Session 7 - How to Test for Trends Part I.docx

6 Pages
Unlock Document

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

HMB325H © Lis| Page 1013 S E S S I O N 7 : H O W T O T E S T F O R T R E N D S ( PA R T I ) LEARNING OBJECTIVES 1. discuss regression & correlation 2. calculate & interpret a regression line & its parameters 3. calculate & interpret a correlation & its parameters 4. complete the statistical procedures relevant to regression & correlation 1. assessing the relationship bw 2 data variables (NOT dichotomous or nominal) that are ordinal (categorical data), but mostly continuous 1. ex. does cardiac output change according to the dose of a drug received 2. allows us to assess whether there is evidence of a ‘dose-response’ relationship DESCRIPTIVE STATISTICS 2. each variable can be described individually (mean, SD, median, percentages) 3. can be described graphically – scatter plot 1. can identify what the best fit of the data is – what line best approximates the observed data 4. most commonly identified best fitting line is a linear relationship (y = a + bx) 1. a: y-intercept 2. b: slope of the line, how the y-values change as the x-values change 5. *assumption: data follow a reasonably linear relationship, ordinal or continuous data 6. ex. is there evidence that ppl’s weight increases according to their height NULL HYPOTHESIS 7. if there truly is no relationship bw the dependent & independent variable, the relationship would look like a flat line 1. distribution of y-values for all x-values are around the same – b = 0 8. H 0 flat line, slope = b = 0 ANALYSIS OF LINES 9. the statistics of lines: 1. calculate the ‘best fit’ slope & intercept of that line – equation of the line that best fits the observed data 1. is where the data values are closest to the lines – the distances bw the observed values and the estimated line are the smallest (the least square method – minimize the square of differences bw the observed data and the line) 2. variability about the regression line – variability of the data around that best fit line 3. stanadard errors of the slope & intercept 4. statistically test the slope and intercept – what they differ from the null hypothesis 2. we want to calculate the slope, the intercept isn’t very imp for testing the null 5. calculate the confidence intervals for the estimated slope & intercept HMB325H © Lisa Zhao 2| Page 2 HMB325H © Li| Page 32013 ‘FITTING’ THE BEST LINE calculate the best fit line 11.y = a + bx 1. a: the y-intercept of the line 2. b: slope of the line 12.the ‘best fit’ line: smallest squared deviations bw the line & the data n(ΣXY)−(ΣX)(ΣY ) b= 2 2 n Σ X − ΣX ) 2 a= (ΣY Σ X − ΣX (ΣXY) n Σ X −(ΣX) 2 13.the mean value of x or y will always lie on the best fit line 14.variability about2the line: Σ Y− a+bX ) S y∗x √ n−2 OR S y∗x n−1 (sY−b s2X2) √n−2 (not as ac
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.