Final Exam PSY301 Saponjic

Tuskegee Syphilis Study

by U.S. government branch, lasting approx 50 years (ending in 70s), where a sample of African American men diagnosed with syphilis were deliberately left untreated, without their knowledge, to learn about the lifetime course of the disease.

The 3 key ethical violations in the Tuskegee Syphilis Study

Participants were: not treated respectfully, harmed, and targeted

Milgram Obedience Studies

In 1960s, ppl administered shocks of increasing voltage for wrong answers, Milgram predicted that most people would stop giving shocks once the "learner" started feeling pain, Predictions were wrong, and 65% of the subjects delivered full course painful of shocks

3 main ethical issues in the Milgram Obedience Studies

anxiety, insufficient debriefing, and harmed

The Belmont Report

ethical principles+guidelines for the protection of human subjects of research, 3 main: principle of respect for persons, principle of beneficence, and principle of justice

The principle of respect for persons

part of Belmont Report, ppl should be treated as autonomous and get the right to INFORMED CONSENT

The principle of beneficence

part of Belmont Report, ppls well being should be promoted and they should be protected from harm

The principle of justice

part of Belmont report, "equals should be treated equally", there should be a fair balance between research participants and the ppl who benefit from it, risks+benefits should be distributed equitably

APA's five general principles "Belmont Report Plus two"

beneficence+Non-maleficence, fidelity+responsibility, integrity, justice, and respect for persons rights+dignity

Fidelity and Responsibility

APAs extra, 2 parts: establish relationships of trust, accept responsibility for professional behavior

Integrity

APAs extra, strive to be accurate, truthful, and honest

What led to the National research Act of 1974?

Public outcry over the Tuskegee Syphilis Study

National Research Act of 1974 (3 main takeaways)

created of the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, led to established ethical standards for research like the Belmont Report, mandated IRBs

IRB (Human Subjects Committee)

committee who review all proposed HS research to ensure safety+welfare of subjects are protected

2 Main analyses the IRB HSC have to consider

risk-benefit ratio and at-risk vs at-minimal-risk

6 APA Ethical Guidelines

Informed consent, confidentiality, freedom to withdraw, protection from harm, deception, and debriefing

Informed Consent

ethical principal requiring participants be told enough info to truly choose if they want to participate in study

freedom to withdraw

Participants can leave the study at any time.

protection from harm

research participants shouldn't experience negative physical or psychological effects/harm

Deception

participants are misled about the purpose of the study or the meaning of something done to them

2 types of deception

Active (commission)- ppl are directly told lies

Passive (omission)- intentional withholding of info

Debriefing

the post-experimental explanation of a study, including its purpose+any deceptions to participants

3 main types of research misconduct

data falsification, data fabrication, and plagiarism

data falsification vs data fabrication

fabrication- inventing data/results that never existed at all

falsification- altering/omitting/manipulating actual research

IACUC (Institutional Animal Care and Use Committee)

federally mandated committee that oversees institution's animal research program, facilities, and procedures to ensure the ethical+humane treatment of animals

Legal protection for laboratory animals

Primarily involves the Animal Welfare Act (AWA, doesn't protect 95% of animals) and the IACUC

Animal Care Guidelines and the Three R's

replacement, refinement, reduction

Replacement (Animal care guidelines)

researchers should find alternatives to animals when possible

Refinement (animal research)

researchers must modify experimental procedures+animal care to minimize or eliminate animal distress

Reduction (animal research)

researchers should adopt experimental designs and procedures that require the fewest animal subjects possible

Control Variables

Factors kept constant between groups, to eliminate confounds

Internal Validity

extent to which we can draw cause-and-effect inferences from a study

3 reasons why experiments support causal claims

they establish covariance, temporal precedence, and well-designed ones establish internal validity

What acts as a threat to internal validity?

confounds

design confounds

when another variable accidentally systematically varies along with the IV, if a study has one it has low internal validity and can't support a causal claim

When it comes to confounds what question should we be asking?

is there a third variable that's associated w/ both a and b independently?

"Not every potentially problematic variable is a design confound"

it's only a confound if it systematically varies

systematic variability

variation in data that is caused by a consistent, predictable factor—either the IV (GOOD) or a confounding variable (BAD, threatens internal validity).

unsystematic variability

the unpredictable "noise" or random differences in data that can't be explained by the variables being studied

What does unsystematic variability do/not do?

it adds noise/makes results harder to detect, but it DOESNT threaten internal validity

IV's must have...

2 or more levels

levels of a variable

The different values a variable is given

qualitative vs quantitative IV level differences

qualitative- dif in type or kind, ex: swimming vs running

quantitative- dif in amount, ex: 1h exercise vs 5h exercise

3 main types of IVs (manipulations)

environmental, instructional, and invasive

Environmental Manipulation

modification of the participant's physical or social environment

instructional manipulation

manipulating the variable by giving differing written/oral instructions

invasive manipulation

create physical changes in the participant's body, through surgery or drugs

simple random assignment

every participant has an equal chance of being placed in any condition

matched random assignment (matched pairs)

matching participants results on a variable known to be relevant to study outcome (after pretest), then randomly assigning the matched participants to dif groups, with one matched participant per group

matched random assignment is an attempt by researchers to...

increase similarity among experimental groups

matched random assignment example

study testing effects of caffeine on memory, give participants memory test, rank them based on scores, then group/pair subjects by score

Independent-Groups Design assignment

use simple random or matched random asssignment

Independent-Groups Design

experimental design where dif groups of participants are exposed to dif levels of the IV, such that each participant is only a part of one group and tf experiences only one level of the IV

Independent-groups design other name

between-groups or between-subjects design

Within-subjects design other name

repeated measures design

Within-subjects design assignment

no random assignment needed, bc all subjects are in all groups

Within-subjects design

experimental design where each participants is exposed to each levels of the IV, essentially uses each participant as their own control

within-subjects design advantages

more powerful and requires fewer Ss/participants

within-subjects design disadvantage

order effects

order effects in within-subjects designs

occur when the sequence of treatments influences participant responses, rather than the treatments themselves, threatening internal validity

Pseudo-experiments

-Test a claim abt variable by exposing ppl to it and noting how they feel/think/behave

Pseudo-experiments have..

no control group

History as a confounding variable

Changes (external) that affect essentially everyone in a large group , event occurs during treatment changing behavior

History as a confounding variable EXAMPLE

drug prevention program+celebrity od

Maturation as a confounding variable

Changes (internal) that occur within a spec person or group over time, that might be real reason for results

Maturation as a confounding variable EXAMPLE

cognitive test that lasts 4 hours, performance will prob be worse in the last hour

regression toward the mean

Tendency of ppl who get high/low scores on a particular measure to score closer to the mean on subsequent testing

regression toward the mean (high vs low)

if initial is high next score will be lower, if initial is low next score will be higher

In regression toward the mean, the change is...

not taking place due to treatement

When does regression towards the mean tend to occur?

when participants are selected based on very high or low scores

The extreme scores are just..

their true score+chance events/ME, the true score is usually closer to mean than results say

OS= TS + ME Example

90 = 96 + sick

Hawthorne Effect

change in behavior simply bc their awareness of being studied

Hawthorne Effect Example

Lighting changes on assembly line didn't impact productivity

testing effects

literally practice effects, "prettest influences posttest", score at T2 will be higher than T1 bc it was practice

How do you reduce testing effects?

add control group (no treatment) see how much score improves

Experimental Mortality aka

attrition

Experimental Mortality

loss of subjects over the course of experiment, failure to complete study

2 types of Experimental Mortality

heterogenous (threatens IV) and homogenous (threatens EV)

Heterogenous Attrition

dif dropout rates between groups, threat to IV bc can't confirm cause-effect relationship

Homogenous Attrition

dropout rate similar across groups, threat to EV bc can't confirm generalizability

7 confounding/obscuring variables

pseudo-experiments, history, maturation, Hawthorne effect, Testing effects, Experimental mortality, and Regression towards the mean

Factorial designs

designs with 2+ independent variables, 2 IVs=2 way factorial, 3 IVs=3 way factorial

main effect

In a factorial design, the overall effect of one IV on the DV, each IV has 1 main effect

Main effect example

talking to a plant makes it grow more

IV Interactions

how the operation of oneIV affects another, # of interactions depends on # of IVs

IV Interactions EXAMPLE

music might be helpful only to plants that have not been exposed to talking

Factorial design notation examples

2x3= 2IVs, 1st with 2lvls, and 2nd w/3 lvls

2x3x4= 3IVs, 1st with 2 lvls, 2nd with 3 lvls, and 3rd with 4 lvls

quasi-experiments

lacks random asssignment bc involves pre-existing groups

When are quasi-experiments used?

when IV can't be manipulated/subjects can't be randomly assigned to dif treatments

Quasi-independant variables

not a true iV bc not manipulated, like gender

4 Quasi-experimental designs

Ex post facto experiments, Pre-test-post-test, time-series/simple interrupted time-series

Ex Post Facto Experiments

Experiment is works backwards from effects to determine IVs

Importance difference between experiments and quasi-experiments is...

the amount of control the researcher has over the subjects

who receive the treatments, think random assignment

Pre-test-post-test design

one of most common QED, to assess if an event (iv) increases or decreases the existing level of behavior (dv)

Pre-test-post-test design Example

observation 1, quasi IV, observation 2

control group wouldn't be exposed to the quasi IV in between

Time-series design

type of pre-test post-test design, measure dv multiple times before and multiple times after quasi-iv, look for changes in trend of data before and after, aka simple-interrupted TS

Simple interrupted time-series design

01 02 03 X 04 05 06, several pretest and posttest measures, X is the IV, only 1 group no control

quasi-experimental research advantages (3)

practical, another perspective to research, and potentially higher external validity

quasi-experimental research disadvantages (2)

no cause and effect and low internal validity