# Search results

1. ### Prediction rankings

That is one particular model: random utility probit model. When I said: "some ... model", I meant: "with a carefully chosen set of predictors"... Here is a link to extensive overview of various random utility modeling frameworks...
2. ### Prediction rankings

That's better. As I said, some random utility probit model could be used.
3. ### Testing different effects of an intervention on different outcomes

If you want to test hypothesis Ha: Expectation[ Y2(experimental) - Y2(control) ] > Expectation[ Y1(experimental) - Y1(control) ] you can use difference in differences analysis.... On a different note, MANOVA does not "construct a composite out of the 2 variables". It tests whether the...
4. ### Testing different effects of an intervention on different outcomes

The answer depends on the number of dependent variables and distributional properties of the dependent variables... If all of them are normally distributed, MANOVA is one path to follow.
5. ### Prediction rankings

You can estimate a random utility probit model, assuming you have observed 1) players' performance, 2) various characteristics of the players (age, experience, success rate, publicity, compensation, etc) over many seasons. 17 seasons is too little for any intelligent inference. Basing one's...

Exactly!
7. ### Compare dummy variables in a regression analysis (not to baseline but among each other) ?

You can compare other categories to one another using the difference in regression coefficients: Delta_km = Beta_category_k - Beta_category_m. Using the covariance matrix of the regression coefficients, you can calculate the standard error of Delta_km and see if Delta_km is statistically...
8. ### How to analyze non-normal DVs from 4x4 repeated-measures design

Since you have only 20 subjects, the normality assumption is essential for the plain vanilla repeated measures ANOVA. You would need to find one transformation which makes the distribution normal at each stage in each group. This is unlikely to happen. So another approach to look at is a very...
9. ### Which test to use?

The suitable test depends on the distribution of Pain Threshold at each stage. If the distribution passes normality tests, you can use repeated measures ANOVA. Otherwise, you would have to program a randomization test. Imperfect approximations with other testing procedures would be too...
10. ### Creating variable (alpha?)

You can make use of Stata commands gen aaa = <...> replace aaa = <...> if bbb == <...>
11. ### Time series analysis. Autocorrelation and differences

Yes, this is normal. Differencing a time series helps with making it stationary and susceptible to subsequent modeling. It does not remove serial correlation though. To capture serial correlation, you should consider an ARMA(p,q) model as applied to the differenced time series. To identify the...
12. ### Can we apply Chi Square goodness of fit test on a survey question having multiple options?

The count for FO and FE is the same. This indicates that the subjects may not have been assigned to cells at random. If that is the case, chi-square test is inapplicable. Chi-square test can be used when 1) you collect a fixed number of subjects and see where they fall, 2) the expected count in...
13. ### Need help with a simple formula i think?

1. To know "how often do you screen", you need to develop a stochastic model for evolution of the decease. Look at the relevant epidemiology literature. Formulas there are far from being trivial. 2. Contracting a decease is random. Therefore, you will never know that you have "picked up ...100%...
14. ### Need help with a simple formula i think?

This statement of yours does not make sense even from the point of view of standard English grammar: "I have to screen to detect say, a 10%, 20% etc of the cases"... I have to ask again: what are you trying to find out? What is your criterion?... What information is available to you for each...
15. ### Need help with a simple formula i think?

You said that already: "screen a population for a... disease"... So I can only reiterate: "What do you want to detect? What is your criterion?... What information is available to you for each member of the population?"
16. ### Need help with a simple formula i think?

The problem is not clear. What do you want to screen the population for? What do you want to detect? What is your criterion?... What information is available to you for each member of the population?
17. ### Cross correlation and t tests

It's all correlated, between the machines and over time. So neither a t-test nor a Wilcoxon test is appropriate. You should model the tasks as counting processes where the intensities are deterministic or stochastic functions of time. The models can be estimated via the method of maximum...
18. ### Good book on markov chains

Chapters 4-7 of Ross, S. M. (2009). Introduction to Probability Models (10th ed). Academic Press. will serve as a rigorous introduction into Markov chains and closely related stochastic processes. Later on, when you solve actual problems, you may fall back on Lawler, G. F. (1995)...
19. ### Good book on markov chains

Noetsi, so are you interested in 1) an introductory book where all major properties of discrete-time and continuous-time Markov chains are nicely laid out or 2) a more advanced book showing applications of Markov chains to real-life processes? Or are you interested in most general Markov...
20. ### Markov chain with covariates

That is why I mentioned: "You would just use the idea for the algorithm, not the functional form." Just apply the algorithm to various defining characteristics of the Markov chain.