Help with sampling method for dissertation project

alex_boothby

In Runtime
Messages
317
As your all such a clever bunch i was hoping you could help.

I am writing my dissertation on basically along the lines of how people learn and react to climate change related knowledge of different education levels and from that how possibley we could change the way people learn about climate change to more effectively reach people with information. To do this i am creating a questionnaire that i need to distribute to many as possible people across the board of education levels.

As this is obviously just a undergrad project and not a post grad 'serious serious' endeavor i do not have resources nor money for sampling methods more reliable such as random phone calling or door to door.

What i have done so far is created the questionnaire online which can be accessed by people if they have the link. To advertise to many as people as possible to complete the survey i was thinking of using social networks sites such as Facebook. I would message as many people as possible ask them to complete the survey and then ask them in turn to ask others to do so. This is the only realistic way i can think of, of reaching a large enough audience size and thus gathering a large enough sample size to which i can decipher some useful stats and conclusions from!

first of all- what do you guys think of this method of sampling- would it effectively be 'respondent driven sampling' or snowballing? And if so is it true to say it is representative of the population that i am seeking to look into aka. education levels? I am worried that the marker may say this is only representative of the people on facebook.

Secondly- if you guys do think this at all appropriate, what is the best way to star the ball rolling? Is it ok to send it too all the people in my friends list or is that not really suitable?

I know this is obviously not the best sampling method but it is the only way i can realistically reach enough people.

thanks in advance for your inputs,

alex
 
Thats a good idea mate, i think i will. Only thing with that is it is still representative of the population of people with different education levels? or would it just be representative of people from this website?! Im just worried ther gonna slate me for my sampling method. Maybe i could just do it, but say i randomly stopped people on the street. (the only reason im not doing this is because i honestly cant see people stopping for ten minutes to do a questionnaire while ther busy, but sat at home is more likely i think.

cheers
 
the only reason im not doing this is because i honestly cant see people stopping for ten minutes to do a questionnaire while ther busy
Few hints with this:

- Go at off-peak times, you'll get less people but those there may be more willing to stop and do it
- Offer something in return, evening if it's just some chocolate
- Keep it as short as it can be.

There's no clear sample that's going to be completely diverse. You could say people on here aren't a representative sample because they're interested in computers, people on Facebook aren't representative because they can afford computers / internet, people on the street aren't representative because they come from a certain area...

Most important thing is that in your dissertation you identify these potential limitations in your analysis, and say in an ideal world how you would've done things. You shouldn't need to lie to get a good mark, as long as you understand the limitations of what you're doing and you make reasonable efforts to collect the best data you can.
 
yeh i agree with you. I like the chocolate idea but then it may be possible to say people who only like chocolate completed it! or they might fill it in how 'they think i want it to be filled in' rather than their honest answers. Its all swings and round abouts! Nightmare! Cheers for the good reply. Would you like to fill it in???? :p
 
Liking chocolate does not specify socio-economic background, or education. -so this is not problematic for the reasons you say.

the fact that a person likes chocolate says nothing else about them. (unless the chocolate that you specify is a particular brand, that is well known for selling "green" chocolate (Rainforest alliance etc).)

Anyone asking such questions (about ingredient sustainability) should be marked as such when answering the questions.
 
I'd happily fill in your questionnaire also if you want a variety of people smart and less smart go on omegle text chat (not video).

If you're not sure what omegle is it's a site that allows you to chat with strangers.
 
Cheers clacker, once ive polished it off i'll pm you the link! I'll Warn you, its extremely exciting......

One more thing, if the main factor im looking into is Education levels, is it still appropriate on the questionnaire to still ask for e.g, age, sex, religious views, Politcal preference? I was thinking taking note of these factors also may able me to locate any anomalies such as people with extreme preference for tory or say people with deeply religious backgrounds which may not be of the norm for people of that education level. I was also thinking what if my data shows that education does not seem to correlate with anything and for example it is politcal preference that matters most, i don't want to really go down the political opinion of climate change- but it may be the only facotr worth talking about? Does that make sense?

Liking chocolate does not specify socio-economic background, or education. -so this is not problematic for the reasons you say.

I agree with you completely Root, just ive read a few studies in which if you give people something e.g prize, confectionery for taking part in a study they often complete the test/questionnaire under false pretences , as if you have given something to someone they believe they should complete the test to how they think 'you' the 'researcher' would be most pleased. Rather than honest answers they would ordinarily put. That what i was getting at, but ultimately bribery may be the only way , ha!
 
I didn't know that.

It's a tough call though, I mean a lot of times without offering a prize you'd be hard pushed to get people in the street to even talk to you.

probably the best thing to do in this case is:
run a survey online, with no prize, write that you assume that only those that care about the subject matter will actually fill it in (broadly true).
run a survey in the street with prizes, make reference to studies that make this point, but the key thing is that you weight your results differently,

you assume that there will be those that fill in the survey online that are weighted either way, (zealots on both sides).
but those in the street you assume will want to perform for their prize and weight their results differently, so five positive street answers carry the same weight as four positive internet answers?

Though clearly you can decide on the weighting given the study. -that you've read but I have not.
 
Back
Top Bottom