Title
gendist -- Generates distances for a battery of spatial placements
Syntax
gendist varlist [, options]
options Description ------------------------------------------------------------------------- respondent(varname) (required) the variable containing the respondent's self-placement in the space (e.g. the issue space) in which items (e.g. the political parties) have been placed. contextvars(varlist) a set of variables identifying different electoral contexts (by default all cases are treated as part of the same context). stackid(varname) a variable identifying different "stacks", for which distances will be separately generated if gendist is issued after stacking. nostack override the default behavior that treats each stack as a separate context. missing(mean|same|diff) plugs missing values on object placements round rounds plugged values to the nearest integer pprefix(name) prefix for generating mean-plugged placement variables (default is "p_") mprefix(name) prefix for generating variables indicating original missingness of either component (item location or respondent location) of a distance measure (default is "m_") dprefix(name) prefix for generating distance variables (default is "d_") mcountname(name) name of a generated variable reporting original count of missing items for each case (default is "_gendist_mc") mpluggedcountname(name) name of a generated variable reporting the count of missing items for each case after mean-plugging (default is "_gendist_mpc") replace drops all party location variables and mean-plugged placement variables after the generation of distances.
-------------------------------------------------------------------------
Description
gendist generates Euclidean distances for a battery of spatial items, where variables in varlist contain the placement of different objects on the spatial scale and the variable specified in respondent contains the self-placement of the respondent on the same spatial scale. Distances between the respondent and each spatial item in the battery are placed in corresponding members of a new battery of items. Only one battery of items can be processed on a single invocation of gendist.
The items in the new battery are given names derived from appending the names in varlist to the prefix established in option dprefix (default d_).
If optioned by missing, gendist also generates a new battery of items with the prefix established in option pprefix (default p_) which is identical to the original battery but with missing values plugged by mean values. These mean values can be mean placements (e.g. of political parties on the left-right scale) by all respondents, mean placements by respondents who themselves have the same position as the placement, or mean placements by respondents themselves having a different position, depending on what is specified in option missing.
Conventionally in published work the plugged value has been based on all placements. However, it might be thought that respondents having the same position would be more knowledgeable about the object concerned. Alternatively it might be thought that respondents having the same position might include individuals who were simply assuming that 'their' party had the same position as they did. Each of the missing options is defensible theoretically so the user should think carefully about which to employ. The default is not to plug the missing data, so that distances are generated only for valid cases.
The gendist command can be issued before or after stacking. If issued after stacking, by default it treats each stack as a separate context to take into account along with any higher-level contexts. However, the nostack option can be employed to force gendist to ignore the stack-specific contexts. In addition, this command can be employed with or without distinguishing between higher-level contexts, if any, (with or without the contextvars option) depending on what makes methodological sense. NOTE that it is unlikely to make methodological sense to employ gendist after stacking along with both the nostack and the mean options, since this would result in missing values being plugged with a mean that combined the values of what were (before stacking) several different variables.
SPECIAL NOTE ON MULTIPLE BATTERIES. Gendist is only aware of the battery it is currently processing. Thus it cannot diagnose an error if that battery is of a different length than other batteries of items pertaining to the objects (eg political parties) being asked about. Yet stacked datasets (the type of datasets for which distances are wanted) absolutely require all batteries pertaining to the objects being stacked to contain the same number of items and have these items in the correct sequential order (gendist will produce stacks in the correct order, padded as needed with stacks that contain only missing values, if the numeric suffixes to all batteries of items are correct). In datasets derived from election studies is is quite common for some questions (eg about party locations on certain issues) to be asked only for a subset of the objects being investigated (eg parties). Moreover, questions relating to those objects may not always list them in the same order. If the user employs tab1 or gendummies to generate a battery of dummy variables corresponding to questions that did not list all parties or listed them in a different order then not only may the number of items in the resulting battery be different from those in another battery but also the numeric suffixes generated by tab1 or gendummies may refer to different objects in the case of items from the different batteries. One part of this problem is alleviated by the use of gendummies which generates dummy variable suffixes from the values actually found in the data, rather then numbering these sequentially as does tab1. But those values do need to be correct, which only the user can check. See also the special note on multiple batteries in the help text for genstacks.
Options
respondent(varname) (required) the variable containing the respondent's self-placement on the battery of items.
contextvars(varlist) if present, variables whose combinations identify different electoral contexts (by default all cases are assumed to belong to the same context).
stackid(varname) if specified, a variable identifying different "stacks", for which distances will be separately generated in the absence of the nostack option. The default is to use the "genstacks_stack" variable if the gendist command is issued after stacking.
nostack if present, overrides the default behavior of treating each stack as a separate context (has no effect if data are not stacked).
missing(mean|same|diff) if present, determines treatment of missing values for object placement variables (by default they remain missing). If mean is specified, missing values are replaced with the overall mean placement of that object, calculated on the whole sample. If same is specified, missing values are replaced with the mean placement of the object, calculated only among those respondents that placed themselves on the same position as the object. If diff is specified, missing values are replaced with the mean placement of the object, calculated only among those respondents who placed themselves on a different position than the object (see discussion under 'Description' above regarding choice between these options). When missing values are plugged, a set of p_varlist variables is generated, and the original variables are left unchanged (the p_ prefix can be altered by use of the option pprefix). NOTE: More sophisticated imputation facilites are offered by iimpute.
round if present, causes rounding of all plugged values to the closest integer.
dprefix(name) if present, provides a prefix for generated distance variables (default is "d_").
pprefix(name) if present, provides a prefix for generated mean-plugged placements (default is "p_").
mprefix(name) if present, provides a prefix for generated variables indicating for each case whether, before mean-plugging of an item in the battery, either the item placement or the respondent placement was missing for that case (default is "m_").
mcountname(name) if specified, name of a generated variable reporting original number of missing items (default is "_gendist_mc")
mpluggedcountname(name) if specified, name of a generated variable reporting number of missing items after mean-plugging, which could still be non-zero (even after all missing values on item positions have been plugged) if the respondent's own self-placement is missing (default is "_gendist_mpc")
replace if specified, drops all party position and mean-plugged placement variables after the generation of distance measures
Examples:
The following command generates distances on a left-right dimension, where party placements are in variables lrp1-lrp10, and R's self-placement is in lrresp; missing placements are replaced by simple mean-plugging, and then rounded to the nearest integer.
. gendist lrp1-lrp10, respondent(lrresp) missing(mean) round
Generated variables
gendist saves the following variables and variable sets:
p_var1 p_var2 ... (or other prefix set by option pprefix) a set of mean-plugged placement variables with names p_var1, p_var2, etc., where the names var1, var2, etc. match the original variable names. Those variables are left unchanged. m_var1 m_var2 ... (or other prefix set by option mprefix) a set of variables with names m_var1, m_var2, etc., where the names var1, var2, etc. match the original variable names of the battery of items. These variables indicate the original missingness of var1, var2, etc., or of the corresponding placement of the respondent on the scale concerned. d_var1 d_var2 ... (or other prefix set by option dprefix) a set of distances from the respondent to each (mean-plugged if optioned) placement variable. These distance variables are named d_var1, d_var2, etc., where the names var1, var2, etc. match the original variable names. Those variables are left unchanged. _gendist_mc a variable showing the original count of missing items for each case. _gendist_mpc a variable showing the count of remaining missing items for each case after mean-plugging.
NOTE that a subsequent invocation of gendist will replace _gendist_mc and _gendist_mpc with new counts of missing values for that invocation of gendist. So the user should save these values after issuing the