Previous: Example:

Up: EIGRAPH

Next: Globals:

Options:

Choose any of the items below in place of name to see the corresponding graph. (Items that use true values from individual-level data will only work if you have vput this information into _Eres prior to running ei under the name truth. These are useful for verifying the method if individual-level data are available, but they are not useful for most ecological inference applications, where this information is not available. The options are included here only because I needed them while writing the book.)

beta

betaB and betaW

betaB

density estimate (i.e., a smooth version of a histogram) of point estimates of $\beta^b_i$'s, with whiskers.

betaW

density estimate (i.e., a smooth version of a histogram) of point estimates of $\beta^w_i$'s, with whiskers.

betabw

lines,Tlines,bivar,Tbivar. See Figure 10.8, page 214.

bias

combination of biasB, biasW, TbiasB, TbiasW. See the points in Figure 13.2, page 238.

biasB

$X_i$ by estimated $\beta_i^b$.

biasW

$X_i$ by estimated $\beta_i^w$.

bivar

estimated $\beta_i^b$ by $\beta_i^w$. See the right graph in Figure 13.6, page 243.

boundX

combination of boundXB, boundXW, boundXB overlaid on TbiasB, boundXW overlaid on TbiasW. See Figure 13.2, page 238.

boundXB

$X_i$ by the bounds on $\beta_i^b$ (each precinct appears as one vertical line), see the lines in the left graph in Figure 13.2, page 238.

boundXW

$X_i$ by the bounds on $\beta_i^w$ (each precinct appears as one vertical line), see the lines in the right graph in Figure 13.2, page 238.

estsims

All the simulated $\beta_i^b$'s by all the simulated $\beta_i^w$'s. The simulations should take roughly the shape of the mean posterior contours, except for those sampled from outlier tomography lines. Homogeneous precincts are excluded.

fit

Combination of xtfit and tomogP

fitT

Combination of xtfit and tomogT

goodman

$X_i$ by $T_i$ plot with Goodman's regression line plotted. See Figure 4.1, page 60.

lines

$X_i$ by $T_i$ plot with one estimated line per precinct. Homogenous precincts do not appear. See Figure 10.8, page 214.

movie

For each observation ( $i=1,\ldots,p$), one page of graphics appears with the posterior distribution of $\beta_i^b$ and $\beta_i^w$ (with whiskers drawn at each simulated value), a plot of simulated values of $\beta_i^b$ by $\beta_i^w$ from the tomography line) and the numerical values of $T_i$, $X_i$, $N_i$, and $i$. After the graph for each observation appears, the user can choose to view the next observation (hit return), jump to a specific observation number (type in the number and hit return), or stop (hit ``s'' and return). See Figure 8.1, page 148.

movieD

For each observation ( $i=1,\ldots,p$), one tomography plot appears (like tomogD) with the line for observation $i$ darkened. After the graph for each observation appears, the user can choose to view the next observation (hit return), jump to a specific observation number (type in the number and hit return), or stop (hit ``s'' and return). See Figure 5.1, page 81.

nonpar

combination of tomogD and a nonparametric density estimation with contour plot and surface plot representations. See Figure 9.9, page 195.

post

combination of postB and postW. See Figure 10.4, page 208.

postB

density estimate of (weighted) district quantity of interest, the posterior distribution for $B^b$. See Figure 10.4, page 208.

postW

density estimate of (weighted) district quantity of interest, the posterior distribution for $B^w$. See Figure 10.4, page 208.

prectB

plot of estimated $\beta^b_i$ by true $\beta^b_i$. See Figure 10.5, page 210.

prectW

plot of estimated $\beta^w_i$ by true $\beta^w_i$. See Figure 10.5, page 210

profile

Profile plots -- that is plots of the posterior of each element of $\phi$, holding constant all other values at their maxima. The maximum posterior point, as found by CML is displayed as a vertical line from the max down to the bottom of the graph. After seeing these plots, you can zoom in by changing _eigraph_pro (as is useful for seeing how curved the likelihood is near the maximum). (Kinks in the profile are due to imprecision in the cdfbvn function necessitating truncation by _EcdfTol.) This option is very useful for verifying whether the maximization procedure did indeed find the maximum.

profileR

Profile plots of the cdfbvn function, (Equation 6.15, page 104). These are plots of R by each element of $\phi$, holding constant all other values at their maxima. The maximum posterior point, as found by CML is displayed as a vertical line from the max down to the bottom of the graph. After seeing these plots, you can zoom in by changing _eigraph_pro. Kinks in the profile are due to imprecision in the cdfbvn function necessitating truncation by _EcdfTol. This option is very useful for verifying whether the posterior is numerical stable near the maximum.

ptile

combination of ptileB and ptileW. See Figure 10.7, page 213.

ptileB

true percentile at which $\beta_i^b$ falls by estimated $\beta_i^b$. See Figure 10.7, page 213.

ptileW

true percentile at which $\beta_i^w$ falls by estimated $\beta_i^w$. See Figure 10.7, page 213.

results

combination of postB, postW, betaB, betaW.

sims

combination of simsB and simsW. See Figure 10.6, page 212.

simsB

simulations of $\beta_i^b$ by true $\beta_i^b$. See Figure 10.6, page 212.

simsW

simulations of $\beta_i^w$ by true $\beta_i^w$. See Figure 10.6, page 212.

TbiasB

$X_i$ by true $\beta_i^b$. See the points in the left graph in Figure 13.2, page 238.

TbiasW

$X_i$ by true $\beta_i^w$. See the lines in the right graph in Figure 13.2, page 238.

Tbivar

true $\beta_i^b$ by $\beta_i^w$. See Figure 13.2, page 238.

Tlines

$X_i$ by $T_i$ plot with one true line per precinct. Homogenous precincts do not appear. See Figure 3.1, page 41.

tomog

tomography plot with ML contours. See Figure 10.2, page 204. Dashed lines are added for observations _Eselect'd out of the estimation stage, and a note appears at the bottom left and top right corners if any observations are included for which $T_i=0$ or $T_i=1$ (since as tomography lines they are represented as barely visible points). The global _EselRnd is ignored.

tomogCI

tomography plot with 80% confidence intervals. Confidence intervals appear on the screen in red with the remainder of the tomography line in yellow. On printed output, which is often easier to see than the screen display, the confidence interval portion is printed thicker than the rest of the line. See Figure 9.5, page 179.

tomogCI95

tomography plot with 95% confidence intervals. Confidence intervals appear on the screen in red with the remainder of the tomography line in yellow. On printed output, which is often easier to see than the screen display, the confidence interval portion is printed thicker than the rest of the line. See Figure 9.5, page 179.

tomogD

tomography plot with data only. See Figure 5.1, page 81. Dashed lines are added for observations _Eselect'd out of the estimation stage, and a note appears at the bottom left and top right corners if any observations are included for which $T_i=0$ or $T_i=1$ (since as tomography lines they are represented as barely visible points). The global _EselRnd is ignored.

tomogE

tomography plot with estimated mean posterior $\beta_i^b$ and $\beta_i^w$ points. Dashed lines are added for observations _Eselect'd out of the estimation stage, and a note appears at the bottom left and top right corners if any observations are included for which $T_i=0$ or $T_i=1$ (since as tomography lines they are represented as barely visible points). The global _EselRnd is ignored.

tomogP

tomography plot with mean posterior contours. Dashed lines are added for observations _Eselect'd out of the estimation stage, and a note appears at the bottom left and top right corners if any observations are included for which $T_i=0$ or $T_i=1$ (since as tomography lines they are represented as barely visible points). The global _EselRnd is ignored.

tomogS

combination of tomog, tomogp, tomogCI, and Tbivar (or estsims if truth isn't available). See Figures 10.2 (page 204), 9.5 (page 179), 7.1 (page 126). Dashed lines are added for observations _Eselect'd out of the estimation stage, and a note appears at the bottom left and top right corners if any observations are included for which $T_i=0$ or $T_i=1$ (since as tomography lines they are represented as barely visible points). The global _EselRnd is ignored.

tomogT

tomography plot with true $\beta_i^b$ and $\beta_i^w$ points. See Figure 7.1 (page 126). Dashed lines are added for observations _Eselect'd out of the estimation stage, and a note appears at the bottom left and top right corners if any observations are included for which $T_i=0$ or $T_i=1$ (since as tomography lines they are represented as barely visible points). The global _EselRnd is ignored.

truth

combination of post, precB, precW (compare truth to estimates at district and precinct-level). See Figures 10.4 (page 208) and 10.5 (page 210).

xgraph

a scattercross with data plotted. See Figure 12.1, page 227.

XgraphC

a scattercross with data plotted, with size proportional to $N_i$.

xt

basic $X_i$ by $T_i$ scatterplot

xtC

basic $X_i$ by $T_i$ scatterplot with circles sized proportional to ${N_i}$ or some other variable defined by the global _eigraph_circ.

xtfit

$X_i$ by $T_i$ plot with estimated ${\text{E}}(T_i\vert X_i)$ and conditional 80% confidence intervals. See Figure 10.3, page 206.

xtfitg

xtfit with Goodman's regression line superimposed