Statistical Fridays J C Horrow, MD, MS STAT Clinical Professor, Anesthesiology Drexel University College of Medicine
Session Review New concepts: Observations vary Observational vs. experimental data Graphing your data Example: 50 patients induced with TPL or PPF Homework: 20 patients given spinals for C-section
Session Summary 1.A statistic is a function of the data. 2.Useful statistics have known distributions. 3.Statistical tests are based on a “null” hypothesis to be disproved. 4.The “Normal” distribution is the most common and useful distribution.
Statistic: function of data Average of all observations (mean) Average of the smallest and largest The middle observation (median) The largest observation The first observation USES: for estimation and testing
Statistical Inference ESTIMATIONHYPOTHESIS TESTING Point Estimators Interval Estimators mean, variance95% Conf. Int.
Useful Statistics Are unbiased (on average, hit the mark) Have minimal variance Have known distributions Sample average ~ Normal ( , ) Sample average ~ t ( , s ) Sample standard deviation = s (N-1)s 2 / ~ (n-1)
Useful Statistics Suggest several useful statistics for the induction data and state the assumptions for each PPF TPL
Concept #3: Null Hypothesis Statistical tests are based on a “null” hypothesis to be disproved. EXAMPLES: TPL = PPL for SBP MEDIAN TPL < MEDIAN PPL TPL = PPL PPF TPL
Concept #3: Statistical Tests Statistical tests utilize “test statistics” (duh?) To test equality of means: TPL – PPL and compare it to zero To test equality of medians: Sum of Ranks TPL - Sum of Ranks PPL To compare sample variances: s TPL / s PPL and compare it to one
The Normal Distribution Most common way numbers distribute Occurs when measurement results from sum of individual parts Sums and averages “Bell”-shaped curve; symmetric Observations clustered in center; fewer occur farther from center. No “cut-off” at either end
The Normal Distribution
Session Review New concepts: Statistics are functions of the data Useful statistics have known distributions Statistical inference = estimation; testing Tests seek to disprove a “null” hypothesis Example: 50 patients induced with TPL or PPF Homework: 20 patients given spinals for C-section