Data in Brief

The data describes Christian mission stations established in Ghana 1752–1932. Data is reported at an annual basis. For all 2,144 mission stations, the data includes station name, de-nomination, circuity, longitude, latitude, year of entry, exit, whether the station is a main or out-station, and whether it had a school attached. For sub-periods the data also includes information on the number of church members, attendance and seat capacity. The data was mainly sourced from ecclesiastical returns provided by the mission societies and published in the British Blue Books of the Gold Coast 1844–1932. The source represents a comprehensive census of missions. Various other sources were consulted to extend the data base to Ghana’s ﬁrst mission (1752), to include missions from German To-goland incorporated into Ghana after World War I, and to account for years, for which no Blue Books have survived. Mission stations were then georeferenced based on the place name where the mission is located. Coordinates were retrieved from NGA place name gazetteer as well as other sources. The data can be used to study patterns in and effects of Christianization in Ghana. The geographic coordinates of the mission stations allow researchers to ﬂexibly link the data to other spatio-temporal databases.

The data has been used in: Jedwab, R., F. Meier zu Selhausen and A. Moradi (2021).Christianization without economic development: Evidence from missions in Ghana.'

Value of the Data
• The data represents an annual, complete panel of mission stations in one African country.
Studies using these data can address concerns that frequently arise from (selective) misreporting by other sources (e.g.mission atlases).• The data describes the type of mission, whether main or out-station, with/without mission school and also record temporary entries and exits.• The data is useful to describe patterns and analyze causes and effects of Christianization in Ghana.It also allows to analyze the full heterogeneity of mission stations that existed.• The geographic coordinates of missions can be flexibly used to add other spatial data.

Raw data
• < church locations.xls> .This is the raw data entered from the ecclesiastical returns of the Blue Books.The data includes geographic coordinates ("lat" and "lon"), mission society ("Church"), whether main or out-station ("mainstation"), the name of the mission ("place" and alternative spellings of the name ("Diff spelling"), years of existence ("yr1828", … yr1932_33), name of priest residing at that station "Priest 1857, … "Priest 1870"), church members ("Members 1846", … "Members 1881"), seat capacity ("Seat Capacity 1861", …, "Seat Capacity 1899") and circuity ("Circuity 1857", … "Circuity 1932").• < Togo_church_locations.xls> .We entered the mission data of German Togoland in this file.The data includes geographic coordinates ("lat" and "lon"), mission society ("Church"), whether main or out-station ("type"), the name of the mission ("place" and alternative spellings of the name ("Diff spelling"), years of existence ("togo_yr1847", … togo_yr1936) and whether the mission was on the territory of the British or French administered part of Togoland ("togo_coloniserpost1918").• < school locations.xls> .This file contains the raw data on mission schools from the Education section of the Blue Books.• < priests.xls> .This data includes background information of the priests stationed in Ghana including the denomination ("church"), duration of residence ("min_year" and "max_year"), and year of death if died in Ghana ("death_yr").• < mission_city_coordinates.xls > .This file contains coordinates as well as grid cell identifiers • < Wesleyan1847_56 > .This files contains the data entered from the Minutes of the Wesleyan Methodist Church Conferences.

Do files
The raw data was processed using Stata 16.To replicate the processed data, open the < master_dofile > in Stata, change the working directory, and run the < master_dofile > in Stata.There are sub do files that broke down the coding by topic.All do files detail input files and how input files are processed to output files.Notes are added within the do files.

Processed data
The raw data was cleaned and processed using Stata.The do-files are included in a separate folder.Every step is described in detail there.The output files can be generated running < master_dofile.do> .There are two main output files.
1. < church_locations_panel > .The file contains the panel of missions ("mission") starting from the year of foundation (yr_startup) to the year of exit ("") and whether in operation in the respective years ("exist")..It includes information of the denomination ("church"), whether main or outstation ("mainstation"), school ("school9"), priests known to be stationed there ("priest").2. < church_grid.dta> .It contains the same information as < church_locations_panel > .However the unit of observation is a grid cell of 0.1 × 0.1 resolution.The shapefile of the grid can be found in the GIS folder.

GIS data
We added polygon shapefiles of the Muslim and Protestant spheres as well as the 1914 German Togoland border in the GIS folder.All projections are UTM 30N.Finally, we added the grid raster of Ghana (0.1 × 0.1 degrees) on which < church_grid.dta> is based.

Materials and Methods
Our primary source of mission church data are the Ecclesiastical Returns that missionary societies submitted to the British colonial administration on an annual basis and that were pub-lished in the Blue Books of the Gold Coast 1844-1932 [1] .Hence, the data refers to "officially" recognized mission stations.
For certain years ecclesiastical returns are unavailable and we used secondary sources to reconstruct the data.
• 1751-1843: The Blue Books of the Gold Coast start in 1844.The early beginnings of missionary work in Ghana are particularly well-documented and we reconstructed the period 1751-1843 from a variety of secondary sources [3][4][5] .• 1862-66: Blue Books were not available in The National Archives (Kew, London).This was a time of a rather stable environment for missionary work.Most of the mission stations in 1861 also existed in 1867.We consulted the Minutes of the Wesleyan Methodist Church Conferences [2] , Schlatter [4] and Schreiber [6] for the Methodist, Basel and Bremen Mission, respectively, to confirm dates of new openings and closures.• 1873-74: During the Third Anglo-Ashanti war 1873-74 many mission stations were abandoned and destroyed.We reconstructed the history of each mission station using a variety of sources [ 4 , 7 ].
• 1917-1919: The Blue Books did not publish ecclesiastical returns.It was impossible to reconstruct the history of each mission station from secondary sources.In 1916, the number of mission stations already exceeded 700.We therefore interpolated assuming that mission stations that existed in 1916 and 1920 also existed 1917-1919.The number of Methodist mission stations stagnated with 311 and 322 in 1916 and 1920, respectively.The assumption seems therefore unproblematic in this case.As for the Basel and Bremen Mission, German and Swiss priests and missionaries were interned during WWI and deported when the war came to an end.The Scottish Mission then took over their mission stations in 1920.We observe a fall in the number of mission stations between 1914 and 1915, from 302 to 215.However, we believe that most of these churches operated 1917-1919, even if under difficult conditions, under the supervision of African personnel.The number of Catholic missions, in contrast, increased from 154 to 256 in 1916 and 1920, respectively.In this case, we ignore the expansion underestimating the number of Catholic missions in 1917-1919.
• Firstly, we set those mission stations to exist that according to the Blue Books existed in the year before and after the gap in reporting.For example, the Methodist mission of Komenda was listed in the ecclesiastical returns for the year 1899 and 1904, but not for 1900-1903, as it was an outstation of Elmina.It is likely that Komenda continued to exist, particularly because the period 1900-3 was a time of expansion.• Secondly, if any mission school was reported in the Blue Books, we assume that the corresponding mission station was also in operation.This assumption is unproblematic, because this is what we overwhelmingly observe: we only found 256 location-years where a school was reported without a mission station out of a total of 6,342 location-years where both school and mission were reported.• Thirdly, we complemented the Blue Book data with information reported in the Minutes of the Methodist Conferences [2] .We did not add more church locations from this source, but rather restricted the coding to those mission stations that were reported in the Blue Books at least once.