Toggle Navigation
Kerem Eryilmaz
Everything
Blog posts
Pages
Home
CV
CV (PDF)
Publications
Blog posts
New regression equation for LeapArticulator experiments
Home
CV
CV (PDF)
Publications
Posted by:
kerem
8 years, 4 months ago
(
0 comments
)
I have just discovered a better model for participants scores in the experiment. This used to be our prime model: ```prettyprint model <- lmer(score ~ nstates_amp_and_freq_n + (1 | id) + (1 |condition:phase) , data=all, REML=F) ``` I have discovered the following one outperforms this for both the original and the discrete datasets, and is much easier to interpret: ```prettyprint model <- lmer(score ~ nstates_amp_and_freq_n:phase_order:phase + (1|id), data=all, REML=F) ``` Of course, we are using MCMC samples, so the model declaration becomes: ```prettyprint model.glmm <- MCMCglmm(score ~ nstates_amp_and_freq_n:phase_order:phase, random=~id, data=all, nitt=100000, thin=100, pr = TRUE) ``` The result for the original dataset is: ```prettyprint > summary(model.glmm) Iterations = 3001:999501 Thinning interval = 500 Sample size = 1994 DIC: -8.19229 G-structure: ~id post.mean l-95% CI u-95% CI eff.samp id 0.0005342 0 3.195e-11 32.48 R-structure: ~units post.mean l-95% CI u-95% CI eff.samp units 0.04708 0.0318 0.06755 526.8 Location effects: score ~ nstates_amp_and_freq_n:phase_order:phase post.mean l-95% CI u-95% CI eff.samp pMCMC (Intercept) 0.63858 0.57880 0.70080 1994 < 5e-04 *** nstates_amp_and_freq_n:phase_order0:phase0 0.06740 -0.02253 0.15539 1994 0.14644 nstates_amp_and_freq_n:phase_order1:phase1 0.17233 -0.01143 0.36640 2316 0.06921 . nstates_amp_and_freq_n:phase_order2:phase1 0.09099 -0.06046 0.22500 1781 0.19157 nstates_amp_and_freq_n:phase_order1:phase2 0.01965 -0.08214 0.13196 2114 0.73220 nstates_amp_and_freq_n:phase_order2:phase2 0.21595 0.06910 0.36329 3173 0.00301 ** --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 ``` The resulting regression for the discrete dataset is: ```prettyprint > summary(model.glmm) Iterations = 3001:999501 Thinning interval = 500 Sample size = 1994 DIC: -25.13271 G-structure: ~id post.mean l-95% CI u-95% CI eff.samp id 0.006968 1.794e-204 0.02561 41.72 R-structure: ~units post.mean l-95% CI u-95% CI eff.samp units 0.03836 0.02075 0.05622 89.19 Location effects: score ~ nstates_amp_and_freq_n:phase_order:phase post.mean l-95% CI u-95% CI eff.samp pMCMC (Intercept) 0.81966 0.76106 0.87812 1994.0 < 5e-04 *** nstates_amp_and_freq_n:phase_order1st:phase1to1 -0.04302 -0.11980 0.03557 1994.0 0.2718 nstates_amp_and_freq_n:phase_order2nd:phase1to2 0.10226 -0.03057 0.25411 1489.9 0.1685 nstates_amp_and_freq_n:phase_order3rd:phase1to2 0.07254 -0.04080 0.18695 313.8 0.2207 nstates_amp_and_freq_n:phase_order2nd:phase2to2 0.14377 0.03020 0.26946 1994.0 0.0191 * nstates_amp_and_freq_n:phase_order3rd:phase2to2 0.07734 -0.02952 0.18401 887.3 0.1374 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 ``` \\( R^2 \\) values are 0.6153981 and 0.616292, respectively.
Tags:
From old blog
Currently unrated
1
2
3
4
5
Share on Twitter
Share on Facebook
← Big poop - Journal paper is wrong!
New regression on original dataset →
Comments
There are currently no comments
New Comment
Name
required
Email
required (not published)
Website
optional
Comment
If you enter anything in this field your comment will be treated as spam
Comments
There are currently no comments
New Comment