Salary Report Documentation

Information Technology Job

Information Technology

United States Location

United States of America
Metropolitan Areas

Competitive Position® Salary Report

Source Documentation

Web Site Want Ads

The Salary Report is based on a sample of Want Ads listed in the career web sites:

  • Craigslist
  • USA Jobs.

Duplicate Want ads found in the same month were eliminated.

Please Note: The career web sites only act as venues for job postings and are not responsible for the content of the want ads. They are not associated with and do not endorse this Salary Report.

Information Technology Professional


Collected Past 3 Years

The sample was collected in the 158 weeks between:

  • Tuesday April 1, 2008
  • Thursday March 31, 2011.
Large Number of Want Ads

235,987 Want Ads were collected.

Each want ad listed:

  • Salary
  • Required Experience
  • Qualifications for an IT Job
  • Location in a US City.

Statistical Documentation

Information Technology Professional


Regression Analysis

The regression equation of the average salary is derived.

The regression equation minimizes the variance of the salaries across the sample of want ads.

The Regression Equation

Salary for the United States of America is lowest at entry level, increases rapidly with the first years of experience and approaches a ceiling as experience matures.

$34,742 Entry Level Salary average
+ ( $27,534 ∗ ln(Number of Years of Experience) )
= Salary Average.

Note: Since the natural logarithm, 'ln', is not defined at zero, +1 is always added to the number of years of required experience.

Collinearity Tested

The variables of the regression equation are determined to be independent by a 95% one-tailed t-distribution test.


Many want ads state a salary that is either greater or less than the Salary Average.

A residual is equal to the difference between the salary offered in a want ad and the salary as calculated by the regression equation for the want ad.

The variance is the sum of the squared residuals for the entire sample of want ads.

R Squared Statistic

The R Squared statistic is a measure of the 'goodness of fit' of the regression equation.

It states the percent of the sum of the squared salaries in the sample of want ads calculated by the regression equation:

  • 16.65% = R Squared.

The remaining percentage is explained by the variance:

  • 83.35% = Sum of Squared Residuals.

An R Squared statistic of 100% would indicate that all want ads offered the average salary. A reasonable degree of variability should be expected due to the many factors influencing individual want ads.

t-Distribution Statistical Tests

The t-Distribution is applied to test if a variable within the salary regression equation is equal to zero.

A variable can be insignificant if its standard deviation is too large.

The t-Distribution multilpied by a variable's standard deviation determines the 95% Confidence Interval and the probability the variable is equal to zero salary.

Significant confidence is placed in a regression equation variable when the low point of the 95% Confidence Interval is above zero. Even more confidence is placed when there is little probability that the variable is equal to zero salary.

1.96 is the factor of the t-Distribution where only 2.5% of the sample of 235,987 want ads have higher values.

The Salary Average, Standard Deviation, 95% Probability Range and Probability of zero salary for each variable:

Entry Level Salary Average = $34,742
Standard Deviation = $220
  • 95% Probability Range = $34,311 to $35,173
  • there is less than a one-hundredth of one percent ( < 00.01 % ) probability that the Entry Level salary is equal to zero
Experience = $27,534 ∗ ln(Number of Years of Experience + 1)
Standard Deviation = $128
  • 95% Probability Range = $27,284 to $27,784
  • there is less than a one-hundredth of one percent ( < 00.01 % ) probability that Experience determines zero salary.
F-Distribution Statistical Test

The F-Distribution probability considers whether the Salary Average regression equation is statistically equivalent to an equation set to zero.

The regression equation can be insignificant if its standard deviation is too large.

The lower the F-Distribution probability the more confidence is given to the regression equation:

  • The United States of America Salary Average equation has less than a one-hundredth of one percent ( < 00.01 % ) probability that it is equal to zero.
Heteroscedasticity Correction

The Salary Average regression equation required a correction for Heteroscedasticity.

The residuals are not uniform for all job characteristics:

  • the residuals are larger with the qualifications of an Information Technology Director

This additional information was factored into the analysis by dividing each want ad by its level of variance found in the heteroscedasticity regression equation:

  • (e^(4.7474 + 1.5344DirectorPosition))^.5.

The heteroscedasticity regression equation is verified to have an F-Distribution probability of less than 1 tenth of 1 percent chance ( < 00.10 % ) of not existing.

Each job characteristic of the heteroscedasticity regression equation is verified to have a t-Distribution probability of less than a 1 percent chance of not existing.

Standard Deviation

The Standard Deviation is the average residual found in a want ad.

$23,588 = Standard Deviation.

Salary Range

The Salary Ranges are calculated by adding ±(Standard Deviation ∗ t-Distribution Statistic) to the Salary Average.

The Salary Range factors are:

$23,588 = Standard Deviation
t-Distribution Statistics for the 235,987 want ads of the sample =
0.1257 for 10% Salary Range
0.3186 for 25% Salary Range
0.4307 for 33% Salary Range
0.6745 for 50% Salary Range
0.9674 for 67% Salary Range
1.2816 for 80% Salary Range
1.6449 for 90% Salary Range
1.96 for 95% Salary Range

The 95% range of Experience is calculated by adding ±(Standard Deviation ∗ t-Distribution Statistic) to the Experience Average:

  • 2 Years and 1 Month = Standard Deviation of Experience
  • 1.96 = t-Distribution statistic for 95% Experience Range.
= and