Explanatory- what causes/explains something Response-what changes as a resut of the explantory Positive assoation-as one variable increases so does the other if one decreases so does the other negaitve as one variable increases the other decrease Residual means do the equation with the predicted then subtract or add the residual Correlation can only be between -1 and 1 correlaction must be mesuring 2 numarical datas if the explan and respons are switched what would happen to least square reggresion line - would be inverted and look diffrent because x and y are swapped correlation coefficient- would be the same cause the points would still vary away from the new lsrl by the same ammount inser the number get the prediction then subtract the actual by the predivtd and the left over is residiual y = 2100 + 400x 2100 is y intercept linear model needs scattered random dots if there is pattern linear model is not apporpriate if there is no control group nothing can be concluded from the data 2237467049248 take the first digets and if its not in range skip and keep going until done completly randomized is totally random mathced pairs is when they do both options randmized block is when there blocked into groups and then randomized sampling methods stratisifed-making strata groups and taking sampels from each strata convencnie-taking whatever is easiest cluster-taking entire groups vote simple ranodm- totally random scatter plot linear,positive,strong relation ship slope - for each increase of ___ the ____ increases by ___ if its not on the chart yout using the least squares regression line r2 = ___ ___ of the variation of _____ can be explaine dby the LSRL s= ____ we would expecr the _________ to be off by ___ when using the LSRL to predict _______ insert number get predicted subtrahcted real number by predicted Real - predict leftover is residual this ____ was ___ lower/higher than expecrted linear is appropriate when random scatter statistical sig the ammount of ___ for this group was so large/small that it could not have be of random variation between subjects under coef is y-int an under the y-int is the slope y=y-int+slopeX power model if log of both exponential if 1