Tutor profile: Daniel P.
Subject: Public Health
What are some possible factors that may bias the findings of a longitudinal study? Assume the study was conducted over several decades and no loss to follow up.
Survival bias- There is a chance that the people who survive to the end of the study share a similar effect which could bias the results. Cohort effect- Individuals born at roughly the same time (usually considered in decades) may have similar life experiences to each other but differ in life experiences to different cohorts. This may lead to results showing possible effects as due to the shared life experience, rather than the effect you are studying To avoid bias or confounding, consider performing a sensitivity analysis to check for possibly biased results.
Imagine this, you are in a large gathering with people talking all talking over each other. Your friend walks in and you notice her at the other end of the room. She strikes up a conversation with someone and you can barely hear her, so you return to your own conversation. From across the room, you hear her say your name. She didn't yell it but yet you still heard it and took notice. What is this effect called? (hint: imagine this happening at a "party")
This is called the "cocktail party effect". It is a psychological phenomenon in which a particular verbal sound attracts your attention to a speaker, despite not being engaged with the person. The speaker doesn't have to yell, as your attention is drawn by the content of the person's words, rather than the tone.
Under a normal distribution, how much of the area under the curve falls between one, two, and three standard deviations (SD) from the mean?
Moving out 1 SD from the mean gives you 68.3% of the area under the curve, 2 SD has 95% of the area under the curve and 3 SD gives you 99% of the area under the curve. These percents relate to the probability that observation will occur.
needs and Daniel will reply soon.