## The Concord Water Study

During g the 1970s the city of Concord, New Hampshire, experienced a growing demand for water, despite a roughly stable population. In late 1979 this rising demand, together with unusually dry weather, led to a shortage in water supply. In late summer of 1980, as the shortage worsened, the Concord Water Department and municipal officials began a campaign to persuade citizens to use less water. Over the next year, water use declined by about 15%, which was naturally interpreted as evidence of the conservation campaign’s success.

The 1981 Concord Water Study examined variation in water savings. Questionnaires went out to a random sample of Concord households, asking about demographic characteristics, opinions, and conservation behaviour. The questionnaires were then matched with Water Department records (from meter readings) of the amount of water each household had actually used during the summers of 1980 and 1981, before and after the conservation campaign.

The following is a description of the variables used in the data file:

**Variable Name Description of Variables:**

WATER81- Water use in summer 1981, in cubic feet ,

WATER80- Water use in summer 1980, in cubic feet,

INCOME- Household income, in thousands of dollars,

EDUCAT- Education of household head (in years),

RETIRE- Retirement, coded 1 if household head is retired and 0 otherwise,

PEOP81- Number of people living in household in summer 1981,

CPEOP- Change in the number of people, summer 1981 minus summer 1980.

The response variables is postshortage (1981) water use and the explanatory variables are the remaining six above variables. Start your analysis of the data with obtaining a scatterplot matrix and a correlation matrix. Then consider a regression model of postshortage (1981) water use on the six explanatory variables. Who tends to conserve more water?

Apply the variable selection techniques in SPSS to the data and compare the corresponding regression models. Moreover, use the appropraite diagnostic tools to check whether the regression assumptions are not seriously violated. Write a brief statistical report summarizing your findings.

Download the Data