To get a good answer, you must write a good question. Answering a statistics question without context is like boxing blindfolded. You might knock your opponent out, or you might break your hand on the ring post.

What goes into a good question?

**1.**Tell us the PROBLEM you are trying to solve. That is, the substantive problem, not the statistical aspects.

**2.** Tell us what math and statistics you know. If you’ve had one course in Introductory Stat, then it won’t make sense for us to give you an answer full of mixed model theory and matrix algebra. On the other hand, if you’ve got several courses or lots of experience, then we can assume you know some basics.

**3.** Tell us what data you have, where it came from, what is missing, how many variables, what are the Dependent Variables (DVs) and Independent Variables (IVs) – if any, and anything else we need to know about the data. Also tell us which (if any) statistical software you use.

**4.** Are you thinking of hiring a consultant, or do you just want pointers in some direction?

**5.** THEN, and ONLY THEN tell us what you’ve tried, why you aren’t happy, and so on.

I specialize in helping graduate students and researchers in psychology, education, economics and the social sciences with all aspects of statistical analysis. Many new and relatively uncommon statistical techniques are available, and these may widen the field of hypotheses you can investigate. Graphical techniques are often misapplied, but, done correctly, they can summarize a great deal of information in a single figure. ** I can help with writing papers, writing grant applications, and doing analysis for grants and research.**

** Specialties:** Regression, logistic regression, cluster analysis, statistical graphics, quantile regression.

You can **click here to email** or reach me via phone at 917-488-7176. Or if you want you can follow me on Facebook, **Twitter**, or LinkedIn.

The article is behind a paywall so I cannot read it. I don’t know of any good way to convert dichotomous items into Likert ones.

Hi Dr Peter,

Can you give me some suggestion what should i interpret where my p-value is 1.248 which is more than 0.05 and my coefficients is -1.0003?

Hope to see your replied soon.

Thank you,

Regards.

Hi Steffenie

The p value cannot be 1.248. P values are always between 0 and 1.

Peter

Dear Peter,

Can to please help me out with this statistical problem:

An independent variable “X”(construct) has 4 factors in it A, B, C, D which are significantly positively correlated with each other. There are two dependent variables ( “Y” and “Z”) which are correlated with these factors. A, B, C, D significantly positively correlated with “Y”. Except for A , rest are uncorrelated with “Z”. “Y” and “Z” are unrelated. The correlation matrix is as follows:

A B C D Y Z

A 1

B 0.67** 1

C 0.58** 0.64** 1

D 0.56** 0.68** 0.67** 1

Y 0.48** 0.41** 0.39** 0.38** 1

Z 0.21** 0.05 0.07 -0.04 0.10 1

Now on moving to the regression model, Y is significantly positively predicted by A.

For Z, A was a significant predictor as well as D (uncorrelated bivariate relation) also predicted it significantly but in opposite direction.

A–>Z (beta= 0.34, p=0.001)

B–>Z (beta= -0.11, p=0.241)

C–>Z(beta= 0.10, p=0.616)

D–>Z(beta= -0.40, p=0.015)

Please help me to clarify how come a nonsignificant bivariate correlation becomes so significant in regression model?

When you do the regression you are controlling for the other variables. When you do the correlations, you are not.

Peter

i did case series where i found out of 11, 10 patients with duodenal ulcer smokes more than 20 cigarette/day and 1 smoked less than 20 cigarette/day can you tell me which stastical test is suitable for this type of study

Fisher’s exact test should work

Peter

peter

In initial testing of the efficacy of a new potential vaccine against a presently incurable infectionous disease is causing great economic loss .

which confident lvl should be use and what should we do

I am validating a test on an chemical analyser. The manufacturer has provided data that shows that the CV is 3% when the sample is replicate analysed 200 times. Is there a way of calculating what CV I should expect to see if I replicate analysed the sample 20 times?

Well, if it was up to me, I wouldn’t use confidence levels at all. I’d look at the effect size and its precision and whether that was important. However, you may have to do something to please someone.

Peter

not that I know of.

I have ordinal/categorical data for 3 groups (control, treatment 1 and treatment 2). My data is cross section.

If I want to compare the mean differences between 2 groups (treatment 1 and control) and (treatment 2 and control), what kind of test should I use? I am checking wilcoxon test.

If you have ordinal data, then the mean is probably not appropriate.