Notes 1/21/16-2/11/16 STA 210
Popular in Intro to Statistical Reasoning
Popular in Department
This 3 page Class Notes was uploaded by Maddi Caudill on Thursday February 11, 2016. The Class Notes belongs to STA 210 at University of Kentucky taught by Dr. William S. Rayens in Spring 2016. Since its upload, it has received 15 views.
Reviews for Notes 1/21/16-2/11/16
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 02/11/16
Mean ---> add all and divide by how many you have Median --> put the numbers in order smallest to largest, then ﬁnd the middle number Standard deviation --> subtract the mean from each observation and square the difference Add all up and divide by n-1 and take square root • S measures spread about the mean. It should only be used when the mean is used to describe the center • S is not food to use (nor is the mean) if the distribution of numbers in the data set is skewed • S is always nonnegative. It can be zero only when all the observations in the data set are the same • If the results have a large enough difference for you to say the results were anything other than chance, they are statistically signiﬁcant. • If the difference is not big enough to say the results are based on anything other than chance, the results are NOT statistically signiﬁcant. Important to Know: • Statistical science can quantify the risk you are taking in saying that an experiment has found a difference in treatments. Sort of. • Statistical science has other roles in experimentation as well - like offering a quantitative way to compare different types of designs We Started More Simply The Human Inference Question • We asked: if controlled studies produce some of the purest data, what can still create problems? • Answer organized in an artiﬁcially simple way on the video by introducing confounding from: • Inadequate or improper comparison • Lack of randomization Example: How far two different designs of paper airplanes ﬂy when thrown Response variable: linear ﬂight distance of airplane Explanatory variable: design type Subjects: different airplane samples Any Obvious Confounding? Lurking variables that could compromise this experiment: • Different "pilots" • Different weight of paper (perhaps) Rats and Cage Preferences (Example) Randomly assigned----> new diet ---> standard diet • Measured weight gain in rats and compare Results: rats on the top shelf tended to gain more weight regardless of which diet (new or standard) they were on • They were unable to make nay conﬁrmations on results of diet due to the confounding in the shelves • The scientists never would of been able to anticipate it *don’t underestimate the power of the placebo Addressing Confounding Double Blindness Neither the person receiving the treatment nor the person evaluating the symptoms knows which treatment has been administered. This eliminates a potential source of blindness. Other Problems Deal With: • Dropouts • Non-adherers • Generalizations • When it is possible to achieve, randomization is critical to experimentation • In some cases, experiments aren't really experiments at all, but are observational studies that compare two groups of data that have already been collected • In other cases, new experimental data are compared to existing data • In still other cases, randomization is simply not possible for ethical or practical reasons • In all of these situations, the potential severity of confounding must be evaluated on a case-by-case basis Quasi Experiments: studies that are unable to use randomization to evaluate effectiveness of interventions • This can make it difﬁcult
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'