                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                   COUNTY LONGITUDINAL TEMPLATE, 1840-1990                   
                                                                             
                                (ICPSR 6576)                                 
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                           Principal Investigators                           
                                                                             
                              Patrick M. Horan                               
                            University of Georgia                            
                                                                             
                                    and                                      
                                                                             
                               Peggy G. Hargis                               
                         Georgia Southern University                         
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                             First ICPSR Version                             
                                October 1995                                 
                                                                             
                                                                             
                                                                             
                       Inter-university Consortium for                       
                        Political and Social Research                        
                                P.O. Box 1248                                
                          Ann Arbor, Michigan 48106                          
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
-                                                                            
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                             BIBLIOGRAPHIC CITATION                          
                                                                             
                                                                             
            Publications  based  on  ICPSR  data  collections  should        
            acknowledge  those  sources  by  means  of  bibliographic        
            citations.  To ensure that such source  attributions  are        
            captured  for  social  science  bibliographic  utilities,        
            citations must appear in footnotes or  in  the  reference        
            section  of publications.  The bibliographic citation for        
            this data collection is:                                         
                                                                             
                                                                             
                 Horan, Patrick M., and Peggy G. Hargis.  COUNTY             
                 LONGITUDINAL   TEMPLATE,   1840-1990  [Computer             
                 file]. ICPSR version.  Athens, GA:  Patrick  M.             
                 Horan,   University   of   Georgia,   Dept.  of             
                 Sociology/Statesboro,  GA:  Peggy  G.   Hargis,             
                 Georgia Southern University, Dept. of Sociology             
                 and Anthropology [producers], 1995. Ann  Arbor,             
                 MI:  Inter-university  Consortium for Political             
                 and Social Research [distributor], 1995.                    
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                REQUEST FOR INFORMATION ON USE OF ICPSR RESOURCES            
                                                                             
            To provide funding agencies  with  essential  information        
            about  use  of  archival  resources and to facilitate the        
            exchange  of  information   about   ICPSR   participants'        
            research activities, users of ICPSR data are requested to        
            send to ICPSR bibliographic citations for each  completed        
            manuscript or thesis abstract. Please indicate in a cover        
            letter which data were used.                                     
                                                                             
                                                                             
                                                                             
                                 DATA DISCLAIMER                             
                                                                             
            The original  collector  of  the  data,  ICPSR,  and  the        
            relevant  funding  agency bear no responsibility for uses        
            of this collection or for interpretations  or  inferences        
            based upon such uses.                                            
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
-                                                                            
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                         DATA COLLECTION DESCRIPTION                         
                                                                             
                                                                             
     Patrick M. Horan and Peggy G. Hargis                                    
          COUNTY LONGITUDINAL TEMPLATE, 1840-1990 (ICPSR 6576)               
                                                                             
     SUMMARY: The County Longitudinal Template is a  tool  that  enables     
     researchers  to  allow  for  temporal  changes  in  the  geographic     
     boundaries of counties in the United States. These data  provide  a     
     decade-by-decade  account  of  the  administrative  status  of each     
     county, starting in  1990  and  tracing  each  census  period  back     
     through  1840.  The first four variables are the county name, ICPSR     
     state code, FIPS code, and ICPSR county code. These four  variables     
     allow  the  researcher  to  select  the  counties  for the state in     
     question. The next 16 variables are ID variables  for  each  census     
     year,  1990 back to 1840. The last 13 variables are boundary change     
     flags for each census year from 1960 to 1840.                           
                                                                             
     UNIVERSE: All counties in the United States.                            
                                                                             
     NOTE: This collection is structured as list, or free  field,  data,     
     with  one  case  in each record. Each record has the same number of     
     variables, but the actual record length of each line varies due  to     
     the differing width of certain variables. The maximum record length     
     in the file is 114.                                                     
                                                                             
     EXTENT OF COLLECTION: 1 data file + machine-readable  documentation     
     (text) + SPSS data definition statements                                
                                                                             
     EXTENT  OF  PROCESSING:   MDATA.PR/   REFORM.DOC/   UNDOCCHK.ICPSR/     
     DDEF.ICPSR                                                              
                                                                             
     DATA FORMAT:  Logical  Record  Length  with  SPSS  data  definition     
     statements                                                              
                                                                             
     Cases: 3,189                                                            
     Variables: 33                                                           
     Record Length: 114                                                      
     Records Per Case: 1                                                     
                                                                             
     RELATED PUBLICATIONS:                                                   
          Land, Kenneth C., Glenn Deane, and Judith R. Blau.  "Religious     
     Pluralism  and  Church  Membership:  A  Spatial  Diffusion  Model."     
     AMERICAN SOCIOLOGICAL REVIEW 52 (1991), 237-249.                        
          Hargis, Peggy G. "In Search of  the  Golden  Fleece:  African-     
     American  Landholding  in  Georgia,  1874-1933"  (Unpublished Ph.D.     
     dissertation).  Athens, GA: University of Georgia, 1994.                
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
             Documentation for the 'County Longitudinal Template' *          
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                 Patrick M. Horan                            
                             Department of Sociology                         
                              University of Georgia                          
                              Athens, Georgia  30602                         
                                                                             
                                       and                                   
                                                                             
                                 Peggy G. Hargis                             
                     Department of Sociology and Anthropology                
                           Georgia Southern University                       
                            Statesboro, Georgia 30460                        
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
      *Research support was provided by the Georgia Agricultural Experiment  
      Station in support of U.S. Department of Agriculture Regional Project  
            S-184: "Local Markets and Labor  Force Differentiation in        
                             Nonmetropolitan Areas."                         
                                                                             
                                                                            
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                             
                                                                            
                                                                             
             The 'County Longitudinal Template' as a Research Tool           
                                                                             
          Students of U.S. social, economic and political organization are   
     becoming increasingly aware of time and space as important dimensions   
     for empirical inquiry.  This awareness takes a variety of forms, but a  
     central premise for many is that in order to understand contemporary    
     configurations in social, economic and political structures, they must  
     be recognized as transformations of earlier structural arrangements.    
     Hence, in the study of social structures and processes which vary across
     geographic areas, both the temporal sequences of events and the         
     characteristics of local areas at different time periods become critical
     for understanding outcomes.  While this recognition is not new, it      
     raises important practical problems for the use of existing aggregate   
     data sources, and requires important departures from the cross-sectional
     research designs of conventional survey or census data sources.         
          At first glance, the preparation of a data matrix for longitudinal 
     comparisons of local areas might appear to be a relatively              
     straightforward task.  After all, the United States, like most Western  
     industrial countries, routinely collects and reports census data for    
     local geographic areas, such as counties.  For any particular decade, a 
     conventional file containing social and demographic variables for each  
     county unit that was in existence at the beginning of that decade can be
     readily constructed from published U.S. Census sources, or obtained from
     machine- readable versions which have been prepared by the              
     Inter-university Consortium for Political and Social Research (ICPSR    
     1995).{1}  Such a file can be conceptualized as a basic counties-by-    
     variables data matrix for every decade, a matrix in which each county is
     represented by a row in the matrix and each variable by a column.  On   
     the surface it may appear to be a simple task to match names (or the    
     associated FIPS codes) across successive decades so that a longitudinal 
     data file can be created.{2}  However, name-matching procedures only    
     offer the illusion of geographic comparability, since problems may arise
     when changes occur in the basic unit of analysis, the county.           
          It is not surprising that the single most troublesome obstacle for 
     the empirical analysis of local areas in a historical/longitudinal      
     framework may be that of changes in the geographic configurations of    
     counties.  In response to changing social, economic and/or political    
     conditions all states have altered county boundaries, eliminated        
     existing counties, or created new ones out of those which already exist.
     While some changes in county boundaries involve the transfer of rather  
     small areas among adjacent counties, the most common form of change in  
     county specifications is the partitioning of existing counties to form  
     new counties.                                                           
          One problem which arises when new counties are created is the lack 
     of consistency of county names across successive decades.  New counties 
     typically appear over time as a result of the partitioning of one or    
     more existing counties, and this has the effect of adding additional    
     rows to the original data matrix, thus creating complications for       
     row-by-row (i.e., county by county) comparisons across decades.         
     Consequently, additions to the matrix can result in a mismatch between  
     geographic areas and temporally relevant information.  For example, new 
     counties will have no counterparts for earlier points in time, and the  
     old counties, which persist in name across successive decades, may      
     change dramatically in geographic definition.  As a result, county      
     reorganization typically remains "invisible" to researchers using data  
     bases organized around county names making this name-matching approach  
     to longitudinal county comparisons clearly unacceptable.  Thus, the     
     problems raised by intertemporal changes in county units must be        
     overcome before longitudinal analysis employing county level data can be
     contemplated, much less undertaken.                                     
          A less common version of the name-matching problem occurs when one 
     or more counties are abolished, counties which exist at an earlier point
     in time have no contemporary referent.  It is also possible for the same
     name to denote totally different physical areas at different points in  
     time.  For example, William County, North Dakota was annexed to Mercer  
     and Stark Counties in 1899 and William County ceased to exist.  Within  
     the year, but in a different section of the state, a "new" William      
     County was established, resulting in the fairly uncommon occurrence of  
     having the same county name appearing in two successive decennial       
     censuses, but with that name referring to completely separate geographic
     areas.  While this is an extreme example, it serves to illustrate the   
     problem raised by using county names for longitudinal matching between  
     empirical information for counties and the geographic areas to which    
     such county data apply.  As one author summarizes ". . . we are         
     confronted with the bothersome task of adjusting reported data at       
     different points in time to allow for such changes, so that             
     comparability may be maintained over time . . . (Tang 1958)."  It is to 
     the solution of these problems that the construction of the County      
     Longitudinal Template (CLT) is directed (Horan, Hargis, and Killian     
     1989).                                                                  
          An optimum strategy for dealing with the variability in the county 
     boundaries across decades would be to somehow impose the current system 
     of county definitions on all earlier time points.  Since most areas in  
     the U.S. have experienced a proliferation of counties over time, this   
     strategy would lead to the most detailed framework for longitudinal     
     analysis of geographic areas.  The problem here lies in the             
     execution--how can data collected under earlier geographic frameworks be
     subdivided according to contemporary boundary definitions?  One         
     possibility, which we considered at the outset, is to subdivide counties
     in earlier time periods by extrapolating on the basis of changes in     
     acreage or population associated with historical boundary changes.      
     Perhaps the main drawback to this strategy, besides the resource and    
     manpower demands required to prepare and implement the computational    
     algorithms, is the problem of finding an appropriate base for           
     extrapolating relevant census variables.  For some demographic variables
     a population base would seem appropriate, while an acreage base seems   
     more relevant for such data items as farm acreage and size              
     distributions.  But there are other important census variables, for     
     example, those concerning industrial development, occupations and       
     employment, for which neither base seems appropriate.  As a result, the 
     standardization strategy for longitudinal analysis would require        
     excluding such variables from the analysis framework and would severely 
     limit the range of applications for the resulting data.{3}              
          Rather than partitioning older county units to approach current    
     county specifications, we have opted to seek longitudinal comparability 
     through an aggregation strategy which groups 1990 counties into larger  
     units on the basis of earlier historical county boundary configurations.
     Rather than using acreage or population change figures to adjust the    
     older census data to the newer political boundaries, the strategy       
     reflected in the CLT is to aggregate counties which have common         
     historical origins.  In order to ensure the maximum possible flexibility
     and detail for longitudinal comparisons, this approach employs a        
     separate aggregation schedule for each decade.  Given a selected earlier
     decade, the corresponding aggregation template can be used to group     
     available county-level data for that decade and all subsequent decades  
     up to 1990.  For counties which have been involved in splits or mergers 
     during the time period of interest, the template specifies an           
     aggregation pattern that defines comparable multi county geographic     
     units, referred to here as "cluster areas."  Thus, once the initial time
     point is chosen, county level data for that decade and all intervening  
     decades can be aggregated to specify the smallest possible geographic   
     units which then can be used for longitudinal comparisons.              
          As we move the point of comparison backwards in time, the          
     aggregates needed to provide comparable geographic areas will tend to   
     get larger.  When using this strategy of aggregating counties on the    
     basis of historical origins, it is necessary to include the clusters    
     which were formed more recently within the parameters of the clusters   
     created at an earlier date.  In some cases, the scale of the            
     historically comparable area becomes unwieldy as counties are aggregated
     across time, while in other cases the comparable areas retain           
     considerable detail.{4}  Sections of the state that have witnessed the  
     proliferation of new counties are the most likely to require these large
     aggregated comparison groups.  While this is undeniably a potential     
     problem inherent to this strategy for defining a longitudinal file, all 
     detail is not lost, and even the aggregated multi county areas offer    
     more detail than the state level data which is the only alternative for 
     longitudinal comparisons in these instances.  The result of the         
     application of this template to successive decades of county-level data 
     will be a series of conformable county-by-variables matrices for        
     geographic areas.                                                       
          The template itself was constructed using information from ICPSR on
     decennial county boundary changes to construct a template (CLT) for     
     comparing geographic areas within the continental United States from    
     1990 back to 1840.  Remember that the extent to which this aggregation  
     procedure is necessary, and the degree to which geographic detail is    
     lost, will vary by historical period as well as by state and region.    
     The mechanism employed in the systematic grouping of contemporary       
     counties into historically comparable units may be thought of as a      
     "current counties-to-past counties" transformation matrix.  The         
     organization of the matrix rests on each 1990 county being assigned its 
     own unique county identifiers.  The county identifiers are designed to  
     correspond to the county codes either on the ICPSR historical           
     county-level data sets or with the U.S. FIPS codes.                     
          The bulk of the CLT matrix consists of separate transformation     
     vectors for every decade from 1840 to 1990.{5}  Each transformation     
     vector indicates how contemporary counties need to be aggregated in     
     order to approximate the county boundary origins of a particular state  
     for a specific past decade.  When merged with the corresponding         
     counties-by-variables data matrices, this transformation matrix will    
     allow the available county level data for each decade to be grouped into
     geographic areas that are comparable across the specified time period.  
     The inclusion in the CLT of these two basic components of the           
     transformation matrix, county identifiers and each decade's             
     transformation vector, provide flexibility in two ways.  First, once a  
     beginning time point is chosen, the decade I.D. can be used to aggregate
     the appropriate counties into comparable geographic areas.  And second, 
     the county identifier can be used to partition an aggregated cluster    
     area into 1990 individual county units.                                 
          In summary, the CLT matrix is designed to merge with and organize  
     the various county- by-variables data files from the U.S. Census Bureau 
     or ICPSR.  County group clusters are identified by arranging the decade 
     I.D.'s on the earliest decade in the analysis.  To use this information 
     for a longitudinal investigation of local areas, three steps are        
     necessary: (1) the transformation matrix, which includes; the county    
     name, ICPSR state code number, FIPS state code number, ICPSR county code
     number and the appropriate decade code, is merged with each of the state
     data files being used (2) once the various files are merged, the        
     county-level observations in each of these files are then aggregated    
     into comparable units using the decade I.D. and (3) an analysis of      
     changes in these areas over time is undertaken.                         
          The most obvious applications for this research tool, in addition  
     to the longitudinal/time series application in political science, will  
     be for students of social structure and social change within local      
     areas.  Here the CLT will provide a valuable tool for assessing the     
     comparability of county-level data over specified time periods, and     
     where direct comparability is not available, for constructing           
     alternative geographic frameworks for longitudinal or time series       
     analyses.  While such a resource should prove invaluable to researchers 
     in a variety of disciplinary settings, it is not the only application.  
     Another interesting application of the CLT involves the construction of 
     empirically grounded, comparative historical contexts for the study of  
     historical survey and U.S. Census data on individuals and households.   
     Consider, for example, a research design in which the characteristics of
     local areas are used to place those areas within a broader,             
     comparative/historical framework, while cross sectional analyses of     
     individuals or households from those areas are used to examine the      
     "micro" implications of such macro-level distinctions.  In the rural    
     South such an analysis might organize areas according to different types
     of agricultural production, land ownership patterns, or credit          
     arrangements, and then examine the implications of such structural      
     arrangements on the family economies of local residents.  In this way,  
     we anticipate that this template for longitudinal analysis of           
     county-level U.S. Census data will open new horizons for theory and     
     research on local social and economic structures.  In addition to the   
     obvious applications to studies of temporal change in the social and    
     economic organization of local areas, we hope that it will provide a    
     basis for theory and research on historical linkages between local      
     social structures and the opportunities and constraints experienced by  
     both individuals and households.                                        
                                                                             
     Format for The County Longitudinal Template                             
          The basic format for the CLT is a county-by-decade matrix for all  
     states within the continental U.S. and all decades from 1990 through    
     1840.{6}  The CLT matrix consists of three primary components: (1)      
     county name, the state ICPSR and FIPS codes (see Appendix A) and the    
     county ICPSR codes (2) the county decade identification codes, which are
     used to link pertinent county areas into historically comparable units  
     and (3) the county boundary change flags.  While the CLT does not       
     incorporate minor boundary changes (those which do not result in the    
     creation of new counties) into its aggregation scheme, it does provide  
     information on the existence of such boundary changes for each county   
     and each decade from 1840 to 1960.  The discussion below will consider  
     each of these parts separately.                                         
          First, each county within a state is identified by name and a      
     unique numerical county identifier.{7}  The county code used in the CLT 
     is a four-digit number which provides the necessary components to link  
     the longitudinal county boundary file with either Inter-university      
     Consortium for Political and Social Research (ICPSR) county data files, 
     U.S. Census data, or any other machine-readable county level data file  
     that uses U.S. FIPS codes to identify land areas.  Below are two        
     examples of CLT county and state codes from Michigan.  The first,       
     Alcona, is a county which is still in existence, the second, Isle Royale
     County, was abolished in 1897.                                          
     For example:V1= Alcona   V2= 23  V3= 26  V4= 0010                       
     where:    V1   County name                                              
               V2   ICPSR state code                                         
               V3   U.S. FIPS state code                                     
               V4   This four-digit code uniquely identifies each county     
               within a state.  For contemporary counties the first three    
               digits correspond to the Federal Information Processing       
               Standards (FIPS) county code number.  The fourth digit is a   
               place holder which allows for the identification of counties  
               which were abolished before the enactment of FIPS codes.  This
               is the same coding system used in ICPSR files.                
     Or:       V1= Isle Royale   V2= 23  V3= 26  V4= 0745                    
     where:    V1   County name                                              
               V2   ICPSR state code                                         
               V3   U.S. FIPS state code                                     
               V4   ICPSR historical county code that is used to identify    
               counties which were abolished before the enactment of FIPS    
               codes.  (The ICPSR historical county codes always ends in 5). 
          Second, the CLT decade identification numbers are used to define   
     groupings of geographic areas between 1840 and 1990.  For each decade   
     within this range, every contemporary and historical county is assigned 
     a decade identification number which places that county within the      
     geographical configuration prevailing at the earliest decade in that    
     range.  That is, these numbers identify counties which must be clustered
     together in order to make 1990 counties comparable throughout the       
     specified period.  Thus, once the initial time point is chosen, the     
     decade I.D. for that particular decade aggregates counties from all     
     intervening decades into the smallest possible geographic units, which  
     then can be used for longitudinal empirical comparisons.{8}  In cases   
     where a county ceased to exist (abolished) or where counties emerged    
     from unorganized territories, the missing information is recorded on the
     template as missing (e.g., -9).                                         
          Third, the final part of the CLT flags those cells in the          
     county-by-county matrix (1960-1840) where a particular county boundary  
     was changed but no new county emerged from the shift in boundaries.  The
     county boundary change flags are a binary code.  For each county,       
     decades in which boundary changes occurred (other than those which have 
     already been taken into consideration) are denoted by a one, while      
     decades in which no changes occurred are denoted by a zero.  This       
     boundary change flag should be used by the individual researcher to     
     check for substantial boundary changes which may violate the assumptions
     of geographic correspondence.  This is of particular importance in the  
     western states where county boundary shifts often entailed large        
     exchanges of land.  In summary, there are 33 variables in the CLT.  They
     are;  county name, state ICPSR code, FIPS state code, county ICPSR code,
     Decade I.D.'s for 1990, 1980, 1970, 1960, 1950, 1940, 1930, 1920, 1910, 
     1900, 1890, 1880, 1870, 1860, 1850, 1840, and zero/one Boundary change  
     flags for 1960, 1950, 1940, 1930, 1920, 1910, 1900, 1890, 1880, 1870,   
     1860, 1850, 1840.                                                       
                                                                             
     Use and Management of the CLT                                           
          The starting point for linking the CLT to county-level data is to  
     convert one or all data sets into compatible system files.  Once the    
     various county-data files are ready for use, each of the cross-sectional
     files needs to be linked to the CLT.  It is not necessary to retain all 
     33 of the CLT's variables, only the county name, one of the state codes,
     the county I.D. code, the decade I.D. code (which identifies the        
     earliest decade of interest).  And, in some cases the researcher may    
     wish for these to be augmented with the boundary change codes for each  
     decade under study.  Remember, these boundary change codes are a binary 
     code where zero indicates no boundary changes, other than what is       
     already contained in the CLT, and one indicates that boundary changes   
     occurred but no new county resulted.  These CLT variables need to be    
     merged with each of the various data sets selected by the researcher in 
     order to provide a common basis for linkage across all the files.  This 
     is very important since all counties do not necessarily exist throughout
     the time period of interest.  In addition, it is recommended that the   
     files be merged by the county codes and not by county names.  The       
     renaming of counties over time and the variation in spelling between    
     files may cause unnecessary duplication of counties between files (as   
     each new spelling would appear as a separate observation).  Also, files 
     must be merged first by state and then by county I.D. since the decade  
     I.D. codes are state specific.  Once each of the individual county files
     has been merged with the CLT the process is repeated for a final time in
     order to merge all combined CLT/county data files into one data set.    
          The data set is now ready to be organized so that individual       
     counties can be aggregated into historically relevant geographic areas. 
     However, missing values within your data require attention before       
     individual county groups can be aggregated.  Missing data is aggregated 
     along with valid data when counties are clustered into historically     
     comparable areas.  If a particular county has missing data because it   
     didn't exist for a time then it doesn't present a problem.  However, if 
     the data is missing because it wasn't reported or because the original  
     data was illegible, etc., then it could be problematic.  When the data  
     are aggregated into county cluster areas, missing data could cause      
     aggregated values to under represent the "true" values if the county    
     existed but the data wasn't reported.                                   
          These work examples are meant to demonstrate one possible way to   
     organize data for analysis using the CLT.  The template is intended to  
     aggregate past county configurations so that they are comparable to     
     present county geography, thus only past to present comparisons are     
     possible (and not past to past comparisons).  The CLT comprises separate
     state files that must be individually merged with historical data (for  
     each selected period), and then all of the state files can be merged    
     across time to construct a single longitudinal file.  At this point data
     can be aggregated into temporally comparable cluster areas.  Once the   
     county units have been aggregated, the data is ready to be customized to
     suit the researcher's needs.{9}                                         
                                                                             
                                                                            
                                                                             
                                   REFERENCES                                
                                                                             
     Brosveet, Jorle. 1980. "Standardization of Longitudinal, Aggregate-level
         Data in the Norwegian Commune Database" Pp. 513-523 in Historical   
         Social Research: The Use of Historical  and Process-Produced Data,  
         edited by Jerome M. Clubb and Erwin K. Scheuch, Stuttgart, Germany: 
         Klett-Cotta.                                                        
                                                                             
     Horan, Patrick M., Peggy G. Hargis and Molly Sizer Killian. 1988.       
          "Longitudinal Research on Local Labor Markets:  The County         
          Longitudinal Template," Pp. 99-121 in Perspectives  on Rural Labor 
          Markets, edited by Thomas A. Lyson and William W. Falk, Greenwich, 
          CT: JAI Press.                                                     
                                                                             
     Inter-University Consortium for Political and Social Research. 1995,    
          Guide to Resources and  Services, 1994-1995, Ann Arbor: ICPSR.     
                                                                             
     Tang, Anthony. 1958. Economic Development in the Southern Piedmont,     
          1860-1950, Chapel  Hill: University of North Carolina Press.       
                                                                             
                                                                             
                                                                            
                                                                             
     APPENDIX A: State Identification Codes                                  
                                                                             
                                                                             
                                                                             
                                                                             
                   ICPSR   U.S. FIPS                    ICPSR   U.S. FIPS    
                   State     State                      State     State      
                   Codes     Codes                      Codes     Codes      
                                                                             
                                                                             
     Connecticut     1         09       Arkansas          42       05        
                                                                             
     Maine           2         23       Florida           43       12        
                                                                             
     Massachusetts   3         25       Georgia           44       13        
                                                                             
     New Hampshire   4         33       Louisiana         45       22        
                                                                             
     Rhode Island    5         44       Mississippi       46       28        
                                                                             
     Vermont         6         50       North Carolina    47       37        
                                                                             
     Delaware        11        10       South Carolina    48       45        
                                                                             
     New Jersey      12        34       Texas             49       48        
                                                                             
     New York        13        36       Kentucky          51       21        
                                                                             
     Pennsylvania    14        42       Maryland          52       24        
                                                                             
     Illinois        21        17       Oklahoma          53       40        
                                                                             
     Indiana         22        18       Tennessee         54       47        
                                                                             
     Michigan        23        26       West Virginia     56       54        
                                                                             
     Ohio            24        39       Arizona           61       04        
                                                                             
     Wisconsin       25        55       Colorado          62       08        
                                                                             
     Iowa            31        19       Idaho             63       16        
                                                                             
     Kansas          32        20       Montana           64       30        
                                                                             
     Minnesota       33        27       Nevada            65       32        
                                                                             
     Missouri        34        29       New Mexico        66       35        
                                                                             
     Nebraska        35        31       Utah              67       49        
                                                                             
     North Dakota    36        38       Wyoming           68       56        
                                                                             
     South Dakota    37        46       California        71       06        
                                                                             
     Virginia        40        51       Oregon            72       41        
                                                                             
     Alabama         41        01       Washington        73       53        
                                                                             
                                                                             
     ENDNOTES                                                                
                                                                             
                                                                             
     1. The Inter-university Consortium for Political and Social Research    
     (ICPSR) at the University of Michigan has prepared a county-level data  
     collection called Historical Demographic, Economic and Social Data: The 
     United States 1790-1970,(ICPSR #0003).  This file contains a wide       
     variety of summary measures of economic and social variables at the     
     state and county levels form each of the decennial censuses, 1790 to    
     1970.                                                                   
                                                                             
     2. The Federal Information Processing Standards (FIPS) code is a        
     five-digit number which uniquely identifies each county (3 digits) and  
     the state (2 digits) within which it is located.                        
                                                                             
     3. Perhaps the most impressive implementation of such a standardization 
     strategy appears in the Norwegian Social Science Data Services (NSD)    
     (Brosveet, 1980: 516-520).  The first author acknowledges the support of
     a 1985 Fulbright Research Fellowship to Norway which allowed first-hand 
     experience with this longitudinal data base.                            
                                                                             
     4. Even though county-level census data is available back to 1790 for   
     some U.S. counties, it may not be practical to trace comparable areas to
     such an early period for three reasons.  One, because of the rapid      
     aggregation of county units as we move back through the earliest time   
     periods, the comparable geographic areas become unwieldy in terms of    
     size and little advantage is gained over using the state as the unit of 
     analysis.  Two, census data becomes less detailed the earlier in time   
     one goes, making the changing definition of data items, and hence their 
     comparability, all the more problematic.  Three, about one half of all  
     states were admitted into the union after 1845 and detailed census      
     information regarding territorial lands are scare.  For these reasons we
     have elected to put a cap (1840) on the beginning time point for the    
     aggregation template presented here.                                    
                                                                             
     5. The CLT file also includes a boundary change file which identifies   
     decades in which boundary changes for a particular county occur, but no 
     new counties were created.  This boundary change file can be used by the
     individual researcher to check for substantial boundary changes which   
     may violate the assumptions of geographic correspondence.               
                                                                             
     6. Thus the CLT does not include state matrices for Hawaii and Alaska.  
     In addition, Virginia and West Virginia must be used as a single matrix 
     if the initial time point of interest is before 1860.  Since West       
     Virginia emerged from Virginia in 1863, West Virginia has some          
     contemporary counties which share common county origins with Virginia.  
                                                                             
     7. A numerical county identifier is important since county names are    
     frequently unreliable.  County names sometime change over the years so  
     that the same county area was referred to by different names.  Less     
     frequently there were counties that were abolished but the "old" name   
     was given to a "new" county so that the same name might appear in       
     separate locations over time.                                           
                                                                             
     8. It is perhaps worth noting that the CLT is a data  organization      
     scheme and was designed to reorganize data for across decade comparisons
     and is not intended to provide historically accurate maps.              
                                                                             
     9. For recent applications of this resource see, Kenneth C. Land, Glenn 
     Deane and Judith R. Blau. 1991. "Religious Pluralism and Church         
     Membership: A Spatial Diffusion Model" American Sociological Review     
     56:237-249.  Also see, Hargis, Peggy G. 1994. In Quest of the Golden    
     Fleece: African-American Landholding in Georgia, 1874-1933. Unpublished 
     Dissertation, Athens: University of Georgia.                            
