Midterm 2 (4, 7, 12, 13) Flashcards
What are the 2 basic categories of ethical responsibility
- Responsibility towards the participants/subjects (ensure their welfare/dignity)
- Responsibility towards the discipline of science to be accurate and honest in the diffusion of results
What are research ethics?
Concerns the responsibility of researchers to be honest/respectful to all individuals who are affected by their research studies or reports
What are ethical principles?
Dictate how studies should be conducted from beginning to end in an ethical way
What are the 3 principles of the Belmont report? Also the 3 politics of the Tri-Council Policy Statement (prevalent in Canada)
- Principle of Respect for persons (autonomy): Individuals should be able to consent to participate in studies. Those who cant give consent should be protected
- Principle of Beneficence: Researchers have to minimize harm/risks, while maximizing possible benefits
- Principle of Justice: Fair and non exploitative procedures for the selection of participants (represent the pop who will benefit from this study)
What is the principle of no harm (APA guideline)?
The researcher has to protect participants from (psychological/physiological) harm, all possible harm has to be identified, minimized and justified. Participants must be made aware of risks and offered post-experiment assistance.
What is clinical equipoise?
Clinicians have to give best possible treatment to their patients. In studies sometimes there are treatments perceived to be better than others (ex: placebo vs actual treatment). The No Harm principle is violated
• Solution: compare only equally preferred treatments (or honest uncertainty about which is best)
What is the principle of informed consent (APA guideline)?
Human participants should be given complete info about the study/their role in it before participating (in broad terms, not to affect the participant’s behaviour)
What is one of the possible problems regarding blind studies and informed consent?
Sometimes we cannot say everything to the participants first hand. The solution is to explain what will be done but not explain why (reveal enough for the participant to be comfortable but not too much to influence their behaviour)
What is assent?
Consent obtained directly from the participants
What is consent?
Obtained from the official guardians of the people who cannot consent to the study
What is passive deception (omission)?
Concealing the true nature of the study
What is active deception (comission)?
Deliberately present false info (confederates)
Ex: not telling participants that their memory will be tested until we test their memory
What is the type of info we cannot conceal from the participants?
Possible harm
What is debriefing? Why is it required
Because the principle of informed consent is being violated with deception, researchers must debrief the participants by telling the participants what was the study about, counteracting or minimizing any negative effects of the study, explaining nature of study, allowing any questions (not always effective)
Can we inform the participants that there will be deception?
Yes, if we use a placebo group
What is the principle of confidentiality (APA guideline)?
All info collected about participants might be deemed personal (attitudes/opinions, measures of performance, demographics), and have to be protected by confidentiality guidelines (= protection of data)
What is anonymity?
Practice of ensuring that an individual’s name/identity is not directly associated with the info/measurements obtained (= protection of identity)
What are the 2 strategies that might be used for confidentiality?
- No names/identification appears on records of data (when participants do not have to be linked with their results)
- Coding system to link participants with data (when necessary) (pseudonyms, code names)
When does research that relies exclusively on publicly available info does not require consent?
- Info is legally accessible to the public and appropriately protected by law
- The info is publicly accessible and there is no expectation of privacy
When is consent not required for the observation of people in public spaces?
- There is no intervention staged by the researchers
- There is no reasonable expectation of privacy (people know that they are in a public space)
- The dissemination of research results would not allow the identification of specific individuals
What are the 2 major problems on collecting data on internet?
• Expectation of privacy
• Persistence and traceability of quotes/info
Internet users do not expect to be subjects; they are most likely to perceive information collection as invading privacy
How can we protect anonym,ity when we took quotes from Internet?
- Do not identify name/provide website address
- Anonymise and paraphrase quotations + use search engines to ensure that they are not traceable
- Use composite instead of direct quotes
What are 3 reasons to choose animal subjects in research?
- To understand animals for their own sake
- To understand humans
- To conduct research that is impossible to conduct using human participants
What are the 2 cases in which animal research is acceptable?
- Only acceptable if it contributes to the understanding of fundamental biological principles or
- To the development of knowledge that can reasonably expected to benefit animals or humans
What are the “3 R’s” when it comes to limiting harm in animal research?
- Replacing / avoiding animals in science
- Reducing the number of animals in science
- Refining care and procedures to minimize pain and distress
What are 3 important guidelines around animal research?
- Must receive proper nutrition / overall care during their lifetime even if they have to be euthanized after the experiment
- Only can be handled by trained professionals
- Research on animals must go through ethics committees also
What is a research error?
honest mistake that occurs in research process (ex while collecting data, entering it into computer, etc)
What is fraud in research?
Explicit effort to falsify/misrepresent data to make it support the hypothesis
What is plagiarism?
to present someone else’s ideas as your own
What are 3 safeguards against fraud?
- Finding cannot be replicated; possible indication of fraud
- Peer review: allows to detect suspicious aspects of study
- Sharing of original data for others to verify (watchdogs)
- Penalties that come with fraud often sufficient to keep researchers honest
What is the goal of descriptive research?
simply describe a phenomenon as it is right now.
○ NOT explain relationships, why, how, underlying causes
What is a behavioural observation (observational research)
• Researcher observes and systematically records the behaviour of individuals to describe the behaviour
What are the 2 problems of behavioural observation?
○ Essential that the behaviours are not disrupted or influenced by the presence of an observer
○ Requires at least some degree of subjective interpretation from the observer (questionable reliability)
How can we make sure that behaviours are not disrupted by the presence of an observer?
- Concealing the observer so that subjects do not know they are being observed (public spaces only)
- Habituate the participants to the observer’s presence
How can we make sure that the observations of behaviours are reliable across different researchers (eliminate subjectivity)?
- Use a list of well-defined categories of behaviour
- Use well-trained observers
- Use multiple observers to assess inner-rater reliability
What are the 3 techniques for quantifying behavioural observations?
• Frequency method: counting the instances of each specific behaviour that occur during a fixed-time observation period
○ Ex: child committed 3 aggressive acts during 30 min period
• Duration method: recording how much time an individual spends engaged in a specific behaviour during a fixed-time observation period.
○ Ex: child spent 18 minutes playing alone during 30 mins observation
• Interval method: dividing the observation period into a series of intervals and recording whether a specific behaviour occurs during each interval
○ Ex: 30 mins observation period divided into 30 x 1 min intervals, child observed in group play in 12 intervals
How can we sample behavioural observations?
- Time sampling: observing for 1 interval, pausing for 1 interval (to note/record observations)
- Event sampling: identifying one specific event/behaviour to be observed and recorded during 1st interval, and changing for a different event during the second interval
- Individual sampling: one different individual at each interval
What is a content analysis?
Using the technique of behavioural observation to measure specific events in literature, movies, television programs, etc
What is an archival research?
Looking at historical records to measure behaviours or events that occurred in the past
What are the 3 types of observations in observational research?
- Naturalistic (nonparticipant)
- Participant
- Contrived (structured)
What is naturalistic (nonparticipant) observation?
• Observing/recording behaviour in natural settings w/out intervening
• Provides insight into real-world behaviour
• High degree of external validity
• Limitation: long time needed to conduct it, potential for observer influence, potential for subjective interpretation
Ex: Jane Goodall’s research on chimpanzees
What is participant observation?
• Researcher interacts w participants and becomes one of them
• Allows observation of behaviours not always open to scientific observation (ex: occult activities)
• High external validity
• Limitations: time consuming, potentially dangerous for observer, may interact w participant’s behaviour
Ex: actors asked to fake mental illness to enter psychiatric hospital and evaluate the conditions of patients there
What is contrieved/structured observation?
• Observation of behaviour in settings arranged specifically to facilitate the occurrence of specific behaviours
• No need to wait for the behaviours to occur
• Limitation: environment is less natural, therefore behaviour might be too
Ex: ask parents and child to come to lab, ask them to perform task and record them
What are the 4 steps of building a survey research?
- Questions must be developed
- Questions must be assembled/organized
- Selection process must be developed
- Determine how the survey will be administered
What are the 3 types of questions in a survey?
- Open-ended
- Restricted
- Rating-Scale
What are open-ended questions? (advantage/disadvantage)
Simply introduce topic and allows participants to respond in their own words
• Advantages: allows great flexibility in answers
• Disadvantages: different participants might have different perspectives, hard to summarize/compute,
What are restricted questions? (advantage/disadvantage)
limited nbr of responses possible (multiple choice question)
Advantages: easy to analyze/compute, compute average is possible
What are rating-scale questions? (advantage/disadvantage)
selecting numerical value on scale (Likert scale)
• Range from 5-10 categories because:
○ Participants tend to avoid extreme categories
○ Hard to discriminate among 10 categories
• Flaw: participants use same category for all questions (called response set)
• Advantage: produce numerical values that can be treated w interval scales, participants find them easy to understand/answer/
What are the anchors in a rating-scale question?
the end points of the scale
Which type of questions should be at the end of a survey?
demographic, because they are boring
Where should sensitive/embarrassing questions be in a survey?
In the middle
What are the 4 possible ways to administer surveys?
• Internet surveys
○ Easy, fast, economical
○ Researcher has great access to participants w certain characteristics
○ Flexibility in presenting questions/responses alternatives (jump ahead)
○ Subject to nonresponse bias (self-selected individuals that do not represent the pop)
• Mail surveys
○ Convenient, nonthreatening
○ Can be time consuming and expensive
○ Response rate of 10-20%
○ Those w more intense feelings about the topic are + likely to answer (bias)
○ Solution: cover letter w survey, include gift
• Telephone surveys
○ Time consuming
○ Can be conducted from home/office
○ Keep questions short
• In-Person surveys/interviews
○ +efficient
○ One on one = key informants (very specific views on the subject)
○ Interviewer bias (body language influencing the expected response)
What is a case study?
In-depth study and detailed description of a single individual (or very small group)
What is the idiographic approach of the case study?
the intensive study of one individual
What is the nomothetic approach of the case study?
the study of whole groups
How is called a case study that observes no treatment?
a case history
What are the strength/weaknesses of the case study
○ Strengths: intense details included, identify specific situations (in which a treatment can work), can be more powerful/convincing than lab studies
○ Weaknesses: simply describes events, not underlying mechanisms, can suffer from bias (selective bias of which cases we present
What is the goal of the correlational research study?
examine and describe the associations and relationships between variables
**NO attempt to explain the relationship and no interference with the variables
What is a positive relationship?
(+X, +Y)
What is a negative relationship?
(+X,-Y)
What are the 2 possible types of forms of a relationship on a scatterplot? (interval/ratio/ordinal)
Linear (straight line) Monotonic relationship (general direction but no line)
What is the tool/calculation used to describe/measure linear relationships?
Pearson correlation (interval/ratio scales)
What is the tool/calculation used to describe/measure monotonic relationships?
Spearman relation (ordinal scores, ranks)
What is the solution possible when 1 of the 2 variables in a correlational study is non-numerical?
Using non-num variables to divide into groups and then compute num v for each group
What is a point biserial correlation?
When either of the 2 variables is non-numerical, and one of them has only 2 categories, we can compute the categories as 1 or 0 and match them with scores of the numercial v.
Then the Pearson correlation can be used (do not check direction because 0 and 1 are distributed arbitrairlly
What is the solution possible when both of the variables in a correlational study are non-numerical?
Organize the variables into a grid and compute the nbr of individuals for each variable, and evaluate data with Chi-Square
(If each v can be coded w 1 or 0, we can compute Pearson correlation, and the result will be called a phi-coefficient)
How can we establish the strength of a relationship?
Coefficient of determination (squaring the numerical value of the correlation)
*Explains how much of the variability of one variable is predictable from its relationship w the other variable
How can we use relationships for prediction?
With a predictor variable and a criterion variable (predictor predicts criterion)
What is called the process of using one variable to predict another?
Regression
What is the third variable problem?
possibility that a third variable controls the 2 variables observed
What is the directionality problem?
Hard to determine which variable might be the cause and which might be the effect
How are called relationships with more than 2 variables?
Multivariate relationships
How can we study relationships with more than 2 variables?
Multiple regression
What are the 2 types of design of the experimental research?
• Between groups (independent groups receiving diff treatment) and Within subjects (all subject receive diff value of the ind v)
What are the 4 basic elements of experimental research?
- Manipulation (of 1 variable)
Allows researchers to determine the direction of the relationship - Measurement (of the 2nd variable)
- Comparison (scores are compared between treatment condition)
- Control (of all other variables)
what is a factorial design (experimental research)
When there is more than 1 ind v
What are treatment conditions (experimental research)
How we manipulated the ind v
What are the levels of the ind v?
The specific conditions used in an experiment
What is a multivariate study? (exp.research)
when there is + than 1 dep v
What are extraneous v? (exp. research)?
all other variables in the study besides ind v and dep v
What are confounding v? (exp research)
special class of extraneous v that can seriously influence dep v (its effects are confounded with the effects of the ind v)
What are 5 ways to control for confounding variables?
- Remove them
A. Ex: remove researcher from music/math experiment - Hold them constant
A. Ex: using the same researcher across the levels of ind v - Use placebo control
- Match them
- Randomize them
What are 3 ways to control for extraneous v?
1-Holding constant
2-Matching
3-Randomize
What are the 2 categories of control conditions?
No treatment and placebo conditions
What are manipulation checks?
Measure to assess how the participants perceive and interpret the manipulation and/or to assess the direct effect of the manipulation
What is mundane realism
superficial (physical) characteristics of the simulation (little effect on external validity)
What is experimental realism
psychological aspects of the simulation (greater effect on validity)
What are the 3 types of generalizations that increase external validity?
1-from sample to general pop
2-from one study to another
3-from a study to the real world
What are research artifacts?
Non-natural feature accidentally introduced into the study
What is experimenter bias?
Inadvertent tendency to influence results in expected direction
Solutions: blind or double blind study
What is participant reactivity?
○ Participants modify natural behaviour in response to being in a study
§ Good subject role: trying to help the experimenter (what they think is the anticipated result)
§ Negativistic subject role: opposite way of what they think the study is about
§ Apprehensive subject role: don’t completely act as they would in the real world bc they’re worried about being observed
Solutions: blind studies, deception
How can confoundgin variables influence internal validity?
1-history: • External events occurring between the 1st and the second tests
The longer time interval btw a pretest and a post-test measurement, the greater the possibility that outside events will influence a particular treatment outcome
2-maturation:Processes within the person producing changes in a subject over time that are not related to the treatment
3-Repeated testing: • Possible effects of the pre-test on the post-test (the same measure taken at 2 different times)
Progress on the test may be due partially to experience (getting better as test)
4-Instrumentation: • Changes in characteristics of a measurement over time
A ruler doesn’t change over time, but a human observer may
5-Regression to the mean: Whenever participants have extreme scores on a pre-test measure, their scores will be more likely to change on a subsequent measure than will scores that are closer to the mean
What is attrition
• Loss of subjects from 1 or + groups of the study
Ex: vacation, no longer interested, too busy, sick, etc
What is differential attrition?
Differences in the nbr of drop-outs or refusals that affects one group more than another
What is compensatory equalization?
Untreated individuals or groups learn of treatment received by others and demand the same treatment or something equivalent
What is compensatory rivalry?
Untreated group learns about the treatment of the other groups and works extra hard to make sure that the superiority of the treatment groups is not demonstrated
What is resentful demoralization?
Individuals in the control group learn others are receiving special treatment and Become less productive, less efficient, less motivated because of feelings of resentment
