Number of Equations and Citations for Evolutionary Biology Publications
EquationCitations.RdAnalysis of citations of evolutionary biology papers published in 1998 in the top three journals (as judged by their 5-year impact factors in the Thomson Reuters Journal Citation Reports 2010).
Usage
data("EquationCitations")Format
A data frame containing 649 observations on 13 variables.
- journal
Factor. Journal in which the paper was published (The American Naturalist, Evolution, Proceedings of the Royal Society of London B: Biological Sciences).
- authors
Character. Names of authors.
- volume
Volume in which the paper was published.
- startpage
Starting page of publication.
- pages
Number of pages.
- equations
Number of equations in total.
- mainequations
Number of equations in main text.
- appequations
Number of equations in appendix.
- cites
Number of citations in total.
- selfcites
Number of citations by the authors themselves.
- othercites
Number of citations by other authors.
- theocites
Number of citations by theoretical papers.
- nontheocites
Number of citations by nontheoretical papers.
Details
Fawcett and Higginson (2012) investigate the relationship between the number of citations evolutionary biology papers receive, depending on the number of equations per page in the cited paper. Overall it can be shown that papers with many mathematical equations significantly lower the number of citations they receive, in particular from nontheoretical papers.
Source
Online supplements to Fawcett and Higginson (2012).
https://www.pnas.org/doi/suppl/10.1073/pnas.1205259109/suppl_file/sd01.xlsx
References
Fawcett, T.W. and Higginson, A.D. (2012). Heavy Use of Equations Impedes Communication among Biologists. PNAS – Proceedings of the National Academy of Sciences of the United States of America, 109, 11735–11739. doi:10.1073/pnas.1205259109
Examples
## load data and MASS package
data("EquationCitations", package = "AER")
library("MASS")
## convenience function for summarizing NB models
nbtable <- function(obj, digits = 3) round(cbind(
"OR" = exp(coef(obj)),
"CI" = exp(confint.default(obj)),
"Wald z" = coeftest(obj)[,3],
"p" = coeftest(obj)[, 4]), digits = digits)
#################
## Replication ##
#################
## Table 1
m1a <- glm.nb(othercites ~ I(equations/pages) * pages + journal,
data = EquationCitations)
m1b <- update(m1a, nontheocites ~ .)
m1c <- update(m1a, theocites ~ .)
nbtable(m1a)
#> OR 2.5 % 97.5 % Wald z p
#> (Intercept) 28.673 20.701 39.716 20.189 0.000
#> I(equations/pages) 0.781 0.656 0.930 -2.782 0.005
#> pages 1.049 1.024 1.074 3.929 0.000
#> journalEvolution 0.947 0.761 1.177 -0.494 0.622
#> journalProcB 1.137 0.905 1.429 1.102 0.270
#> I(equations/pages):pages 1.016 0.997 1.037 1.636 0.102
nbtable(m1b)
#> OR 2.5 % 97.5 % Wald z p
#> (Intercept) 20.931 14.781 29.639 17.135 0.000
#> I(equations/pages) 0.732 0.606 0.884 -3.244 0.001
#> pages 1.049 1.023 1.076 3.692 0.000
#> journalEvolution 1.070 0.848 1.351 0.573 0.567
#> journalProcB 1.215 0.952 1.551 1.565 0.118
#> I(equations/pages):pages 1.010 0.989 1.032 0.937 0.349
nbtable(m1c)
#> OR 2.5 % 97.5 % Wald z p
#> (Intercept) 6.137 4.173 9.025 9.219 0.000
#> I(equations/pages) 0.966 0.791 1.180 -0.338 0.735
#> pages 1.050 1.021 1.080 3.379 0.001
#> journalEvolution 0.702 0.543 0.908 -2.692 0.007
#> journalProcB 0.926 0.707 1.212 -0.561 0.575
#> I(equations/pages):pages 1.029 1.006 1.052 2.443 0.015
## Table 2
m2a <- glm.nb(
othercites ~ (I(mainequations/pages) + I(appequations/pages)) * pages + journal,
data = EquationCitations)
m2b <- update(m2a, nontheocites ~ .)
m2c <- update(m2a, theocites ~ .)
nbtable(m2a)
#> OR 2.5 % 97.5 % Wald z p
#> (Intercept) 29.109 21.018 40.315 20.287 0.000
#> I(mainequations/pages) 0.724 0.572 0.918 -2.673 0.008
#> I(appequations/pages) 0.993 0.649 1.519 -0.032 0.975
#> pages 1.048 1.023 1.074 3.856 0.000
#> journalEvolution 0.943 0.759 1.173 -0.523 0.601
#> journalProcB 1.127 0.897 1.416 1.029 0.303
#> I(mainequations/pages):pages 1.022 0.995 1.051 1.572 0.116
#> I(appequations/pages):pages 0.998 0.956 1.043 -0.073 0.941
nbtable(m2b)
#> OR 2.5 % 97.5 % Wald z p
#> (Intercept) 21.237 15.001 30.065 17.230 0.000
#> I(mainequations/pages) 0.648 0.502 0.837 -3.330 0.001
#> I(appequations/pages) 1.074 0.679 1.698 0.305 0.760
#> pages 1.048 1.022 1.075 3.608 0.000
#> journalEvolution 1.068 0.846 1.347 0.553 0.580
#> journalProcB 1.205 0.945 1.538 1.503 0.133
#> I(mainequations/pages):pages 1.024 0.994 1.054 1.534 0.125
#> I(appequations/pages):pages 0.973 0.928 1.019 -1.161 0.246
nbtable(m2c)
#> OR 2.5 % 97.5 % Wald z p
#> (Intercept) 6.143 4.177 9.036 9.221 0.000
#> I(mainequations/pages) 0.958 0.729 1.258 -0.311 0.755
#> I(appequations/pages) 0.983 0.602 1.607 -0.067 0.946
#> pages 1.050 1.021 1.080 3.365 0.001
#> journalEvolution 0.702 0.543 0.908 -2.691 0.007
#> journalProcB 0.926 0.707 1.212 -0.562 0.574
#> I(mainequations/pages):pages 1.030 0.997 1.063 1.788 0.074
#> I(appequations/pages):pages 1.027 0.976 1.079 1.022 0.307
###############
## Extension ##
###############
## nonlinear page effect: use log(pages) instead of pages+interaction
m3a <- glm.nb(othercites ~ I(equations/pages) + log(pages) + journal,
data = EquationCitations)
m3b <- update(m3a, nontheocites ~ .)
m3c <- update(m3a, theocites ~ .)
## nested models: allow different equation effects over journals
m4a <- glm.nb(othercites ~ journal / I(equations/pages) + log(pages),
data = EquationCitations)
m4b <- update(m4a, nontheocites ~ .)
m4c <- update(m4a, theocites ~ .)
## nested model best (wrt AIC) for all responses
AIC(m1a, m2a, m3a, m4a)
#> df AIC
#> m1a 7 6211.517
#> m2a 9 6214.349
#> m3a 6 6209.166
#> m4a 8 6204.002
nbtable(m4a)
#> OR 2.5 % 97.5 % Wald z p
#> (Intercept) 13.321 7.869 22.551 9.640 0.000
#> journalEvolution 0.872 0.684 1.111 -1.111 0.266
#> journalProcB 1.203 0.940 1.540 1.471 0.141
#> log(pages) 1.762 1.438 2.158 5.469 0.000
#> journalAmNat:I(equations/pages) 0.928 0.791 1.089 -0.914 0.361
#> journalEvolution:I(equations/pages) 1.083 0.922 1.273 0.970 0.332
#> journalProcB:I(equations/pages) 0.814 0.743 0.893 -4.368 0.000
AIC(m1b, m2b, m3b, m4b)
#> df AIC
#> m1b 7 5847.891
#> m2b 9 5850.161
#> m3b 6 5844.457
#> m4b 8 5835.272
nbtable(m4b)
#> OR 2.5 % 97.5 % Wald z p
#> (Intercept) 10.736 6.131 18.799 8.304 0.000
#> journalEvolution 0.925 0.715 1.197 -0.591 0.554
#> journalProcB 1.237 0.951 1.608 1.585 0.113
#> log(pages) 1.714 1.381 2.127 4.888 0.000
#> journalAmNat:I(equations/pages) 0.770 0.648 0.915 -2.972 0.003
#> journalEvolution:I(equations/pages) 1.007 0.848 1.197 0.082 0.935
#> journalProcB:I(equations/pages) 0.712 0.643 0.789 -6.513 0.000
AIC(m1c, m2c, m3c, m4c)
#> df AIC
#> m1c 7 4226.051
#> m2c 9 4230.046
#> m3c 6 4225.965
#> m4c 8 4225.754
nbtable(m4c)
#> OR 2.5 % 97.5 % Wald z p
#> (Intercept) 2.359 1.252 4.443 2.656 0.008
#> journalEvolution 0.694 0.520 0.925 -2.488 0.013
#> journalProcB 1.031 0.769 1.383 0.204 0.839
#> log(pages) 1.895 1.485 2.418 5.138 0.000
#> journalAmNat:I(equations/pages) 1.369 1.138 1.647 3.331 0.001
#> journalEvolution:I(equations/pages) 1.398 1.159 1.687 3.500 0.000
#> journalProcB:I(equations/pages) 1.121 1.008 1.246 2.113 0.035
## equation effect by journal/response
## comb nontheo theo
## AmNat =/- - +
## Evolution =/+ = +
## ProcB - - =/+