Title
tabxml - Save results in XML format for use in Microsoft Excel and OpenOffice Calc
Syntax
tabxml [, options]
options Description ------------------------------------------------------------------------- variables dv(varname) dependent variable ivd(varlist) independent discrete variables ivc(varlist) independent continuous variables split(varname) split data into subgroups
Output save(string) name of the output file options(varlist) select output for independent variables order(varlist) order of independent variables display display output
Conditions cond(string) condition - modifies total population size subcond(string) sub condition - creates a sub population
Table Formatting percent(string) select percent type for discrete variables tformat(string) select decimal places for output justify(string) data alignment rotate(string) rotate dependent variable labels labrown(string) determines the way the 'n' counts are displayed for the independent variables labcoln(string) determines the way the 'n' counts are displayed for the dependent variable bold make independent variable names bold olines omit lines between independent variables otab omit indentations from indepedent variable labels sp show the % symbol for all percents chi2 display chi2 statistic for all independent variables chidec(integer) number of decimal places for chi2 statistic pval display p-value for all independent variables pdec(integer) number of decimal places for p-value
Excel/Calc Paths excelpath(string) select directory for Excel calcpath(string) select directory for Calc
svy options osvy data is not svy
Description
tabxml export Stata output directly into an XML file that could be opened with Microsoft Excel or OpenOffice Calc. The program is relatively flexible and produces journal quality tables in Microsoft Excel or OpenOffice Calc; and allows users to apply different formats to the elements of the output table and > essentially do everything Microsoft Excel or OpenOffice Calc can do in terms of formatting from within Stata as oppose to using an external package like LaT > ex. A key feature includes displaying trailing zeros after decimal places.
tabxml can create formatted tables of frequencies and/or percents for discrete independent variables, means and standard errors / confidence intervals for con > tinuous independent variables. Two-way tables can be ceated by selecting a dependent va > riable. Multiple independent variables are permitted.
Options
+--------+ ----+ Output +-----------------------------------------------------------
save(["]filename["]) specifies a name for XML file where tables are outputted. Files are saved in a folder named "Excel" which is created (if missing) in the > current directory. If save(["]filename["]) is omitted, the output will be saved in tab_ > out.xml located in the current working directory.
options(varlist)} modify output type for discrete and continuous independent variables. Output type options for discrete independent variables include (n) d > isplayed "n" or (per) displayed "per". Note default output is "n (per)". Output type opt > ions for continuous independent variables include (mean) displayed "mean", or (mean ci) > displayed "mean (CI)". Note default output is "n (SE)". If split is specified option(noto > tal) only outputs the subgroups of split and omits the totals.
order(varlist)} rearrange order of independent variables. Default order of independent variables is ivd(varlist) followed by ivc(varlist). To rearrange the order of the independent variables write the varlist of discrete independen > t variables and continuous independent variables in the order you wish them to ap > pear. display displays the output in Stata.
+------------+ ----+ Conditions +-------------------------------------------------------
cond(string) reduces the population size of the dataset to meet the condition. For example cond(panel==1) reduces the populations size such that the variable > panel equals one.
subcond(string) creates a sub population of the total population. For example subcond(age>30) creates a sub population such that the age of the respondents is greater than 30.
+------------------+ ----+ Table Formatting +-------------------------------------------------
percent(string) tables for independent discrete variables can be either row or column (col) percents. Default is column percents.
tformat(string) modify the number of decimal places for table percents for disc > rete independent variables; and means, confidence intervals and standard errors for > continuous independent variables. Default is two decimal places.
justify(string) justify the output of the data. Left (l), right (r) and centre > (c) justify. Default justify is left align.
rotate(string) rotate the variable labels of the dependent variable. For exampl > e, rotate(60) produces a pleasing effect. Default is horizontal variable names.
labrown(string) if this option is specified counts are now included in the variable labels of the independent variables. For example, label(n=) produces the variable label followed by (n=#) where # is the count.
labcoln(string) if this option is specified counts are now included in the vari > able labels of the dependent variable. For example, label(n=) produces the variable label f > ollowed by (n=#) where # is the count.
bold makes the independent variable names bold face.
olines omit lines between independent variables.
otab omit indentations from indepedent variable lables.
sp show the % symbol for all percents.
chi2 display chi2 statistic for all independent variables.
chidec(integer) number of decimal places for chi2 statistic. Default is 2 decim > al places.
pval display p-value for all independent variables.
pdec(integer) number of decimal places for p-value. Default is 2 decimal places.
+------------------+ ----+ Excel/Calc Paths +-------------------------------------------------
excelpath(string) select the full directory for Excel to open the file from Sta > ta. Default is "C:/Program Files/Microsoft Office/office*". Where * represents the > current version of office. This option is particulary useful when Office is not on the > C:\ drive.
calcpath(string) select the full directory for Calc to open the file from Stata > . Default is "C:/Program Files/OpenOffice*/program". Where * represents the curre > nt version of OpenOffice. This option is particulary useful when OpenOffice is not > on the C:\ drive.
+-------------+ ----+ svy options +------------------------------------------------------
osvy data is not svy. No survey characteristics are set
Examples
The following uses the sample data set multistage. To access this dataset and d > eclare the dataset is survey data use the following code:
.use http://www.stata-press.com/data/r10/multistage .svyset county [pw=sampwgt], strata(state) fpc(ncounties) || school, fpc(nschools)
+--------------+ ----+ Basic Syntax +-----------------------------------------------------
.tabxml, dv(county) ivd(race) ivc(weight)
In this example, tabxml display a cross tab of county against the discrete independent variable race and the continuous independent variable weight in tha > t specific order. Variable labels for independent variables have indentations and > there are lines between each independent variable. The output for the discrete independent variables is "n (per)", similarily the output for the continuous independent variable is "mean (SE)". The percents are column percents and "per" > , "mean" and "SE" are all rounded to two decimal places. All data is left aligned > . The file is saved as "tab_out.xml". The output is omitted in Stata.
Extensions: options split percent tformat save subcond justify
.tabxml, dv(county) ivd(race) split(sex) save(temp) tf(1) subcond(weight>200) justify(r)
In this example, tabxml display a cross tab of county against the discrete independent variable race split by sex (female, male and total are the subgroup > s) for the sub population where the weight is greater than 200lbs. Variable labels > for independent variables have indentations and there are lines between either side > of the independent variable. The output for the discrete independent variable i > s "n (per)". The "per" is rounded to one decimal place. All data is right aligned > . The file is saved as "temp.xml". The output is omitted in Stata.
Extensions: options options order display bold olines otab
.tabxml, dv(county) ivd(race) ivc(weight) split(sex) cond(height>=400) order(weight race) options(n mean nototal) display bold olines otab
In this example, tabxml display a cross tab of county against the continuous independent variable weight and the discrete independent variable race in that specific order split by sex (female and male are the subgroups). The dataset is > reduced where race equals one (white). Variable labels for independent variable > s do not have indentations and there are no lines between each independent variab > le. The independent variable names are bolded. The output for the discrete independ > ent variables is "n", similarily the output for the continuous independent variable > is "mean". The percents are col percents and "per", "mean" are all rounded to t > wo decimal place. All data is left aligned. The file is saved as "tab_out.xml". The output is omitted in Stata.
Extensions: options pval chi2 pdec sp label options excelpath
.tabxml, dv(county) ivd(race) options(per) display pval pdec(3) chi2 sp labrown(sample count=)
In this example, tabxml display a cross tab of county against the discrete independent variable race. Variable labels for the independent variable have indentations and there are lines either side of the independent variable. The independent variable name is not bolded. The output for the discrete independen > t variable is "per". The percents are col percents and "per" are rounded to two decimal places. The percents are followed the % symbol. All data is left aligne > d. The file is saved as "tab_out.xml". The independet variable labels are followed > by (sample count=#). The p-val statistic is displayed and rounded to 3 decimal places. The chi2 statistic is displayed and rounded to 2 decimal places. The ou > tput is displayed in Stata.
Authors
Richard Ryall ARCSHS (Australian Research Centre in Sex, Health and Society) Email: R.Ryall@latrobe.edu.au
Jason Ferris ARCSHS (Australian Research Centre in Sex, Health and Society) Email: J.Ferris@latrobe.edu.au