Chapter 1 Flashcards
data
Collection of observations
population
everything/everyone being studied
census
collection of data from an entire population
sample
sub-collection of data from a population
context
data being analyzed MUST have a goal
source
source must be unbiased and trustworthy
sampling method
sample must be representative of the entire population
analyze
graph and explore; visually represent data
conclude
find statistical significance
practical significance
whether a result is practical, despite its statistical significance
[PITFALL] misleading conclusion
use FDR’s so the results are clear
[PITFALL] small samples
sample must account for all possibilities
[PITFALL] loaded questions
worded to lead the participant to a desired result
[PITFALL] order of questions
may lead to participant’s answer being bias (e.g. “cats or dogs?” vs “dogs or cats?”
[PITFALL] non-response
participant refuses to answer
[PITFALL] missing data
data necessary to represent the ENTIRE population is missing
[PITFALL] precise numbers
precise numbers used as an estimate can be misleading
parameter
measurement computed using data from a population
statistic
measurement computed using data from a sample
quantitative data
numbers
qualitative data
categories/qualities
discreet data
quantitative data found by counting (e.g. students in a classroom, pages in a book)
continuous data
quantitative date found by measurements (e.g. rainfall, temperature, weight)
nominal level of measurement
qualitative data with no order
ordinal level of measurement
qualitative data with order
interval level of measurement
quantitative data with no significant zero (e.g. Fahrenheit scale)
ratio level of measurement
quantitative data with a significant zero (e.g. miles traveled)
lurking variable
impacts the result of a study, but is not included in the study
simple random sample
allows the entire population the same chance at being selected
systematic sampling
every kth item is selected after 0 (e.g. every 3rd person, every 100th subject)
convenience sampling
easy to select (e.g. family members)
stratified sampling
population is categorized into at least 2 groups, then a sample of each group is selected
cluster sampling
population is divided into groups, and one of the groups is randomly selected to participate
multistage sample design
multiple sampling techniques used
cross-sectional study
data collected during a specific point in time (e.g. Fall semester students)
retrospective study
data collected from the past
longitudinal study
data actively being collected into the future (e.g. the height of a child as they grow)
randomization
subjects assigned to either a treatment or control group by random methods
replication
repeating an experiment on more than one subject in order to control external variables
blinding
participant is unaware wether they are receiving treatment or a placebo
double blind
both the participant AND the researcher are unaware of who has treatment vs. a placebo
placebo effect
seeing results in a study because the individuals are aware of what they are being studied for
non-sampling error
typically caused by human error when working with data
non-random sampling error
sample was selected systematically
