回归方程中的数据转换问题请教

现在碰到如下一个案例, 在分步回归过程中看起来”composite index” and “warrants volume” 之间有很强的相关性(P<<0.05), 但是看散点图他们之间似乎一点关系都没.

原题如下:
The stock market database contains eight variables on the New York Stock Exchange. There are three observations per month for nine years yielding a total of 324 observations per variable. The variables are: Composite Index, Industrial Index, Transportation Index, Utility Index, Stock Volume, Reported Trades, Dollar Value, and Warrants Volume. Dollar value is reported in units of millions of dollars. Recognizing that time of the month may make a difference in the value of an observation, each variable contains an observation from on or near to the tenth of the month denoted in the database as 1 under the variable Part of the Month, an observation from on or near to the twentieth of the month denoted as 2, and an observation from on or near to the thirtieth of the month denoted as 3. This database was constructed from data displayed on the Internet by the New York Stock Exchange. The original data can be accessed by ww.nyse.com/public/market under the title "NYSE Statistics Archive."

Question:
Get scatter plots between the major random variables. Check if they meet the assumptions of linear regression model. If the assumptions of linear regression model do not meet, you may need to conduct some transformations for some random variables. Get the scatter plots for these new random variables again. Continue your work until you think these new random variables satisfy the assumptions of linear regression model. If yes, conduct the linear regression for these new transformed variables.

晕, 不能传excel文件....