Student Survey

[cluelessmedstudent]

Hi guys,

I'm part of the Student Union in my college and I'm trying to push hard to extend library hours this year. So I conducted a poll on the student body in my school but I'd like to synthesize that in a statistically satisfying format for the school board. BUT I have no idea how and I've forgotten all my high school statistics. So I was wondering if you guys could help me out...

Anyway, my null hypothesis is that: Students do not want extended library hours.
Research hypothesis: Students want extended library hours.

There are 2000 students in my college. 237 students responded to the poll.
Of the 237 students that responded, 228 voted YES for extended library hours. 9 voted NO for extended library hours.

So my question is this: Is it possible to calculate from the above data the statistical significance of the poll results or if the population sampled (237/2000) is of sufficient power?

Thanks in advance!
Confused College Student

[Outlier]

To estimate your probable success in this project, look up "utility function".

The school's utility for having extended hours may be radically different from the students' utility. You should then try to find ways of increasing the school's perceived utility.

[CowboyBear]

First off, I'd agree with Outlier - there's more to take into account here then whether the students want extended hours. Most students (except those who work at the library) would not experience any cost if hours were extended, so any perceived benefit (no matter how marginal) would likely result in them answering "yes". You're probably going to need to provide more evidence to your school that the benefit of extended hours would actually outweigh the (probably considerable) financial costs.

A more stats-specific comment: With a simple one-variable poll question like this it'd be unusual to frame a sharp null hypothesis and use a significance test. The convention would be to report the percentage of students answering "yes" along with a margin of error or 95% confidence interval. You can calculate these here. This is more informative - instead of just rejecting a null hypothesis that fewer than half of students support extending opening hours, you are actually providing a reasonable estimate of the actual level of support in the overall student body.

[cluelessmedstudent]

great replies! thanks v much. back to the drawing board