal subjects
The Institutional Animal Care and Use Committee (IACUC) oversees animal research as much as the IRB oversees human research
Belief perseverance
The bias humans have to maintain their belief Humans don't like being wrong over and over again, don't like to have beliefs that are seen as unusual or different to other people
Ceiling effect
The clustering of scores at the high end of a measurement scale, allowing little or no possibility of increases in value Occurs when the measurement is too easy; below the skill level of the sample
Floor effect
The clustering of scores at the low end of a measurement scale, allowing little or no possibility of decreases in value; a type of range effect Occurs when the measurement is too difficult; above the skill level of the sample
Empirical method and how it's used within the scientific method
The empirical method involves the use of observation or direct sensory experience to obtain knowledge. Within the scientific method, it has a place in data collection as it is a matter of observing things in a careful, systematic, and structured manner
Target population
The entire set of individuals who have the characteristics required by the researcher
Independent variable
The experimental factor that is manipulated; the variable whose effect is being studied.
Internal validity
The extent to which a research study produces a single, unambiguous explanation for the relationship between two variables.
External validity
The extent to which the results of a study can be generalized to other situations and to other people
Alternative hypothesis
The hypothesis that states there is a difference between two or more sets of data.
Null hypothesis
The hypothesis that there is no significant difference between specified populations, any observed difference being due to sampling or experimental error.
Reliability
The level of consistency or stability of the measurements
Availability heuristic
The likeliness of remembering events that are only significant, meaningful, impactful Less likely to remember mundane occurrences; our thinking is driven by the unique situations
Interval
The sequential order of a series of equal intervals Mathematical relations between points on the scale Equal intervals between points on the scale Arbitrary zero
Sum of squares
The sum of each score's squared deviation from the mean
Confirmation bias
The tendency of human beings to seek out and give more attention to evidence and experiences that support what we belief, and a tendency to ignore, avoid, and stay away from topics that challenge what we believe
Deduction
The use of a general statement as the basis for reaching a conclusion about specific examples.
Induction
The use of a relatively small set of specific observations as the basis for forming a general statement about a larger set of possible observations.
Dependent variable
The variable that is measured for each variable
2 primary rules for hypothesis sets
They must be mutually exclusive - Cannot say the same thing They must be exhaustive - Must account for all possible outcomes - Don't want to conduct research and realize that the results don't support either
If the null is true and the researchers reject it...
Type I error (α) (researcher rejects the null when it was true; thought they found something when they actually didn't)
If the null is false and researchers accept it...
Type II error () (researcher retains the null when it was actually false; there was something to be found that the researcher failed to show or demonstrate, data just doesn't show it to be true)
Confounding variable
Type of extraneous variable Variables that could potentially affect the DV If it systematically varies with the level of the IV Eg. child's need for attention
Violation of expectation paradigm
Violation expectation is a method which was focused on whether infants' understand the principle that a solid object cannot move through the space which is occupied by the other solid object.
Interpreting alpha
When α = 0.05, it means that we are willing to reject the null in error 5.0% of the time; 5% chance of error is acceptable If we ran the experiment 100 times, and even if our treatment had no effect, we would make this error 5 times
Factorial notation
i.e., 2x3 factorial notation = The first variable has 2 levels (i.e., male or female) and the second IV has 3 levels (age, height, weight)
Sample equations/variables
x̅ = ∑ 𝑋/𝑁 𝑆𝑆 = ∑(𝑋 − 𝑋̅)^2 𝑠^2 = 𝑆𝑆/(𝑛 − 1) 𝑠 = √𝑠 2
z-score
z = x (score we want to ask something about) - μ (mean) / σ (standard deviation) (x-μ)/σ If above the mean, (+) z-score If below, (-) z-score If the mean, 0 z-score Use z-table!
Alpha level (a priori)
α = 0.05 significant (typical threshold) α = 0.01 significant (stricter) α = 0.10 significant (trend)
Population equations/variables
𝜇 = ∑ 𝑋 /𝑁 𝑆𝑆 = ∑(𝑋 − 𝜇)^2 𝜎^2 = 𝑆𝑆/𝑁 𝜎 = √𝜎 2