A Summary of OSU Extension Internal Salary Study

Introduction and Definitions

In Summer 1996, the Ohio State University (OSU) Extension Administrative Cabinet appointed a committee to examine the compensation system of OSU Extension county agents and district specialists. The committee was asked to examine the data for any differences between different demographic groups and report findings with recommendations to the cabinet. The committee was also asked to determine if other issues exist within the current compensation system, report those findings, and make recommendations.

Given limited resources, compensation of employees continues to be a difficult issue. Equitable distribution of available resources is very important to maintaining a highly motivated workforce. OSU Extension has examined its salary structure in the past to see if systematic differences exist between demographic groups. This study was an attempt to continue that process as well as to identify how existing agents' salaries compare to newly hired agents.

The population for this study consisted of all county and district administrative and professional Extension agents and faculty Extension agents and specialists employed as of July 1, 1996. The study included all agents and was considered a census. The Frame was established from personnel data maintained by Extension Administration. The subjects were examined separately in two groups: (a) administrative and professional agents, and (b) faculty agents and specialists.

Data for each of the variables examined were obtained from personnel data maintained by Extension administration. The data were checked for errors and corrections made by the district directors. Two methods were utilized to examine the data. Those methods were Formula and Regression.

The Formula method was utilized to calculate the salaries of newly hired agents and to determine if current agents with satisfactory or above performance would be receiving a higher or lower salary than they would if they entered the organization today with the same education and experience levels. The Regression method utilized a multiple linear regression model to determine the predicted salary of each employee based on the compensation of their peers. The nature of the Regression method places 50% of salaries above the predicted amount and 50% below the predicted amount. Findings from the Regression method were used to look for systematic compensation differences and not for making recommendations to adjust the salaries of individuals.

Definition of Variables

Variables studied included performance, years of experience, base salary, leadership positions, education, title or rank, current salary, gender, race, program area, and district.

Performance scores were assigned by district directors and ranged from 1-to-4 each year and was based on the individuals' 1993, 1994, and 1995 evaluations. A score of 2 was considered to be acceptable performance.

Years of service was a credit value for experience. This value included experience with OSU Extension and any previous experience credit given at the employee's date of hire.

Leadership positions studied were district specialist and county chair. Credit was given for county chair and district specialist assignment.

Title or rank used was Agent 1,2, 3 and 4 for administrative and professional agents and Instructor, Assistant Professor, Associate Professor and Professor for faculty agents and district specialists. Credit for promotions was given to agents above the Agent 1 level for administrative and professional agents and above the rank of Instructor for faculty agents.

Salary was the actual salary of faculty and administrative and professional (A&P) agents. All salaries were adjusted to 100% full time equivalent (FTE).

Gender was gathered from existing data. Data concerning race were also gathered from existing data and was considered Asian, Black/African American, White, unspecified or missing.

Program area was obtained from existing data. The program areas considered in the study were Agriculture and Natural Resources (ANR), Community Development (CD), Family and Consumer Sciences (FCS) and 4-H Youth Development (4-H).

Analysis of Data using the Formula Method

The predicted salary of agents using the formula method was used to compare the salaries of currently employed agents to the formula that would be used if they were to be hired into the system today. The formula for A&P agents was Base Salary + Education Credit + Service Credit + Agent 2 + Agent 3 + County Chair = Formula predicted salary. The formula for faculty agents was faculty base salary + Education Credit + District Specialist + Assistant Professor + Associate Professor + Professor + County Chair = Formula Predicted Faculty Salary.

Current salaries were then compared to the Formula predicted salary. The committee hypothesized that only agents with below acceptable performance would have current actual salaries below their Formula predicted salary. However, the committee did find that a significant number of agents were below their predicted salary based on the formula used to determine starting salaries for new employees.

Analysis of Data using the Regression Method

Alpha levels for significance were established at the .05 level for all tests. Due to the high level of multicolinearity between years of service and title or rank the decision was made to delete title or rank from the regression analysis.

A one-way Analysis of Variance (ANOVA) was conducted to determine if salary differences existed between districts on the variables salary, years of experience and performance score for either A&P or faculty. ANOVA was used in this census with nominal variables to look at the variables in isolation in order to reduce the number of dummy coded variables. No significant difference was found between any of the districts. The committee concluded that the mean salary differences between districts could be directly linked to years of experience.

A multiple linear regression analysis was used to determine the best predictors of salary from the variables studied.

Administrative and professional agents

Seventy-seven percent of the variance in salary could be explained by the linear combination of the variables Gender, Race, Performance, Years of Experience, Leadership Position, Education, and Program Area. The best predictors of salary were found to be Years of Experience (51%), Education (20%), Leadership Position (2.5%), and Performance(2.7%). The variables Gender, Race and Program Area were not significant predictors of A&P agents' salary.

Faculty agents and specialists

Sixty-eight percent of the variance in Faculty agents and District Specialists Salary could be explained by the linear combination of the variables Gender, Race, Performance, Years of Experience, Leadership Position, Education, and Program Area. The best predictors were found to be Years of Experience (50%), Leadership Position (12%), Performance (5%), and Gender (1.4%). Program area and race were not significant predictors of faculty salary. When the program area Agriculture was added to the regression model, the gender variable was no longer significant. A t-test was conducted to determine if there was a significant difference between agents with ANR program assignments and agents with other program assignments. There was a significant difference between the salary of ANR agents and agents with other program assignments. However there was also a significant difference based on a t-test in the number of years of experience between ANR agents and agents with other program assignments.

In an attempt to examine why gender was a significant predictor, t-tests were performed on the variables Performance, Years of Experience, and Salary with Gender to determine if a significant difference existed between the means of each variable for males and females. No significant difference was found between males and females on Performance, or Years of Experience. There was a significant difference between males' and females' salaries.

A Contingency Table was used to determine if there were differences in the distribution of leadership positions (county chair or district specialist) based on gender. Sixty-two of the 96 individuals in leadership positions were male. A Chi Square examining differences between groups was significant (p<.05). A contingency table was also used to examine if there was a difference in the distribution of male and female faculty with county chair assignments. Forty-four of the 68 faculty chairs were male indicating that males were over represented as county chairs. A Chi square indicated there was a significant difference in the number of males who were chairs and the number of female faculty who were chairs.

The difference in gender is thought to be caused by a difference in the number of males in leadership positions as district specialist or chair and higher compensation for agents in the ANR program area as opposed to the 4-H and FCS program areas. Much of the variance in salary is explained jointly by Gender and Program Area.

Conclusions and Recommendations

The current compensation system used by Ohio State University Extension rewards experience more heavily than any of the other variables studied. Leadership positions among faculty and A&P agents is an important predictor of salary. Whether or not an A&P agent has a master's degree is an important factor influencing their salary. Performance is also an important predictor of salary among both Faculty and A&P agents, but to a lesser extent than experience, leadership positions, or education level.

Gender was not a significant predictor of salary among A&P agents, but was a significant factor among faculty agents. The committee believes that this is due in part to two factors, leadership position, and the ANR program area. An examination of salaries among 4-H agents revealed no significant difference between the salaries of males and females. This, and the fact that when the ANR program area was controlled in the Regression, the gender variable was no longer significant led the committee to conclude that there was no systematic salary discrimination of females.

Among faculty, the leadership positions are heavily male. This is due in part to the number of district specialists in the ANR program but also the number of male county chairs. The ANR program area is dominated by males and is believed to be influenced by market factors affecting starting salaries of qualified ANR agents salaries, and the fact that several individual ANR agents and specialist are significantly above their predicted salary level.

Race was not a significant predictor of salary among either the faculty or A&P agents. Therefore, no systematic salary discrimination is taking place based on race.

Recommendations

The committee challenged OSU Extension's Administrative Cabinet to reexamine its compensation philosophy and adopt a system that recognizes market values or make adjustments to the salaries of agents to ensure equity. However, we recommend that agents be examined separately from district specialists before any adjustments be made based on program or gender.

The committee recommended adjusting the salaries of all agents who are acceptable performers and who are below their predicted salary based on the Formula method to their predicted salary. Additional adjustments may also need to be made to the salaries of those agents who have exhibited performance well above acceptable. The committee examined the effects to the regression equation that would be caused by adjusting the salaries of all A&P and faculty agents that were below the predicted salary based on the Formula method. The variables that were significant predictors of salary were affected only in their percentage of variance explained and not which variables were significant.

Salary adjustments were made in December 1997. The adjustments were made within University policies and affected 131 agents and district specialists. Over $200,000 was added to the base salary of these agents and specialists.

The organization should examine the issues surrounding county chair assignments to ensure an equitable division among males and females. The committee believes there is no systematic discrimination and chair assignments are more a factor of staffing of individual counties, but cannot make this conclusion without further study.

A review of the compensation of OSU Extension employees compared to other state Extension organizations and other employers who compete with OSU Extension for agents should be conducted. The results of this study should provide a better understanding of how OSU Extension's compensation compares to its workforce competition and how it should be positioned to successfully compete in the future.