FAQs - Conducting In-School Surveys
Written by Tasha Snyder, Agricultural Sciences

The school setting is an optimal location to survey large numbers of students for research projects. Before planning a data collection study, however, it is important to be aware of federal guidelines and policies that contain regulations relevant for researchers conducting in-school studies of youth. These guidelines have implications for sampling, the informed consent process, and survey content.

Q: What federal regulations should I be aware of before planning an in-school study of youth?

A: The two key federal regulations you need to be familiar with are the Family Education Rights and Privacy Act (FERPA) and the Protection of Pupil Rights Amendment (PPRA). Both federal regulations apply to any institution receiving funds from the U.S. Department of Education.

FERPA
Initially signed into law in 1974, FERPA regulations are periodically updated, most recently under the No Child Left Behind Act of 2001. The main function of FERPA is to protect parent's rights regarding inspection and modification of their child's educational records. Under FERPA regulations parents have the right to inspect their child's educational records, request a correction to any errors in those record, and parent's written permission must be sought before releasing their child's educational records. Schools may release limited contact information for students (name, address, phone number, attendance record, date and place of birth, honors and awards) without written parental permission, but parents must be notified of these requests and given the opportunity to request that their child's information not be released.

PPRA
Part of the FERPA and No Child Left Behind legislation, PPRA is designed to protect parent's and pupil's rights regarding inspection of any materials - as part of the instructional curriculum or study instruments - to which students are exposed. Instructional and survey materials must be made available for parents to review, and schools must obtain written parental consent before their minor students participate in any study that asks about the following:

• political affiliations or beliefs of the student or student's parents;
• mental and psychological problems that are potentially embarrassing to the student and/or his or her family;
• sexual behaviors and attitudes;
• illegal, antisocial, self incriminating and demeaning behavior;
• critical appraisals of family members;
• legally protected relationships, such as those with lawyers, clergy and physicians;
• religious practices or beliefs of the student or student's parents; or
• income

Under PPRA schools are required to develop and adopt policies in conjunction with parents regarding their rights to inspect research surveys and instructional materials, protect students' privacy related to the eight items noted above, administering physical exams to students, and collecting student information intended to be used for marketing purposes. In addition, parents must be made aware in advance of dates when data collection studies will occur and of their rights to withdraw their children from participating in any study that asks about the eight items listed above, any marketing surveys, and any non-emergency physical exam or screening.

For a complete description of the FERPA and PPRA regulations, recent Supreme Court rulings, and a history of the legislation, see the following urls:

http://personalinfomediary.com/FERPA_info.htm
http://www.ed.gov/policy/gen/guid/fpco/hottopics/ht10-28-02.html

Q: What are the implications of the FERPA and PERPA regulations for the design, sampling, and implementation of my in-school study?

A: There are two main implications for study design and implementation. First, the content of your survey largely determines the level of parental consent required. In-school studies with surveys that contain sensitive questions related to any of the eight items described above must first provide parents full information about the study and survey content and then obtain written parental consent before a student can participate in the study. Surveys that contain more benign questions might not require active written parental consent but rather "passive" consent where parents are notified of the study, usually at least 2-3 weeks in advance, and reply to deny consent, rather than provide it. This type of scenario is addressed on a case-by-case basis by the IRB. Regardless though, all parents must be notified of the dates when data collection will occur so they can choose to withdraw their child from the study, even after giving written parental permission to participate.

Second, the level of consent required has important implications for the representativeness of your sample and can introduce sample bias into your study. Findings from several recent studies consistently highlight how sampling bias is introduced into research involving youth when active written parental consent is required (Dent, Galaif, Sussman, Stacy, Burtun and Flay 1993; Ellickson and Hawes 1989; Esbensen, Miller, Taylor, He and Freng 1999; Henry, Smith and Hopkins 2002). Two studies in particular, Esbensen et al. (1999) and Henry et al. (2002), compare samples from the same population involving active written parental consent with those involving passive parental consent and document how response rates are affected and bias is introduced.

Esbensen et al. (1999) collected data from 7th grade students in six U.S. cities and employed a sampling design that used passive parental consent for their pre-test wave of data collection, and then were required to obtain active parental consent for a subsequent wave of data collection on the same sample. This design provides a unique opportunity to compare the two samples of the same population, and determine how passive and active parental consent procedures differentially impact sample selectivity. Regarding response rates, in their pre-test survey, where passive parental consent was used, only 13 of the 2,496 eligible 7th grade students (0.4%) could not participate because their parents denied consent. When active parental consent was required for their first wave of data collection the response rate and sample size of their study were considerably reduced. Between 23% and 45% of youth did not participate in the study, depending on the site, due to non-response from the parents. Extensive follow-up of non-respondents found that 78% of the parents subsequently provided consent for their child to participate in the study, and 22% refused to provide consent. This suggests that non-response indicates passive acceptance rather than being synonymous with refusal. Thus, the active written parental consent process that is required by PPRA could result in an unnecessarily large non-response rate, preventing students from participating in studies.

Regarding sample bias, Esbensen et al. (1999) provide strong evidence that the characteristics of the students whose parents did not respond were different from those whose parents responded, thus introducing bias into their sample. Comparing the pre-test sample with the sample of responders to the active parental consent, including those who allowed their children to participate and those who did not, revealed that the sample of responders was more likely to be white, to come from intact homes (two married parents), and to have parents with more than a high school education. In addition, parents of "at-risk" youth (positive attitudes towards and engagement in delinquent behaviors) were less likely to return consent forms at all. This study documents a selection bias introduced by the active parental consent procedure, and recommends that a mailing be sent home to parents notifying them of the study, and that the parents be given 2-3 weeks to deny their child's participation in the study, after which time period the consent is implied.

More recently, similar findings were reported when Henry et al. (2002) conducted an in-school study of 7th grade students in nine school districts in rural Pennsylvania. Active written parental consent was required for all students participating in their study, however, the project was able to access secondary data on several student characteristics for all eligible students--including GPA and absence from school--which allowed them to determine whether and how their sample was biased on these important characteristics. This study further made comparisons between the consent, non-consent, and non-response groups and determined not only if bias was introduced into their sample, but whether it originated from the non-consent group, the non-response group, or both. Findings revealed that the students whose parents declined their consent did not differ significantly from those whose parents provided consent, but that the students with non-responding parents did differ significantly from those who consented on two key educational variables. The students whose parents provided consent had fewer days of school absence and higher grade point averages compared to the students with non-responding parents. Thus, comparisons of the three groups (consenters, decliners, and non responders) find that the students eliminated from the study because of the lack of permission due to non-responders is the source of sample bias, not the presence of decliners. The result is that the Henry et al. (2002) sample represents students who are less "at-risk" for poor outcomes, thus introducing sample bias.

Both of these studies employed the numerous strategies suggested to increase return rates for active written parental consent. These strategies include designing effective informational and consent forms, working with key school personnel, multiple mailings to parents, providing rewards for classrooms that have high return rates (such as pizza or ice cream parties), and following up with non-responders (Esbensen et al., 1999; Fletcher and Hunter, 2003; Henry et al., 2002). Even though these strategies were implemented, both studies report biased samples due to the high degree of non-response associated with written active parental consent procedures.

Q: Yikes! So, what's a child and youth researcher planning to use in-school data collection to do?

A: Well, there really isn't much you can do to completely avoid sampling bias with in-school youth surveys of sensitive topics where active written parental consent is required. The FERPA and PPRA regulations were developed and instituted with the goal of protecting the rights and privacy of parents and students, not to promote easier access to students for research purposes.

Following the suggestions outlined above will help reduce bias in your study. Also keep in mind that sampling bias is an issue for most primary data collection studies.

References

Dent, C.W., Galaif, J., Sussman, S., Stacy, A., Burtun, D. and Gley, B. 1993. "Demographic, Psychosocial and Behavioral Differences in Samples o f Actively and Passively Consented Adolescents". Addictive Behaviors, 18: 51-56.

Ellickson, P.L. and Hawes, J.A. 1989. "An Assessment of Active Versus Passive Methods for Obtaining Parental Consent." Evaluation Review, 13: 45-55.

Esbensen, F., Miller, M., He, N. and Freng, A. 1999. "Differential Attrition Rates and Active Parental Consent." Evaluation Review, 23: 316-335.

Henry, S.L., Smith, E.A. and Hopkins, A.M. 2002. "The Effect of Active Parental Consent on the Ability to Generalize the Results of an Alcohol, Tobacco, and Other Drug Prevention Trial to Rural Adolescents". Evaluation Review, 26: 645-655.