Chapter 1: Measurements Flashcards
Accuracy
The closeness of a test result or empirically derived value to an accepted reference value.
Bias
The difference between the expected and the experimental result.
Precision
The reproducibility of a series of replicate measurements obtained under comparable analytical conditions. Precision is often measured by %RSD.
Random Error
An inescapable error, small in magnitude and equally positive and negative, associated with any analytical result.
Systematic Error
Analytical errors that are the same every time (i.e., predictable) and that are not random.
Grubbs Test
Is the suspect data point too far outside the edges of the distribution to be a legitimate data point?
-Assumes the underlying data distribution is normal.
What is the formula? (mean-questioned value)/standard deviation
Type 1 error
Null hypothesis incorrectly rejected
Type 2 error
Null hypothesis incorrectly accepted
Hypothesis Testing
provides a statistical measure to make a decision (confidence level)
Pedigree
The systems, people, and procedures that stand behind forensic chemistry testing and results.
Metrology
The science of making measurements.
What is a good measurement?
-One that answers the forensic question at hand and is:
-Fit for purpose
-Reasonable
-Defensible
-Falsifiable
What is a forensic question?
It is the pertinent question.
-What did this person die from?
-Is this white powder a controlled substance?
What is a legal question?
It is a question dealing with the law and crime.
-Does the soil in this shoe link a person to a crime scene?
-Did that person commit a crime?
What does NUSAP mean?
Number
Unit
Spread
Assessment
Pedigree
Mean
The average
Standard Deviation
Average deviation from the mean
Sq. rt of (Xi-Xmean)^2/n-1
Where n is the sample number
Variance
CV or %RSD
standard dev/Xmean *100
Probability distribution factor
Confidence Interval
Student t distribution
Population vs sample statistics
Population is greater than 30
Sample applies if a subset are tested or there is any doubt about the subset being representative of the parent population