Survey for College Course

[linux.leep]

Hi all,

I am in a course in my bachelor's degree in Information Systems and currently am taking a course in "Research Methods." In this class I have to produce a survey and have people answer it, review the results, and write a paper and produce a presentation.

I decided to research the "red pen" controversy that teachers should not be correcting students papers with red ink. I do not know how big it is throughout the United States and the World so here are some links:

http://www.signonsandiego.com/uniont...s_1m4pens.html

http://www.msnbc.msn.com/id/7374218/...ews-education/

http://www.telegraph.co.uk/news/news...-students.html

Now, I am asking people to please go to my survey and fill it out. I will only take a few minutes. There are only 10 questions total.

Here is the link to my survey:

http://www.surveymonkey.com/s/RHM7GGT

I chose to ask this community instead of handing it out to random people on the street, at my school, or at my job because I want some truthful answers from an intelligent group of people.

Thanks all, for your time and our opinion!

[rigao]

I chose to ask this community instead of handing it out to random people on the street, at my school, or at my job because I want some truthful answers from an intelligent group of people.

That hurts me a lot.

This will be enough reason to fail the course, in my view at least.

You are not supposed to select the people who will fill your survey, because you are introducing biases. The world is not full of smart geeks, hence this community, although great, won't represent properly this world.

If you want to conduct a serious research, you must ask people randomly. Note that going to a street and ask people walking by is not a good way either, because you are only going to survey people who happen to have the time or job to let them be walking this street at this hours.

The proper way would be to select randomly houses from a house directory, if you want to make a personal interview, or select phones randomly from a telephone directory. Although it is impossible to account for the whole population, with this methods you will only ignore a few, hence your bias will be small.

If you are studying researching methods for a private firm, probably you want to look at quota surveys. This method is far cheaper and it has been shown to led to the same results as truly random sampling in social surveys. The basics of this method is that you, before starting the survey, decide how many people you will be asking from a quota: 100 men and 100 women, of which 80% will be white, 20% black, or 50% with income less than xxx, 25% with income between xxx and yyy, and 25% with income bigger than yyy.

Obviously the quota sampling is a biased method of sampling, and is just semi-random, but allows you to sit in a street waiting for your quotas to fill. It is not as if I like this method, but everything will be better than what you are trying to do.

Anyway, I'll answer your survey, just for fun.

[NovaAesa]

I really think you need to scrap your survey are write a better one...

[hebotick]

It's best to let not only intelligent people answer your survey, but everyone who can think.

[andymorton]

Another point is that people on here (I think) tend to be male, so by asking us you're excluding 50% of the population. Although there are some women on here, there doesn't seem to be that many. Sadly, computer geekiness seems, generally, to be a male pastime. Apologies to any members of the fairer sex.

andy

[NovaAesa]

Okay, I'll give you some constructive criticism:

Q1:
"1. Does the use of red ink by teachers, while grading work, affects a student's self esteem?"

This question is vague. Does it want my opinion on whether there is an effect or not on the average student? Does it want my opinion on whether there is an effect on me if I was previously a student? Does it want me to use information that I have previously gathered myself in a previous study? Are you referring to a positive effect on the student's self esteem or a negative effect? Or any effect?

Q2:
Good question.

Q3:
What would someone answer here if they answered "cannot remember" for question 2?

Q4:
Again, are you talking about a positive effect, negative effect, or any effect? It's ambiguous.

Q5,6,7,8:
Same problems as question 1. Are you asking for my opinion? My experiences when I was a student? What exactly?

Q9:
The fourth option is redundant and breaks your survey. What if someone checks both none of the above as well as parent? How will you interpret this data? It's always better to design a robust survey than have to throw away invalid data.

Q10:
Seems to be okay, although I find it amusing that the question is so heavily biased towards people living in America when you are primarily getting data from an international forum..

[linux.leep]

Thanks for the constructive critique all. My questions were different originally. The teacher reviewed my questions and told me to change them. Glad I was not the only one who thought the questions were bad.

Thanks everyone for your help, and sorry that I did not add more international approach to the last question.

Again, thanks all!

The survey is gonna be up for another few days if anyone is still interested in answering it!