Week 11 & 12 Exam Notes PDF

Week 11 & 12 Week 11: ethics Week 12: comparing research strategies Week 11 Ethics Tuskegee Syphilis Study (1932-1972) 600 African-American agricultural workers Promised health care as...

Week 11 & 12 Week 11: ethics Week 12: comparing research strategies Week 11 Ethics Tuskegee Syphilis Study (1932-1972) 600 African-American agricultural workers Promised health care as an incentive No diagnosis, advice, treatment or cure was given Penicillin cure widely available by 1947 128 died  Nuremberg Code 1947 10 guidelines for the ethical treatment of human participants 1. Free and informed consent is essential 2. Research needs to be fruitful — benefit, increase knowledge, supports theories etc 3. Should have a scientific basis (e.g., based on the results of animal experimentation) that justifies the experiment 4. Avoid unnecessary physical and mental suffering and injuries 5. No research if death/disabling injury is likely 6. Risk of participation should not exceed research benefits 7. Adequate facilities must be provided to protect against risks 8. Researchers must be skilled and qualified 9. Freedom to withdraw 10. Researcher’s duty to stop if injury is likely But…. Not legally binding  Bad researchers Dr Chester Southam (1963) geriatric hospital patients were injected with live liver cancer cells telling patients “they were getting human cells grown in test tubes.” Stanley Milgram (1963) Participants believed they were administering shocks but it was a fake person pretending to feel pain Psychological distress Deception/informed consent?  Declaration of Helsinki (1964) ethical principles for medical research involving human subjects “The design and performance of each experimental procedure involving human subjects should be clearly formulated in an experimental protocol which should be transmitted to a specially appointed independent committee for consideration, comment and guidance” Australian Code for the Responsible Conduct of Research providing a basic reference for the development of appropriate policies and procedures Written specifically for universities and other public sector research institutions. Principles and practices to encourage responsible research Management of research data and primary material Supervision of research trainees Publication and dissemination of research findings Authorship Peer review Conflict of interest Collaborative research across institutions  National Statement on Ethical Conduct in Human Research standards for design, review and conduct of human research. promote ethical human research, which: 1. participants be shown respect and protection 2. fostering research that benefits the community Research involving any human participants requires ethical approval Taking part in surveys, interviews or focus groups Undergoing psychological, physiological, medical testing/treatment Being observed by researchers access to personal documents Collection and use of body organs, tissues or fluids, exhaled breath Access to their information  Use of Animals use of live vertebrate animals must apply for ethics approval The Australian Code of Practice for the Care and Use of Animals for Scientific Purposes Principles Replacement: Reduction: Refinement: Where possible, Reducing the Minimising the use alternative methods number of animals used pain, suffering & distress (cell cultures, computational models)  Conflict of interest Any situation in which financial or personal considerations have potential to compromise scientific or professional conduct Examples: Researcher may financially benefit from specific research outcome Research psychiatrists paid more than they disclosed Peer-reviewer wants to get research funded/published that is very similar to the one he/she works on Peer-reviewer has personal relation to author or applicant Examiner has personal relation to examinee  Breaches “breach” is used for deviations from the Australian Code for the Responsible Conduct of Research A complaint that a researcher has not acted responsibly requires a response that may include the following steps: a discreet investigation a formal inquiry actions to remedy the situation A complaint or allegation relates to research misconduct if it involves ALL of the following: an alleged breach of the Australian Code deliberation, recklessness and persistent negligence serious consequences, such as false information on the public record, or adverse effects on participants, animals or environment It includes fabrication, falsification, plagiarism or deception failure to declare/manage a serious conflict of interest.  Diederik Stapel Dutch social psychologist Suspended for fabricating and manipulating data in 2011 misconduct affected at least 57 publications (now retracted)  Week 12 Research Strategies Descriptive research Describing how things are, rather than explaining why they are like that Observational & Survey Research RQ: What is the typical number of hours spent studying each week? Survey the participants then perform descriptive statistics high ecological validity Correlational research Goal: describe the strength and direction of relationships between 2 (or more) variables researcher simply measures the variables Analyse whether they show any consistent pattern of relationship use of correlational results to make predictions NO manipulation positive relationship change in the same direction —> i.e. both increase or both decrease negative relationship change in opposite directions —> 1 increases while other decreases Linear/non-linear relationships Monotonic relationships  Pearson’s correlation p = probability of results happening by chance if there is really no relationship r = correlation coefficient, details about strength and direction of the relationship r2 – the “coefficient of determination” the proportion of variation in 1 variable that is accounted for by another variable Does r = 0.219 represent a strong relationship? p = 0.023* significant r2 = 0.048, i.e., model accounts for 4.8% of the variance General rule for reporting correlations: Small — 0.1 – 0.3 Medium – 0.3 – 0.5 Large — 0.5 – 1 Directionality problem Third-variable problem Terms to use: Does A cause B Relationship may be Association or does B cause A? caused by C Relation Correlation Prediction  Experimental research strategies Identifying cause-and-effect relationships Must have 4 elements to be considered “experimental” Manipulation (IV) — 1 variable creates 2 or more treatment condition Measurement (DV) – The dependent variable is measured for participants to obtain a set of scores in each treatment condition Control – All other variables are controlled so they don’t influence the IV or DV Comparison – scores are compared from each treatment condition  Definitions Between subjects Different participants in each condition Vulnerable to individual differences threats to validity Not vulnerable to time-related threats to validity Within subjects Same participants in each condition Not vulnerable to individual differences threats to validity Vulnerable to time-related threats to validity Extraneous variables: all variables in a study other than IV & DV Only becomes a confounding variable if it influences the DV Confounding variable: changes with the IV & has the potential to influence the DV  Controlling for extraneous variables Holding a variable constant Matching values Restricting the range Balancing levels of the variable across treatment conditions eliminates potential to become a confounding variable. alternating order of two treatments (matching through time – i.e. counterbalancing) testing only females – gender held constant age is balanced across treatments Testing only 30 year olds – Age held constant (average age is the same for all conditions) Randomisation Use of a random process to avoid a systematic relationship between variables E.g. random assignment of participants using a coin toss ( likely that individual participant variables (age, gender, height) are distributed randomly)  Quasi-experimental strategies Typically compares non-equivalent groups/conditions Pre-existing participant variables e.g. Sex; smoker vs non-smoker IV the variable that differentiates the groups or conditions not manipulated Disadvantages lack random assignment difficult to control for confounding variables. can lead to selection bias where differences between groups may influence the outcomes instead of the intervention itself. Without the ability to fully control extraneous factors, it becomes harder to establish causality, as other variables could be responsible for the effects observed. Negatively impacts internal validity Unsure whether the intervention directly caused the changes in the dependent variable.