Chapter 7: Statistical AnalysisConfidence LimitsPowerPoint PresentationSlide 4ExampleHypothesis TestingDetecting Gross ErrorsSlide 8Slide 9Dealing with Outliers in a Small Set of DataSlide 11Chapter 7: Statistical AnalysisEvaluating the DataConfidence Limits•Confidence limits define a numerical interval around x, experimental mean, that contains u, the mean for a population, with a certain probability.•Confidence interval is the numerical magnitude of the confidence limit. It is computed from the standard deviation and depends on how accurately we know s, the sample standard deviation, compared to o, the population standard deviation.In the absence of bias and assuming s, the sample standard deviation, is a good estimate of o, the population standard deviation,The confidence limit (CL) for a single measurement can be expressed by CL = x+ zo The confidence limit for N measurements can be expressed by CL = x + zo N square rootIn determining the confidence levels, expressed as a percent, we can examine the normal bell curve. These concepts also apply when o is not known, and we are dealing with a very small sample.ExampleExample 7-5 page 173a) 90% Confidence Limit8.53 ug Cu/mL + 1.64 X 0.32 ug Cu/mL / square root of 1 = 8.53 + 0.53ug Cu/mL99% Confidence Limit 8.53 ug Cu/mL + 2.58 X 0.32 ug Cu/mL / square root of 1 = 8.53 + 0.80ug Cu/mL b) 90% Confidence Limit 8.53 ug Cu/mL + 1.64 X 0.32 ug Cu/mL / square root of 4 = 8.53 + 0.26ug Cu/mL 99% Confidence Limit 8.53 ug Cu/mL + 2.58 X 0.32 ug Cu/mL / square root of 4 = 8.53 + 0.40ug Cu/mL c) 90% Confidence Limit 8.53 ug Cu/mL + 1.64 X 0.32 ug Cu/mL / square root of 16 = 8.53 + 0.13ug Cu/mL 99% Confidence Limit 8.53 ug Cu/mL + 2.58 X 0.32 ug Cu/mL / square root of 16 = 8.53 + 0.21ug Cu/mLHypothesis TestingNull Hypothesis – A claim that a characteristic of a single population is equal to some specified value or that two or more population characteristics are identical.If the observed difference is greater than or equal to 5%, then the null hypothesis is questionable.Detecting Gross Errors•Outliers – data points that differ excessively from the mean in a data set.Retention or Rejection of Outliers•Q Test – the absolute value of the difference between the questionable result and its nearest value is divided by the spread of the entire set.Q exp = xq -xn wQ exp is then compared with rejection values Q critExampleExample 7-17 page 174 a) Q exp = | 85.10 – 84.70| = 0.83 Q crit = 0.970 85.10 – 84.62 Q exp < Q crit RETAIN b) b) Q exp = | 85.10 – 84.70| = 0.83 Q crit = 0.829 85.10 – 84.62 Q exp > Q crit REJECTDealing with Outliers in a Small Set of DataReexamine all data relating to the outlying results to see if gross error could have affected its value. This demands a properly kept lab notebook containing careful notations of all observations.If possible, estimate the precision that can be reasonable expected from the procedure to be sure that the outlying result is actually questionable.Repeat the analysis if sufficient sample and time are available.If more data cannot be obtained, apply the Q test to the existing set of data to see if the doubtful result should be retained or rejected on statistical ground.If the Q test indicates retention, consider reporting the median of the set rather than the mean. The median has the greater virtue of allowing inclusion of all data in a set without undue influence from an outlying value. In addition, the median of a normally distributed set containing three measurements provides a better estimate of the correct value than the mean of the set after the outlying value has been