Statistics Help @ Talk Stats Forum - Blogs

Wannabe Doc

Hockeyboy — Sat, 14 Nov 2009 03:40:09 GMT

I need some guidance on one of my proposal questions.
I want to know how to apply t-tests and chi-square for the following:

Are there significant differences between youth who are released from a juvenile corrections facility and successfully reenroll in secondary community schools and those who do not successfully reenroll with respect to: gender, age, ethnicity, residing community, grade attended, last 2 school attended, type of school attended, and type of educational program?

Hockeyboy

Linear regression - Please help.

bulu — Wed, 11 Nov 2009 11:54:45 GMT

Hi - Much appreciate your help

Question - Should I accept Ho in this case?

Ho = There is no significant correlation between liked commercials and purchase intent for the advertised brand.

Ha = Commercials that are liked will lead to higher level of purchase intent for the advertised brand.

The test is to examine the strength of a causal relationship between the independent variable � commercial likeability and the dependent variable � purchase intent.

Results : r = .304, r2 = .093
interpretation :Approximately 9% of the variance of purchase intent was accounted for by its linear relationship with commercial likeability. The overall regression was ----??? with F (1, 43) = 4.384, p = .042. Alpha was set at .05 (95%).

Other values = Beta = .304

Ideally, I would I reject H0 in favour of H1 simply because p<.05. However given that only 9% of the variance of purchase intent is account for by its linear relationship with commercial likeability, I am not sure if H0 should be rejected.

Please let me know urgently what should I do.

Classical Approach To Probability question

David_langstaff — Fri, 30 Oct 2009 13:11:57 GMT

Hey all, I am having a real problem whit a probability question that looks easy but is giving me a real pain.
The question goes, "A coustomer randomly selscts 6 doughnuts form 58 doughnuts of which 8 contain jelly. What is the probability that she selscts none of those containing jelly?
I am using the classical approach to probability to answer the question.
What I have done is to figure out the probability of her picking 1 nonjelly doughnut if she randomly selects using 50(nonjelly)/58total doughnuts= .862069
given that the probability that if she picked one doughnut and it would be none jelly is .862069 I figure that to find out what the probability of her picking 6 none jelly doughnuts I shoud raise .862060^6=.410442.
Does that sound right to anyone or am I making this to complicated

Help Please!

Silver Frog — Fri, 23 Oct 2009 05:40:53 GMT

Hi There,

This is part of my final assignment for the compulsory Research Methodology subject for my Master's degree. I am studying on line and basically that means teaching myself at this particular institution. I had to design a research project and explain how I would carry it out - we don't actually conduct the research.

My aim is to find out what kind of writing adults do in their daily lives after they finish formal education.

I chose a qualitative simple descriptive design with a survey. My question is probably so basic as to be embarrassing but I am really having trouble deciding which 'things' I do with the results. For instance the lecture notes talk about descriptive statistics - a written 'summary' of the data - is this ALL I would give with my analysis because I chose a simple descriptive design? :confused: The lecture notes also talk about things like chi squares - but I don't have a 'change' to discuss - t-tests; but I don't have 2 groups to work with (and it is not relevant to divide the group into male and female - or is it!!!:eek:).

Then I start wondering if I should consider the kinds of writing they do as the 'groups' - and look at the ANOVA.:shakehead

Obviously I am NEVER going to conduct my own research but I really have to get through this assignment. I have answered the other 3 parts to the question easily enough and I'm prepared to put in the hard work to finish this bit but I honestly don't understand how to figure out what I should do with the data.

Even a teensy clue would be appreciated:o

spss aggregate data

Hanna — Wed, 21 Oct 2009 15:09:28 GMT

Hi,

I have a quite large dataset, with cases represented by individuals. One individual may be represented many times, but with different data each time. I'm now trying to aggregate this data, so that each individual is only one case. For some variables, I can use the "sum" or the "first" function, but some variables are such that I would like to keep all of them as separate entries.

For example, individual x has five times had some kind of event, and I want to keep all of the different kinds of events for individual x.

I'm using spss.

Does anyone know if this is possible?

Best regards,
Hanna Edberg

I really need help with this question can anyone help me

kaylita — Tue, 20 Oct 2009 02:09:44 GMT

Test scores on the ACT exam are normally distributed with a mean of 21.1 and a standard deviation of 4.8.

Using the 68-95-99.7 rule, find the percentage of ACT scores within 14.4 of the mean of 21.1.

Help needed for linear and logistic regression

Dimitris — Fri, 09 Oct 2009 19:39:05 GMT

Hi i would appreciated most if anyone would be kind enough to help me with the following!

1. Is it possible in a forward LR logistic regression analysis, the explanatory variables identified as independent predictors in the final model (variables in the equation) have a p value (sig for Wald) > 0.05?

2. How do i test for interactions in a regression model?

Thank you in advance

I need help on a really simple problem

swharton — Fri, 02 Oct 2009 06:10:21 GMT

K so the problem is:

there are 10 problems, 6 of which a student gets right. 5 of the problems are chosen for a test. What is the probability that the student gets all 5 right.

So I know that the student has 60 percent chance of getting 6 out of 10 right. But then do I half that in order to get the answer or do I multiply by 1/10 because that is the probability of the one of the answers being chosen.

I think I'm over analyzing it but I need the help:shakehead

help!! it's simple but i dont get it! im just learning..

bittersweetb — Wed, 30 Sep 2009 15:59:21 GMT

"If an IQ scale test has a mean (p) of 100 and a standard deviation of 15, and one student's IQ is 125 -- then what is the percentage of the students that have a higher IQ than he has? and what is the IQ below which the lowest 10% of the student population falls?"

---
so... am i supposed to use the z score formula and get -1.66 or is it something to do with 4.75 and 80.8%? and how do i get that?

Need Help with this Probability Question

mercer0407 — Wed, 30 Sep 2009 00:59:34 GMT

Suppose that a classroom has 10 sets of lights. The probability that one of the lights work is 0.6. Suppose that each light works independently of the others. What is the probability that all of the lights work? Round your answer to four decimals places.

Factor Analysis and Liketer Scale

Diala — Sat, 19 Sep 2009 22:16:25 GMT

Hi,

I needed a help in performing factor analysis on a likert scale survey.

I have a survey of 30 questions, the answers are based on a likert scale (1,2,3,4,5) which corresponds to (Strongly agree to Strongly Disagree)

I know the assumptions for Factor analysis is
1-Normality
2- Homocedasticity
3- Linearity

I grouped the question into variables and summed up the questions related to one variable and overcome the Normality problem, also I conducted the Bartlett's test to test for Homocedasticity, and i was able to reject the null that says there is no correlations between my variables.

My problem is that neither my variables or questions are lineary related to each other.

And it doesn't make sense that the variables values (summation of questions) be lineary related for a survey answers based on a likert scale?

Is linearity critical in this case ?

Any Help..

Thanks

help please

tanu_bathla — Sun, 23 Aug 2009 15:18:42 GMT

i am a new member of this forum. i have done a research on four groups having 10 participants each. i want to compare all the groups but due to small sample size i am afraid of using ANOVA. Should i use Kruskal wallis test for comparision.
thanks,
regards,
tanu

panel data random effects

confused24 — Sat, 15 Aug 2009 09:00:32 GMT

could anyone help me with my regression ? I use STATA for balanced data and it came out as random effects results. It is my first time when I use stata for panel therefore Im not sure what to do next. I know the best results should be fixed effects and then, xtgls for heteroscedasticity and autocorrelation, but I m not sure what to do with random effects. Could anyway help me with that??

Normality test-Split plot CRD

georgesote — Fri, 14 Aug 2009 10:59:27 GMT

I would like to learn how to perform a normality test on my data using JMP 7.0.

I am already familiar with the proccess however I am not sure whether I should perform the test on the residuals or the actual values of the experiment.

Furthermore I am not sure whether I should perform the test on each sub-plot individually or all together.

I would appreciate any help.

Logistic and Stepwise Regression

cmac123 — Thu, 30 Jul 2009 18:23:33 GMT

I was told by another that Logistic regression could be performed with a continuous DV? I was not taught this and want to find extra feedback. Also, I was told that SPSS will provide information regarding the ordering of the IVs for a Stepwise regression. I was told that steps were arranged by theoretical information. Does anyone have anymore feedback on either?