1. Introduction
In the early 1900s, following the postbellum reconstruction era, education and school building quality were incredibly low for poor rural children throughout the southern United States [
1]. Disenfranchisement, segregation, and unfair treatment in this region worsened after the emancipation of slaves and the end of the Civil War [
2]. White schools and institutions were given a higher percentage of the total taxes raised in the south, leaving Black institutions and schoolhouses derelict [
2]. In one report from the Board of Education in Columbia, South Carolina, the schools were described as “in the most instances wretched, the terms short, and salaries low, practically no equipment, and the preparation and fitness of the teachers generally very inferior” [
3]. Many other school agents and workers described the schools in other states in a comparable manner.
One person who worked to improve schooling for African Americans in the south was Julius Rosenwald. Rosenwald was a philanthropist and part-owner of the large American cooperation Sears, Roebuck, and Co., and he donated funds to small southern communities to build “common schools” for southern Black schoolchildren [
4,
5]. The story of the Rosenwald fund and its schools starts with Booker T. Washington and his Tuskegee Normal and Industrial School. Washington, after graduating from the Hampton Normal and Industrial Institute in 1875, founded his own school in Tuskegee, Alabama, USA, to further the moral and educational improvement of Black people in the south [
4]. Both Rosenwald and Washington believed that to improve the schooling of the Blacks throughout the southeastern USA, it would be best to focus on physical capital and resources rather than just donating money. If they relied on cash donations, it was feared that the primarily White leadership in the school boards would inappropriately divert funds away from Black schools [
1]. To carry this out, Rosenwald and Washington came together to work on building brand-new schoolhouses across all the counties of the south, and to give them the best equipment while keeping costs as low as possible by supplying grants that were to be matched by the local community to build the schools [
4].
Rosenwald’s venture into philanthropy first began as a large donation towards building YMCAs, or Young Men Christian Associations, for Black men [
6]. Rosenwald continued donating with a large 25,000-dollar birthday gift to Washington [
4]. This donation was to be spent on the Tuskegee Institute and its associated educational enterprises, not building brand new schools initially. Clinton J. Calloway, who worked for the Tuskegee Department of Extension, worked with small Black rural communities and put in the request to Rosenwald, through Washington, to use leftover donations to build some new schools nearby as an experiment [
4]. The experiment proved a success, and after Washington’s death in 1915, Rosenwald donated 100,000 more dollars in Washington’s name for the building of more schools [
6]. Aid, donations, and building continued until 1932 with Rosenwald’s death and the ever-strengthening effects of the Great Depression [
2]. Many local communities struggled to raise money to match the Rosenwald grant and therefore could not afford the building of new schools.
Recent research shows the significant impact Rosenwald schools had on education in the Black south. By the end of the school-building program, 92% of Black children lived in a county that had at least one Rosenwald school [
7]. Between 1914 and 1931, the creation of Rosenwald schools accounted for a 40% decrease in the literacy and enrollment gap between Black students and White students [
8]. When compared to the Jeanes fund (another philanthropic institution for Black education), Rosenwald schools returned higher dollar amounts based on the cost of investment, anywhere between 17% and 42% [
8]. This is mostly due to the lack of physical resources prior to the building of schools. In fact, the reason the return on investment was so high in comparison was that the resource was nearly absent in most communities to start with.
Even with the positive benefits that Rosenwald schools brought to the south, the exact location of many of these schools and their full impact are not officially known. It is estimated that of the 5357 schools that were built, only 10–12% stand today and are often not used as schools anymore [
9]. The National Trust for Historic Preservation placed Rosenwald schools in its 11 Most Endangered Historic Places list in 2002 [
9]. However, slowly but surely, many local communities have been reinvesting in extant Rosenwald schools and restoring them for other purposes [
10,
11,
12].
The substantial impact of Rosenwald schools has warranted research in several domains. For example, Black student achievement was measured by comparing Rosenwald school statistics with World War II enlistment test records [
1]. They found that a higher exposure to the Rosenwald school program led to increases in the amount of school years completed, increased high school attendance, and higher Army General Classification Test (AGCT) scores. Overall, as mentioned earlier, the Rosenwald school program accounted for closing the Black–White education gap by forty percent [
1]. The impact of Rosenwald schools was also measured and compared with incarceration rates, which suggested that those with higher exposure to the Rosenwald school program reduced the likelihood of future incarceration due to increased opportunity cost of criminal activity [
7]. Other studies focused on specific states like North Carolina [
13], Georgia [
14], and preserving schools in Texas [
11]. This study seeks to continue this work from a geographic perspective by mapping and identifying the geographic locations of Rosenwald schools in the state of South Carolina using public records and geographic insights. Once the significant research portion is complete and the school locations mapped, census data for townships in South Carolina, along with ancillary information about the schools themselves, are used to provide a spatial analysis of school locations, cost, and spatial patterns observed in the state. In a first-of-its-kind spatial investigation of Rosenwald schools within a state, this paper seeks to address the following research questions:
How were Rosenwald schools spatially dispersed through the state of South Carolina as they were built from the 1910s to 1932?
Were all African American populations provided similar opportunities through the program, or were there still communities left with difficult educational circumstances?
Are there spatial differences in White contributions, Black contributions, and public contributions for the schools in addition to the Rosenwald funds? If so, why?
2. Materials and Methods
2.1. Study Area
South Carolina, USA, is situated in the heart of the southeastern United States of America, between neighbors Georgia and North Carolina (
Figure 1). South Carolina can be separated into four distinct regions, including the Lowcountry to the south along the coast where most plantations and slavery were present antebellum, the Pee Dee region to the northeast along the Pee Dee rivers, the Midlands at the center of the state with the rolling hills and red clay soil, and the Upstate where the state meets the Appalachian mountains. Each of the regions encompasses a beautiful environment with significant cultural and historical relevance. The regions are often used to divide the state in public records and data collection (
Figure 1). Racial relations in South Carolina during the years Rosenwald schools were built were tense but also varied by region. Some of the most complex relations were in the lowcountry, where a majority of plantations and plantation owners were in the years prior.
Though one of the wealthiest of the 13 original colonies, the reliance on slavery and indentured servitude for labor resulted in a struggling economy following the defeat of the Confederate States of America, of which South Carolina had been a part of throughout the Civil War. The cultural racism present throughout the state created barriers to reconstruction, limiting new African American citizens access to a proper education. However, the first schools were brought into the South Carolina Sea Islands by northern philanthropists as early as 1861, still during the war, when the area came under federal control [
15]. As Bartels (2004) stated, the state stumbled into the 1900s with no real system for public education, contextualizing the tremendous challenges faced by the Rosenwald foundation to building schools for Black students [
15]. The era of time in which we investigate Rosenwald school buildings, from 1900 through the 1930s, saw 30% growth in South Carolina population, growing from 1.3 million to over 1.7 million people [
16]. The percentage of African American population actually declined during this period due to the underlying northward migration patterns of many Blacks for work, leaving a White majority in the state for the first time in 120 years [
17]. Those trends reversed in the following years, eventually resulting in being the state with the third highest percentage of Blacks in the country [
17].
2.2. Mapping SC Rosenwald Schools
Rosenwald school data for the state of South Carolina were compiled in the form of PDF school lists that contained details about the school costs, including White, Black, Rosenwald, and public contributions, and teacher count, acreage, date, and often a short memo pertaining to the school location or history. Sometimes the files include a photograph of the original “Rosenwald cards” that were maintained to keep records. These data are maintained by the South Carolina Department of Archives and History. A total of 500 Rosenwald schools are identified in the public records shared by this department.
Few records were archived with sufficient geographic information for mapping the exact locations of each of the schools. To improve mapping efforts, a multi-step approach was utilized depending on how much information was available in the dataset. The end goal was to provide an approximate location for as many schools as possible (
Figure 2). First, each card was thoroughly investigated for any amount of spatial or geographic information. These geographic data, if extant, were recorded with the other variables in the database. If the data provided were sufficient, latitude and longitude coordinates were derived from the address given or historical monument location. Only 20 of the 500 documented Rosenwald schools, or 4%, are recognized as South Carolina historical markers, suggesting that a significant number of schools have been forgotten and are currently unrecognized. Even fewer are still standing. Several others included a previous address of the school that was used to then determine latitude and longitude coordinates of its approximate location. However, most other schools in the registry only provided an old description of the location. Often, this was at the cross of two particular roads or in a certain part of a township. This information was used, along with other historical information about road name changes and changes in the township over time, to estimate each school’s location.
The most difficult ones to locate were the remaining schools that lacked most spatial information and only provided a place or school name. In these instances, significant time and resources were spent scouring historical sources and gazetteers for spatial information about local historical schools, where certain schools were torn down and replaced, and how communities had changed over time. Through this sifting of information, several school locations were mapped with minimal error. However, often the only known information about a school was the township within which it resided. In these instances, the centroid for the township was used as the approximate location for the school because we could not find more specific spatial data. Unfortunately, even after using the multi-step method for identifying school locations as specifically as possible, a number of school locations still could not be identified. In the end, 446 of the 500 schools received at least an approximate centroid location with spatial coordinates and were mapped for further spatial analysis.
2.3. Spatial Analysis of Rosenwald Schools
The geolocated school data were analyzed spatially using ancillary and contextualizing data from the time period. Much of the school data printed on the cards, including school acreage, number of teachers, and monetary contributions by Rosenwald and others, were mapped with the locations of the schools to investigate spatial patterns. Visual spatial interpretation was the primary method of analysis. Cartographic visualization techniques, including heat maps, were also used to enhance visual patterns for efficient interpretation. Heat maps were generated using the ESRI ArcGIS Pro 3.1.1 heat map visualization tool in the symbology options.
Census Township Data for 1920 and 1930 were incorporated to provide context on population and especially racial data [
18,
19,
20]. Acquired from the U.S. Census Bureau, the population data were tied to specific latitude and longitude coordinates for townships rather than polygonal geographic units. Therefore, Thiessen polygons were generated using the Thiessen polygon tool in ArcGIS Pro 3.1.1 to represent the areas closest to each township on record. Thiessen polygons, named for the late US meteorologist Alfred Thiessen, define areas of influence around a set of points, where any location within the polygon itself is closest to the point within that polygon to any other point. Therefore, analysis in terms of number of Rosenwald schools per township, African American population per township per Rosenwald school, and racial populations and their contributions to nearby Rosenwald schools could be performed as well. This allowed us to overlay Rosenwald school bivariate point data with a choropleth map behind it to investigate multiple variables at a time. When generating choropleth maps, statistical breaks for symbolizing data were consistently generated using the Jenks or Natural Breaks designation in order to maintain higher congruence for comparisons.
3. Results
3.1. General Information About South Carolina Rosenwald Schools
The 446 mappable Rosenwald schools across South Carolina were distributed fairly evenly across the state’s four regions except for the Lowcountry region. The Lowcountry, representing 11 of the state’s 46 counties, had 64 mapped Rosenwald schools, while the other regions each had 124 (Midlands), 121 (Upstate), and 136 (Peedee) Rosenwald schools (
Figure 3). More schools are found in regions closer to where the Rosenwald funds come from (the principal office was in Chicago, IL) and further from the original locations of plantations in the South Carolina Lowcountry.
Of the 484 townships represented in the 1930 census data as Rosenwald school building slowed down into the Great Depression, Simpsonville Township, Greenville County, in the Upstate Region had eleven Rosenwald schools and within its Thiessen polygon, the most by five schools of any township in the state (
Figure 4). Over 48% of the townships in SC had no Rosenwald schools, while 28% had at least one. Interestingly, each region varied greatly in the most Rosenwald schools nearest to any given township. For example, the eleven schools within the Simpsonville Township were in the Upstate Region. The Pee Dee and Midlands regions both had townships with a maximum of six schools. The Lowcountry had the fewest, with four schools being the region’s maximum number within a township.
The Upstate saw the earliest influx of Rosenwald schools, with seven being built before the dawn of the 1920s, the most in the state (
Figure 5). The Midlands came in a close second with six, while the Pee Dee and Lowcountry regions recorded the first built Rosenwald schools in 1920–21. On the other hand, the Lowcountry was the slowest to grab on to the idea of Rrosenwald schools. The years when the most schools were built in each region differed slightly. For example, the Upstate Region saw the greatest percentage of schools built between 1924–26, while the Lowcountry was 1923–24, Pee Dee was 1920–21, and again 1923–27, and the Midlands recorded the most growth between 1923–27 as well. Just over 37% of all Rosenwald schools were built before the end of 1924, indicating significant effort at the start of the philanthropic campaign but slowing as the Great Depression began five years later.
3.2. Spatial Distibutions of Contributions, Size of School Land, and Number of Teachers
Monetary contributions to Rosenwald schools were balanced between four entities as registered on the Rosenwald cards for the state: Rosenwald fund contributions, public contributions, Black contributions, and White contributions. The spatial patterns of contributions, as well as general financial patterns by contribution source, provide context for the racial and political climate of South Carolina, especially when investigated through a spatial and regional perspective.
The entire state received a total of USD 2,850,215 (all dollar references here are in 1920–30s dollars) to build the Rosenwald schools. This averaged to USD 5887.63 per school. Of that amount, the Rosenwald fund provided USD 444,300 to construct school buildings, while the public contribution provided the highest amount of all of the sources at USD 1,655,503. Black citizens provided a significant USD 511,810, or over USD 1000 per school, and Whites also contributed statewide, but at a smaller scale, providing USD 219,135. All of these funds were significant investments by the populations in South Carolina and the Rosenwald fund at that time and deserve praise. However, when we investigate these data regionally, interesting patterns arise (
Table 1).
In each region, total investments as well as percentages of contribution sources varied greatly. In the Upstate, 23% of all funds were provided by Blacks, greater than the 17% provided by the Rosenwald fund and 5% provided by Whites. As we move south/south east in the state, the Midlands showed a similar pattern, but to a lesser degree. The Rosenwald fund provided 15% of total funding, Blacks contributed 19%, and Whites donated 9%. In the Pee Dee Region, the Rosenwald fund provided 18%, Blacks 20%, and Whites contributed 7%. While these three regions are similar, the Lowcountry provides interesting insight as it was different than expected. The Rosenwald fund provided 14% of the needed cost for schools in that region, while 13% came from the Black population and 12% came from the white population. Each of those three funds only provided a small fraction of the total costs compared with the public funding and were the most similar to each other in amounts of any of the regions.
In fact, White contribution was highest on average (per school built) in the Lowcountry than any other region. Likewise, Rosenwald fund contributions per school were highest per school in the Lowcountry than anywhere else, while the Black contribution total and average were the lowest of all the regions. A specific example of this is Robert Smalls School, within the Beaufort Township in modern Beaufort County and in the Lowcountry Region, where a total of USD 19,000 was contributed by White citizens to aid the building of the school. This was 76 percent of the total building cost and shows the significant aid that Whites gave there, which is in line with the higher average White donation percentage in this region.
If investigated at a more local level, there is much apparent randomness in the cost of all of each of the schools (
Figure 6), in the size of the school property, and in the number of teachers per school. In each region, there were several more expensive schools interspersed with many lower-cost ones, which invites more investigation into the reasons behind this wide variability. In terms of the lot sizes, the total area of land purchased for these schools in the state was 622.30 hectares. Generally, the average size of land purchased was just over 1.2 hectares, with the maximum size being 16.19 hectares in the lowcountry region. The largest average land sizes were found in the Lowcountry (1.6 hectares) and Midlands (1.42 hectares) regions, while the lowest was in the Upstate Region (0.98 hectares). Pee Dee was a bit higher with a mean land size of 1.09 hectares. Finally, the number of teachers per school that was included on each school card shows a variability by region too. The average number of teachers per school was 3.2 for the state, and once again the Lowcountry averaged the highest number of teachers, at 3.75, but most other regions were not far behind, with the Midlands Region at 3.31, Pee Dee at 3.18, and the Upstate coming in last at 2.93. The most common number of teachers found were 4 per school house built, but the number got as high as 14 teachers in the largest schools built.
3.3. Spatial Distribution of Rosenwald Schools with Associated Spatial Patterns in 1920s and 1930s Census
Population data from both the 1920 Census and the 1930 Census were used to provide context for the racial changes and challenges of this time (
Table 2). The Black population was larger than the White population in 1920, but that trend ended for the first time in 1930, when the White population edged the Black population in the state. This is due, in large part, to what is termed “the great migration”, referring to the exodus of Black Americans from the southern states to the Midwest and northern states in search of better living conditions and gainful employment. Because of the small numbers of other races recorded in the census, only Black and White Americans are studied in this paper.
The relative statewide racial statistics vary when compared by region. In 1920, each region other than the Upstate had more Black residents than White residents. This reflects the Upstate’s agricultural and historically more White cultural background. Furthermore, the closer you move toward where slave quarters and plantations were present (to the south and east), the greater number of Black Americans that were present. In 1930, the White population in the Upstate increased by almost 80,000 people while the Black population declined by 14,000 people, increasing that divide. The Midlands followed suite with more White Americans than Black Americans in 1930, and the difference in the race populations for the other two regions decreased as well despite being more Black. Generally, we find more schools in township areas where the Black population was higher in 1930 (
Figure 7). However, fewer schools are found in the Lowcountry region despite being predominately Black when we look regionally, indicating an apparent greater lack of access to education for Black children in the region. This is somewhat expected with its proximity in time and place to slavery and racial discrimination well documented in the region [
21,
22,
23].
Rosenwald schools were found to be within 249 of the 484 township Thiessen polygons as recorded in the 1930 census. The mean populations of the townships with Rosenwald-funded schools were 3153 people, with 61% of the total population of the state within these townships. Racially, these townships averaged more White population (1847) than Black population (1303). In the 235 townships that had no schools within their Thiessen polygons, there was an average population of only 2106 people. In these smaller townships, there was a larger average Black population (1077 per township) than White population (1028). This indicated that many of the targeted school locations were in areas of greater populations, but unfortunately still left many smaller populations without access to the Rosenwald school infrastructure.
4. Discussion
Rosenwald schools are an important part of the history of American education, especially in the southeastern USA. Research regarding Rosenwald schools has focused on their impact on educational outcomes where they were built but has not as often considered their relative spatial patterns. Few projects have even investigated the locations of the schools, and those that have do not analyze spatial patterns or spatial context of the school locations [
24]. This study presented Rosenwald schools in a way that has not been investigated before and was able to provide location-based data for 89% of all of the Rosenwald schools in South Carolina.
Other historical locations have also been rigorously mapped, despite limitations in geographic information availability. Historical documents and resources have been used in many efforts to map lost sites as accurately as possible, like what was determined concerning the abandoned medieval town on the Sekanka promontory [
25] and identifying other lost cultural heritage assets using GIS and old planning maps [
26]. Some have required extensive historical research, as we carried out, in order to map a phenomenon accurately [
27]. Others have even tried to recreate detailed building footprints of lost historical buildings using a deep learning model and historical map data [
28]. These studies show that geospatial techniques, like those involved in this study in the form of GIS and cartographic representation, provide new tools for advancing historical mapping [
29,
30,
31]. However, few have investigated historical building locations at a state or provincial scale of analysis from a geographical perspective. This work presents the historical context of Rosenwald schools in 1920s and 1930s South Carolina in a spatial framework, providing insight into that complicated time American history.
Our research has answered the three research questions originally postured in the introduction. First, we documented that Rosenwald schools were not equally distributed throughout the state. Temporally, most schools were built between the years 1923–1927, and spatially more were built in the northern parts of the state than the southern parts. In fact, fewer Rosenwald schools were found in the Lowcountry region, where the percentage of Black population was higher than White population by a significant margin. This is well represented in the ratios of schools to Black population from the 1930 census. The Rosenwald School to Black Population ratio for the Lowcountry in 1930 was 0.00033, while the ratios for each other region were 0.00060 and above. Additionally, the Upstate, Midlands, and Pee Dee regions each shared a near equal proportion of the Rosenwald schools even as the proportions of Black residents to White residents declined over the 10-year period insomuch that the White population became the majority in the Midlands Region (Blacks continued as the majority in Pee Dee and Whites in Upstate). This aids in addressing the second question as well regarding educational opportunities for Black individuals in South Carolina. Unfortunately, the data showed areas of need still where no schools were located in areas of significant Black populations (
Figure 7).
Finally, we also addressed the contributions made by the Rosenwald fund, public funds, Whites and Blacks. As part of the philanthropic donations by Rosenwald, he required contributions to be made by others to support his original donations. This, in fact, allowed his funds to go further and assist more individuals in the pursuit of education. Statewide, South Carolinians contributed significant amounts of funds through public tax dollars and private donations from both Black and White individuals. One of the more interesting findings is the proportion of dollars donated by the White population in the Lowcountry region, where slavery was prevalent and racism has been documented throughout history. This tells another side of the story, a more positive and uplifting one to juxtapose against deplorable racist history. Despite being a minority in the region, the White population worked together with the Black population and public funds to supplement the Rosenwald donations to build the few schools within the region.
Some limitations of the analysis include the lack of detailed spatial information or outdated spatial information on the Rosenwald school cards as recorded by the South Carolina Department of History and Archives. A significant number of these cards lacked any real direction, so it was necessary to rely on primary and secondary historical sources to provide geographic context for the school, along with when and why it was demolished or removed. Despite the care taken throughout the research process, there are also potential biases and errors in the data researched for each of these schools through online and book resources. We attempted to mitigate these issues by using a review process of investigating other sources to find related information, but errors may still be in the data. Other limitations include the lack of other census-related datapoints and surveys that might provide further context for the Rosenwald school locations.
Further research is required on Rosenwald schools throughout the entirety of the southeastern United States. Some resources are still being procured and prepared through a recent National Park Service project involving several entities in the state of Louisiana. The project includes digitization through remote sensing technologies of the schools still standing, which should be expanded to South Carolina. Furthermore, crowdsourcing geospatial data, or using volunteered geographic information, is a good opportunity for finding the remaining Rosenwald schools in the state that have yet to be located. There may be individuals in the state who experienced these schools as youth and who could provide better context for their locations.