STAT 1000 Study Guide - Midterm Guide: Random Variable, Categorical Variable, Percentile
Document Summary
Sampling designs: srs, systematic, stratified (same type) groups. Cluster (random) groups, select some entire groups to test, multistage, census, volunteer/self select, quota sampl (a pre-set number from non overlapping groups but not random (gerrymander), snowball. Experiment: control, treatment, ethics, placebo, doubleblind, hawthorne effect (alter behavior b/c being watched), lack realism. Two-sample (2 indep groups), paired/matchpair: same subject before/after treat, or compare twins/sibs/ couples etc. graphs: bar+pie=cat, stem+histogram quant. Cumulative frequency distributions: number of cases falling at/ below a certain level. Cum percent: at or below point in distitrbuion. Add up % at interval and below it. Percentile rank: look at percent, at th (30%-30th), corresponding number is score at that percentile. Central tendency: mean: affected by outliers/add/subtract data point and if +/-/x// by constant do same to mean, +/- not change mean, x// does. Mode: ideal for skewed, but if can"t calculate, media next best. interval of ratio: mean, median, or mode (q).
