topic 3 Flashcards
Units of observation
are the specific entities or subjects from which data is collected and observed in a study
> Units we get data from
Units of analysis
are the entities or levels at which data is analyzed and conclusions are drawn > Units we are interested in
ecological fallacy
is an error in statistical reasoning where conclusions about individuals are made based on group-level data. It occurs when one assumes that a relationship observed at the group or population level holds true for individual members within that group. This assumption can lead to inaccurate or misleading conclusions about individuals.
Theoretical variables
Theoretical variables are concepts or ideas that researchers study and seek to measure in their research (that describe the data).
what should the theoretical variables be
complete and mutually exclusive
Complete
(exhaustive) means that data point or observation can be categorized or described using one and only one category of the variable (variable applies to a unit)
Mutually exclusive
means that the attributes or values of the variable don’t overlap or share characteristics. Each observation falls into one category (unit) without simultaneously belonging to another.
Attributes and Values
The words ‘values’ and ‘attributes’ refer to the same thing, however
- ‘values’ often refer to ‘numerical’ attributes (age or weight)
- ‘attributes’ often refer to ‘non-numerical’ attributes (colors, religions)
Five levels of measurement of variables
dichotomy, nominal, ordinal, interval, ratio
dichotomy
If a variable has two attributes only
- E.g. Male/Female; Correct/Wrong, age (two age groups, for example over 50 and under 50)
Nominal
A level of measurement describing a variable that has attributes that are merely different but are
not ordered
- E.g. Martial status
Ordinal
A level of measurement describing a variable with attributes we can logically rank-order along some
dimension but with unknown intervals
- E.g. Rankings based on raw scores, rankings in percentile range, age (age in categories, where you
don’t know the difference (age difference) between someone in group 1 and someone in group 3)
Interval
A level of measurement describing a variable whose attributes are rank-ordered and with known
intervals (the distance)
- E.g. Temperature, length, age
Ratio
A level of measurement describing a variable whose attributes are rank-ordered and with known
intervals (the distance) AND a meaningful zero point
- E.g. Height, Income, age (Trick: you can say ‘twice as old’: 30 is twice as old as 15) etc.
Categorical
measurements involve data that can be sorted into distinct categories or groups, like colors or types of cars. These data have no inherent numeric value.
