Spatial Analysis GEOG 5670
Popular in Course
Popular in Geography
This 6 page Class Notes was uploaded by Dr. Cristopher Deckow on Wednesday September 30, 2015. The Class Notes belongs to GEOG 5670 at Western Michigan University taught by Staff in Fall. Since its upload, it has received 36 views. For similar materials see /class/216847/geog-5670-western-michigan-university in Geography at Western Michigan University.
Reviews for Spatial Analysis
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: 09/30/15
Hypothesis testing and basic inferential statistics GEO G567 Hypothesis testing Twotailed test 7 Rejectlon region located in both tails of distnbution e No reason to aipect sample statistic to be geater or less than population value Onetailed test 7 Rejectlon region located in one tail or other Type I error on 7 Probability offailing to accept a true hypothesis Type 11 error 3 7 Probability offailing to reject a false hypothesis Power ofthe test 1 ater as difference between the hypothesized mean and true ean increases Testing for normal distribution I Chisquare X2 goodness of t test 7 Used to determine if data follow particular distr39bution 7 Compare sample data with expected distribution ifnull hypothesis is true 7 Data should be placed in categories classes 7 We can determine expected frequency for each category assuming null ypothesis is true 7 Null hypothesis data are distributed as hypothesized 7 Can only be used with frequencies 7 Assumptions Sample is random Sample size is large gt30 Expected frequency in each cell should be at least 5 Determining if two categorical variables are independent I X2 contingency tables 7 Uses same chisquare statistic as above 7 Ifdifference between observed and expected Wlues is small differences could occur by chance a Ifdifferences are large reject null hypothesis and conclude there is relationship between two Variables ofslze n drawn Born a large population Sample size is relatively large Expected cell equencles should be at least 5 l izlodgjz The t distribution I Used with samples in which We must estimate the standard error of the mean IfWe are dealing with alarge sample We canuse the normal distribution Z I tdistribution is useful with smaller samples 7 As sample size gets larger the tdistribution resembles the normal distribution 7 tdistn39bution also must take into account degrees of freedom One sample t test I Used to compare a sample mean to a knomitheoretical population mean Assumptions 7 Variable is normally distributed or sample size is large 7 Population std deviation is unknown sample used as estimate Twosample t test Assessing relationship between two vaIiables Testing means obtained from two samples Do samples come from same population Assumptions 7 Independent random samples 7 Population nonnally distributed XI NNM 0i X2 NltH2102gt 7 Null hypothesis 1 p or Jam 0 7 Use modi cation ofttest because 52 is unknown ZEFEHM MJ 0R i 6m Basic inferential statistics 11 GEO G567 Twosample ttest 0 Assessing relationship between two variables 7 Testing means obtained from two samples 7 Do samples come from same population 7 Use modi cation of ttest because 62 is elude Ga ZEK 1 OR Twosample ttest 0 Must determine if sample variances are equal 7 Use Ftest dfn1 7 1 n2 71 largavanance smalls variance Pooled variances quotinfr m4 mm mm 0 Separate variances A1 A 1 A L L 01 m m Paired ttest I Samples are NOT independent 7 Populations normally distributed XINN 17012a nd X2 NUquot 622 e d equals mean ofpaired differences 7 SDquotn is esumaie of sid error of mam diflerence Nonparametric tests I MannWhitney U test 7 Equivalent of two sample ttest I Wilcoxon test 7 Equivalent of paired ttest I KruskalWallis test 7 Equivalent of ANOVA 925 MannWhitney U test I Nonparametric test for difference of means I Independent samples I Rank all variables keeping group identity U n1 n2 2n1n117R1 7 where R equals sum ofranks for group 1 MannWhitney test can be convened Zstatistic U 7 2n1n2 fiiln2nln2112 Wilcoxon Signed Ranks Nonparametric equivalent of paired 1651 7 Rank paired differences from lowest to highest 7 Use absolute diiferenee not positive or negative 7 Use average rank in 39 event oftle T Sum ofpositive or negative ranks e Ifdifferences are random sum ofpositive ranks T e out the sane as sum ofnegative ranks TN Twotailed test 39 Use smaller of T and TN Onetailed test Test appropnate rank Sum
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'