Class Notes (1,020,074)
CA (585,730)
UVic (3,044)
CHEM (208)
Lecture 5

CHEM 212 Lecture Notes - Lecture 5: Talc, Null Hypothesis, Observational ErrorPremium

2 pages38 viewsFall 2016

Department
Chemistry
Course Code
CHEM 212
Professor
Hamilton M.Coreen
Lecture
5

This preview shows half of the first page. to view the full 2 pages of the document.
Chem 212!
Lecture 5!
Sept. 16/2016!
!
Graded assignment, identical question in class. Sometime next week!
Finish worksheets this weekend****!
!
Stats continued...!
-Stats test testing if there is a dierence or not!
-at higher level confidence we need much stronger evidence!
-Testing the null-hypothesis Ho!
#-xbar and mu are not significantly dierent and can be explained by random error alone!
#-testing whether the dierence is real or not!
#-t value is the value expected only from only random error!
-Setting up the t-test!
#1. Calculate t-cal= |38.1-33.56|*(5^1/2/3.15)=3.2!
#2. Look up table: for d.f.=5-1!
#3. t-table=2.78 at 95% conf & t-table=4.60 at 99%!
#4. Conclude:!
##There is a significant dierence at 95%!
##There is not significant at 99%!
!
The stronger evidence is saying there is something else going on rather than just random error.!
Your tcalc tells you if you can reject the null hypothesis that says there is no dierence besides random error!
!
Steps for conducting a t-test:!
calc xbar and s!1.
Calculate talc!2.
look up t-table based on d.f. and desired level of confidence!3.
compare talc and table!4.
make conclusion!5.
if t-calc>t-tbale there is a significant dierence at the selected confidence interval!A.
t-calc<t-table there is not a significant dierence at the selected confidence interval!B.
!
Comparison of means: Unpaired: xbar, s!
###########Secondary calculation!
######t-test EQN 1###S pooled!
-do an F test to choose !
!
######t-test EQN 2###degrees of freedom!
!
####do t-test!
#!
Comparisons of dierences Paired:!
###don't average results!
!
###t-test on dierences!
!
-dont expect the results to be the same therefore do not average them. Do our t-test on our dierences!
!
-The hardest part of statistics for students is choosing paired or unpaired**!
Do not assume that whenever you have a list of numbers it is valid to average them!
-average only if the data points are replicate values for the same measurement!
!
Comparison of means t-test:
-must have unpaired data!
-first use an F-test to select appropriate t-test and then do the t-test!
-F test compares s.d of one data set (Sa) to that of second data set (Sb)!
You're Reading a Preview

Unlock to view full version

Subscribers Only

Loved by over 2.2 million students

Over 90% improved by at least one letter grade.