# 课程代写：家庭支出数据

The analysis uses the household expenditure data for the year 1996 as taken from ODESI. The data takes a total of 10203 observations across 10 provinces of Canada. Table 1 shows all the variables used in the analysis and their definitions. The model identifies sixteen relevant variables that are likely to impact or determine the expenses incurred by the households in Canada for recreational purposes. Table 2 shows the descriptive statistics of all the variables.The dependent variable used here is the expenses on recreational and related activities. These variables include expenses in recreational vehicles, visual recreations like stage plays, audio recreations like music systems, other recreational systems like newspaper and magazines and so on. Figure 1 shows the distribution of the amount spent on recreational expenses. To ensure that the dependent variable is smooth and there are fewer outliers, log of the recreational expenditure is used instead of using the variable directly. The average of log of recreational expenditure is 7.996 with a standard deviation of 1.502. It is evident that the data is smooth and that there are no outliers.The income of the household is obtained by taking a sum of the income after tax of the household head and his or her spouse.

Those observations which corresponded to negative income have been filtered out as negative income makes no economic or econometrical sense. The mean income of the households is CAD 37094.805. To ensure scalability and to make the data smoother, log of income is used instead of income. The average of the log of income is 10.296 and the standard deviation is 0.984.These variables include “sex”, “educ”, “marstat” and “social assistance”. The variable on gender “sex” takes a value of 1 for male respondents. The data contains 45.9% males. The variable “educ” takes numeric values ranging from 0 to 16 depending on the number of years of education taken by the respondents. The average education of the households is 12.851 years. The variable age takes the mean of the age group the respondents belong to. The average age of the respondents is 48.722 years. The marital status of the respondent is covered in the variable “marstat” which takes a value of 1 for married respondents. 61.7% respondents are married. The variable “social assistance” is also a dummy variable. It takes a value of 1 if at least one member in the family receives social assistance, else it becomes 0. 12% of the households receive social assistance.