2006 AND 2007 data included!

Librarian Caveats

Caveats should be observed in reading the tables. The intent of the survey is to collect and present a statistically valid report of region- and state-level data for each position and library type. This was not possible with an overall response rate of thirty-one percent. Although we received at least one response from either a public or academic library in forty-nine states (49) and the District of Columbia, separating those responses by library class and region reduced the significance of individual library responses.

· Regional- and state-level salaries are reported in this survey for each position and by type of library, acknowledging that in most cases, especially for state-level data, there was not a statistically significant response rate. These data are not representative, so use caution in reviewing them or re-using them in any way. We look forward to increased participation in future editions (See Discussion).

· Very Large public libraries had the highest response rate at sixty-six percent. ARL member libraries had a fifty-nine percent response rate. Response was particularly low for Two-Year Colleges (21 percent) and Very Small Public Libraries (23 percent).

· The higher the number of cases (N), the more reliable the results of the sample in giving a true picture of the total population. For three regions, the response rate for Very Large public libraries (serving over 500,000) was sufficient: Great Lakes & Plains at 73 percent, N=8 libraries; Southeast at 56 percent, N=10; and West & Southwest at 80 percent, N=28. The response rate was above 40 percent for Large public libraries (serving 250,000-499,999) in the Great Lakes & Plains at 53 percent, N=35 and West & Southwest at 54 percent, N=61. Response rates are defined as the percentage of responses divided by the surveys sent by category. For example, eleven surveys were sent to Very Large public libraries in the Great Lakes & Plains region and eight responded (8/11 = 73 percent).

· Another caveat is that when the mean and the median are not close together, the mean is being influenced by some unusual values. When the mean is much higher than the median, there are several very high salaries. When the mean is much lower than the median, there are several very low salaries.