Week 2: Observation and Measurement Flashcards
two criteria for a good variable
- exhaustive: all possible answers/values
- mutually exclusive: you cannot have two answers
measurement levels: 1. nominal
nominal is the lowest level, we can only name categories. values are categories that have arbitrary (non-numerical) values. we could also just assign letters (‘a’, ‘b’)
measurement levels: 2. ordinal
values are ordered categories, but steps between categories are not necessarily equal.
1(very bad) - 5(very good)
measurement levels: 3. interval
values are ordered and steps between values are equal but there is not an absolute zero (e.g., degree c)
measurement levels: 4. ratio
values are ordered and steps between values are equal and there is an absolute 0 point; we can say one thing is twice as much as the other
reaction time
measurement levels: 5. absolute
unit is fixed, only one possible unit
percentage
discrete variable
can only take on an integer value
intelligibility: 1, 2, 3, 4…
continuous variable
can take on infinite number
reaction time: 29, 29.95, 29.99995…
typically metric/scale/quantitative variables are continuous
pros and cons of questionnaire
+ cheaper
+ anonymity
+ not usable for children, illiterates
pros can cons of interview
+ verify whether question was understood
+ possible to ask follow up question
- more expensive, need to train interviewers
self-report: response bias
- social desirability
- acquiescence response style / Nay saying
disadvantage archival data
no control over data by researcher
simple random sample
every study object has same chance of being selected
systematic sample
select every 20th student that walks through the hall
stratified random sample
take random subset of each stratum (males vs. females)