Stat
Terms for my stat midterm
Terms
undefined, object
copy deck
- p-value
- probability of getting a value of the test stat as large or larger than the one calculated in the sample
- central tendency / spread
- two types of univariate stat
- residual
- vertical distance between a point and a reqression line. squared so positives and negatives do not cancel one another out
- quota / snowball
- nonprobability sample
- intercept + slope coefficiant * x + residual
- bivariate equation
- one tailed
- a greater than b
- 99.7
- within 3 std. devs
- interval / ratio
- continuous data
- standard error
- measures the inaccuracy of a stat
- summarizing data / hypothesis testing
- two things stat can be used for
- std. deviation
- square root of variance
- inadequate control / measurement error / sampling error
- 3 problems with stat
- confidence interval
- % likely the range would contain the true value
- mean
- average
- declines with square root of sample size
- how does sample size effect standard error
- inferential
- draw conclusions from incomplete or error-ridden data
- nominal
- qualitative data
- correlation
- measure of strength of linear relationship between two variables
- ordinal / interval / ratio
- quantitative data
- ratio
- income
- 68
- within 1 std. dev
- median
- middle of data set
- 0
- no correlation
- alternative hypothesis
- what we are trying to prove
- null hypothesis
- what we are trying to disprove
- regression
- causation
- descriptive
- summarize data
- interquartile range
- between 25th and 75th percentile
- variance
- sum of squared absolute difference between mean and given values
- + or - 1.5 * 75th percentile - 25th percentile
- whiskers
- - 1 to 1
- range of correlations
- mean + or - (1.7*standard error)
- 90% confidence interval of mean
- boxplot
- 5 number summary
- ordinal
- hs rank
- interval
- temp
- probability sample
- known chance of being chosen
- two tailed
- a not equal to b
- nominal / ordinal / interval / ratio
- 4 types of data
- 95
- within 2 std. dev
- standard error
- standard deviation of that statistic across hypothetical repeated samples
- ordinal
- likert scale
- correlation
- no causation
- mean + or - (2.5*standard error)
- 99% confidence interval of mean
- as r squared increases, standard error decreases
- regression and standard error
- mean + or - (2*standard error)
- 95% confidence interval of mean