Session 2 - How to Test for Differences Between Groups.docx

4 Pages
Unlock Document

University of Toronto St. George
Cell and Systems Biology
William Navarre

HMB325H © Lisa| Page 113 L E C T U R E 2 : H O W T O T E S T F O R D I F F E R E N C E S B E T W E E N G R O U P S ( C H A P T E R 3 ) LEARNING OBJECTIVES 1. describe the difference & the relationship bw a sample & a population 2. explain how each indiv sample provides a single estimate of the quantity being estimated and how (& why) the estimates arising from multiple samples are normally distributed around the ‘true’ (population) value 3. state what the standard error of the mean (SEM) measures & which 2 factors affect its size 4. describe & carry out the processes necessary to complete an analysis of variance (ANOVA) 5. describe the rationale for ANOVA 6. describe the assumptions of ANOVA & confirm that each has been met SAMPLES PROVIDE ESTIMATES 1. ex. an anesthesiology researcher studied how diff anesthetic protocols affect patient heart rates 1. the initial part of the study required the measurement of baseline resting heart rates of 64 consenting volunteers 2. research goal – to determine the avg heart rate prior to surgery 3. practical limitation – can’t measure & collect the heart rate of every patient prior to surgery 4. compromise – a sample (representative subgroup) is selected, w/ its mean being used as an estimate of the ‘true’ (unobserved, population) mean 5. but what would happen if a diff sample had been studied? how do we know if our estimate is close to the ‘true’ (population) mean? 6. many possible samples can be selected & even when rigorously measured, these will vary simply by chance 7. how good are the estimates? 1. if this study was repeated many times, w each selecting & using a diff sample, then we could measure how these diff study means vary from each other 2. the standard deviation of these different study means is called the standard error of the mean (SEM) and it provides an estimate of the potential variability bw the same means 3. interpretation – recognizing that other samples from the same pop will differ from one another, we won’t be surprised if the true pop mean is as high as 80.5 or as low as 78.5 bpm 1. for the distribution of the true population mean, use ±x SEM 8. distribution of the study means should follow a normal distribution w: ~68% w/in ±1 SD, ~95% w/in ±2 SD, & ~99.7% w/in ±3 SD 1. which value is the best estimate of the ‘true’ (population) mean? STANDARD ERROR OF THE MEAN (SEM) 9. we don’t need to conduct multiple studies to estimate the SEM 10. instead, the SEM is estimated by the formula: sample¿¿ √ ¿ SEM= SDof sample ¿ 11. do we have more ‘confidence’ in means arising from large or small studies? why? 12. is it probable that a randomly selected sample would have a mean that is ‘extreme’? 1. ~68% is w/in ±1 SD of the mean  32% chance that data is outside of ±1 SD 2. ~95% w/in ±2 SD  5% chance outside ±2 SD 3. ~99.7% w/in ±3 SD  3/1000 chance outside ±3 SD (possible, but much less probable) HMB325H © Lisa | Page 33 RESULTS BY CHANCE 13. ex. a clinical trial is conducted to compare a novel drug against the standard drug & a placebo in the tx of high bp 14. explanations for results of studies: 1. random variation (chance) accounts for the results 2. flawed study design (bias, confounding) 15. if the drugs have no effect (null hypothesis), then each of these means is simply an independent measure of the same underlying ‘true’ (unobserved, population) value 1. any observed difference is simply due to chance (random and/or biologic variation) 2. if that is true, then the differences among the observed group means should be consistent w the amount of variability w/in the set of data VARIABILITY 16. the variability (variance) of the underlying pop can be estimated in 2 ways: 1. by the ‘average’ variability (variance) of the observed groups: within-groups variance 2. by the amount of variability (difference) among the group means (the SD amongst the diff group means = SEM): between-groups variance 1. if there truly is no effect, then these 2 variances will generally be similar to one another (their ratio should be ~1.0) 1. average of the gro
More Less

Related notes for CSB345H1

Log In


Don't have an account?

Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.