TY - JOUR
T1 - A Diamond-Shaped Equiponderant Graphical Display of the Effects of Two Categorical Predictors on Continuous Outcomes
AU - Li, Xiuhong
AU - Buechner, Jennifer M.
AU - Tarwater, Patrick M.
AU - Munoz, Alvaro
N1 - Funding Information:
Xiuhong Li is Senior Statistical PrograrmD/ta Analyamst, Joehns Hopkins Bloombeg Srchool of Public Health, 615 N. Wofe StrleeRotom,E-7005, Baltimore, MD 21205. Jnneifer M. Buechner is Prograr mAnalyst,mUniveersity of Wisnsinc–Madoison, Dpartmente of Biostatistics and Medical Informatics, Room 239WA–BuRilding,F610 WautlStnreeMadt,ison, WI 53726. PaiktcMr. Tarwis AssistanatteProrfe, UnsivseoofrTrexsasiScho,tolyof Public Health, 1100 N. Stanton, Suite 110, El Paso, TX 79902. Alvaro Mñoz isuProfeor, ss Johns Hopkins Bloomberg School of Public Health, 615 N. Wolfe StreeRotom, E-7008, Baltimore, MD 21205 (E-mail: amnoz@juhsph.edu). This work was supported by grants UO1-AI-35043 and UO1-AI-42590 from the U.S. Ntionala Institute of Agylanld Infectioerus Disaesand ePO1,-ES-06052 from the National Institute of EvironnmetalnHeaSciences.lthThe authors are grateful to DvidGeorgea foreditorialassistance.
PY - 2003/8
Y1 - 2003/8
N2 - Three-dimensional (3-D) bar graphs and their current 2-D alternatives have certain limitations when the primary objective is to provide an equal representation of the effects of two predictors on an outcome. This article proposes a graphing methodology (in the shape of a diamond) that projects 3-D bar graphs into 2-D whereby the third dimension is replaced with a polygon whose area and middle vertical and horizontal lengths represent the outcome. The proposed graphical representation is invariant to rotations and avoids outcomes in categories being concealed by others. This article shows several applications of our proposal to a variety of data types (e.g., proportions, incidence rates, relative risks), which can be easily implemented with available software.
AB - Three-dimensional (3-D) bar graphs and their current 2-D alternatives have certain limitations when the primary objective is to provide an equal representation of the effects of two predictors on an outcome. This article proposes a graphing methodology (in the shape of a diamond) that projects 3-D bar graphs into 2-D whereby the third dimension is replaced with a polygon whose area and middle vertical and horizontal lengths represent the outcome. The proposed graphical representation is invariant to rotations and avoids outcomes in categories being concealed by others. This article shows several applications of our proposal to a variety of data types (e.g., proportions, incidence rates, relative risks), which can be easily implemented with available software.
KW - Epidemiology
KW - Graphical methods
KW - Statistics
KW - Three-dimensional bar graphs
KW - Two-dimensional projections
UR - http://www.scopus.com/inward/record.url?scp=0043075798&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0043075798&partnerID=8YFLogxK
U2 - 10.1198/0003130031883
DO - 10.1198/0003130031883
M3 - Article
AN - SCOPUS:0043075798
SN - 0003-1305
VL - 57
SP - 193
EP - 199
JO - American Statistician
JF - American Statistician
IS - 3
ER -