Stats 503 Terms with Complete Solutions

sampling error - ANSWER-variation in results from multiple replications of same study because of chance only way to avoid is a census every time you gather different groups of people get diff average even if same topic different between paramters true value and out estimate that results from chance model - ANSWER-focus on details that are important and ignore the rest sampling theory - ANSWER-organizing data collection unbiased, efficent study population - ANSWER-collection of every person study is applicable to like every human for height sample - ANSWER-subgroup of study population like cancer cells in a study population of all cells sample/study unit - ANSWER-individual members of population individual observation 1 specimen is a study unit distribution - ANSWER-connections of different values in a group describes the data what values the data can take and how often they occur associated - ANSWER-if knowing variable tells us about the other human height - men are statistically taller parameter - ANSWER-quantities that summarize a distribution average, proportion, variation estimate - ANSWER-single value of best guess standard error - ANSWER-our uncertainity biological model - ANSWER-physical and casucal relationships in a living system statistical model - ANSWER-equation to describe distribution and how it changes in response to other variables make inferences compare this to bio model to draw conclusions want to figure out what is rare and common care about: location - clutter vs spread out spread - variation around centerpoint shape - spread in more than 1 direction multiple peaks? to test a prediction (will data follow what we think or something else) - ANSWER-1.) collect data 2. bulid stats model to interpret data have to take into account sampling error (everything has variation) and determine how strong the relationship is data - ANSWER-info collected in a consistent way has to be representative of a pipliation random sample units equal chance of being selected independent of all other units check for bias variable - ANSWER-traits of sample units convience sampling - ANSWER-individual easy to observe issue if it effects study variable volunteer bias - ANSWER-differ from general population precise - ANSWER-estimates that have low sampling error are very precise repeadable and consistent increase precision you increase sampling size random sampling error dont cause bias but decrease precision measurement error - ANSWER-inaccuracy for single point or random calibration is not sampling error independent - ANSWER-sample units if knowing 1 doestn tell you the other and cant affect the probability of another family that goes hiking around independent because i know that about 1 i know it about all of them goal of data collection - ANSWER-decrease bias increase precision (low variability) random sample good sample size precise measurement random assigning of sample units to remove pattern response variable - ANSWER-measures an outcome of a study graph shows distribution of response variable explanatory variable - ANSWER-a variable that we think explains or causes changes in the response variable explanatory variable describes a condition and given this condition want to know the prob something will happen casual relationship - ANSWER-if manipulating explanatory variable alters distribution of response variable causation requires statistical association its like cause and effect observational - ANSWER-values of explanatory variable not randomly assigned can show assocations but not cause and effect cant randomize smoking so it is observational unless it was an animal study numeric - ANSWER-counted continuous (measure rate, ratio) classified into bins that variable into discrete categories the vairbale is measured in infinitely divisible units probability distribution of X is the prob density funtion follows rules of probability always use intervals discrete (counts) - any data that is measured in indivisible units counts or cateogires like flipping a conin counting number of times category came up if a category is counted its discrete its the count we look at can observe each value exactly categorical - ANSWER-classification nominal (no rank or order) ordinal (inherent rank/order: baby and adult) visualize absolute freuqency sample proportion - probability proportion - what is the fraction of sample unites do we expect to be in each category in future samples univariate - sample proportion tells us probaility a randomly selected member of popuation will belong in a specific category of response variable classify variable - ANSWER-numeric or categorical frequency distribution - ANSWER-what values of variables are possible number of times that each value occurs in the table more data increases approximation