Skip to main content

New answers tagged

2 votes

Two-sided t-tests: Why do we need to test two-sided if your estimate is telling you in what tail to look already?

It's more complicated than you might expect, and more complicated than the current answers tell you. The number of tails for a test has been argued back and forth many times, but as far as I can tell, ...
Michael Lew's user avatar
  • 15.3k
0 votes

Two-sided t-tests: Why do we need to test two-sided if your estimate is telling you in what tail to look already?

Peter’s answer is very good but I’ll add that the reason that people will say it’s cheating is that if you change the “sides” you’re altering the hypothesis and also changing the size and location of ...
Rick Hass's user avatar
  • 358
1 vote

Two-sided t-tests: Why do we need to test two-sided if your estimate is telling you in what tail to look already?

Let's suppose your procedure is: 1- Look at which side points the estimate. 2- Perform a single-sided test using that side with significance alpha (for example 5%). It should be noted that this is ...
Pere's user avatar
  • 6,633
1 vote
Accepted

What is the difference between a one sample t test of the difference between two groups versus a two sample t test comparing the two groups?

Let's say the true (in the population) covariance $covariance(x_1, x_2)$ of $x_1$ and $x_2$ equals $C$. It can then be proven that the covariance of the sample means $\bar{x}_1$ and $\bar{x}_2$ is ...
BenP's user avatar
  • 1,169
12 votes

Two-sided t-tests: Why do we need to test two-sided if your estimate is telling you in what tail to look already?

Look at your hypotheses: H0: kids in a town sleep 7 hours on average (mu= 7) HA: kids in a town do not sleep 7 hours on average (mu=! 7) Your alternative hypothesis is two sided, so you need a two ...
Peter Flom's user avatar
  • 120k
1 vote

Why does the one-sided T-test reject the Null Hypothesis while the two-sided doesn't?

First, they are two different null hypotheses. In one sided tests $H_0$ is $\mu \leq 0$ and in two sided tests $H_0$ is $\mu = 0$. However it is still somewhat paradoxical since you can reject that $\...
Enrique Saint-Pierre's user avatar
1 vote

selecting the most appropriate statistical test for comparison of multiple independent variables belonging to two groups

Multivariate Analysis of Variance (MANOVA) should do this. Whether this is "most appropriate" depends on how the data behave regarding model assumptions, which I of course can't know without ...
Christian Hennig's user avatar
3 votes

Replicate t or F test from regression using regression likelihoods

I've heard that the t-test and F-test we use to get the significance of our regression results are derived from the likelihood ratio test, but I'm having trouble replicating the p-value of the t/F ...
Sextus Empiricus's user avatar
1 vote

Does a 2x2 study mean there are 4 groups or 2?

The language here does get confusing! You have two groups, in that very few people would think of "sex" as a group. On the other hand, you have a 2x2 design because you are, apparently, ...
Peter Flom's user avatar
  • 120k
4 votes

Replicate t or F test from regression using regression likelihoods

The short answer is: you should NOT expect these two tests are strictly equivalent (or mutually derivable). To corroborate my point above, we need to review some theory. To this end, let's denote the ...
Zhanxiong's user avatar
  • 18.9k
0 votes

What test should I use to compare the means of two groups in this case?

Unless you have a good reason, e.g. something like data is bi-modal, I would suggest modelling your task outcome as either beta or logit-normal distributed. e.g. something that respects form of your ...
Cryo's user avatar
  • 596
2 votes

Calculate $t$-statistics given coefficients + correlations

You don't have information about the variances of the variables. Let's pretend we do and see how that affects the results given at https://stats.stackexchange.com/a/135240/919, which shows how to ...
whuber's user avatar
  • 323k
0 votes

Planned comparisons: t-tests vs multiple regression analysis

A one-way ANOVA or Kruskal-Wallis test would be a good first approach for testing differences between treatment groups followed by pairwise t-tests or Wilcoxon rank sum tests. Both can be done with ...
Jack's user avatar
  • 344
5 votes

Best statistical approach for determining a significant cutoff value?

A dataset which is comprised solely of infected individuals will not answer your question. In short, you can not determine correlation without variation, and since your disease status does not vary (...
Demetri Pananos's user avatar
1 vote

Test for significant difference between two groups measured across time

Mixed effects models can still have a nonlinear function of time, e.g., a spline function in time as a fixed effect. But random effects usually don't capture the correct correlation pattern in ...
Frank Harrell's user avatar

Top 50 recent answers are included