,
5 tweets,
5 min read
Read on Twitter
The R syntax for mixed models, "y ~ 1 + a*b + (1|id)", has two origins. The fixed part is Wilkinson notation from 1973: jstor.org/stable/2346786. The random part was introduced in the nlme package in the late 90s: stats.stackexchange.com/a/285026. It doesn't have a name. 1/5
History confirmed by nlme- and lmer-author Douglas Bates (@BatesDmbates): . So we may call it *Wilkinson-Bates notation* 2/5
It's now being massively extended in #brms. Check out this crazy mixed-effects meta-analysis w. censored y predicted by missing + monotonic + measurement error:
y | se(y_se) + cens(y_cens) ~ mi(a) + mo(b) + me(c, c_se) + (1|study)
Bayes --> infinite power! @BayesDose 3/5
y | se(y_se) + cens(y_cens) ~ mi(a) + mo(b) + me(c, c_se) + (1|study)
Bayes --> infinite power! @BayesDose 3/5
brms is being developed by Paul Bürkner (@paulbuerkner). So 10 years from now, we will call it Wilkinson-Bates-Bürkner notation or WBB for short. So the history lesson is this:
1973: W notation
1999: WB notation
2016: WBB notation
@Rbloggers @fusaroli @sharon000 4/5
1973: W notation
1999: WB notation
2016: WBB notation
@Rbloggers @fusaroli @sharon000 4/5
Eternal glory to whomever implements a text field to specify models using W, WB, or WBB syntax in a mainstream stats GUI! @jamovistats (@jjjjonathon?) or @JASPStats (@EJWagenmakers?)
END
END
