Because "success" is not an objective term, quantifying it is not entirely intuitive. However, objective measures for success still exist. Measures such as salary, position in the corporate hierarchy, speed and frequency of promotion and discrepancy in salary, when compared to coworkers of equal level, are all objective or can be written numerically, and relate to an individual's success.
Surveys, while not objective, display important information about employees. Researchers use surveys to gain subjective estimates of an individual's success in a company as well as understand how an individual's supervisors evaluate him or her. The survey has the advantage of being able to be formed in an open-ended manner so that information hard to objectify may be accounted for. Researchers who wish to use analytical means to examine the data can later quantify survey data by giving rankings or scores to specific comments. Alternatively, researchers can use Likert scales -- scales that use numbers to represent feelings -- in the surveys.
Mathematical models are suitable for researching processes where there is a clear beginning and end, such as climbing the corporate ladder. The advantage of a mathematical model is that it can include several variables, allowing the researcher to input many measurements. The researcher can create separate models for men and women. After creation, a model is analyzed through mathematical means to determine the overall success of a gender. At completion, the researcher can compare the results between the models.
Statistical hypothesis testing is a scientific application of statistics that helps researchers answer well-formulated questions. The process begins with a hypothesis. In the case of the research at hand, there are three possible hypotheses to choose: men and women do not differ in successes, men are more successful or women are more successful. After choosing a hypothesis, the researcher then collects data on the subjects. Depending on the type of data, statistical processes such as ANOVA, F-tests or paired t-tests play the roles of determining whether the original hypothesis is valid or should be rejected.