PRIVACY POLICY PROVISION PATTERNS AMONG THE TOP ONE HUNDRED NONPROFIT WEBSITES
Janet Stephens
Gundars Kaupins
Boise State University
ABSTRACT
Nonprofits in need of developing or refining their website privacy policies might benefit
from reviewing the content of other nonprofits’ privacy policies. They could do “one-stop-
shopping by looking for the longest policies, highest website rankings, most recent policy years, or
highest reading grade level. These four characteristics might increase their chance of finding
desirable provisions beyond only using keywords and key phrases. The use of only keywords or
phrases might not be beneficial in some cases because illegal, awkwardly worded, and out-of-
context phrases might be found. This study incorporated human and machine learning content
analyses of 100 nonprofits’ online privacy policies to find the association between 71 nonprofit
privacy policy provisions and the four characteristics.
Although the number of words in the policies positively correlated with the sum of
privacy policy provisions, the number of words did not have many more specific provisions than
policy age, reading grade level, and website rank. The number of words has the most significant
relationship with liability and fraud-related provisions. Other significant relationships with the
number of words include the personally identifiable information collected such as emails and
photos. Longer privacy policies do not significantly increase the chances of finding specific
provisions.
Keywords: Privacy policies, reading level, nonprofits, policy length, websites, confidentiality