Having had an amateur interest in the operation of the stock market for a long time, I chose to perform my study on the relationship between US tech companies (which seem to have had a constant rate of growth/increase over the past two decades) against the composite score of the New York Stock Exchange, which includes corporations from many more industries, to see while well the performance of the business economy on the whole, represented by the composite, compared to the performance of specific, notable tech companies during the same time. During this time, the significant regression of 2009 would become one of my primary testing as this would most likely show exactly what I wanted: if the composite of the NYSE decreased significantly, would technology companies mirror this trend, or remain unaffected due to previous consistency in their growth patterns?

Figure 1.

Figure 1.

All quantitative data, the historical values of closing stock prices were collected in the period of Jan 1, 2008 to Dec 1, 2011, was gathered via the historical values tool of the Yahoo Finance search engine. Total data gathered was 48 data points for the 10 companies which gives a total of about 480 values, not including dates. This data was arranged in an excel document for the following procedure.

PROC IMPORT 
DATAFILE="GoleyS_FP"
OUT=FPlib
DBMS=xlsx 
  REPLACE;
  SHEET="Sheet1";
  GETNAMES=YES; 
RUN;

The means by which this data was brought into SAS was through the creation of a new library “FPlib” followed by the import of an excel sheet “GoleyS_FP” into that library. A second data set, “stock2” was created with a calculation of the percentage of change through use of a difference and lag function.

Figure 2.

Figure 2.

Interpreting the above early estimation, it does appear that trends of the market are universal as peaks such as Sep, 2008 or dips like Jan, 2010 are shared by all companies.

Figure 3.

Figure 3.

In statistical testing, the p-value of a test is what allows one to keep or reject the hypothesis of the test. As this was tested for in each variable, we can conclude whether each of these companies has a statistical significance on an individual basis simply by looking at the value of the “Pr > F” column in the Analysis of Variance table for each company. From this, we find that At&t and Verizon share an identical p-value of .0002 while all others fall in the range of >.0001. The standard of testing for statistics is generally accuracy to either 1 or 5 percent. Both of the values still significantly outside of that range and from this a decision can be made about the null hypothesis. In this case, the null is rejected for all variables which leave us with alternative hypothesis: “There is a direct correlation between the dependent and independent variables.” Consequently, we can safely say that there is enough evidence to conclude that the US technology companies have a significant linear correlation to the changes in the Composite Value of the New York Stock Exchange. When the general economy has ups and downs, technology companies are just as affected by these movements.