William Revelle

[William Revelle @WilliamRevelle]

@MatthewBJane psych has a number of ways to graph sem type models. See the help page for structure.diagam, and the various diagram options.

[William Revelle @WilliamRevelle]

Celebrate the summer solstice with psych 2.3.6 and psychTools 2.3.6 now on CRAN. lmCor for linear models from correlations and raw data. Several new functions in psych tools allow for easier data manipulation and merger (see vJoin, selectBy and splitBy).

[William Revelle @WilliamRevelle]

@aaronjfisher In bigCor, I compare the timings for various ways of speeding up large correla
tion matrices. Look at the help file for bigCor where I document the various mc.core options.

[William Revelle @WilliamRevelle]

@aaronjfisher For examples of doing this, look at some of the functions in psych. e.g. tetrachoric, polychoric, bigCor, bestScales, etc.
Psych will assume (if you have a mac) that parallel will use two cores. The biggest advantage is going from 1 to 2. See the documentation for bigCor /2

[William Revelle @WilliamRevelle]

@aidangcw This was the brilliance of Lew Goldberg's International Personality Item Pool and the follow up by David Condon on the Database of Individual Differences Survey Tools. Thousands of open source items measuring many constructs. Open to adaptation and addition.

[William Revelle @WilliamRevelle]

Celebrate the equinox with the latest versions (2.3.3) of psych and psychTools. Now on CRAN. Some improvements (see news(package="psych"), some inevitable bug fixes. Although suggestions for improvements are always welcome, only some will be implemented.

[William Revelle @WilliamRevelle]

@EmorieBeck Science insisted the title should be "Introversion/extroversion, time stress, and caffeine: Effect on verbal performance". With the exception of AAAS, most publishers think that extraversion is the appropriate spelling. 2/2

[William Revelle @WilliamRevelle]

@EmorieBeck The same question happened to me in 1973 when the grad school at U.M. questioned my dissertation title: "Introversion/extraversion, skin conductance and performance under stress." I convinced them that was correct. Unfortunately, Science (1976) did not agree 1/2

[William Revelle @WilliamRevelle]

@dovesandletters @ben_j_todd @GarettJones Francis Galton discussed "reversion to mediocrity" in the late 1800s. Daniel Kahneman gives a lovely example in Thinking Fast and Slow. Instructors think poor performance is enhanced by punishment, good performance diminished by reward. We need to remind people of basic stat.

[William Revelle @WilliamRevelle]

@BrentWRoberts @martin_gell @ajshackman @FullanaMa @shannon_grogans Don't confuse reliability with internal consistency. That was the contribution of the empirical school (eg. MMPI). Fred Lord has a nice paper discussing the relationship between increased validity at the expense of internal consistency.

[William Revelle @WilliamRevelle]

@ianhussey @dingding_peng The proper way to report a bug in any package in R is not to tweet it, but to send a report to the author. Include the version number of the package as well as a short example of the failure. In this case, send the report to me at [email protected]

[William Revelle @WilliamRevelle]

@ianhussey @dingding_peng Please give me the specific example of the function doing this. If it is doing it, I will fix it.

[William Revelle @WilliamRevelle]

@Jodie_Boisvert @KinglaKing If you want to make the assumption that the dichotomous variable reflects an underlying latent normally distributed variable, you can find a biserial correlation. The point biserial is just a Pearson applied to the data. See biserial in the psych package in R.

[William Revelle @WilliamRevelle]

@ISSIDOfficial @BSPID @SPSPnews @personalityEAPP @ISIRonline @PsychScience @EAPA @ACPIDCongress @centringafrica @BehaviorGenetic Looking forward to it.

[William Revelle @WilliamRevelle]

@MattMotyl @PsychRabble As we discuss in the article, this is not just the effect of increased reliability with aggregation, but is rather the effect of increased validity by adding variables that are not that highly associated. I recommend reading the original article. It is open access at POPS, 22.

[William Revelle @WilliamRevelle]

@AndrewRAConway @AnnaLSchubert @Kane_WMC_Lab @JProtzko @Kristof__Kovacs Andrew, please add me to the mailing list for this. It is great news that psychonomics has discovered individual differences. ISIR will be meeting in the bay area sometime in the late summer as well.

[William Revelle @WilliamRevelle]

@JoshuaGrubbsPhD We originally just used dumb phones and text messages, have since switched to smart phones. How do students think about serious life events (e.g. graduation). The ability to capture data about life as lived is a feature of EMA.

[William Revelle @WilliamRevelle]

@JoshuaGrubbsPhD EMA allows one to study individual differences in the structure of emotions over time. Josh Wilt and I have a number of studies (and tutorials) that suggest that the math is not that complicated. Individual differences in the correlation of positive and negative affect 1/

[William Revelle @WilliamRevelle]

@KouMurayama I tell them I am a psychometrician. That stops the conversation. Thus, nothing better to read on an airplane that "Factor analysis at 100" to stop people from wanting to talk with you. If I want to talk, I say that I study personality. That works as a starter.