SOCY 201 Study Guide 05 03 2012 given a visualization have to describe distribution pay attention to location variance mean median skewness dispersion Q1 Descriptive Statistics 7 points a duplicate from a earlier exam use numbers to punctuate words tell if its homo heterogenous a histogram normal distribution Q2 3 parts Confidence intervals and difference of means t test 12 points given a confidence interval and asked to interpret confidence interval 1 ask to interpret it in words understandable through words understandable for people who haven t taken stat 2 given a graph with 2 confidence intervals and asked to compare A WHISKERS WITH NO BOX o if overlapping then not if not overlapping then they are statistically significantly different o are those two means significantly different Interpret graphic o magnitude direction no calculation 3 calculate a difference of means t test o do a hypothesis test for the significance of difference between two means o find t critical o t should be stated o like exam 2 number 5 o Q3 Cross tabulation Chi squared 10 points USE SENTENCES FOR NULL REG HYPOTHESIS 2 discrete variables cross tabulation given percentage corrected chi2 is calculated gamma is calculate compare across columns p value 1 do a hypothesis test chi 2 describe relationship interpret gamma no calculations gamma gives us a sense of the magnitude of the relationship Q4 Anova 10 points 1 continuous 1 discrete variable conduct a hypothesis test find f critical TALK ABOUT P VALUE part two given a table with the means and standard devs for the groups given the anova table don t have to fill it in explain what it is and why its important o look for rachel s email look at what prof says in examples with p o the likelihood of getting a value o if p is less than alpha then reject null Q5 from last homework 2 check slides on correlation print them HW 5 4 Correlation problem 10 points given a scatterplot two continuous variables given correlation coefficient conduct a hypothesis test about the correlation coefficient significance of difference null hypothesis r 0 alternative r is not equal to 0 t distribution to find critical value t table confidence interval z t SE o Hypothesis test step 4 calculations be able to convert to Fisher s Z be able to put a confidence interval around the Fisher Z convert fisher z back to R p 471 convert both upper and lower bound does zero fall in your interval outside fail to reject Direction neg positive magnitiutde how strong association linear Describe relationship outliers Q6 from last homework 2 part c Regression question 6 points BRING THOSE SLIDESSSS print out bivariate regression pages given the correlation coefficient given the means of two variables standards devs beta hat sub 1 beta hat sub 0 t statistic associated with beta hat sub 1 write regression model interpret the relationship between these two variables based on the regression model direction magnitude statistical significance from slides Bivariate Regression An Example o correlation cofficient direction positive places with a high rate of crime crime tend to have high rates of the other crime etc magnitude range of possible outcomes 1 to 1 you can use gamma table to interpret this statistical significance small sample size error determined largely by the number of cases few cases bigger standard error harder to reject the null beta hat sub 1 look up formula o part B go back to the simplest case look at coefficient for every 1 unit increase in x there is an expected increase in Y of 203 Beta sub 1 regression coefficient is scaled to be measured in y Regression Example water o Means o Correlation coefficient 83 implies direction positive strong relationship o Beta hat sub1 Beta hat sub 0 o Part D Model With interpretation for every 1 percentage point change in acess to potable water there is an increase in life expectancy of 2383 years 3 months o Y Bo B1x e Hypothesis Diff of means mu1 mu2 Anova mu1 mu2 mu3 05 03 2012 05 03 2012