5: Levels of Measurement and Distributions Flashcards
What are the types of variables?
Constant, categorical, numerical, continuous, discrete.
What is a constant variable?
Any characterisitc that is observed only take on one, single value.
What is a categorical variable?
Qualitative in value. Describes categories of a characterisitc, e.g. party affiliation. Special case - dichotomous variable.
What is a numerical variable?
Quatitative variable. Further classification - continuous vs discrete.
What is a continuous variable?
Can take any value in a specific range. Varies between a smaller and larger number. Examples - GDP, age, weight, etc.
What is a discrete variable?
Can only take on certain values, many are not possible. Arises from a counting process. Examples - no. of children in a family, etc.
What are the levels of measurement?
Nominal scales, ordinal scales, interval scales, ratio scales.
What are nominal scales?
Categorical scales or qualitative scales. Assign people or objects into qualitatively diff categories. Assumes that all items in a category are equal with respect to that category. No intermittent categories. Examples - type of student.
What are the descriptives for nominal scales?
Mean - no. Median - no, requires ordering. Mode - yes, measure for mode for a particular religion, for example.
What are ordinal scales?
Rank people or objects on some variable. Requires classification. Requires ranking. Equal diff between the ranks does not imply equal distance between characterisitcs.
What are the descriptives for ordinal scales?
Mean - no (understanding as categorical variable, most often the case), yes (understanding as numerical variable). Median - yes. Mode - yes.
What are interval scales?
Most commonly used score in stats. Gives info about ranking of people or objects. Provides info on how far apart people or objects are on that variable. Examples - temp, year, etc.
What do interval scales assume?
Equal diff between all points on the score.
What are the descriptives for interval scales?
Mean - yes. Median - yes. Mode - yes.
What are ratio scales?
Most commonly used score in stats. Gives info about ranking of people or objects. Provides info on how far apart people or objects are on that variable. Has a natural point of 0 - indicates the total abscence of a charactertisitc. Treated a numerical, can be discrete or continuous. Examples - height, weight, age, blood pressure, etc.
