Groundwork
A student survey is the one piece of evidence you can create that nobody else has. It is also the easiest thing on this site to do badly, and a survey with leading questions or no sample size gets dismissed in one sentence, taking your credibility with it.
New Jersey has its own student survey statute, and most people have never heard of it because they know the federal rule instead. It is stricter than the federal one in the way that matters here: it has no funding condition at all.
any academic or nonacademic survey, assessment, analysis or evaluations.1.a
That phrase is the whole reach of the statute. A district may not administer one revealing information in eight listed areas without prior written informed consent from a parent or guardian, requested at least two weeks in advance. Those areas include mental and psychological problems potentially embarrassing to the student or family, illegal or self-incriminating behaviour, political affiliations, sexual behaviour, income, and a social security number. Districts face monetary penalties for getting it wrong, which is why the answer to an unannounced survey is often simply no.
A separate 2021 provision lets a district use notification rather than consent for an anonymous, voluntary survey, but only on four health topics, and a phone-policy survey is not among them.source
The federal Protection of Pupil Rights Amendment adds a parent’s right to inspect “a survey created by a third party before the survey is administered or distributed by a school to a student”.s.1232h(c)(1)(A) Whether a student counts as a “third party” is not defined anywhere in the statute. See below.
This is not a formality, and it is the step people skip. The school controls access to students, and the moment a survey is handed out in a classroom, emailed on a school list, or announced over the PA, it stops being a private project and becomes something the district is administering. Get a yes in writing, from a real adult, before a single student answers.
New Jersey has a survey statute of its own and it is stricter than the federal one. A district may not administer any academic or nonacademic survey that reveals information about political affiliations, mental or psychological problems, sexual behaviour, illegal or self-incriminating behaviour, critical appraisals of family members, privileged relationships, income, or a social security number, without prior written parental consent, requested at least two weeks ahead.source
No names. No email addresses. No “what grade and what lunch period” together, because in a small school that is a name. Ask what you actually need and nothing more. If you cannot explain why a question is on the form, take it off.
This is the test. If every question is written so that only one answer is reasonable, you have not measured anything and everyone who reads the questions will know. A survey whose result you cannot predict is the only kind worth running.source
Every question, in order, exactly as asked. How many people you asked, how many answered, and how you distributed it. A percentage without a denominator is not a finding, and the first thing any serious reader looks for is the number underneath it.source
You surveyed your school. You did not survey New Jersey, and you did not survey students who chose not to answer, who may differ from those who did in exactly the way that matters. Say “180 students at one high school in Bergen County” and let the number speak.
This is the part that decides whether anyone believes your number. The left-hand column is what an advocate writes on the first draft; the right-hand column is the same question asked honestly.
Leading: Do you agree that banning phones at lunch is unfair to students?
Better: Should students be allowed to use their phones during lunch? (Yes / No / Not sure)
The first tells you the answer it wants and attaches a word, “unfair”, that is hard to disagree with. AAPOR calls this leading: an opening phrase that makes one answer awkward to give.
Leading: Given that the state already allows phones on the school bus, should it allow them at lunch?
Better: Should students be allowed to use their phones during lunch? (Yes / No / Not sure)
Putting your argument inside the question measures how persuasive your argument is, not what people think. Supplying information about only one side of an issue biases the result in that direction.
Leading: How much has the phone policy hurt your ability to focus and your mental health?
Better: Since the policy started, has it been easier or harder to focus in class? (Easier / About the same / Harder / Not sure)
Two problems. It asks about two different things at once. Pew calls this double-barrelled, and the answer cannot be interpreted. And “mental health” walks straight into a protected area under New Jersey's survey statute.
Leading: Most students think the policy goes too far. Do you agree?
Better: Which comes closest to your view? (The policy goes too far / It is about right / It does not go far enough / Not sure)
Agree-or-disagree questions attract agreement regardless of content. Pew describes this as acquiescence bias. Offering a choice between stated alternatives avoids it.
Both of the sources behind this section are worth reading in full before you write a question: AAPOR’s four-page note on question wording, and Pew’s explainer on writing survey questions.sourcesource
Attach the full question list. An administrator saying no is almost always saying no to an unknown: a survey they cannot see, about topics they cannot predict, that they would be answerable for. Handing them every question up front removes most of the reason to refuse.
Change this before you send it. Send it to the person who actually knows you (a club advisor, a teacher, an assistant principal) rather than cold-emailing the superintendent. Fill in how you plan to distribute it, because that is the question they will ask first. And mean the last paragraph: if the answer is no, the reason is worth more than the survey was.
Subject: Permission to run a short student survey Dear [principal / assistant principal / advisor], I would like to ask permission to run a short, anonymous survey of students at [school] about the district's internet-enabled device policy, and I wanted to check with you before doing anything. Here is what I am proposing: - [Number] multiple-choice questions, taking about two minutes. - Neutral wording, with no question that pushes toward an answer. I have attached the full list of questions so you can read every one before deciding. - Completely anonymous. No names, no email addresses, no grade-and-class combinations, nothing that could identify an individual student. - No questions about any of the areas listed in N.J.S.A. 18A:36-34: nothing about mental or psychological problems, illegal behaviour, family relationships, religion, income, or anything similar. - Distributed [how: e.g. a paper form at a club meeting, a QR code on a poster, a link shared by students to students] outside instructional time. - I will share the full results with you before I share them anywhere else. I know the district has a policy governing surveys of students, and I would rather follow whatever process applies than find out afterwards that I should have. If this needs to go to the superintendent's office, please tell me and I will submit whatever is needed. If the answer is no, that is fine and I would still be grateful to know why: it would help me understand the concern. Thank you for your time. [Your name] [Grade]
We run the same instrument statewide, with neutral questions, no names, no school name, and county-level geography only. Every question is published on our methodology page exactly as asked, and results are held back until enough people have answered that nobody is identifiable.