36-402/608 Homework #11 due 10:30AM 4/81. Dyads (60 points)You must use SAS for this problem! Use the DDFM=SATTERTH option.This problem is a study of income in married couples in Massachusetts. Use thiscode to load the data in dyads.dat.DATA dyads;INFILE "dyads.dat" FIRSTOBS=2;INPUT agentNum dyadNum Agender Acollege Aage Pgender Page Pcollege income;Aage30 = Aage-30;Page30 = Page-30;RUN;Use only the adjusted age variables to obtain a more interpretable intercept.The explanatory variables are gender (1=female, -1=male), college (indicator vari-able of college graduate), and age (in years). “A” indicates “agent” and “P” indi-cates partner. First do some EDA to get familiar with the study variables. Thenselect an appropriate dyadic random effect model. Justify the need for a randomper-dyad intercept. Don’t try any random slopes (or serial correlation). Start withthe largest possible fixed effects model without interactions. Remove unneeded fixedeffects.Turn in a brief description of what you learned about the study from the EDA, theBIC values you used in your model selection, the output and SAS code for the finalmodel including a residual plot and your interpretation of it, and a brief interpre-tation of the estimated parameters. Also state any one interaction that would beworth studying, and what a small p-value would indicate for that interaction.2. Schools (40 points)You must use SAS for this problem! Use the DDFM=SATTERTH option.File fundses.dat has data from a study of the relationship between a test score andschool socio-economic status and funding level. Schools were randomly selected, andone randomly selected 7th grade classroom was tested from each school. The vari-ables are school id number, school funding level, school average SES, and individualtest score in that order with one line per student. The funding is in thousands ofdollars per student. Adjust the funding variable so that the intercept will corre-spond to $10,000 per student. The per school SES variable is a z-score (mean 0,sd=1).Run the mixed model with a random per-school intercept and fixed effects for theother explanatory variables without interaction and using REML. (You don’t needto do EDA. Don’t do model selection. You don’t need to make residual plots.) Turnin the SAS code, the SAS output, and your interpretations of the