Levels of Data - Lecture 2 Flashcards
How many types of data are they? What are they called?
Two Types:
1. Continuous Data
2. Discontinuous Data (or Discrete Data)
Define Continuous Data
Data that represents positions along a continuum and can be broken down into smaller units of measure
Give examples of types of Continuous Data
millimetres, centimetres, kilometres, metres
Define Discontinuous Data (or Discrete Data)
- Data that has distinct values and are bound by the perimeters of the category
- cannot be broken down
- can be either present or absent
Give an example of a type of Continuous Data
categorical data like colour
What are the Four Levels of Data?
- Nominal
- Ordinal
- Interval
- Ratio
What is another name for the Four Levels of Data?
The Stevens’ Data Types
What is Nominal Data?
- means “in name only”
-measures discrete data only - used to identify observations
- CANNOT be ranked
- have to be mutually exclusive (meaning they can only occupy one of the categories)
- are exhaustive (meaning all data points have a category)
- the way the data is placed is NOT a ranking system and is random
Give examples of Nominal Data
colour and sex estimation since there is no greater or ranking system between different colours or genders
What does Mutually Exclusive mean in terms of Data?
that something can only fit into and occupy one of the categories
What does Exhaustive mean in terms of Data?
that all data points have a category
What is Ordinal Data?
- measures discrete data
- mutually exclusive and exhaustive categories
- ordered in a way that is logical to the data
- can be ranked
- the ranking is asymmetrical because it is going in one direction
- amount of change between categories CANNOT be (accurately) assessed
Give examples of Ordinal Data
Size: small, medium, large
Cranial Suture Closure: open, partial closure, significant closure, obliterated
- a ranked system but unclear how much of a difference is between the sizes
What is Interval Data?
- can be either discrete or continuous data
- equal and known difference between data points
- measurement between ranks has a standardized unit of measure
- lacks a true 0 point (meaning 0 does not mean an absence of something)
- amount of change between two variables CAN be assessed
Give Examples of Interval Data
- temperature since 0 degrees does not mean an absence of temperature
- time since death
What is the last bodily function to stop after death?
auditory function (hearing)
What is Ratio Data?
- continuous data
- equal distance between all variables
- has an absolute 0 (0 means absence of something)
- values can be compared with one another
Give examples of Ratio Data
money, length, weight, volume
What system is used to summarize data and why?
The Models of Central Tendency because they give a picture of the basic trends of the collected data
What three strategies are included in the Models of Central Tendency?
- Mode
- Median
- Mean
What is a Mode?
- the mode of data is the most frequently occurring score
- used for discontinuous data to see which categories has the most number of observations
- can be nominal, ordinal, interval, and ratio
- not influenced by the outliers (ie the extremes of the categories)
What Models of Central Tendency can Nominal Data fit into?
only the Mode
What is a Median?
- the median is the exact central point of the data
- can be ordinal, interval, and ratio
- true model of central tendency (50% of data is on one side and 50% is on the other)
- outliers can impact where the median is
What is a Mean?
- the average unit of data
- can be interval and ratio data
- significantly influenced by the outliers (especially in small sample sizes)
What Models of Central Tendency can Ordinal Data fit into?
the Mode and the Median
What Models of Central Tendency can Interval Data fit into?
The Mode, Median, and Mean
What Models of Central Tendency can Ratio Data fit into?
The Mode, Median, and Mean
In normal distribution are the mode, median, and mean the same, similar, or different to each other?
Either the same or similar
What defines a normal distribution?
where the average or mean is right in the centre, the median would also be at the centre with the exact same number of values on either side of it, and the mode (the frequently occurring value) would also be right in the centre
Why is a normal distribution important in statistics?
So confidence intervals can be calculated
What is a Standard Deviation?
a very specific unit of measure away from the mean
What is a Standard Error?
The same as a standard deviation which is a very specific unit of measure away from the mean
Can a Standard Deviation be used when it is not a normal distribution?
No
What percentage of the collected data or population does 1 SE/SD away from the mean account for?
68.26%
What percentage of the collected data or population does 2 SE/SD away from the mean account for?
95.26%
What percentage of the collected data or population does 3 SE/SD away from the mean account for?
99.73%
What is a population?
the number of samples used to create a very specific method
When using someone’s method and you want to increase the accuracy of that method what would you do regarding the SE/SD?
you would need to increase the SE/SD to be further away from the mean
How many Sources of Error are there? Name them.
- Random Error
- Systematic Error
- Negligent Error
What is Random Error?
- unknown error
- mistakes made by people
Give an example of Random Error
accidentally hitting the wrong number on a calculator or reading a measurement wrong
What is Systematic Error?
- a consistent bias or flaw in the tools being used
- easier to correct since it is easier to find what the error was and correct it
- flaws between observers
Give an example of Systematic Error
tools aren’t properly calibrated and are giving an incorrect measurement
What is Negligent Error?
- aka Observer Error
- doing the procedure wrong
How can a Negligent Error happen?
- either were trained incorrectly
- or using a tool you weren’t trained on
What are all the methods used for Reducing Error?
- specific to forensic anthropology
- measure something 3 times
- compare your answers with others - “data cleaning” (regularly checking through your data for mistakes)
- regular maintenance and calibration of equipment
- standardized training or instruction on the operation of equipment or execution of a method
- conducting studies to identify negligent error in measurement (intra/inter observer error)
- identifying “acceptable” levels of error
What is an acceptable level of error?
2-3%
What are the two ways to assess error?
- to see if the measurement is accurate (the correct answer)
- to see if the measurements are precise (the same answer)
How do you test for Inter-Observer Error?
- compare measurement results from multiple observers on the same specimen
- then calculate the difference between measurements from a known measurement and present as a percentage
What is Intra-Observer Error?
Error from the same person (ie. the error is caused by one person making mistakes)
What is Inter-Observer Error?
Error from the variation of accuracy of data recorded from different people
Why is it necessary to have all these rules about error and standardization?
it helps to convey a sense of confidence and reliability when a forensic anthropologist is giving evidence in court
What is the Daubert Case?
A ruling from 1992 that allowed for more cutting-edge methods to be admissible and for the judge to be the gate keeper of what is and isn’t admissible
What is the Daubert Criteria?
- certain guidelines that a judge can use to see if they want to allow a forensic anthropological method to be sustained in court
What are the guidelines that make up the Daubert Criteria?
- The technique has been or could be tested
- The technique has been through the peer review process and published
- The technique has a known error rate, or at least an error rate that can be determined
- The technique is standardized and able to implement reliability
- The technique is generally accepted within the relevant scientific community
What criteria do you need to meet to be considered an “Expert” witness?
- must have a level of knowledge that an ordinary person would not
- evidence presented must be necessary and relevant to help the judge understand what is being presented
What defines “Opinion?”
the interpretation of data or evidence
What are the 3 types of Opinion?
- Speculation: a statement based on little or no data
- Possible: offering an opinion on a characteristic or event occurring from unknown parameters
- Probable: Opinions based on known parameters. The highest level of certainty
