I suck at statistics - please help. . .

[episodic]

Ok, I'm taking a class that the instructor assumes we are stats gurus - and we are not. . .

I had to conduct a survey, then analyze the data. I thought this was going to be easy till I sat down.

I had 100 respondents.

39 of them were not college grads
61 were



My hypothesis was that college graduates were NOT more likely to be prepared for a disaster (earthquakes, floods, etc).

I had several questions on the survey.

1. Are you a college grad? Yes/No
2. On a scale of 1-5 - five being the best - rank your knowledge of Earthquakes and what a large New Madrid earthquake will mean for the region. 1 - 2 - 3 - 4 - 5
3. Do you have a family disaster plan? Yes/No
4. Do you have a 3-5 day non-perishable supply of food/water that does not require refridgeration? Yes/No
5. Do you or someone in your family know how to shut off water mains, gas lines, and electricity? Yes/No
6. Do you have a radio and batteries or a crank radio on hand? Yes/No


Ok - so how do I analyze these results? I never took a stats class - and don't really have a clue.

I basically want to prove that a college degree does not = prepared for a disaster.

I know my independent variable is education. I only have 2 values though - college grad or not - 0 o 1. . .


I knew I should have done stats. ..

Help!

 

[KLin]

err 42%?

 

[DaWhim]

spss?

 

[Semidevil]

I"m just guessing here but maybe get a pink marker and a blue marker. blue = grad, pink = non grad.

just start counting how many of the grad said yes or no, and how many of the non grad said yes or no. and then take the average of grad, average of non grad, standard deviation of grad, standard deviation of non grad, etc etc? and then just compare your % and maybe do a correllation coefficient?

 

[episodic]

Ok standard deviation and correlation coefficient - these lost me. I head them all the time - have no idea how to derive them. I've looked up how to 'read them' before. . .

 

[homercles337]

Simply calculate the binomial probability of college grad and preparedness (ie. if college grad == 1 & preparedness 3-6 == 1. 1 being yes. Sum and divide by N. #2 is extraneous). Correlation would work also, but you will get a similar answer. PM if you want me elaborate. Basically, if your hypothesis is accurate then you should find something near 1 (ie, grad == preparedness), if youre guessing it will be around 0.5, and if youre dead wrong near 0. This should be fine for a class like this, but if you want to address the unequal sampling it gets a little more involved. Was this part of the requirement?

EDIT: oops, i thought your hypothesis was that college grads are more prepared. My post still applies though just do the necessary null hypothesis, hypothesis swap. Sorry.

 

[imported_michaelpatrick33]

Originally posted by: episodic
Ok, I'm taking a class that the instructor assumes we are stats gurus - and we are not. . .

I had to conduct a survey, then analyze the data. I thought this was going to be easy till I sat down.

I had 100 respondents.

39 of them were not college grads
61 were



My hypothesis was that college graduates were NOT more likely to be prepared for a disaster (earthquakes, floods, etc).

I had several questions on the survey.

1. Are you a college grad? Yes/No
2. On a scale of 1-5 - five being the best - rank your knowledge of Earthquakes and what a large New Madrid earthquake will mean for the region. 1 - 2 - 3 - 4 - 5
3. Do you have a family disaster plan? Yes/No
4. Do you have a 3-5 day non-perishable supply of food/water that does not require refridgeration? Yes/No
5. Do you or someone in your family know how to shut off water mains, gas lines, and electricity? Yes/No
6. Do you have a radio and batteries or a crank radio on hand? Yes/No


Ok - so how do I analyze these results? I never took a stats class - and don't really have a clue.

I basically want to prove that a college degree does not = prepared for a disaster.

I know my independent variable is education. I only have 2 values though - college grad or not - 0 o 1. . .


I knew I should have done stats. ..

Help!

You have no chance unless we create everything for you. You can't prove anything with this study you can only draw a limited correlation.

I took two stats classes and I have no idea how to help you. :laugh: Sad but true. You could assign a value of p and q to each variable and create a binomial analysis?

I would snare a senior psych or stats major with access to a good stats program. You have a number of variables that would do well with a stats program.