Experiments of others
One hundred to three hundred library books are selected from the collection, usually a portion of a collection: fiction, non-fiction, biographies, books that have not circulated in four years, etc.
The selected books are randomly divided into:
- A “control group” of books which remains on the shelves throughout the experiment. The control is a baseline to clearly reveal differences the experimental treatment makes on circulation.
- An “experimental group” of books which receives a treatment that, if is works, increases circulation. Some experiments will create two or more experimental groups.
Experiments begin with a three to six month “pre-test” measuring the circulation of both groups while the books are still on the shelves. This establishes a baseline to clearly see the increase in circulation caused by the experimental treatment.
After the pre-test, the experimental group receives the “treatment” intended to increase circulation. “Experimental treatments” in this document are putting books on display, orienting them face-out on shelves, putting red dots on their spines, listing their titles on booklists, etc.
The test period for three to six months measures the circulation of the experimental and control groups.
Researchers’ reports show the results in tables.
| Experimental group pre-test | Experimental group test |
| Control group pre-test | Control group test |
The first column shows the circulation of the three to six month pre-test. Since the books in both groups are still on the shelves, there is no difference between the groups. The circulation of both groups should be very close.
The second column shows the circulation for the test period. Since the control group remained on the shelves, the circulation of the test control group should be similar to both pre-test groups. If the experimental treatment worked, then circulation of the experimental group should be much larger.
Some experiments have two experimental groups. These experiments also have two test periods. One experiment below displayed books in a prime display location near the circulation desk and in a second non-prime location behind the fiction stacks at the back of the library. The second test period switched the books in the two display locations.
In this Web site, results are converted into graphs to more clearly and dramatically presents the data. In most experiments, the experimental condition dramatically increases the circulation compared to circulation of the same books during the pre-test period and the circulation of the control group.
Our reader-friendly libraries experiments
Since the experiments of others clearly demonstrate that moving books from shelves to displays or listing titles on booklists, etc. dramatically increase circulation, we do not run pre-tests to compare with tests to replicate again this well-demonstrated fact.
Most of our experiments run only test periods for comparing the circulation between two or more conditions, e.g., displays vs. booklists, displays with props vs. displays without props, etc.