html5-img
1 / 23

Statistical Techniques I

Statistical Techniques I. EXST7005. Linear Combinations. This is a function of random variables of the form S aiYi where ai is a constant and Yi is a variable. Generic Example: We want to create a score we can use to evaluate students applying to LSU as freshmen.

adamma
Télécharger la présentation

Statistical Techniques I

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Statistical Techniques I EXST7005 Linear Combinations

  2. This is a function of random variables of the form SaiYi where ai is a constant and Yi is a variable. • Generic Example: We want to create a score we can use to evaluate students applying to LSU as freshmen. • Score=a(VerbalSAT)+b(MathSAT)+c(GPA) • where a, b and c are the constants and • VerbalSAT, MathSAT and GPA are the variables (they vary among students). Linear combinations

  3. Score=a(VerbalSAT)+b(MathSAT)+c(GPA) • we need to choose values of a, b and c • if a=1/3 and b=1/3 and c=1/3 then we have an average (a+b+c)/3. • If a=1 and b=1 and c=1 we have a simple sum (a+b+c). • But VerbalSAT and MathSAT are values in the hundreds and GPA is around 2 or 3. So we might choose a = b = 1/100 and c = 1. • Any of these is a linear combination. Linear combinations (continued)

  4. Mean and Variance • So what is the mean value of our linear combination, and can we put a variance on it (to get a confidence interval)? • Linear combination: SaiYi • Expected value: SaimYi • Estimate of mean: Sai`Yi

  5. Mean and Variance (continued) • Variance of the linear combination. • The variance of a linear combination is the sum of the variances of the individual variances (with squared coefficients) plus twice the covariances of the variables (with both coefficients). • Estimate of the variance: • Sa2iSi2 + 2aiaj(Covariances)

  6. Mean and Variance (continued) • For example, with the Score we calculated previously, the variance might be • VAR(Score) = a2VAR(Verbal) + b2VAR(Math) + c2VAR(GPA) + (2abCOV(Verbal,Math) + 2ac(Verbal,GPA) + 2bcCOV(Math,GPA)

  7. Mean and Variance (continued) • HOWEVER, if the variables are independent the covariance can assumed to be zero. • The linear combination reduces to the sum of the variances of the individual variables (with squared coefficients. • VAR(Score) = a2VAR(Verbal) + b2VAR(Math) + c2VAR(GPA)

  8. As the course progresses we will see applications of linear combinations to almost everything. • Two sample t-test: H0: m1 - m2 = 0, estimated by `Y1 -`Y2 = 0. This is a linear combination! Suppose you want to test 0.5*`Y1- 2*`Y2 = 0. • Regression: The model, Yi = b0 + b1X + ei is a linear combination. Utility of linear combinations

  9. Utility of linear combinations (continued) • Analysis of variance: We will look at contrasts to test for differences between means similar to a two sample t-test (but with more means). For example, test the hypothesis that H0: 2m1 - m2 - m3 = 0

  10. An application: Stratified Random Sampling • Suppose we want to estimated the number of ducks in an area on the Louisiana coast. The area of interest is 300 acres. We fly 9 transects, counting ducks for 1/4 mile on either side of the plane, and from each transect we estimate the number of ducks per acre. We can then calculate an estimate and a confidence interval for that estimate.

  11. Sampling results and calculations An application: Stratified Random Sampling (continued) Sample value Statistic Value 8 n = 9 19 sum = 5103 30 mean = 567 23 var = 684349.25 56 std dev = 827.25 89 std error = 91.917 2024 t-value = 2.306 1732 acres = 300 1122

  12. Estimates and confidence interval An application: Stratified Random Sampling (continued) Sample value Statistic Value 8 mean 567 19 acres 300 30 est total 170100 23 variance 61591432500 56 std error 82725.40 89 91.917 2024 half width 190765 1732 CL-lower -20665 1122 CL-upper 360865

  13. An application: Stratified Random Sampling (continued) • Where the true number of ducks is mDUCKS, we can state our results as a probability statement. • P(-20665£Total Ducks£360865)=0.95 • Note that this calculation is for the total number of ducks on the 300 acres of interest.

  14. Now lets suppose we noticed that the ducks species of interest were primarily a fresh water species, and occurred only infrequently in the brackish and saline areas of the 300 acres of interest. • The 3 habitat types would be called "strata", and we could estimate the number for each strata separately (and INDEPENDENTLY). Stratification

  15. Results for separate areas Stratification (continued) Obs Saline Brackish Fresh 1 8 23 2024 2 19 56 1732 3 30 89 1122 n = 3 3 3 sum = 57 168 4878 mean = 19 56 1626 var = 121 1089 211828 std dev = 11 33 460.248 std error = 3.667 11 153.416

  16. Linear combination of independent means. Stratification (continued) t-value 2.306 acres 300 est total 170100 variance 2130380000 std error 15385.347 half width 35478.70 CL-lower 134621.30 CL-upper 205578.70

  17. Old and new estimates. • P(-20665 £Total Ducks£ 360865)=0.95 • P(134621 £Total Ducks£ 205579)=0.95 • Why does stratification give a smaller interval width? Because we replaced one sample with a very large variance, 61591432500, with 3 samples with smaller variances: 121, 1089, and 211828. Stratification (continued)

  18. You may recall that one way of increasing power is to reduce the variance. We did not dwell on this previously because the only mechanism I could suggest at the time for reducing variance was "improving measurement error". Now we have another method of reducing variance, stratification. This involves sampling smaller homogeneous units instead of one large heterogeneous unit. Stratification (continued)

  19. Stratification (continued) • Is the fact that the 3 variances were not similar a problem? No, nowhere in working with linear combinations did we state that the variances had to be similar. Later we will find that this can be advantageous, but it is not necessary for this type of analysis.

  20. The use of "Linear Combinations" is a rather generic technique with many applications throughout statistics. We will see them again in the two sample t-test, regression, and ANOVA. Sampling is another example of the application. • It is important to determine if the variables in the linear combination are independent or not. If they are, the covariances can be considered to be zero. Summary

  21. Summary (continued) • The linear combination and it's variance is calculated as • Linear combination = SaiYi • Variance = Sa2iSi2 + 2aiaj(Covariances) • Where if we can assume the variables are independent, 2aiaj(Covariances)=0. This will be very important later. We will assume independence in t-test and ANOVA, and parts of regression, but not all of regression!

  22. A final note on linear combinations • We have been assuming "independence" for a while. We sample at random to obtain independence, and to get a good representative sample. • But do "linear combinations" have anything to do with the simpler calculations we talked about earlier when we assumed independence, say some hypothesis test of the mean?

  23. A final note on linear combinations (continued) • I'm glad you asked. As a matter of fact the mean is calculated as • mean = (Y1 + Y2 + Y3 + ... + Yn) / n • and that is a linear combination. Fortunately for us we do not have to consider the COVARIANCES of individual observations because they are sampled INDEPENDENTLY!

More Related