RMC, W2 Flashcards
What is a rationale for research?
- A rationale is an explanation or justification for the reason of doing something > this is usually presented in the introduction (why is what we are researching important?)
- Explain why your particular study was carried out > could do a summary paragraph at the end before the hypotheses or develop rationale throughout the intro
Types of rationale:
- Methodological problems: Your study might be motivated by methodological problems in previous research > question how were other studies conducted? + how you address what they did
- Gaps in theories/explanations: If you are considering how well different theories explain datasets, you may emphasise where there are gaps in the research and explain how you would address them
- Regardless of the type of rationale, the rationale needs to link past research (lit review) to the research you are doing (your report)
Why is the research important?
• Need to see if your study will extend on previous research (does variable A effect variable B in the same way?) > may replicate with new ppts or stimuli
• Address a previous methodological issue > see if effects are the same
• Literature review may show inconsistent findings so you may want to try understand it better by doing your own research > ties in with if there are theoretical questions to resolve or if you want to try build a better theoretical model
• Existing rival theories may be built on to see which one is giving a better explanation of data > your research will support this
- There might not be much research in this area so your research might be exploratory for the future > other people can use it to review later
Writing a rationale:
- A rationale needs to give a purpose > should state what is to be achieved > the rationale for doing X is to achieve Y
- Have to be clear on what you plan to do and why this is important
Planning research: design
- After writing a lit review, rationale and hypothesis you will need to figure out how to do the experiment + analyse it
- First create your hyp then create the conditions in which the hyp will be tested aka IV and also create a metric/measure which determines the outcome of the condition aka DV
- This will yield the design you choose > design considers type and set up
- Design type could be if you are looking at differences or relationship
- A design looking at differences will require statistical analyses such as T-test or F-test
- A design looking at relationships is usually measured by statistic R
Simple comparisons:
• Simplistic experiments involve only one IV with two conditions which are either between (Different subjects) or within groups (same subjects)
- E.g. if you are creating a drug for headaches you may -use condition 1 as a control or baseline for the exp while condition 2 will have a given dose where you can compare the effects > the big question here is what dosage is right to help headaches
One-way designs: a bit more complex
• Sometimes comparing two conditions within one IV of drug dose is not enough > the given dose may not be as good as a larger dose for instance
• We still would keep one IV which is drug dose but better exp could have 3 conditions > placebo condition, small dose condition and larger dose condition > 3 conditions within IV so we can make more comparisons
- So we can compare between condition 1 (placebo) + condition 2 (small dose), compare condition 1 and condition 3 (larger dose) and compare condition 2 to condition 3 > this is an additional comparison giving us more insight (shows if the larger dose is beneficial at all)
Factorial design: more complex
• Includes more than 1 IV > similar to two experiments being done simultaneously
• Here you may for instance want to see whether a small dose works better than no dose (2 conditions) but equally you may want to see if this is different between males and fem > Males so simultaneously you are testing two different IV’s
- First factor = dose and second factor = gender > 1A = No dose effect on females, 2A= Small dose effect on females, 1B = No dose effect on males, 2B = Small dose effect on males
Statistical interaction:
• Definition: when the difference between two conditions in factor 1 is different than the difference between the same 2 conditions in factor 2
• The example shows that there is no difference between the pain experienced by males when they have the placebo vs small dose, whereas females experience more pain with the small dose
- Thus there is no main effect of dose for males but a main effect of dose for females, so this is a statistical interaction
Adding more conditions
- You can add more conditions in one or both factors so for instance one factor could still be placebo, small dose and large dose while another factor could be the time of day
- E.g. so 1,1 would be no dose in the morning, 2,1 would be small dose in the morning, and 3,1 would be large does in the morning while 1,2 would be no dose in the afternoon (factor 1 = 1 (no dose), 2 (small dose) and 3 (large dose) (factor 2 = 1 (morning), 2 (afternoon) and 3 (evening)
- Because there are 2 factors there can only be 2 main effects (factor 1 main effect and factor 2 main effect) + because there are only 2 factors there can only be 1 interaction (factor 1 x factor 2)
More factorial designs: 2x3
• Factor 1 can have 3 conditions within it while factor 2 has 2 conditions > this leads to 2 main effects because there are still only two factors
• There is also only one interaction factor 1 x factor 2 > imagine each factor as a person and the conversation between them is the interaction, so 2 people = 1 conversation aka 1 interaction, 3 people = 4 different possible conversations > aka 4 different possible interactions.
- More factors = more interactions
3 factors/IV’s:
• 2 factors may not be enough to explain what dosage is right you may need to combine the dose level with sex with the time of the day the drug is consumed
• So 3 factors leads to this cube where factor 1 = 1 dose, factor 2 = 2 gender and factor 3 = 3 time of day > all combinations can be seen below for instance 221 = small dose in females in the morning
• This would give you 3 main effects (1 from each factor) but would give you many more interactions
- Two factor interaction is known as a two-way interaction while a three factor interaction is known at a three-way interaction
Why do we need ethical principles?
• These principles inform research, practice, our behaviour & teaching.
• This research impacts people, policies & organisations. Adherence to ethical principles is essential for…
• The continuation of the discipline. People will not participate in our research or follow our recommendations if they do not trust or respect the discipline.
• The credibility of the subject matter. People will not believe our findings or recommendations if our behaviour has given them a reason not to.
• Our credibility & safety. There can be very serious ramifications for violating ethical principles. This can include being fired, losing funding, being unemployable, being ‘struck-off’ if you are a practitioner or even facing criminal charges.
- Ethical approval processes and committee’s are your safety net – if we say no to your ideas it is to protect you, your participants and others.
Guiding Principles:
Respect
- Valid consent (this is not always straight forward!) > for instance, if you are working with a child, you need consent from the child’s parents but if the child says no then you have to stop to avoid causing the child distress
- Confidentiality (links to GDPR here – data must be stored securely) > if there is anything which could be potentially harmful to the ppt then it should be disclosed to appropriate bodies but generally confidentiality should be kept
- Right to anonymity (this is not always straight forward either!) > need to be able to not track the person back and if your study is unethical, anonymity doesn’t resolve it
- Fair treatment (respect the participants views, identity, culture, choices, behaviours etc) > your opinions stay private
- Due process (avoid unfair, prejudiced or discriminatory practices) > e.g. you cannot say you want only men or women unless that is what the research requires
- You need to put the effort in to find participants. > may take time to find people to recruit > contacting one person twice is ok but more than this can seem a bit like harassment > you cannot force them to partake if they don’t want to
Guiding Principles: Integrity
- Scientific integrity refers to a commitment to high quality research at all stages (from design to completion/dissemination).
- Not adhering to this principle wastes everyone’s time, including your own
- Unreliable findings/recommendations can lead to harmful/negative outcomes > your work if published is used in the real world to make changes + has real world impacts so it is important to ensure it is reliable and of good quality
- The value of your work and your professional reputation/credibility will be seriously undermined if you produce and publish poor quality research.
- No study is ever perfect, but have you done the very best you could with what you had and have you acknowledged the strengths/limitations? NEVER COMPROMISE YOUR SCIENTIFIC INTEGRITY!!! (e.g. making up data or manipulating it could ruin your career)
