Frequency Distributions

DOE-HDBK-1078-94

6.2.1 Frequency Distributions. Frequency distributions should be used for

organizing, summarizing, and displaying data. They can be constructed using

simple counting, averaging, and graphing procedures that show how often

particular events have occurred. They are normally used as the first step in

analyzing responses to surveys and trainee progress test results.

After all data from the indicators is collected, responses are tabulated. Totals

are then entered into the corresponding spaces on a blank survey. The average

(mean) response for each item is calculated and displayed on a bar chart. The

bar chart presents survey information in a simple visual form. It highlights high

and low values and permits easy comparison with acceptable performance

standards or sets of previous data.

6.2.2 Exception Analysis. Exception analysis is used for reviewing data to detect

unacceptable variations from a predefined standard. Facility operating,

maintenance, and industrial safety experience should be analyzed using this

method. Increases in the frequency of accidents, injuries, personnel errors,

rework or unscheduled maintenance, or increases in overtime above normal

levels may indicate a need to provide additional training or improve existing

training. Acceptable levels should be established for each of these parameters

as criteria for comparison. If any observed value deviates from the criteria, the

cause should be investigated.

Feedback from employees and employee exams, supervisors, and instructor

and trainee critiques should be analyzed to indicate if any training problem

needs to be investigated.

6.2.3 Content Analysis. Content analysis depends primarily on the expertise and

professional judgment of the individuals performing it. Content analysis should

be considered for use with all types of information and may be used in

conjunction with exception analysis. Interview responses should be analyzed

using content analysis. The following guidelines should be used when

performing content analysis:

Look for agreement. If respondents provide the same or similar

answers, these answers are more likely to be valid;

Do not disregard responses. Do not attempt to "second-guess"

employees, supervisors, or subject matter experts. If a response

appears erroneous or exaggerated, follow up with observations and

additional discussion; and

Focus the analysis on discovering specific tasks or subject areas in

which training refinements seem necessary.