From March 15 to May 8, data were collected using online survey software (SurveyMonkey) from private schools across the state of Michigan. Emails inviting individuals to participate in the survey were sent to all members of the Michigan Association of Nonpublic Schools. The initial email invitation was sent from MANS officials to MANS member schools on March 15. MANS sent follow-up and reminder emails to MANS member schools on March 22, March 31 and May 5.

MANS also distributed email survey invitations to all non-MANS members with private school contacts listed in the Michigan Department of Education’s Educational Entity Master, which is managed by the Center for Educational Performance and Information. The initial email invitation for non-MANS members was sent on March 22. Follow-up and reminder emails were sent on March 31 and May 6. In total, 659 emails were sent from MANS to private school contacts.

Upon the closing of the survey on May 9, 2016, 475 unique responses were entered into the survey. Just over 100 responses were duplicative (23 percent). Duplicate responses were identified using both school name and IP address. If a duplicate response was identified, the response with the most data available was retained. For those duplicate responses that had the same amount of data, the most recent entry was retained. The total number of nonduplicate responses was 366. In addition to duplicate data, 32 schools entered the survey but did not provide any data other than their school name. These responses were also removed from the final data set. Thus, the total number of nonduplicate, complete responses was 334, for a total response rate of 51 percent.

Data within the survey was cleaned on an individual basis. Data cleaning procedures consisted primarily of investigating abnormal responses, such as percentages greater than 100 percent. When a percentage response was greater than 100, the response was investigated further to better understand the nature of the response. For example, when asked to enter the percent of students that were in shared time programs, an individual entered “254.” While 254 percent was not possible, further investigation revealed that the total enrollment for the school was 259 students; as a result, the data point was manually changed to 98 percent. Other data cleaning procedures included shifting spelled-out responses to numerical responses (e.g., “eight” to “8”) and identifying data entry issues (e.g., manually changing a response of “0.05” to “5%”).

In addition to survey data, data from CEPI were collected from the MDE Nonpublic School Database. This database includes data on private schools’ geographic location, affiliation, teacher counts and characteristics (e.g., education level and certification status), enrollment counts by grade, and basic course offerings.