ࡱ> IKHi@ ;bjbj.. 6DD=2#G , , , @ d,d,d,8,L,d@ LX-~.(../5"5 .5CCCCA5DGuK$NR0QK-, f64@4@f6f6K ./tK>>>f6t Z., /C>f6C>>:nC,B , C/L- Ngd,8,C CK0 LC Q;QC@ @ Q, C(65>t5,>5$5656565KK@ @ $d,=@ @ d,t-test and One-Way ANOVA Purpose: In this lab, we will compute an Independent t-test and a One-Way Analysis of Variance (ANOVA). Background: Most commonly employed biostatistical procedures is the comparison of two samples to infer whether differences exist between the two populations sampled are t-tests. Independent sample design The samples are independent from each other. The observations from one sample are not related to the observations from the other. Need to check for skewness and equal variances using the F-test. If variances are unequal use the Mann-Whitney test. The null hypothesis for the Independent t-test is  EMBED Equation.3   EMBED Equation.3 . tcal = observed differences between sample means/pooled standard error, or  EMBED Equation.3 , where  EMBED Equation.3 . Paired sample design The samples are not independent from each other. Sample 1 is in some way correlated with sample 2, so that the data may be said to occur in pairs. Measurements are taken on a single subject; one common example is before and after experiment. A second type of pairing occurs when an investigator matches the subjects in one group with those in a second group so that the members of a pair are as much alike as possible with respect to important characteristics, such as age and gender. This minimizes environmental effects but also reduces inferences limits. The null hypothesis for the Paired t-test is  EMBED Equation.3   EMBED Equation.3 . The test statistics for the null hypothesis is  EMBED Equation.3 , where  EMBED Equation.3  = mean difference between pairs.  EMBED Equation.3 = (# of pairs 1), standard error Test does not have the normality and equality of variances assumptions of the independent t-test, but assumes instead that the differences, dj, come from a normally distributed population of differences. Thus test by:  EMBED Equation.3  (standard deviation of differences) /  EMBED Equation.3  (mean difference); if the calculated value < 2 do paired, > 2 Wilcoxon paired-sample test. One-way ANOVA The 1-way ANOVA is used to determine if there are differences among means. The analysis is used in situations where there has been a variable (Independent variable) that has been manipulated (e.g. treatment and control) and one measured variable (Dependent variable). If there are only two levels, sometimes individuals use a 2-sample Independent t-test. However, the test statistics for the two procedures are mathematically equivalent. The validity of ANOVA depends on whether or not certain conditions (assumptions) are true; if the assumptions are not true, the results are not valid. The assumptions are: 1) data in each level are normally distributed, 2) variances of measurements within each level are equal for all levels, 3) observations/individuals are randomly selected for the experiment and randomly assigned to the levels, and 4) the observations/individuals are independent among levels (i.e., different individuals/observations in each level). General procedure for conducting a One-Way ANOVA 1) Determine the appropriate alpha for the experiment. 2) Develop planned (a priori) comparisons. The ANOVA will tell you if there are differences among the level means but not which means are different and which are not. The planned comparisons are used for determining where those differences are. 3) Obtain random observations for each level of the independent variable. Make sure that there are different observations/individuals for each level. 4) Before you do an ANOVA, you MUST check the assumption of equality of variance. Use a statistical test (i.e., Burr Foster Q-test) to determine if the variances of individual measurements are equal between levels. If the assumption is TRUE, proceed to computing the ANOVA. If the assumption is NOT TRUE, check for skewness, apply the appropriate transformations and retest the assumption. If the variances ARE EQUAL AFTER TRANSFORMATIONS, use the transformed variable instead of the measured variable as the dependent variable in the ANOVA. If the variance are NOT EQUAL AFTER TRANSFORMATIONS, you must use a different type of test, a non-parametric test. For a 1-Way ANOVA, the appropriate non-parametric test is the Kruskal-Wallis test. 5) Compute the ANOVA to test for differences among means. The null hypothesis is:  EMBED Equation.3  If you ACCEPT HO, you conclude that there are NO DIFFERENCES among level means. If you REJECT HO, You conclude that there ARE SIGNIFICANT DIFFERENCES among level means. Compute the planned comparisons to find out where those differences lie. If you reject an Ho for a planned comparison, produce a bar chart showing the values of the two means that were compared. 6) Conduct an unplanned multiple comparisons test (a posteriori) to see if there are differences among means that were not examined during the planned comparisons. This procedure is meant only to help you identify potentially interesting results for further experimentation. DO NOT use this instead of the planned comparisons. If you reject an Ho for an unplanned multiple comparisons test, you should design another experiment to actually test that result. Example of a 1-Way ANOVA PROBLEM: You wish to determine if beetle density on leaves is related to plant height. Your original idea was that beetles probably would select shorter plants because they would be easier for them to get to shorter plants. For this experiment, you decided to use a specific species of plant on which the beetles are normally found. You developed racks to place them at ground level, shrub level, and tree level. Question: You decide to guard against Type I error and set alpha to 0.025 (WHY?). DESIGN: You will conduct the experiment at a local county park. The independent variable (factor) is Height with three levels: Ground (< 0.2m), Shrubs (0.2 TO 1.0m) and Trees (>2.0m). The dependent variable is Beetle Density (number of beetles per 0.25 kg of plant biomass measured after five days. You will use plants of the same species that are approximately the same size. The plants will be placed on racks at the three different heights. After five days, you will count the number of beetles on the plant and weight the plant. Question: Figure 1 illustrates two different rack designs: Which would be appropriate for a completely random design and why?  Develop planned comparisons - These are comparisons between levels that are planned in advance. Each comparison will be between TWO GROUPS, but a group may contain more than one level. The comparisons must be orthogonal (independent). Comparisons made in this manner are referred to as CONTRASTS. A) Determine Maximum Number of Comparisons (Contrasts) Max Contrasts = a-1, where a = number of groups For this example, the maximum number of contrasts would be 2. B) Specify Contrasts. After you choose your first contrast, your choices for further contrasts will be limited so make sure that this contrast addresses your main question. The original idea is that beetles would not prefer tree height because of the difficulty in getting to the leaves. The two groups for this comparison would be: Mean tree height vs. the mean of lower heights combined. The levels would be: Tree height (T) vs. Shrub height (S) & Ground level (G) The second contrast must be INDEPENDENT of the first. This means that once a treatment is used by itself on one side of a comparison, it cant be used again. For example, if T vs. G & S and is the first comparison, the tree height treatment (T) cannot be used in any other comparisons. Therefore, in this example, there is only one possibility left, G vs. S. Question: Figure 2 illustrates the park in which you plan to conduct your experiment. How will you randomly locate the sites for the racks? Do so on the figure.  In the first part of this lab, we will compute an Independent 2 sample t-test. For the second part of the lab, we will conduct a one-way Analysis of Variance (ANOVA). COPY the file, SUGAR DATA, to your diskette from the BIO 156 data folder in the BIOl 156 folder on the file server or enter the data from below. PART 1: Independent t-test You are interested in comparing cholesterol levels between males and females aged 20 to 29, height 5 4 to 5 9 weighing between 120 and 150 lbs. to determine if one sex may have higher risk of arteriosclerosis heart disease. You selected 10 males and 10 females at random from patients with the appropriate characteristics at Kaiser Permanente. Your results were: Male Female 164.76 185.61 139.67 107.92 158.38 165.43 189.20 187.34 168.29 222.57 222.12 170.56 207.81 205.69 150.56 174.81 194.09 150.39 153.93 186.55 Use SYSTAT to compute the mean and variance for each group. Enter the data in two columns, SEX$ and CHOLEST Set BY GROUPS in the DATA menu to SEX$ Select STATISTICS from the STATS submenu from the STATS pull down menu and compute the summary statistics on CHOLEST Check assumption of equality of variance. What is Ho? What is the critical F-value in the table? Compute the F-test. What were your results? What did you conclude about the validity of using a t-test? Use SYSTAT to compute the t-test. Turn the BY GROUPS OFF!!!!!!! Select Two Groups from the t-test fly-out menu from the STATISTICS pull-down menu. Select CHOLEST as the measured variable and SEX$ as the grouping variable. Click on OK. Did you accept or reject Ho? Interpret your results. PART 2: One-Way Analysis of Variance You are interested in determining the effect of auxin on pea section length when sugars are present. You have set up 30 cultures of pea sections with auxin in each culture. Ten of the cultures received a 0% sugar solution, 10 received a 2% fructose solution, and 10 received a 2% glucose solution. You then measured the length of the pea sections after three weeks. Biological Questions: 1. Do sugars cause changes in pea growth in the presence of auxin? Ho:  EMBED Equation.3 Control vs Fructose+Glucose  EMBED Equation.3  EMBED Equation.3 error 2. Does Fructose cause different changes in pea growth in the presence of auxin than Glucose? Ho:  EMBED Equation.3 Fructose vs Glucose  EMBED Equation.3  EMBED Equation.3 error Measurement variable: length of pea sections grown in tissue culture. preliminary experiments showed that sample size (n) should be10. Test for equality of variance A Burr Foster Q-test shows that the variances are considered to be equal. Inference limits: There are three different seed sources in California. We want to infer over all peas in California for the current crop year. The Data: Download the SUGAR DATA file onto your floppy diskette. CONDUCT THE ANOVA Use SYSTAT to compute the analysis of variance. From the STATISTIC pull-down menu, select ANOVA and Estimate Model from the fly-out menu. In the next screen, you need to specify the dependent variable (GROWTH) and the Factor (SUGAR$). Click on OK The program will then compute the ANOVA for the main Ho. Results of Hypotheses tests: Main Ho:  EMBED Equation.3 SUGAR  EMBED Equation.3  EMBED Equation.3 ERROR Decision - REJECT Ho Conclusion: Average growth is not the same for all treatment levels. DO THE A priori comparisons: USE SYSTAT to compute the a priori comparisons. Do the following for EACH comparison: From the STATISTICS pull-down menu, select ANOVA and then Hypothesis Test from the fly-out menu. Enter SUGAR$ in the Effects box and select Contrast The next thing we have to do is specify the contrast formula. Select Custom. The contrast tells the computer which treatments you are going to compare. For the first comparison, we want to compare the control to the two sugars combined. The contrast formula consists of coefficients, with one coefficient per treatment level. In this case we have 3 levels, so we will need 3 coefficients. For the first comparison, we are going to compare the Control to two other treatment levels. So the coefficient will be 2 for treatment level 1 (Control). Give each of the sugars a coefficient of -1. To do this, enter 2 -1 -1 in the box. When you click on OK, the computer calculates the a prior comparison. For the first a priori: Ho ________________ Accept or reject Ho? Compute the average length for the combined sugars (the mean of the two means) _________ __ How did it compare to the mean for the control? Conclusion? For the second a priori: Ho _________________ Accept or reject Ho? How did the mean for Fructose and Glucose compare? Conclusion? Overall conclusion? 12-PAGE 8 12-PAGE 1  Figure  SEQ Figure \* ARABIC 1: Rack Designs "NO) + { |      , - . / vijv@ hzv2CJUVjthzv2EHUjq\v@ hzv2CJUVjhzv2EHUjf[v@ hzv2CJUVhzv26H*]jLhzv2EHUj2v@ hzv2CJUVjhzv2EHUjv@ hzv2CJUVjhzv2Uhzv26OJQJ]hzv2OJQJhzv2 hzv26]&"3 4 N H I 1 2 G } ~  $^`a$$a$ !gd=;`;; "#$%-.ABCDgh{|}~-./|}Ż鮤闍~qgjhzv2EHUj1z@ hzv2CJUVhzv26H*] hzv26]jhzv2EHUj;v@ hzv2CJUVjhzv2EHUjv@ hzv2CJUVj}hzv2EHUj1z@ hzv2CJUVj hzv2EHUj֤v@ hzv2CJUVhzv2jhzv2Uj hzv2EHU'4F `c|&+J/0CDEFNUVWuLMNrv䯥䯥鞖鞖hzv26H*] hzv26]hzv256H*]hzv256]jthzv2EHU!j> hzv2CJOJQJUV hzv26 hzv2>*hzv2OJQJ hzv25hzv2jhzv2Ujhzv2EHU5-.<1h_ /G;p^p^^@ :!A!B!C!^!!!"#/#0#?$$i&''''W(k(l(r()wl^lwhzv256CJOJQJ]hzv25CJOJQJhzv2CJOJQJhzv2OJQJjhzv2OJQJUjhzv2CJUmHnHu/jhzv2B*CJOJQJUmHnHphuhzv2B*OJQJphhzv26B*OJQJphhzv256B*OJQJphhzv2B*OJQJphhzv25B*OJQJph!?@ :!;!!?!@!A!C!o"p"""###  h8$`'0*-#/##<$=$>$o$$$$$$`&a&b&c&d&e&f&g&h&i& '  h8$`'0*-^  h8$`'0*-h^h  h8$`'0*- ' ''''''''''''''''U(V(W(r()) h8|$]^  h8$`'0*-  h8$`'0*-h^h))) **(*6*D*R*`*n*|*}***++++++ 8T|$]^  |$]^$  |$]^a$$  |$]^a$))+++++,,^,_,,,w-x-y-----5/M//////////////////ꢒscj1=hzv2CJEHOJQJU&j @ hzv2B*OJQJUVphhzv2CJH*OJQJj=;hzv2CJEHOJQJU&js @ hzv2B*OJQJUVphjhzv2CJOJQJUhzv25CJOJQJhzv26CJH*OJQJ]hzv26CJOJQJ]hzv2CJOJQJhzv2>*CJOJQJ&++ ,&,',(,f,g,h,i,,,-M-\-{--------- |$]^ 8T|$]^ 8T|$]^---4/5/M///^0000 1O1P1n111L2M22222333 h8|$]^ |$]^/////_0`0a0d0e0x0y0z0{000000000000000P1n11ùιùιzjιZùOOhzv25CJOJQJjDhzv2CJEHOJQJUjChzv2CJEHOJQJU&j @ hzv2B*OJQJUVphj@hzv2CJEHOJQJUhzv26CJH*OJQJ]hzv26CJOJQJ]hzv2CJOJQJhzv2CJH*OJQJjhzv2CJOJQJUj>hzv2CJEHOJQJU&j)@ hzv2B*OJQJUVph112233344444>4@4B4D4N4R4T4z4|4~44444444444x5z5|555ꭝ~n^jJhzv2CJEHOJQJUjHhzv2CJEHOJQJU&j @ hzv2B*OJQJUVphhzv2CJH*OJQJjFhzv2CJEHOJQJU&j)@ hzv2B*OJQJUVphjhzv2CJOJQJUhzv26CJH*OJQJ]hzv26CJOJQJ]hzv2CJOJQJhzv25CJOJQJ$3333444x5z5|5526666!7_8T9W999999 8T]^ h8|$]^5559999999999D:E:::::::;;=;>;A;B;H;I;J;K;L;N;Q;R;X;Y;Z;[;\;];_;ܶܮrfjhzv20JU*hKe0JmHnHu*hzv2hKe0JmHnHujhzv20JU hzv20Jjhzv2CJUmHnHuhzv2OJQJhzv2>*CJOJQJhzv26CJH*OJQJ]hzv26CJOJQJ]hzv2CJOJQJhzv25CJOJQJhzv256>*CJOJQJ(98:9:H:I:}:~::::::::::::;;;";#;$;%;&; 8T]^ 8T]^&;';;;<;=;L;M;N;\;^;_;`;a;b;c;d;e;f;g;h;i;j;k;l;&`#$ hh]h`h $&`#$a$ 8T]^_;`;|;};;;;;;;hzv2OJQJjhzv2U*hKemHnHu*jhzv2Uhzv2hKe l;m;n;o;p;q;r;s;t;u;;;;; .00 BP/ =!"#$%```!sf# uu&"TE:O.$@: x͙WpUR!qOpL 7@$O0:&tH ]tL4#zBHtI *y2OIfȷ7]L ̌c{9{_k7D<i" M=ޗX{GG^\y[b>;짇?O~/z@˧(o>_&|n}>Yz h4>Q5yz&#c96c}O&v k;';:[fJ:!{^d2tC@gM>iܶ򵴑oh:wsa6%G->/ET4Bn(xCB#N 0q?ӗxtC}zr s՗1}m] zҝ=5XK?Yy#`3uo|Kb-hmFsl:0oA?t=vx'G?$3q#}-_u5=,A`c1GMhF+0x y817hf<#\ >5Dq/?x_v<vSe e/v<]MgX?] \mjp}SHn~2\B?<5N2j3fkfs,ˮcƣmصsar&s3w:'4v4xIAScDHH !2F&zK]2dMD?< ԙ@֓o6R}j-1O,ı;YV*&nq!~ ODvP;/̉an qx71sB4 ~ =qћzT@_WѯBj-k$x\Jr+' .$>Tz㷎^}kB 4o:9yYW;~x!r'\jZb _v2ѱE?Spk3v~(F 't3:inrT{JQeko ݡЮESNi[9Fth;'Y_#/sh @/0kuGA DP! 2!<'OL?㏞?LJ:}>>Ӌ͖Vt$6mS?EͻR~?W&+bh]}'}&{ӣdx5OO6Лh#}DPzU?CFHrE1QG46|M 0PgnqA+`<#\ >5Ds/OBFs=ڎB6\Q\7W}?[uৌu|J"~ʘ{d$õ;NEN&`$ϗX2O6|`%&Fg0wtxEL"ɒ8͒PXMtd4 ,𐍇lPw\sC݆,Xx d,TtRKA7 $$Rg"&XOn9KmWz1AL@SrL$DMfN|y2TiI2$ b; ihMEs L^46,Qs. ƙDУ1u,OfTWqy'W>]ux**o2C㕰9,▢V&x^DEJ'o鮡{}нa׻uszNs\>=zv% qm ⯏uz "d9!zNI9 D"چpNY8#c=&!zTt5sOVꥡV/mGVDs&w\Gf^^4c/ x;pҤ[M"h8#2޽䳥wSnǎHնVi>X+oXGNy7g' ['x.QFNǞ6R5RvWkEꏩ?T=\zت=GZ?'E*l(f9 w:<}?SOƾJ ?-S(dIr;Jz^3T*Ws)e<4qݾR(Rg-m?a/Sw9Gy p2r mpR,CE m9'NrXhbl %H.r^Fhgxt'?<2/;Y (&.BJ Pq CotD}x0^wƭ==£E;wϲRN (e(sL<F2.73lW6޲ . ^)!4G >~-Aa̫d32Y zut>YnlL4u9k:9L|'o<:rMi0d ys%k6>n :nٟ)M돹KHgo-Lݴ~^WIt iumzǻڂVg^%iwexJE{+lM7O6'$nq- YtI ;/d__@|A.ek/dvm>"K/c%tSoѝod@BAenw$VvyZ /O9p2\] N NMNl'>*Tr+'m.<7yNִI RR3bOU5P@=Jw>@Gyxi$o#+ xjD<~;-:ՠ[~5yo'Gxhʡ \Gpx8)՜N-M-'6'-|@5'SxSx:mүXgs_lб z-s/pSo`{O)~ /\/ ߄&4lw^ xNY"E:Dރ\}ʁ=Ƨc}[߃VA]Jd%gxoooTWWߘݾKuBfe相?O=UKpU\`!0TUp p^U9mIa:%0XH"l4wBp=du_Vd 0k DkџHIx 1i8E|N愌@ϻqG_マ^{Ǔz0n-iJ8E\&vUx#Z7G3R21}^੎G:(c/J{~?aбG2m`X0^岎OϚNEc*^#\ kpǁxAN 0?F#= x(KOx .w3>6u"Ο xSzH$`Jx+ :֟ )jbR6&a:c}3Y(4v5$Ng# ?Yxۅ]ҧ\;o7k} 'RX,"5>}(Z+ &3T͒##ea?>~>1|^y*KM7L%男|wyVɔa\*H@$x~S8Rci;|:\W*-isTM%`jFTM3)omJ=UShP}?&O#^OZi-ϻ{POC[B<5qkOvhN/qP;R £oMvbTh AwBm06#(/zၖ^^L>x>;'[֯ xX̫ k"x}IWzE&6ik<&g3 ^ {/$;yG< Nj+ p1> Zwh)A}}Nh pI{$~RaY$|i CA1ߡ">?2X>4iCxs>& Hzaų͢?Bl1TrlbWz&/Ԍ.ry#?e(LDd hJ  C A? "2N\Ҍo6h:GD%`!N\Ҍo6h:G @ |Pxcdd``dd``baV d,FYzP1n:&&! KA?H1Z ㆪaM,,He`HI? @201d++&19(b Xt9\vVJߞ7<= K@FXa HI9 @Qn.\7 dnn% cq?=u)+Jpg$1Ql?p(@&xTrAS8v=4Ĥ\Y\ q1(2tAB N' \'LDd TJ  C A? "2{>z$LT=%`!{>z$LT=   XJPxcdd``dd``baV d,FYzP1n:&&! KA?H1Z ㆪaM,,He`I? @201d++&19(b Xt9 vVJ߄7<= 3@FXa HI9 @ы\\ 1nnJ?~&4{ ؉u+ȈHbl …vcQM=\Fh``$#RpeqIj.ŠQ s> 1h:(G3X:WcDd (J  C A? "2>삢Vٺ`%`!삢Vٺ`*@ xڥSKAfv\O(ApHS,ʡi;X ++. VH8U4VEH]Y޼{,9,10BHA0T(GFm0K^Hezi?Jgpo t$Y EIZKQ,%9DkgBA#[֎l7ա˵m5b6MnyR.pŎ[%+AYH^gۮU>3MDd tTJ  C A? "2QЯf <Ɂ'%`!QЯf <Ɂ'   XJQxcdd``dd``baV d,FYzP1n:&&v! KA?H1Z ㆪaM,,He`c&,eBܤr5b'Xr.P56ceiը$< .3cmg 5\! ~ Ay 1jnMV0<&=870>f ' 0R蟫f. !rAS8v=4Ĥ\Y\ q1(2tAB N' η\^Dd hJ  C A? "2D]ΈǶaޖ5 `!D]ΈǶaޖ52 @F|bxcdd`` @c112BYL%bpu 1h:(G3X^Dd |TJ  C A? "2٬,ySw{ NUVc `!٬,ySw{ NUV2` XJbxcdd`` @c112BYL%bpu?ABCDEFGjJMNOQPRSTVUWYXZ\[]_^`abdcegfhklmnopqrstuvwxyz{|}~Root EntryP F@+GOgLData @LWordDocumentO6ObjectPoolRNg@+GOg_1081517056 FNgNgOle CompObjfObjInfo !$'*-27<?@ABEJOTXYZ[\]^`abcdf FMicrosoft Equation 3.0 DS Equation Equation.39qFD H O : 1 = 2 FMicrosoft Equation 3.0 DS EqEquation Native b_1081517106 FNgNgOle CompObj fuation Equation.39qF. H A : 1 `" 2 FMicrosoft Equation 3.0 DS Equation Equation.39qObjInfo Equation Native  b_10814984706FNg)NgOle  CompObj fObjInfoEquation Native _1081498737F)Ng)Ng$ t=2X 1 "2X 2  2s p2 n  FMicrosoft Equation 3.0 DS Equation Equation.39q04 =2n"1Ole CompObjfObjInfoEquation Native L() FMicrosoft Equation 3.0 DS Equation Equation.39qO+< H O : 1 " 2 =0_1081517249"F)Ng)NgOle CompObjfObjInfoEquation Native k_1081517270'F)Ng)NgOle CompObj f FMicrosoft Equation 3.0 DS Equation Equation.39qO H A : 1 " 2 `"0 FMicrosoft Equation 3.0 DS Equation Equation.39qObjInfo!"Equation Native #k_1081749954,;$F)Ng)NgOle %CompObj#%&fObjInfo&(Equation Native )S_1081517538)F)Ng)Ng7$ t=2ds 2d FMicrosoft Equation 3.0 DS Equation Equation.39qD 2dOle +CompObj(*,fObjInfo+.Equation Native /-_10815176271.F)Ng)NgOle 0CompObj-/1fObjInfo03 FMicrosoft Equation 3.0 DS Equation Equation.39q `  FMicrosoft Equation 3.0 DS Equation Equation.39qEquation Native 4)_10817499133F)Ng)NgOle 5CompObj246fObjInfo58Equation Native 9:_10507171928F)Ng)NgOle :8 S 2d FMicrosoft Equation 3.0 DS Equation Equation.39q``mI\yI  Observed among meansCompObj79;fObjInfo:=Equation Native >0_1082788979E@=F)Ng)Ng2 d" Expected among means by chance alone2 FMicrosoft Equation 3.0 DS Equation Equation.39qnl  2Ole CCompObj<>DfObjInfo?FEquation Native G6_1082788991JBF)Ng)NgOle HCompObjACIfObjInfoDK FMicrosoft Equation 3.0 DS Equation Equation.39qn | d" FMicrosoft Equation 3.0 DS Equation Equation.39qEquation Native L)_1082730909GF)Ng)NgOle MCompObjFHNfObjInfoIPEquation Native Q6_1082789023LF)Ng)NgOle R  2 FMicrosoft Equation 3.0 DS Equation Equation.39qn | d"Oh+'0CompObjKMSfObjInfoNUEquation Native V)1TableQv+/ Y,-e^QޔV9d#iO$~cXZkөFzgIqvIjWynC{^!9)|&4DQmٟ26~n!8B>l]ʷinSnR=ekP3_!nt4aV7G2_k{ZDd @J  C A? "2SfPy`x2-L/ `!'fPy`x2-LfR xcdd``$d@9`,&FF(`TQRcgbR Рm@ UXRY`,@1[)!Ar؅,L ! ~ Ay b'Xr3vq-`bL䂺 jF;=`221)Wx\ ]` ]> 1h:(.f~9;Dd J   C A ? "2<I nQuVJ `!I nQuVJ:@`!xcdd`` @c112BYL%bpu.P16b`޼J.-\`0M!I)$5bPd+YAD9F]`#46Dd @|J   C A ? " 2Yގ3IǓ+5 `!-ގ3IǓ+` 0xcdd``$d@9`,&FF(`T)IRcgbR Al@P5< %!  @_@YN`b]F"L 0L !*'] ZZ\ f22bcL<  \PrC b;~LLJ%  `S [ v@iDd @J   C A? " 2SfPy`x2-L/ `!'fPy`x2-LfR xcdd``$d@9`,&FF(`TQRcgbR Рm@ UXRY`,@1[)!Ar؅,L ! ~ Ay b'Xr3vq-`bL䂺 jF;=`221)Wx\ ]` ]> 1h:(.f~9;Dd 0XJ   C A ? " 2쥦Vѹ" `!쥦Vѹ"+ oCxmRJ@hkKPBZ^O>|/V:D>[LϚv@#o1 };ɹ9{i#V7vkY?#%iI Ddpx px0   # A  2~^i )z8 `!~^i )zB/lЉF/x]l8EP0۔D-0`njjˇ>&AMXe>0 R.CatڌtE`Ȟ ḶL)FnL+K8;HQSЋsϹT[P!}oM0BSB_|F}`򩰭lPt[ _doߪu^3 [_U9n;E'IJ:Ū5 ]OQ$Y~eoۂ ʿw|T1=؝?H|Ti]&σfYM?xʟ7x 4"#LCxW1:]8[lH@r,a:nCTRjh5?Q΢{Lgܽ9%PiG.>LϽmԹ踛nD|VcNfO?=> #=k_'c/8 SA0?s=Xqm_GLE-D?Qœ]r|DW~xIH&dj;>8jB7RT*um?*ܻ3;Տ&1cަ/)D$%EQ$EIT-8car6;]7;׾o&ӂ*k}󅧭H r '5ʫNuG:'`ذ_ T* Xk.oZugA$oZ\CD>`SuE4<*.tk<y{&UC\E,\O/ĵS`@,9 ?1~V&qۢv =Qx$uy~rչNx}]f]7g@(xob|_{sJs ^eKnr1/y :V:Ag,K`?S'XMyƤ+ƿNy\\ۤa`ۍoQ)yk/+N7q,nXӜ2 6Iy~~.)Y}[t-QO氂/D"tyq O^٨Mo_O4Duq}x+eޱnX _82_GnOICWX˷XPoCߴ0mayYj}wOnEb>v}9?߀{%jgk,B`䔟ׁɆJA0I{n k?}DC||@Ò@}q%WP'N',(a8^gab%ӕZO:_(1{-T_,<'V{G&ݸ j -y.2x_`Ct֝ˌƼ;gitX-Bv?/Vv2BKZ`dNǚy-ä%寐x_KEkc8G9҂=J,BI PcG_ -x;* t;0reܤ>x[܂}/?-1E`ȷrzcWT4hu>OU}yf>g ";QT|Okk;Рff.~\.T޶L Ew /T7 nuFKR{<uLa|Q[;ӎԝm57헟j婍A(c+1˖B?{4'ÉC TJ.t%QH1`L#v^JG"nnI9Ɇ*LP, cs S$^UtPưs暬 g{9.˱l(0NhXeHE|/(mD TF+.<'`q@J0}O JgyL:3x=bαAq \ٟR2_̀ I ð,oXc#guǢS@8!_dXx*k4kYVL? w#s|b-f,c{eVH/(HŘQUՅyTuADt^"MMQ,uQ }%eqrT HB F "Q\_"d - :!Ϫ*$~W:u#)D@,҉}F {NY,`0˚._Zl4&-:/~] t%Inr핹$} OA:IP>?(cT-;=ɑҁs*Oޏ@]LH|tXkcĿ+lWP|z%rDv8x,/_};&}+n@ XlR cB1u|đO :3~sUb`}cxڏD#5 h/`;.+}-{zh;4}R'>.UBuYW}륪Clkyݧ})=DڎÛ^ \H(ޏV=VGSފ#FL0F G[yۡf68%ʻi2U4lu|un<:a= s~Y]P<rCׅv1~thЉF/;x}X]luIQ;\2BZCy$+(8FY*D(* h.e)PA_ !A !dIiT][+QRHr|{ιwnsG"zŮ]vi#tt%7|1m)Wv[Q|{>PzjC=k'u5D1 %+>o|n6k('lCdwMrm@m-igՑ*ZYVۻ]l-#ju(AYbd1rSApqTv%*FyJOjЭrx59Q=+h(g+|V#tR)[Wr;0>zaL>{zmTo ͥrX$RB3XqX-VoDdYv;Z@ǹuT Ȑ^|ށTBh&vc@RxJcr`0WaԮ) ,\tkw}✝RSo"vu'Jӥw,n D$&R<764ĄgmIq59` P4 ۈa#iUuPev¢! \dYzaڕ!$A4 6mHHch].G<9*euLZ%B^ "z T`~~ >WK6El:ZnK w,,5x9m%Ӆ7B=eҵ1GZ݁ 6DaP29'c C9)ݜԼۄW^9JOo׏_(DV09YЃĂ4Vr佸7w>,"hSSz]0/? Ӊ{ c gtZ9>hnbiA|FdGM- ځ=Ëɚn钳7OZ ]gQK,u6,CZtc59 ]MATl=q9䣊Ve5fQ b8Fe: :.WڏRv]ɐe֚XΗ8:귃Hƾ22 |]HQ{|B$K֋YyoRw?lޠ i|,9ҏ;qVd_XY2dB)#KmչegَVfz|^t\rՆ<ۀ彩ٸ4T@HFfבFly] 0{Jf,'cc@ywex3̶=|=#m޻!gص}eR!#~ww~ 8$;0k;X3CU~2Zm kHKS~5K~FO>{sww}<7y|I_nt[aDd T@J  C A ? " 2Vrot͕?2; `!*rot͕? XJ xcdd``> @c112BYL%bpu @c112BYL%bpu @c112BYL%bpuT3G~ FC `!gGBE>T3G~:@Rxcdd`` @c112BYL%bpu @c112BYL%bpu @c112BYL%bpuT3G~ I `!gGBE>T3G~:@Rxcdd`` @c112BYL%bpu @c112BYL%bpu?@ACop/<=>o`abcdefghi  U V W r !!!! ""("6"D"R"`"n"|"}"""####### $&$'$($f$g$h$i$$$%M%\%{%%%%%%%%%%%4'5'M'''^(((( )O)P)n)))L*M*****3+++++,v,w,,,,,2----!._/T0W0000008191H1I1}1~111111111111222"2#2$2%2&2'2;2<2=2L2M2N2\2^2`2u222 0 00 000( 00000000000( 00000000000?0?0?0?0?0?0?0?0?0?0?0?0?0?000000000 0000000000000000000000000000000000000000000000000000000000000000000000000000000000000 0 0 0 0 0 00 0 0 00 0000000000000000000000000000 000 000 0000 0 0 00 000 0000000 0 0 0 0 0 0 000(0(0(00000(00(0(0(0000000000000000000000(0(0(0000000(000000000000000000000000000000000000000000000000000000000000@0@000@0@00v?08<00j"""% )/15_;;!"$),-/2# ')+-39&;l;;#%&'(*+.013; { ,."$-ACg{}|/CE'''''''''d(x(z((((((( ,,!,),=,?,@,T,V,2::::::::::::::::::::: %!!35I 4  ll2$sf# uu&"TE0hR$}ԧJ;%Җ&<b$yS֟C.,b$?BZ2,ȼQ0qA.b$zO