AN EXAMINATION OF A SOCIAL ACTION PROGRAM WITH EX POST FACTO METHODS AS EXHIBITED IN AN EVALUATION OF THE FARMERS HOME ADMINISTRATION By WALTER ERWIN BOEK A THESIS Submitted, to the School of Graduate Studies of Michigan State College of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Sociology and Anthropology 1953 ProQ uest Num ber: 10008263 All rights reserved IN FO R M ATIO N TO ALL USERS The quality o f this reproduction is dependent upon the quality o f the copy subm itted. In the unlikely event that the author did not send a com plete m anuscript and there are m issing pages, these will be noted. Also, if m aterial had to be rem oved, a note will indicate the deletion. uest ProQ uest 10008263 Published by ProQ uest LLC (2016). C opyright of the D issertation is held by the Author. All rights reserved. T his w ork is protected against unauthorized copying under Title 17, United States Code M icroform Edition © P roQ uest LLC. ProQ uest LLC. 789 East E isenhow er Parkw ay P.O. Box 1346 Ann Arbor, Ml 4 8 1 0 6 - 1346 Acknowledgements The research presented in this dissertation has several phases, all dependent on the assistance and cooperation of many people. The dissertation was a methodological treatise, and I am in debted to Dr. Duane L. Gibson who provided stimulation to explore new methods for securing and analyzing the data. Dr. Charles R. Hoffer*s many suggestions throughout the planning and carrying out of the project and the preparation of the manuscript were exceedingly valuable. The assistance of these two men as co-directors of this dissertation was sincerely appreciated. The research project of which this Is one report owes its initia tion to Dr. Charles P. Loomis and Dr. Paul A. Miller as leaders of a committee which included Dr. Edgar A. Schuler, Dr. Duane L. Gibson, Dr. Charles R. Hoffer, Dr. Solon T. Kimball, Dr. Asael T. Hanson, and Dr. Orden C. Smucker. These scientists drew up a plan for a project and worked out arrangements whereby the Michigan Rehabilitation Corporation made available necessary financing through the Social Research Service of Michigan State College. It was a privilege to begin with their think ing and to consult with each of them as the work progressed. I appreciated the freedom allowed by this project committee to pursue methodological deviations. Dr. J. Allan Beegle, Dr. Lawrence T. Hitt, and Dr. Raleigh Barlowe gave guidance during the progress of the project and the preparation of the report. Jean K. Boek expertly handled the difficult job of field work. My deepest thanks are due her for this as well as for continued interest and aid in analyzing data and preparing the report. Acknowledgement is due Mr. and Mrs. Charles R. Procter for field work and to Mrs. Evelyn P. Horton for typing assistance. Mr. Roswell G* Carr, Director of the Michigan Rehabilitation Corporation and Administrator of the Michigan Farmers Home Administration had a strong desire to assess the role of his organization in its work with farm families. It was this progressive attitude that made possible the use of Michigan Rehabilitation Corporation funds. The Farmers Home Administration supervisor in the county of the field work was Mr. Fred Boyd. Fred, as he was known throughout the county, gave us excellent assistance.. His keen memory enabled him to describe families, many of whom had completed their participation in the Farmers Home Administration program as many as 15 years before this research. The cooperation of Mr. Patrick J. O rMalley, State Field Representative, Farmers Home Administration, and all other Farmers Home Administration personnel was appreciated. Thanks a.re extended to Hillsdale County families we had the opportunity of visiting. It was a pleasure to work closely with Mr. Fred Boyd and his colleagues who opened their files and their minds to aid us in our attempt at an objective appraisal of their lifeTs work. Contents Chapter I II III IV V VI Page Introduction. 1 Background of Farmers Home Administration: The Social Action Program Studied in this Research. 8 The Role of the Ex Post Facto Method in Experi mental Research. 40 Data for this Study. 50 Selection of Variables to be Controlled. 64 Findings. 72 Part Part Part Part A. B. C. D. Part E. Part F. Part G. Part H. VII Introduction. Comparison ofTotal Groups. Comparison ofMatched Total Groups. Comparison ofNon-FHA Families •with FHA Families who Named Them. Comparison ofNon-FHA Families Matched with FHA Families who Named Them. Comparison of Closely Paired FHA and Non-FHA Families. Comparison of Precisely Paired FHAand Non-FHA Families. Summary of Six Methods. 72 72 84 97 105 113 125 129 Summary and Conclusions. Part A. Part B. Part C. Introduction Generalizations About the Farmers Home Administration. Generalizations About Ex Post Facto Research Methods. 136 136 138 143 Appendix A. Bibliography. 147 Appendix B. Rate of Change Analysis. 151 Chapter I* Introduction Human society is dynamic. Most of the changes occurring con tinually are toward the objective of satisfying the needs of the total culture, and the individuals within it are usually relatively unim portant entities • However, with the accumulation of knowledge about his physical and social environment, man has increased his ability to consciously influence change. This was indicated by Ward when he saids "It is man that is active [Environment is passiv§Q . His will guided by his intellect is ever pressing against the en vironment. In proportion to the development of the guiding faculty man removes the obstruction presented by the environ ment. In the more advanced stages he transforms it, utilizes it, subjects it to his service, and compels the very powers that at first opposed his progress to serve his interests and supplement his own powers. It is this that constitutes civi lization, and to the original natural environment there is now added am artificial environment (social^ of his own cre ation. This... is of far greater vital importance to him than his natural environment, the physical world into which he is born. Yet to this human action the environment opposes its re action, and it is this interaction of man and his environment, or synergy, that accomplishes the results Civilization] This attempt of an individual or a group of individuals to pur posely initiate change may be called social planning. Ward has applied another term, "collective teles is", to this phenomenon.^ Social planning exists when man is not willing to accept the fu ture consequences of present behavior* Successful planning requires the ability and tools to predict so that future tendencies seen as undesirable may be changed. Alexander Leighton pointed out that suc cessful planning in international human relations must utilize m o d e m social science to collect and analyze data on the problem for which a 1 2 Lester F. Ward. Applied Sociology. 1906. P. 1 3 1 . -----------Ibid. P. 12. New York. Ginn and Company. certain change is desired.^ Ronald Freedman and his colleagues define planning as: "The systematic application of the best available knowledge to the control and direction of current trends of change for the purpose of securing stated objectives.”4 A good plan projects existing trends into the future, according to O g b u m and Nimkoff, just as if no planning were contemplated* After this has been done, the extent to which modifications can be accomplished is determined. These two authors continue: "Planning has the virtue of looking ahead, which is e— ssential in a changing society. Unless there is what Comte called 'prevision*, it is idle to talk of control."5 .Planning is preventive and constructive. A plan is laid out as an achievement to be made in a certain length of time. The emphasis is on the practical rather than on aspi rations of the fantasy type* ...A plan is usually not con cerned with eternity or infinity."® The acceptance of the role of planning in the world and es pecially in the United States is demonstrated by the large number of agencies whose entire function is to draft plans. Community planning organizations for slum clearance, hospital construction, road build ing, and school reorganization are well known among civic leaders. Sometimes plans developed by visionaries in these groups are carried out voluntarily by the members of the society and other times laws and regulations are required such as zoning ordinances or the power of eminent domain •^ 3 4 5 6 7 Alexander H. Leighton. Human Relations in a Changing World.. New York. E. P. Dutton and Company, Inc. 1949. P. 7* Ronald Freedman and others. Principles of Sociology: A Text with Readings. New York. Henry Holt and Company. 1952. P. 659. William F* Ogburn and Meyer F. Nimkoff„ Sociology. New York. Houghton Mifflin Company. 1950. P. 585. Loc. cit. Solon T. Kimball. A Case Study in Township Zoning. East Lansing, Michigan. Michigan State College Experiment Station Bulletin. 1947. -3- Many plains involve much more than single communities; some are extended over the whole country. A few of the early nation-wide plans in the United States were designed to influence trends in agriculture. A significant one concerning the education of farm youth was the Morrill Act of 1862. Its purpose was to provide unsettled western land to states for the purpose of securing funds through their sale or use for: "...The endowment, support, and maintenance of at least one college where the leading objects shall be, without ex cluding other scientific and classical studies and including military tactics, to teach such branches of learning as are related to agriculture and the mechanic arts,..., in order to promote the liberal and practical education of the industrial classes in the several pursuits and professions of life."® The central problem of many other agricultural plans concerned the future of the family-type of farm, since it was felt that family farm people were one of the most important stabilizing factors in democracy.^ Family farm policy has been described by Kolb and Brunner in this manner: "The announced objective of a number of laws dealing with agriculture enacted by Congress in the last two decades has been to strengthen the family fairn."^ The authors go on to say that: "...One early effort in this direction was that of the Homestead Act to enable settlers to acquire farms on gener ous terms. A more recent nation-wide farm program which similarly had as its 8 "College-Aid Land Appropriation. Agricultural-Mechanical Colleges." United States Code Annotated<, Title 7_. Agriculture. St. Paul, Minnesota. West Publishing Company. 1939. P. 157. 9 Some questions of the relative importance of the family farm as a democratic force have been raised^ however, See for example A. Whitney Griswold. Parming and Democracy. New York. 1948. 10 John H. Kolb and Edmund deS . Brunner. A Study of Rural Society. New York. Houghton Mifflin Company. 1952. P. 92. 11 Loc. cit. •4- central them© the developing and preserving of the family-farm type of agriculture was the Farm Security Administration established by Congress in 1937* The significant function of this selected tenants purchase family size farms or farm agency was to help supplies by loan- ing them the necessary money which they could not get elsewhere. Low interest rates and a long repayment period were provided with some guidance in farm and home management. The necessity for the Federal government to provide this means for rural families to purchase farms or farm supplies can be seen in agricultural trends. The gradual commercialization of agriculture has modernized it, bringing many conveniences for rural people. ^ At the same time this process has intensified class differences in the farm population.13 it has become increasingly difficult for upward mobil ity to occur, although sustained high agricultural and industrial in come has in part alleviated this condition for some rural low socialeconomic families Barriers to such mobility, which in rural areas usually means moving up the tenure ladder to farm ownership from laborer or renter, are both economic and social, with the factors involved being inter dependent. i'he large amount of capital needed to enter and. carry on commercialized agriculture is one economic barrier. The years of 12 Carl C* Taylor, Helen M. Wheeler, and E. L. Kirkpatrick. Disad vantaged Classes in American Agriculture. Washington, D.C. Social Research Reports. No, VIII. April 1938. 13 For a discussion of hierarchial strata in rural society see Charles P. Loomis and J. Allan Beegle. Rural Social Systems. New York. Prentice-Hall, Inc. 1950. Part ll'l.' 309-391. See also Arthur F. Raper. In. Carl C. Taylor and others. Rural Life in the United States. New York. Alfred A. Knopf. 1949. 14 This is especially true in areas where it is possible to have sub stantial non-farm income and/or where it is feasible to start with small ihrm enterprises and to enlarge them slowly over many years. training and experience necessary to conduct a farm as a successful business enterprise is another. Therefore when farms become available through death, retirement, or mobility of the owner-operators, several alternatives exist if the farm is to remain in agriculture. It can be operated by a tenant and maintained with absentee ownership} it can be combined with other farms under a managerial system interested in highest profits or smallest losses from land, labor, and machinery; or, it can be sold to another farm family. If it is to be sold to another family, there is high probability that the prospective buyer family will not have enough capital to pay the cost of the landand/or Thus it can be seen that the equipment and livestock to operateit. it is necessary for social machinery such as the Farm Security Administration to exist which will overcome these problems if familiesareto remain the primary production units in American agriculture and if the ideal of family type farms is to be continually realized. Charles P. Loomis and J. Allan Beegle in writing about this government loaning program saids "It is the authors* belief that few inventions in rural administration have been more important than the supervised rural rehabilitation loan. ...Borrowers had the advantage of community status as parts of the free enterprise system or the contractual Gesellschaft• They also had the advantage of pro fessional guidance which had to be followed if they were to continue as borrowers The effort to reinforce the family farm in American agriculture continued when Congress set up the Farmers Home Administration in 1946 to replace Farm Security Administration.16 It carried out about the 15 Charles P. Loomis and J. Allan Beegle. Rural Social Systems. New York. Prentice-Hall* 1950. P. 688* 16 The Farm Security Administration and Farmers Home Administration are discussed in detail in Chapter II. same activities as its predecessor. In Michigan the administrators of the Farmers Home Administration felt strongly about doing the right kind of a job with their agency. They believed in the plan of providing the means of helping the Ameri can democratic value of upward mobility to remain a reality among farm people. They wished, however, to view their program objectively in order to determine what it had accomplished and how it could be im proved in the future. To do this, these administrators encouraged the Board of Direct ors of the Michigan Rehabilitation Corporation to finance an evalua tive study of the Farmers Home Administration program in Michigan. The Board requested members of the Social Research Service at Michigan State College to consider the possibility of conducting such an appraisal. The attitude of the research group toward this project was favor able since it was felt that it would provide an empirical case for studying planned social change. Thus a grant was made by the Rural Rehabilitation Corporation to the Social Research Service for measur ing the influence of the Farmers Home Administration program on the participating families. This overall evaluative project of this agency of change required the developing of and experimenting with methods for measuring social change which could be attributed to the Farmers Home Administration’s program in a complex dynamic society. The fact that the program had already been in operation many years before the research was initiated further necessitated the utilization of methods which would allow measurement of such changes that had already occurred. This report in the form of a dissertation, then, has two object ives j (i) to determine the effects on families of participation in the Farmers Home Administration program, and (2) to describe and discuss research procedures which were used in the attempt to quantitatively measure these effects. Carrying out the first objective provided the context within which development and testing of the research pro cedures of the second objective could be handled. -8- Chapter II# Background of Farmers Home Administration - The Social Action Program Studied in this Research Many sociologists, especially those concentrating their attention on phenomena in rural areas, recognized the need for research which would contribute to solving the problems during the depression years of the 1930fs. Their interest was encouraged by financial support from adminis trators in Government programs which made possible nEiny analyses of their records and of other data. The resulting scientific publications along with the more popular type articles on rural rehabilitation and programs developed by Federal and state governments to meet rural depression problems were reviewed for information which might improve this research design. Those related to this dissertation are reported in this chapter. Much of what has been written about the FHA is not pertinent to the general purpose of the present study since many articles and publications merely describe the rehabilitation process or projects. were not useful in designing this research. Most of these V/hen this was true they were listed and marked as not related to this project in the Bibliography appearing as Appendix A. Many of the articles and research reports pub lished during or immediately after the depression seem to lack the perspective of those written since 1942. Traditionally, farmers have been thought of as independent and as capable of absorbing any catastrophe within their own systems of relationships, such as family, community and church associations. Because farmers were thought to be too proud to accept any charity, broad social machinery had not been developed in farm communities to discover -9- or care for poverty,*^ The existence of rural poverty was pointed out by Taylor, Wheeler and Kirkpatrick when they said that within agriculture, as within every other great economic enterprise, people are continually operating at all levels of success and failure. Agriculture is one of the most diversified economic enterprises in the world (for the individual family involved} and it is carried on under extremely varied and changing con ditions, many of which cannot be c o n t r o l l e d , B e c a u s e of this, some of its entrepreneurs and many of its non-entrepreneurs are at all times operating so near the margin of economic safet3^ that a major economic depression would thrust them immediately below the poverty line, American rural life has always had its full quota of disadvantaged families, although it is almost certain that the number has increased in each succeeding generation,*^ Moreover, the commercialization of ag riculture has thrust practically all farm families into the price and market system and has subjected them much more directly than in the past to economic depression,^ The public at large has not been aware of the poverty that has constantly existed among farm people. Merely living on a farm, especially on a farm one does not own, is no proof of economic security, much less proof of economic and social well-being.^ The great number of farmers requiring relief began to come to public attention in 1932, By the next year one million farm families were on relief, and within two years about two and a half million 17 Taylor, Wheeler, and Kirkpatrick, P. 1. k°°• cit, 19 Ibid, Pp. 1-2, 2° Ibid, P. 2. 21 Ibid. P, 1, Disadvantaged Classes. Op. cit, -10- rural families, totaling more than 10 million people, depended on some form of relief for their s u r v i v a l F a r m e r s not able to make a living on their farms and unemployed urban workers were given the same help in the Federal Emergency Relief Act of 1933, In 1933 a few states made loans of tools and workstock to impover ished farmers in place of direct or work relief in an attempt to help them become self-sustaining* Thus the basic elements of a new idea of ^ rehabilitation— loans plus supervision— was established by the states •2 ° These state attempts at rehabilitation attracted national attention and suggested the organization of a Rural Rehabilitation Division in the Federal Emergency Relief Administration, This division, which began functioning April 1, 1934 under Federal financing, was organ ized on a state basis with State Rehabilitation Corporations to handle the new program. 04 The objective of this program was to enable destitute persons eligible for relief to sustain themselves through their own efforts. Help could have been either financial or advisory, but usually it 22 Joseph Gaer. Toward Farm Security, Washington, D.C. United States Department of Agriculture, 1941, P, 50. 23 Olaf F, Larson. MRural Rehabilitation— Theory and Practice•” Rural Sociology. Vol. 12, 1947. P, 226, 24 The language of the statute concerning financing by the Federal Government was; ”To meet the emergency and necessity for relief in stricken agricultural areas to remain available until June 30, 1935, $525,000,000 to be allocated by the President to supplement the appropriations heretofore made for emergency purposes and in addition thereto for making loans to farmers f6r purchase, sale, gift, or other disposition for seed, feed, freight, summer fallowing and similar purposes.” This was taken from: The Statutes at Large of the United States of America from March 1933 to June 1934. Vol. 58. Part 1. (Under Title II Emergency Appropriations, 73rd Congress. Session II. June 19, 1934). -11- was both. Loans made to farm families enabled them to buy the nece ssary livestock, feed and supplies. These were to be repaid in cash, in kind, or in work on self-liquidating projects as soon as circum stances permitted. Farmers receiving loans from these corporations had to work out a farm and home management plan designed to provide the family with most of its own food and to help build up the soil and improve general farm practices.^5 YJithin one year more than 250,000 families had received such rehabilitation loans.^ The law setting up the Rural Rehabilitation Division made provision for its operation only until June 30, 1935. From its beginning in April of 1934 to the start of its replacement, the Resettlement Administration on June 30, 1935, some $49,000,000 was advanced to about 397,000 families The Resettlement Administration included State Rehabilitation Corporations, Federal Emergency Relief Administration resettlement projects and subsistence homestead projects, the Farm Credit Adminis tration, debt adjustment activities, the submarginal land program, and new functions to aid tenants, sharecroppers, and migratory 25 Gaer. Op. cit* Pp. 51-52. Also see Joseph W. Eaton, ploring Tomorrow1s Agriculture. Few York. B rothers • 1943• 26 Loc. cit. 27 Larson. ’’Rural Rehabilitation.” Op. cit. Harper and P. 226. Ex -12- workers,^® Many of the rehabilitation families were overburdened with debts which were adjusted before they worked out satisfactory farm and home managed plans. As part of its program the Resettlement Adminis tration assisted groups of low-income fanners who needed equipment or services and arranged for a community or a cooperative service loan to the group. This enabled neighboring farmers to obtain tractors, combines, veterinary services, spraying equipment and other goods and services which no single farmer could buy for himself. In 1936 President Roosevelt appointed the Special Committee on Farm Tenancy to report on a long-term program of action to alleviate 28 This is the content of the Executive Order establishing the Re settlement Administration: By authority under Emergency Relief Appropriation Act of 1935, approved April 8, 1935 (Public Resolution Ho, 11, 74th Congress) the President established an agency within the Government to be known as the Resettlement Administration, and appointed R. G, Tugwell, Under Secretary of Agriculture as Administrator without additional compensation. The duties of the Resettlement Administrator were: a. "To administer approved projects involving resettlement of destitute or low-income families from rural and urban areas, including the establishment, maintenance, and operation, in such connection, of communities in rural and suburban areas” , b, "To initiate and administer a program of approved projects with respect to soil erosion, stream pollution, seacoast erosion, reforestation, forestation and flood control", c, "To make loans as authorized under the said Emergency Relief Appropriation Act of 1935 to finance, in whole or in part, the purpose of farm lands and necessary equipment by farmers, farm tenants, croppers, or farm laborers". In the performance of such duties the Administrator is authorized to employ services and means in subdivision (a) of Section 3 of the said Emergency Relief Appropriation Act of 1935, To the extent it is necessary to carry out the provisions of this Executive Order, the Administrator is authorized to acquire by piirchase or by the power of eminent domain any real property or any interest therein and improve, develop, grant, sell, lease (with or without privilege of purchasing) or otherwise dispose of any such property or interest therein. For administrative expenses the Emergency Relief Appropriation Act of 1935 has $250,000 allocated from it. Source of this statement was: "Establishment of the Resettlement Administration". Executive Order 7027. April 30, 1935, -13- the shortcomings of the farm tenancy system.^ One type of action the committee recommended was modest loans, with the necessary guidance and education both to prevent small owners from slipping into tenancy and to help the masses of tenants, croppers, and farm laborers at the bottom of the agricultural ladder increase their standards of living, achieve greater security, and begin the upward climb toward land ownership. 30 w On September 1, 1937, the Farm Security Administration, organized under the Department of Agriculture, became the successor of the Resettlement Administration• Its purposes, changed somewhat from those of its predecessor, were to aid low-income farm families to re-establish themselves on productive land and to become self-supporting through a threefold program of rural rehabilitation, farm purchase by tenants, and rural homesteads. This aid was given to help alleviate immediate needs and to ultimately eliminate rural poverty and insecurity on 29 Farm Tenancy, Special Committee on Farm Tenancy. Washington, D.C., Government Printing Office. Superintendent of Documents. 1937* Two paragraphs from this report of special interest here are: ’’For the past 55 years, the entire period for which we have statistics on land tenure, there has been a continuous and marked decrease in the proportion of operating owners and an accompanying increase in the proportion of tenants. Tenancy has increased from 25 percent of all farmers in 1880 to 42 percent in 1935. Because of debt the actual equity of operating owners is far less than these figures indicate. In some of our States, among them a number settled under the homestead system little more than a generation ago, it is estimated that the equity of operating farmers in their lands Is little more than one-fifth. Dearly four-fifths is in the hands of landlords and mortgage holders.” ’’The committee’s examination of the agricultural ladder had indicated a series of groups of farm families whose insecurity is a threat to the integrity of rural life. The families comprised within these groups constitute fully half the total farm population of the country. Approximately one farm family out of four occupies a position in the Ration’s social and economic structure that is precarious and should not be tolerated.” The principal groups it listed were: Tenants, croppers, and farm laborers. 30 Larson. ’’Rural Rehabilitation” . Op. cit. P. 226* -14- the f a r m . ^ One type of aid was given to farmers 7/ho had accumulated debts they could not pay* The FSA cooperated with groups of local farmers, businessmen and townspeople who acted as farm debt adjustment com mittees to whom any farm debtor or creditor could bring their problems. Farm debts could be reamortized, interest rates could be changed, or an extension of time could be arranged to allow a reasonable chance for the debtor to repay. Another type of help was a grant given to farmers in areas such as those stricken by drought, flood, or other natural catastrophes, or in areas where the land had been seriously eroded in which aid could not be deferred until a rehabilitation plan had been developed. In those cases and in instances of illness or other factors which reduced families to a low financial level, an immediate subsidy 7/as made to tide them over to the time when they could request loans. Another part of the program was the loaning of funds to tenants, sharecroppers, or farm laborers selected and approved by local committees in the county to buy their own farms, livestock, machinery or supplies. 31 The function and duties of the FSA which were to be exercised by the Administrator as stated in the law were: 1, To administer approved projects involving rural rehabilitation, relief in stricken agricultural areas, and resettlement of destitute or low-income families from rural and urban areas, including the establishment, maintenance, and operation, in such connection, of communities in rural and suburban areas, 2. To initiate and administer a program of approved projects with respect to soil erosion, stream pollution, seacoast erosion, reforestation, forestation, flood control and other useful projects, 3. To make loans as authorized under the Emergency Relief Appropriations Act of 1935 to finance in 'whole or in part the purchase of farm lands and necessary equipment by farmers, tenants, sharecroppers or farm laborers. The source of this was: The Code of Federal Regulations of the United States of America Having GeneraX Applicabi 11 ty and Legal Effeot in Force June 1_, 1958. Title 6, Agricultural Credit. Washington, D.C. United States Government Printing Office, 1939. 15- Loans for buying farms were called ’'tenant purchase”, bore three percent interest and were to be paid within 40 years. They were to be settled on a variable payment plan, meaning that in poor crop years or when the borrower could not make due payments in full for any other reason, he could pay according to his means and then make proportionately higher payments in good years. Loans for buying machinery, livestock, or other supplies were termed "standard loans" and were usually of about five years duration. It is these two types of loans, tenant purchase and standard, which are of concern in this study* A farm family desiring a loan from the FSA would visit the county administrator's office where they filled out an application. be discussed by a committee of farmers from their county. This would The family and the plans for use of loan funds would be investigated to determine whether or not there was a lack of credit, whether or not the family had ability to rehabilitate themselves with the assistance, and If the farm which they wished to buy or operate could support them so they could repay the loan. If the initial decision was in favor of the committee, a farm and home plan would be worked out by the family and the agriculturally trained or experienced county FSA administrator. After a loan had been received by a family, they would oontinue to receive guidance from the adminis trator and his staff, which usually Included a woman with home economics education for assisting the wives. The Farm Security Administration was replaced by the Farmers Home Administration on August 14, 1946. The purpose of the Act creating the FHA was to improve credit services to farmers and promote farm owner ship by functions of the FSA, Farm Credit Administration, and national Housing Agency in one division. It further authorized government insurance of loans to farmers, created preferences for loans and insured 16- mortgages ho enable veterans to acquire farms, and provided specific authority with respect to liquidation of resettlement projects and rural rehabilitation projects for resettlement purposes. 32 The Farmers Home Administration in Michigan at the time of the field work of this study in 1950 was administered in this manners In all counties where a large number of loans had been made there was a county supervisor. A supervisor administered more than one county when the number of participant families was low. received guidance from a district supervisor. This county supervisor Michigan counties were divided into several districts, each having a district supervisor in charge who was responsible to the state administrator. The county supervisor could obtain specialized help in such fields as home economics or home remodeling from the state office through the district supervi sors• A study of Federal, state and county FSA activities and other rehabilitation programs was made by Olaf F. Larson, assisted by Bureau of Agricultural Economics staff members. The published report was; Ten Years of Rural Rehabilitation in the United States, which summarized research in the rehabilitation of low-income rural f a m i l i e s .^3 It was in part based on an unpublished BAE study made with the assistance of the Works Project Administration. This included a study of 39,295 *32 ’’Farmer's1 Home Administration Act of 1946.” United States Statutes at Large, Containing the Laws and Concurrent Resolutions Enacted During™ the Second Session of the 79th Congress of the United States of America. 1946. Vol. 60. Part I. ^Public Laws and Reorganization Plans Washington U. S. Government Printing Office, 1947. P. 1062. 33 Olaf F. Larson and others. Ten years of Rural Rehabilitation in the United States. Washington, D.C. United States Department of Agriculture. Bureau of Agricultural Economics. July, 1947# -17- standard. rural rehabilitation loan borrowers, distributed among the 12 continental Farm Security Administration regions and covering all such borrowers accepted during the period from March 1, 1936 to February 28, 1939* Other studies of the BAE and state agricultural experiment stations were drawn upon* Files and reports of the FSA were available for analysis.34 Emphasis of Larson’s publication was on the prewar standard rural rehabilitation loan program, but war-induced changes were indicated. This intensive statistical analysis of rehabilitation agencies’ records is probably the most complete analysis in this field. It, along with the article by Larson entitled ’’Rural Reha bilitation - Theory and Practice” already cited, presented many suggestions for the Michigan study.35 sively in this review. Both of these were used exten Most of the other references in the Bibliography which is Appendix A contained data from these two publications. One purpose of his research was to record significant experience accumulated in the United States in conducting the rural rehabilitation loan program. Another was to evaluate this experience in terms that would be useful for any continuing or new programs* This evaluation covered the basic lessons learned vri.th respect to principles, tools, and techniques of rviral rehabilitation from the standpoint of public policy* Obstacles encountered in making the study were that certain aspects of the program such as supervision, tenure improvement, forms of cooperative activity, ITegro communities, and special areas had not been studied adequately as of that time. In some of these instances 54 Ibid". P. 12. 35 Larson. ’’Rural Rehabilitation.” Op * cit. -18- particular activities had been discontinued, the personnel acquainted with such activities had been widely dispersed, and Farm Security Administration files contained little information* In those cases the procedure followed was to report what was known rather than to take time for further research, although it was realized that this gave little basis for evaluation* Data on programs auxiliary to the program of financing machinery, livestock, supplies and other farm and home purchases — the standard loan program, such as grants and debt adjustment, were not always reported so that other types of borrowers or non-borrowers could not be separated from standard loan borrowers* For these reasons a complete picture of the rehabilitation activities for standard loan families was impossible. Larson, spea Icing primarily about the standard loan program, said that the whole program has rested upon a set of assumptions. The first was that causes of poverty for the majority of those affected lie outside the direct control of the individual. Except for certain physical disabilities, accidents, and natural calamities, most poverty has a social origin, being grounded in the nature of our society, its culture and the relationship of the individual to his social environment. The second assumption was that poverty originating in social causes is subject to prevention and correction by social action. Third, for the most part there is nothing fundamentally wrong with poverty-stricken rural families, ^hey are willing and able to earn their own living and can do so if given the right kind of a chance. social action to help them. They will respond to proper Fourth, society has a responsibility to alleviate and prevent poverty since the security of our society, not simply the welfare of an individual family, is at stake. The fifth basic assumption was that the resources for lasting rehabilitation rest -19- within the people themselves. For this reason the program must be directed toward tapping and mobilizing these resources. This last premise is opposed to paternalistic and long-continued supervision#*^ According to Larson the objectives of the standard loan program at first were to relieve the suffering of individual families and to restore them to permanent self-support. As the program continued and an under standing of rural poverty developed, the objectives broadened. The goals, stated in somewhat philosophical terms, were for the individual family to obtain a physically healthful and a socially desirable level of living and to acquire the skills needed to manage their own farm and home successfully and independently. Other goals for the family were to achieve security, to obtain land enough for an economic unit of the family-farm type, to become a full participant in a democratic way of life, and later, to have maximum employment of family labor in the production related to the w a r . ^ The program has also had broad objectives for agriculture, rural society, and the welfare of the nation. They are: to save the tax payers* money, rehabilitation being viewed as more economical than either direct or work reliefs to preserve, reinforce, and perpetuate the family-type farm; to bring about a better adjustment of population and land resources and to conserve land resources; to achieve full utilization of manpower in the low-income farm population— this being a war measure; and to make available to all America the opportunities of a stake in democracy. The whole program is an instrument of policy# To some degree loans have been made or denied to enforce policy, 36 Ibid. 37 Ibid# Pp." 230-231. Pp. 229-230. -20- as, for example, to promote better land use, to encourage population movement out of agriculture, to secure maximum national production and most efficient use of the nation’s manpower in wartime, and to discourage migrati o n . ^ These examples were specific uses to which the program or parts cf it were put in order to exert a force of change or a force of stabilization in special areas. Policy goals have not been concisely stated in official pronounce ments but are implicit in the methods used to carry out the program. Means such as these: live at home methods, diversified farm, tenure improvement, debt adjustment, farm and home plans, and pressure cookers achieve policy ends. Larson deals primarily with the standard loan in his report and in his article. It was the most significant of the publicly sponsored rural rehabilitation activities, as measured by number of families reached, amount of money Involved, geographic coverage, or duration of program contacts with families.^ Chief characteristics of the standard loan program include (1) credit for normal farm and home operating expenses to farm families unable to obtain satisfactory financing from any other private or federal source; (2) supervision, or advisory assistance; (3) and rehabilitation in place, that is, without moving families to resettle ment areas. The standard loan has not been authorized for the purpose of purchasing real estate. is the tenant purchase loan. 38^TbId.“ 39 Ibid. PT~228 . P. 224. The loan authorizing real estate purchases Financial aid for both types of loans is -21- of a high-risk hyp© from a banking point of ■view. The county adminis trator and his staff supervises the loans which include developing farm and home plans and bringing them up to date each year and teaching of improved farm and home practices. The administrator, known as a supervisor, ideally performs an educational function designed to improve the status of the supervised and to develop the families1 skills for their own welfare. Larson says that supervision is the tool which ideally serves to coordinate and unify all the other tools and techniques applied to meet the problems of the individual family. A Farmers Home Administration supervisor performs the functions of a farm or home management advisor, family case worker, group organizer, and banker. Periodic visits are made to the farm and home to carry out these functions. A farm and home plan developed jointly by family and supervisors is supposed to serve as a working guide for the family and to be used as a basis for processing loans, determining repayment schedules, adjusting debts, and planning the amount and kind of supervision. Credit is the tool which gives the program agency a degree of control over the individual families, since the granting of credit is conditional upon the family*s agreement to carry out the farm and home plan worked out with the supervisors. Techniques other than supervision, adapted or developed to promote changes which would bring about rehabilitation include: (i) improvement of tenure; (2) group services established with or without loans to provide needed services or facilities, such as sires or machinery, where family circumstances did not justify individual ownership! (3) group health plans to provide medical and dental care, hospitalization, and drugs; (4) adjustment of debts; (5) environmental sanitation; (6) special types of loans such as for water facilities and 4-H club ac tivity; and (7) specialized programs in limited areas to develop new methods or to meet localized n e e d s . ^ TJhile other techniques have been adapted or developed to aid families in rehabilitation, it is the combination of credit with super vision -which is the unique feature that has resulted in the standard loan program being called a ’’social invention” of high significance.^ Larson contends that an understanding of the rehabilitation process teaches the need for flexibility in the type and use of rehabilitation tools. It forces consideration of those social, economic, and physical factors which stand as obstacles to real rehabilitation. A basic pro gram must operate on a long-run basis, even though administrative considerations may compel a relatively short-time limit on rehabili tation aid. Even during drought and depression years some families were apparently rehabilitated In a year or two. Of families receiving their first loan during 1935, one in four was still active on the program during the ninth year following acceptance, and two of the four had failed during that period in the financial aspects of their rehabilitation. 4.0 In analyzing the way these several tools and techniques have been used to help families with a wide range of characteristics and problems, the conclusion is reached by Larson that the real meaning of rehabilitat ion or habilitation is the attainment of a cluster of interrelated econom ic and social goals which may be defined in terms of the program objectives 40 Ibid. 41 Ibid. 42 Ibid. F. 232. P. 224. P. 234. -23- for the individual family. The content of these goals may he expressed in terras of practices, behavior patterns, attitudes and material possessions. Eligibility requirements for the standard loan have changed from time to time. At first being eligible for relief was a major prerequisite. Later the upper limit was set by the inability to obtain credit at reasonable rates or terms from other sources. has been quite subjective. The lower limit Control of some land resources has been one of the qualifications, thus automatically barring from the program anyone who could not secure the useof a piece of land. In general the loans have been reserved for those whose usual occupation has been that of farm operator, including those recently displaced, and for sons of farmers wanting to start out for themselves. The program has not been used in general to encourage unemployed' industrial workers to enter agriculture. Emphasis has been on assisting full-time farmers, although some allowance has been madefor off-farm work which doesn’t interfere operations,43 with farming j\_s a rule the presence of an able-bodied man, evidence of industry, and apparent capacity to profit from supervision and financial aid have been among the primary requirements. According to Larson, a complex of problems, rather than a single problem, is characteristic of the situation of the typical rehabili tation family. These major problems include lack of (1) credit, (2) physical resources to produce Income and protect health, (3) skills to use resources, (4) institutional arrangements, (5) satisfactory social relationships, and (6) adequate personal traits. For many families the status and the duration of poverty was such that the term 45 Ibid. Pp. 227-228. 24- "rehabilitate" was not appropriate. They were being helped to a position they had never lost because they had never held i t . ^ An analysis of the purpose for which standard loans were made during 1936 through 1939 showed that 47 percent of the amount was authorized for capital goods, 30 percent for current farm expenses, 15 percent for debt settlement and refinancing, and seven percent for family exi-)enses*^5 Repayment schedules varied, depending in part upon loan purposes and borrowervs ability to repay, but generally the original loan was for a maximum of five years • In general the amount of the loans has not varied directly with the amount of economic resources the borrowers had available at acceptance, but it has been related to some minimum amount deemed necessary to provide working capital and meet current expenses. The average size of both the original loans and supplemental loans, made later to the same family, has tended to move upward primarily because of rising costs. The mean size of original loans went from $240 in 1936 to $1,007 in 1944 while the average supplemental loan moved from $67 to $316 during the same period. The average borrower active oil the program in 1943 had received a total of $1,310. Supple mental loans have been used extensively, accounting for 57.5 percent cf the number and 47.3 percent of the amount of standard loans during the whole period of 1935 through 1944, but in 1944 accounting for 85.4 A/> percent of the total number. By the end of 194-5, the number of standard loan borrowers had risen to 770,000 which was equivalent to one out of ©very eight farm operators 44 Ibid. 45 Ibid. 46 Ibid. P. 229. P. 232. Pp. 231-232. -25- reported by the United States Census. The x>roportion would be consider ably higher if the number of those operators for whom farming Is not the main source of livelihood were deducted from the census figures. At this time under the combined effect of good times and the collection policies, 45 percent of the borrowers were paid in full, 7 percent were “dropped", 22 percent were "collection only" and 26 percent were still active. Out of a total of 801.5 million dollars loaned up to the end of 1945, repayments on loans to both individuals and groups, excluding corporation trust funds, already amounted to 68.5 percent of the principal involved plus interest equivalent to 9.6 percent of the cumulative principal and much of the amount outstanding was not due. 4 7' These data on the financial situation did not reveal what happened to the families while they were paying back the loans. A comparison of the number of borrowers with several estimates of the number of farm families potentially eligible for loans makes it clear that by no means all eligible farm families have been reached.-® Significant comparison of standard borrowers with all farmers is handicapped by the fact that the data for all farmers reported by the United States Census includes those who because of age, health, or occupation are essentially rural residents rather than active farm operators. Compared with all farmers, however, borrowers at the time of acceptance were less likely to own the farm they operated. owned they were likely to have a poorer farm. If they If they were operators they had about the same size of farm, on the average, although not likely as fertile nor well-equipped.-^ 47 Ibid. 48 Ibid. 49 Ibid. Pr~235. P. 227. P. 229. -26- Borrowers were younger, had larger families, and were more likely to have received some form of relief aid during the thirties. They tended to have a lower level of living, smaller income, and to do some work off the farm. The average active standard borrower in 1943 had a net income from all sources which would place him just below the mid point in the distribution for all farmers. Comparison with those for whom farming was the major enterprise, rather than with all farmers, would place the rehabilitation group in a lower relative position on the income scale. The average borrower accepted in some areas would have to make tremendous progress to be as well off as were borrowers in other areas even before they received rehabilitation assistance, since poverty Is relative to standards of an area.®® The borrowers accepted under the Federal Emergency Relief Adminis tration rehabilitation program had the lowest socio-economic status, and the least amount of physical and financial resources of those accepted at any time. An upward trend in selection of borrowers was underway by the time the Farm Security Administration was started, and, except for a downswing in 193 9 resulting from administration pressure to dig deeper, continued into the war years. This trend developed from the fact that there were more applicants than money, larger case-loads than supervisors could handle adequately, pressure to make a good collection record, and generally unfavorable loan experience with the very low-incorne group. These forces resulted in the tendency to select applicants with the best prospects for repayment and requiring the least supervisory effort. War time objectives of maximum food production and labor utilization accen tuated this trend. 50 Loc. cit# 51 Ibid. P. 228. -27- The number of cases -which the supervisors have had were heavier than the 125 families per rural supervisor and 20 0 families per home supervisor which had been set up as desirable by the FSA. In October 1942, the average rural rehabilitation supervisor had the equivalent of 200 standard borrowers, and the home supervisor the equivalent of 286 families. During 1939 the average active borrower was reported to have received five supervisory visits. In view of heavy case loads and routine duties Larson points out that a real question may be raised as to how much supervisory assistance of an educational type has actually been received by the average borrower. He also stated that, although he had little proof, the speed of the rehabilitation process may be conditioned by (a) family characteristics, (b) physical and economical resources, (c) rehabilitation aids, (d) the culture as expressed by attitudes, social values, class structure, social and economical institutions and facilities within which the relationships of the family and the program are conducted, and (e) more or less impersonal forces outside the control of families or program agency, such as depression or natural calamities„®^ An analysis of the rehabilitation process suggested these generali zations to Larson. The setting of a definite time limit within which all families must achieve rehabilitation was incompatible with the na ture of the rehabilitation process, since, in general, the lower the economic and social level at the time of starting, the longer and the more difficult it was for a family to achieve rehabilitation. foreign were the basicassumptions andmajor to the local culture,the moredifficult vidual families within that culture.®® 52' TbTcT. 53 Ibid. PY P. ~23&. 234. The more objectives ofrehabilitation was therehabilitation of indi -28- Surveys of families, after acceptance and while 011 the program, indicated improvement for many in such observed items as health, tenure status, production of food for family use, household facilities, and equipment, Income net worth, working capital, land reso^irces, and leasing arrangements. Changes in such significant factors as attitudes, moral© and work habits, however, have not been adequately measured for oven active borrowers. Also, gains made while on the program by families now paid up, dropped or on a collection-only basis is largely an open question. Measuring change after acceptance Is impossible for many Important items unless the "at" or "before" acceptance situation is known.®^ According to Larson there has been almost no testing of the persistence of any progress after families have left the program. There has been virtually no measurement of the "spilling over" effects of the standard program in influencing neighbors of borrowers, local groups and institutions, or the effects on the general welfare. Little Is known about the comparative results of different amounts and types of super vision.®® There have been many other articles and publications describing the different kinds of rehabilitation experiments, some of which are valuable to this study. Harness reported a study in which it was attempted to find reasons for variations in the success of Farm Security Administration borrowers In five Missouri counties.®® Fifty farms 54 Ibid. Pp. 234-235. - Attempting to determine the situation of the families before acceptance and at the time of the loan was a major problem, in developing the design and conducting the research reported in this dissertation. ®® ko°» oit. 56 G. Harness. Farm Security Administration Hehabilitation Loan Experience in Five Missouri Counties. Missouri Agricultural Experiment bulletin. Vol. 376. June 1942. Pp. 1-34. -29- sole c-fced. by random method from each of four counties and 33 records from a fifth county were used in his study. The findings show that the reason FSA borrowers could not use existing lending agencies was that banks would not loan much over 50 percent of the farmers1 assets, and most farmers* assets averaged $165* Also, the general policy of banks was to make loans at eight percent and for not more than six months# The FSA could make loans to those farmers with greater likelihood of success because of the closer supervision. Harness found a tendency for some farm owners to ask more rent of FSA tenants. However, there was no doubt that FSA influence did help tenants obtain a more reasonable agreement in spite of some inclination on the part of farm owners to inflate rent values• In this Missouri study large credit extensions produced more improvement than smaller credit extensions. Harness says this was probably because these borrowers were the best farmers to begin with. Original net worth, size of farm and age of borrower, did not seem to be associated with financial progress, but the productive capacity of the farm definitely was associated with the improvement in net worth of the borrower. Use made of the loan had a significant influence on progress. The less the loan was used for productive livestock, the less was the progress made by the borrower. Size °£ livestock enterprise also had a direct relationship to rate of improvement of the borrowers’ economic status. The advantage iri having several enterprises was greater than the comparative increase in amount of credit extended. Harness said that the number of years the borrower was in the program seemed to be associated with the rate of progress. The longer -30- he remained in the program the less seemed to be his yearly accumulation, partly because the more capable operators left the program# In this Missouri study It was found that, in the main, borrowers owning small land acreages were generally handicapped rather than helped toward economic independence because the size of their enterprise could not carry the cost of the loan. Thus the loan added to their financial problems. In his article, Solem reported that 35 tenant farmers were provided with greater security after receiving loans from the Farm Security Administration than before the loan.®''7 In their previous situation they were handicapped by farm owners wanting annual return on land instead of long-term, planned production. The borrowers had not been able to buy land before because all their money was needed for operating capital# Under the FSA 20 out of the 35 became full owners; the others became part owners# Henderson and Shaw describes how the FSA loaned $1,226,530 to the Rio Farms Incorporated, formed by seven farmers to help put 20,000 acres of Texas land Into production.^ One hundred-fifty sharecroppers were rehabilitated by this program, which Included opening of cooperatives, experimental research, and the buying up of an abondoned FSA housing project for tenants# The managers were farmers who were elected by the tenants. These authors said that the land was worth $1,000. per acre when they wrote 57 l.I. Solem. "nThe Affect of Land Purchase by Farm Security Administration Borrowers, on Agricultural Production in the Northern Great Plains.11 Journal of Land and Public Utility Economics. Vol. 19# May 1943. Pp. 231-233. 58 H. Henderson and S* Shaw. ’’Million-Dollar Sharecroppers.” Vol. 117. June 12, 1946. Pp. 18-19 f. C oilier 1s . -31- the article in 1946. When this land had been bought, its price was below an amount which would be determined by its actual productive capacity* In their book, published in 1941, Raper and Reid reported that four-fifths of southern rehabilitation families under FSA were repay ing their debts.^9 These authors stated that the FSA was doing much to free the South of its greatest handicap, which they said was lack of faith in its own people. It was doing this by demonstrating that the South's landless people would respond to sympathetic guidance and that the money used in such programs was a public investment rather than a public expenditure. The point was raised that farm-tenant families, termed useless by plantation owners, repaid FSA loans and raised their living standard through the education about f a m i n g methods, better diet, and other factors achieved by the loaning program* Heilman in 1940 said that two years after establishment of the Resettlement Administration, one-half of the loan failures were due to bad health.^ jn 1935 a medical care plan was developed which involved 80,000 families, or 400,000 people in 634 counties of 32 states• The privilege of summoning medical aid, Heilman continued, was not abused which was demonstrated indie fact that there were only two or three isolated instances when the plan had to wind up in an 59 Arthur F. Raper and Ira DeA. Reid* Sharecroppers All* Chapel Hill, North Carolina. University of North Carolina Press* 1941* 60 R. Heilman. "Farmers Try Group Medicine." Harpers. Vol. 182. December, 1940* Pp. 72-80. -32- area and start over because more medical care was required than what had been anticipated* Families often called on the doctor the first week of the plan to see if it worked; then the orientation was over* Some families continued not to see a doctor as often as they should have * In this report the author said doctors were paid the same regard less of the amount of sickness, and for this reason they placed their emphasis on preventive medicine. Prior to the FSA many families could not pay any doctors' fees* At the time of this report, doctors collected 65 percent of their fees. One Kansas doctor, according to Heilman, from January 1936 to January 1939 collected 11 percent of the total money owed him by 42 families. *^hen these same 42 families joined the FSA medical plan, that doctor collected 61 percent* Ninety—five percent of the doctors who had experience with the medical program decided to continue with it. 26,000 FSA families were in the plan. dollars per year. In Georgia 14,000 of the Each family paid about fifteen Doctors and hospitals in the plan received about two- thirds of their bills. The average family visited the doctor four times per year, doctors visited families twice. The hospital case average was 14 per year per 100 families, seven days average per case. There were nine pregnancies per 100 families and most of them were home deliveries* Beecher in an article in New Republic reported that up to 1943, #700,000,000 had been loaned to FSA borrowers.®1 Half of these loans had matured, and more than 90 percent had been paid. He said the FSA converted to war needs before big farm interests did — from cotton to 61 John Beecher. "Save Farm Security*" April 26, 1943* Pp. 561-563* Vol. 108. New Republic* -33- soybeans — from tobacco to milk* Its one-half million borrowers, representing 7*6 percent of all United States farmers, in 1942 pro duced 36 percent of the national increase in milk, 27 percent of the increase in dry beans and 10 percent of the increase in eggs, chickens and peanuts* Harding said with reference to a report of 343,000 families who had ccsnpleted at least one year as standard borrowers by the end of 1942, that these families had increased their net income 80 percent, canned 114 percent more fruits and vegetables, and had large increases in milk and meat products for home use*®^ In 1942 all farmers increased pork production by 13 percent over 1941, while the FSA cooperators in creased pork production by 30 percent* Similarly, the increase in beef was 11 percent for all farmers and 38 percent for borrowers, while the increase in eggs was 15 percent for all and 31 percent for the borrow ers • According to Harding it cost about 75 dollars to entirely reha bilitate an FSA family, while it cost about 375 dollars each year to maintain a family on relief. Wehrwein raised an issue about the program when he said that the orders from Washington called for equipping the houses in a resettle ment area with all modern conveniences, surpassing those to be found in thousands of farmhouses in our better f a m i n g areas, and yet ex pecting the settlers to pay for t h e m . ^ This would be a task, he said, for good farmers on good land, to say nothing of fair farmers on second 62 T. S. Harding. "Our Underemployed Farm Labor." Tomorrow* July, 1943. Pp. 25-28. 63 Gr* S. Wehrwein. "Appraisal of Resettlement." Journal of Farm Economics. Vol. 19. February 1937. Pp. 190-205. -34- grade land as was often characteristic of projects located on the edges of submarginal areas • Kirkpatrick analyzed 11,600 rehabilitation families in Michigan, Wisconsin and Minnesota, of which 3,750 were in Michigan.^ However, since characteristics of these families were not compared to those of non-rehabilitation families in the same areas, what was reported may be due to variables other than the rehabilitation program* Witt studied the economic problems of low-income farmers in Iowa, through interviewing a random sample of 782 farmers from which was selected the lowest third with respect to the distribution of monetary incomes. This presented insight into individual problems of these operators in obtaining and in using human and physical resources efficiently in the production process.®^ The major portion of farmers in this low-income third had capital assets, including land, valued at less than 10,000 dollars. The debt load for those reporting was 25 percent of assets owned, approximately the same proportion as that for higher-income faraers. Low-income farmers were usually willing to borrow to purchase land but not to purchase livestock, machinery, buildings or other farm supplies. According to Witt, the available labor supply was larger relative to other resources on the low-income farms. As a consequence there had been an expansion of some enterprises requiring more labor and 64 E. L. Kirkpatrick. Analysis of 70,000 Rural Rehabilitation Fami lies . Washington, D.C. United States Farm Security Adminis tration. Social Research Report No. IX. August, 1938* 65 Lawrence W* Witt. Economic Problems of Low Income Farmers In Iowa Ames, Iowa. Research Bulletin 307. October, 1942. 66 Ibid. P. 208. ■35- less substitution of machinery and other resources for labor* Low-income farmers were handicapped in various ways. Some were hindered by serious health problems; others by insufficient capital resources; many by lack of knowledge of good practices and insuffic ient imagination to reorganize their enterprise for greater returns; and some by too little land; too low productivity, or both# ^ot all low-income farmers were attempting to maximize financial returns from resources# A smaller proportion of low-income farmers followed approved practices, particularly those relating to feeding practices# Very few organized their productive efforts to take advantage of the early markets • Iowa county agents had virtually no contact with the farmers in this lowest third#®*^ Witt stated that the Farm Security Administration did much toward helping low-income farmers, but it had not concentrated all its efforts on the Iowa low-income group, since about 10 percent of its clients were in the high-income group# It had a larger propor tion of its borrowers in the younger age groups, under 40 years, a larger proportion as tenants, and a much smaller proportion among the single and semi-retired operators than was true of all of the people in the low-income group# He said the FSA supervisors may have found younger operators easier to work with, more amenable to suggest ions, more anxious for help and quicker to search out their offices than were older operators# Older operators were less likely to apply even though in serious need of assistance#®® FSA borrowers operated farms somewhat larger in size than those 67 Ibid# 68 Ibid* P. 209# ifc.209-210 -36- of other low-income operators# Tn all areas the Farm Security bor rowers ran their farms more efficiently than did other low-income operators* Witt said this could have been attributed largely to the work of the county supervisor and that it constituted an accomplish ment of their objectives# This higher efficiency was illustrated by the high volume of business handled by the borrowers, relative to their resources, and by the low level of operation expenses. This author considered the sample of FSA borrowers too small for intensive analysis However, the suggestions and facts reported were valuable in planning the present research. Larzelere’s study, based on data obtained through the FSA Milwaukee Regional Office, analyzed 523 debt-adjustment cases in 65 counties of Michigan handled through FSA county debt adjustment com mittees between 1937 and 1940.^ In this the policies and procedures followed by these county committees in dealing with individual debt adjustment cases brought before them were described. The assembled data with respect to financial condition of applicants, nature and extent of debt scale-downs affected by committee mediation between creditors and debtors, and refinancing arranged through the FSA were analyzed. *^2 Larzelers indicated that need for some farm debt adjustment existed during periods of normal economic conditions as well as dur ing economic depressions# Age and tenure were of importance in 69 Loc* cit. 70 Ibid* P. 253# 71 H# E# Larzelere. Farm Debt Adjustment in Michigan Through Farm Security Administration* Michigan Agricultural Experiment Station Quartsrly Bulletin# Vol. 25# August 1942# Pp. 47-65# 72 Ibid. P. 47# -37- determining sound adjustment procedure, because age served as an indication of the maximum repayment period, and the tenure determined to a considerable extent the basic credit requirements of the farmer* Low farm prices in the period when the repayment difficulties arose seriously affected the debt-paying capacity of many farmers; yet debt adjustment through the FSA committees was found to be necessitated in most cases by either the uneconomic use of loan proceeds or the mak ing of loans with a shorter maturity period than the natural liquid ating period of many of the production uses**^ New loans were used either to refinance all or part of the origi nal indebtedness and/or to provide additional capital for a more efficient farming unit* The extent to which new capital was provided depended largely on the proportionate equity of the borrower* larger the equity, the greater the possibility of additional In another study, Larzelere cooperated with The l o a n s *^4 the Michigan FSA in comparing the organization and operation of approximately 500 FSA borrowers with Michigan farmers included in the farm accounting pro ject of the Agricultural Extension Service of Michigan State College.75 The average labor income of FSA farmers was about 250 dollars, which was approximately 400 dollars less than that of farmers in the Exten sion project* He said this had been due principally to smaller size businesses, and to smaller and less productive crops and livestock programs * 75 Ibid". P. 64. 74 Loc* cit * 75 H* E* Larzelere. A Comparison of Income and Expenses on Farm Security Administration Rehabilitation Farms and Other Farms * Michigan Agricultural Experiment Station (Bien.) Report. 1939— 40. Pp. 26-27. -38- FSA farmers operated smaller farms, 115 acres compared with 179, and 84 tillable acres compared with 128 for those in the farm account ing project# They had less productive animal units, 13 as compared with 27, and they provided a smaller number of productive days work per man, 203 as compared with 253# However, despite the smaller size and less efficient enterprises operated by borrowers, the evidence indicated that on the whole their net worth had been increasing from year to year. There have been many research reports useful in developing the design for this project, such as those by Loomis, on rehabilitation programs concerning agricultural village-living people or resettlement projects where families were moved to new a r e a s T h e lessons learned and generalizations made are also important to rehabilitation work, especially the "whole community approach” through already existing organizations and the consideration of cultural influences# The type of rehabilitation observed in this project differs from those studied by Loomis in that the present research is primarily of rehabilitation through loans to individuals on the farms where they were living rather than to individual families moved to new f a m i n g areas or to groups for purchasing livestock or equipment that was used cooper atively. The sources of references reviewed in preparing this chapter are listed in the Bibliography, Appendix A, as well as both the references 76 Charles P7 Lc^omis • Studies of Rural Social Qrganization in the United States, Latin America and Germany. Bast Lansing, Michigan. State College Book Store. 1945# Charles P. Loomis and Glen Gresham. "The New Mexican Experiment in Village Rehabilitation." Applied Anthropology# Vol. II. No. 3# June 1943. Pp. 13—37* 39- which were abstracted and those which were not deemed appropriate enough to justify abstracting. The objectives of this dissertation as discussed in Chapter I were to determine effects on families of participation in the Farmers Home Administration program and to describe and discuss research procedures which were used in the attempt to quantitatively measure these effects, in order to appreciate the relationship of the prob lems faced in this research in perspective, it was felt necessary to present in this chapter an understanding of the background of the FHA and some knowledge of its mechanics of operation* - 40- Chapter III. The Role of the Ex Post Facto Method in Experimental Research Previous research has indicated that the critical element in investigating the FHA (Farmers Home Administration) is the isolation of its effect on families from other variables operating at the same time* How many of the changes in the families, as reported in pre vious publications, were due to the rising price level or to movement through the family life cycle rather than to the supervised loan pro cedure was not known* Actually, the question of whether or not the family would have made the same achievements without the program was unanswered* The ideal design for overcoming difficulties inherent in the problem would be one in which basic data were recorded on a group of families who then would be divided into two matched groups, one of which would receive the FHA assistance. Later the families in these experimental and control groups would be studied to determine whether or not the program had affected change* design is diagrammed in Figure 1. This type of research This projected type of plan re quires a certain amount of time to achieve, however, Figure 1* Experimental design utilizing groups matched in the pre sent as a base line for changes in the future* Future Present Experimental Families Groups matched s o they are equal\ at this j time i Control Families FHA Program ^[Experimental « ^[Families » » i > / J--- -------- “ * i *Control Families J i _ _ J J Any difference might be due to FHA Supervised Loaning Program 41- since the program has ho be operating long enough for any resulting change ho hake place. Another hype of experiment, hhe one already employed in mosh research on rehabilihahion, describes families ah the time of hhe study and then compares them ho what they were like before they enter ed hhe program as determined by records. This "ex posh facto” hype of research design is illustrated in Figure 2. Figure 2. Experimental design utilizing "ex post facto” technique. Past Present *"1 !Experimental j s iFamilies 1 ^ i 1 ..... Expe riment al Families FHA PROGRAM This plan has two weaknesses# It has to depend on existent re cords, and it does not determine what would have happened without the program since no control families were involved. A somewhat similar plan, shown in Figure 3, is to record informa tion on a group of families before they participate in a program and then compare this to data on them afterward. This allows the collect ion of more adequate base data than was permitted in the "ex post facto” research just indicated. Figure 3. But there is still one serious Experimental design in which families are followed after they receive FHA loans. Present Expe riment al Families Future FHA PROGRAM v. Experimental Families limitation inherent in the method which must be considered. Families involved with specific causal factors, particularly those participating in social action programs exist in a complex of other causal factors 42 - usual ly functionally related to the tendency under consideration. One could ask if the families would have made the change regardless of whether or not the specific causal factor was introduced into the situation* One other research plan which takes into consideration this need for control, although it has not been used extensively, is a ”ex post facto” type utilizing experimental and control groups of families. As shown in Figure 4 it has two possibilities. The first matches the families in the present as in Example 1 , tracing them back before their participation to see how they differed. The second secures information about the past and uses this for matching at that stage and then compares them in the present, as in Example 2 . This pro vision for a control group attempts to overcome the possibility of false relationships between the FHA program and factors used to Figure 4. Ex post facto research utilizing matched groups* Example 1 Present Past J Experimental |>_ 1 Families FHA PROGRAM Exper iment al J Families C ompared for \ differences Matched J Control j Families Control Families —43— Example 2 Present PROGRAM Experimental Families Matched Control Families Control Families Compared for Differences measure change as well as to reduce the probability of unknown causa tive factors from being the determining forces* Since the present work required the development of an experimental design which would produce results within a reasonable length of time, the ideal projected type could not be employed* Thus it became nece ssary to develop a suitable "ex post facto" plan* F* Stuart Chapin and Ernest Greenwood have written several pub lications investigating what has been called "ex post facto" research* Greenwood, referring to research in general, stated that an experiment is the proof of a hypothesis which seeks to place two factors in a causal relationship through the study of contrasting situations con trolled on all factors except the one of interest, the latter being either the hypothetical cause or the hypothetical e f f e c t . T h e approach to the cause-and-effect relationship in ex post facto research begins, if we follow Chapin’s conception, "with a description of the present situation as an effect of some previous acting causal factors and attempts to trace back over an interval of time to some assumed 77 Ernest Greenwood* Experimental Sociology* A Study in Method New York* Rings* Crown Press* 1945* P* 28* -44- causal complex of factors which began operating at an earlier date*"*^® Lazarsfeld has called ex post facto studies "experiments-inreverse”, or ”mental-equivalent-of experiments ."79 in the Foreword to Greenwood*s book, Experimental Sociology, Lazarsfeld said that the meaningful way to phrase the question as to whether or not a corre lation is a causal relationship was: "To what degree is a given corre lation equivalent to a controlled experiment? What Greenwood calls the ex post facto experiment is thus only a special case of the broader problem of distinguishing between causal relationships and other types of associations which may be found in any kind of empirical social research*"80 There are two basic variations in the method of ex post facto research according to C h a p i n * T h e first is a comparison of data on individuals and families secured in the present with data on the same persons obtained for some past date* The second is a comparison of populations in the same areas at the present date and at some past date* He considered the first type to be more accurate because of limitations of the additional variables of migration in the second approach* (In some experiments, of course, there would be no oppor tunity for the population to change.) Most ex post facto studies reported in the literature gave little consideration to migration tendencies of the people observed* When describing Christiansen*s St. Paul study of relationship of education 78 P. Stuart Chapin. Experimental Designs in Sociological Research* New York* Harper & Brothers. 1947* P. 95* 79 Lazarsfeld suggested this to Greenwood when he was writing Experi mental Sociology. 80 Greenwood. Experimental Sociology. Op. cit. P. ix. 81 Chapin. Experimental Designs* Op. cit. P. 95. -45- to occupational success, Chapin did not raise the question as to the possible bias resulting from loss of migrants* Out of 1,130 graduates and 997 drop-outs, 42 had moved out of town and 576 could not be trac ed* It was not stated whether or not the same proportion of drop-outs and graduates were lost because of each reason given*8** Chapin was concerned, however, with the migrant problem in his discussion of the social effects of good housing*8® He reported that the net effect of losses was to increase the homogeneity of the resi dual groups from which results of the experiment were inferred# As a consequence the magnitude of absolute scale differences between ex perimental and control groups upon measures of effect were small, and hence the critical ratios were diminished* In discussing this Green wood said, "Had no losses of families talcen place, the terminal groups would have exhibited a much more significant difference in the endfactor# If this tendency which Chapin has observed is a character istic one, it furnishes us with a clue which can be employed in esti mating what probable influences the lost cases might have had upon the results of ex post facto experiments* 84 Greenwood also referred to migration in the statement, •••"often it may be that the missing persons are not representative of the found group for the very reason that they are missing* We are therefore thrown upon our own resources for making judgments and here again insight into the situation comes into play*" 85 One limitation of ex post facto research, according to Chapin, 82 83 84 85 Ibid* F* 101* Ibid* F. 136. Greenwood* Experimental Sociology* Ibid. P* 140* Qp* cit* P. 142. ■46' is that analysis is restricted in choice of causal factors and in selection of controlling factors to information existing in accessible records of past behavior of subjects studied# ^his is in contrast, he continued, to research measuring projected social action projects, since it can use interview schedules and set up adequate planned re cords #88 Greenwood agreed when he stated that ex post facto experi ments occur after the cause has achieved its effect# must reconstruct his experiment from written records# The scientist 3e is almost totally dependent upon them for data on relevant factors, on the nature of the hypothetical cause, and on the extent of the hypotheti cal effect. "Without such complete records on the salient facts, there can be no ex post facto experiment# Only where adequate records are available, is the ex post facto experiment possible#8*^ Both Chapin and Greenwood eontradict themselves in the same pub lications where they said records were the only source of data for ex post facto research* Greenwood’s contradictions are more obvious# Two pages later he reported that, "Should accurate records be unavail able, it is often, though not always, possible to gather and piece together the information afterwards, but the job is rather expensive in time and effort, and the results might still turn out to be meager# In the course of interviewing Christiansen might very well have been able to inquire after measurable data on those factors which needed control but which were not already part of the existing records#"88 An examination of records on Farmers Home Administration families 86 Chapin# Experimental Designs * Op# cit# P# 96# 87 Greenwood# Experimental Sociology# O p # cit» P# 1S4# 88 Ibid# P. 136# -47- in the county studied in the present research revealed them to be much more complete in terms of having loan information than would be expected, since the length of time the records were kept was 15 years« Adequacy of record-keeping was probably due to the fact that the ad ministrator at the time of the study was the one who started the pro gram in that county* However, most record-keeping is necessarily selective in the type of information registered* The selection usu ally does not fit requirements for a penetrating evaluation of the program’s effects several years later* This was found to be true in the present study* The limitations of the Farmers Home Administration records in Hillsdale County were such that the level of validity of the findings which could be expect ed to be achieved made it inadvisable to depend on them for the entire evaluation* In developing the design for the Farmers Home Administration pro ject, it was hoped that interviews with families in the present would provide valid data both about the present and the ’’before" situation labeled the preloan year* Questions were formulated and pre-tests conducted on families about whom there was a considerable amount of available historical information* After many of these formulations and pre-tests, a satisfactory method was worked out and the schedule was completed. This schedule dealt primarily with areas and types of questions where problems of memory could be minimized to make possible a high degree of confidence in the data. The schedule was also con structed to lead the interviewee into a mental set favorable to good memory in the year designated as the preloan year. This procedure as well as the whole interview was conducted without the respondent realizing the fundamental objective in collecting the data* The study was supposed to be interested just in community levels of living rath er than in results of the FHA program# It was pointed out above that reliable information about the past was to be obtained by interviewing the families. It was similarly hoped that data remembered by the FHA families could be used for matching them with non-FHA families. ESA, families would also be obtained from their memories. Data for nonComparisons between the two groups of families, as shown in Figure 5, could then be made with data secured at the time of interviewing, since the time required for change to take place would already have passed# Figure 5 . Experimental design for the present study of ex post facto combined with matching techniques# Present Fast 1 {Experimental ! j FHA {Families ! ^PROGRAM Two groups matched in the prepro gram sit !Control uation Families Expe riment a 1 Families wJ 1 \ X -----t The question of how to obtain a Control Families Comparison made after expe rimental families participated in program control group of families with similar characteristics as the experimental in the preloan year still needed to be answered. One solution suggested was to take a cross- section of the farm families of a county, eliminating those not match ing the FHA group# This was not considered feasible since it was known that FHA families were economically lower than many of the other families. It would have been too expensive to interview enough people in a cross-section to give the desired number for matching. This was 49— especially true because FHA loans were made to some farmers every year beginning in 1936* For this reason, each family in a cross-section survey would have had to furnish data about each year of its existence from 1936 to the time of the study in order to determine which year it would best match one of the FHA families in the experimental group# It was also suggested to compile a list from families furnished by county agricultural agents, soil conservation agents, and other county professionals who would know lower income families. This was not used, because of the bias which might have resulted from the manner of re— ce iving name s . A third suggestion was to have FHA families name other families similar to them during the year before they received a loan. Results of pretesting seemed to indicate that this method would secure a group of families in about the same economic condition as the FHA sample# Families named by FHA respondents were found to be similar to the FHA families in the past, although it was recognized that there may have been tendencies for some FHA families to name only those higher in the social scale than themselves and to designate only families still known to them. It was felt that matching would reduce these variables, however, and this method was used. The methodology of this research project then utilized the ex post facto experiment with experimental and control groups# -50- Chapter IV Data for this Study The geographic region in which data were collected is Hillsdale County, located near the center of the southernmost tier of counties in Michigan* Like parts of Ohio and Indiana, which form its southern boundary, this county has the c o m and livestock production character istic of the north central portion ofthe cornbelt* However, dairying is replacing some of the other types of agriculture The county seat is the city of Hillsdale, located at the approxi mate center of the county. Its population of 7,297 makes it the largest and for many the most important trading area in this county of 31,916 persons*®0 Seven other communities, ranging in population from 380 to 1,594, provide many of the services required by farm peo ple *91 ^ few facilities such as grocery stores, gas stations, and post offices are present in several smaller centers. Some of the population living near county borders trade in villages just outside of the boundaries* The 1945 Census of Agriculture lists 3,329 farms in the county, a decrease of 272 from the 1940 r e p o r t i n g , T a b l e 1 gives the 1945 distribution of the farms by size* 89 E. B, Hill* Types of F a m i n g in Michigan* East Lansing, Michigan* Michigan Agricultural Experiment Station Special Bulletin 206* (Revised June 1939). * 90 1950 United States Census of Population. Michigan. Washington, D.C., United States Department of Commerce. Bureau of Census* Pp. 16-22. 91 Loc* cit* 92 United States Census of Agriculture^ 1945* Vol. 1 . Part 6 * ^Michigan Statistics for Counties1.” Washington, D.C. United States Department of Commerce. Bureau of the Census. 1946. P. 119* -51- Table 1 * Farms in Hillsdale County, Michigan, in 1945 distributed by size** Size in Acres Under 3 3-49 50 - 69 70 - 99 100 - 139 140 - 179 180 - 219 220 - 499 500 and over Total Farms Number Percent 7 757 305 735 612 399 216 285 13 0 23 9 22 18 12 7 9 0 3,329 100 ♦United States Census of Agriculture: 1945* Vol. 1, Part 6 * "Michigan Statistics for Counties*** Washington, D*C*, United States Department of Commerce. Bureau of Census# 1946. P. 126* Table 2 shows that the proportion of tenancy was 21 percent for the farms in 1945# Table 2. Tenure of farms in Hillsdale County, Michigan, in 1945.* Type of Tenure Farms Number Percent Full owners Part owners Managers All tenants 1,983 644 16 686 59.6 19.3 0.5 20.6 Total 3,329 100.0 ♦United States Census of Agriculture% 1945. Vol. 1, Part 6 . "Michigan Statistics for Counties .11 Washington, D.C., United States Department of Commerce, Bureau of Census. 1946. P. 119. -52- Hillsdale County was selected because a large number of loans had been made to farmers by the Farmers Home Administration and its prede cessors and because the administrator who started the program was still conducting it* The large number of loans was necessary to provide an ample base for sampling and the fact that a single administrator had charge of the county’s program from the beginning would reduce varia tions due to differences in skill and personality on the local level. The presence of only one supervisor from the beginning of the program also increased the value of the files since he could remember details of the records as well as the history of the families before, during and after they had participated in the program. One of the first steps in this research was to list all families who had obtained loans in Hillsdale County from the FHA or who had received them in other areas and moved into this county. Six hundred and thirty—six families were found to have had such financial and supervisory assistance. These families were alphabetized and numbered so that a sample could be selected for the research since all of them could not be interviewed. Figure 6 shows the flow chart for the experimental sample of FHA families. In order to be sure that the final sample for interview would be adequate, a random selection of 410 families was made from the universe of 636 families. It was necessary to eliminate 304 of these because they did not meet criteria of the sample. Eighty-eight of the 410 had received only a grant from the FHA and not a supervised loan. The 44 families who received their first loan after 1946 were not included because it was felt that several years were needed for measurable change to take place as a result of the loan. —53** Figure 6 . Sample flow chart for FHA.# Universe of families associated with Farmers Home Administration in Hillsdale County, Michigan 636 Non-supervised loan 88 Left area 30 87 Randomly selected 44 Not farming Loan 1946 or later from Deceased Loan made to bachelor 25 24 Location Unknown Universe 2 ^ 4 Broken family 410 &L Refused Those meeting criteria for being interviewed 2 Illness prevented interview 106 Broken Ifamily Number of Interv iew Records Taken 10 0 *f Successfully interviewed 97 1 Knew reason for interview ------------- — s. 2 Incomplete interview ■54' An attempt was made to trace the remaining 258 families to their present addresses, largely with the help of the county supervisor who was able to remember where most of them were now living# Of these, 87 were not farming, 30 had left the area, 25 had lost either the husband or wife through death, 24 could not be located, four had broken families, and in two instances the loans had been made to bachelors* The families not farming at the time of the study and at the time of receiving their FHA loan were not included in the sample because the instruments for measuring changes, such as level of living, were designed for farm families* Families broken through death, separation, or divorce were excluded to reduce problems of matching the sample with a control group• Figure 6 further indicates that of the 106 families who qualified to be interviewed, three refused, two had illness preventing secural of data, and one knew that the interview concerned the FHA and thus he might have been biased in his responses* Two interview records from the 100 families visited were found to lack essential information, and one was found to be a broken family* Therefore, 97 were successfully interviewed and they constitute the FHA sample. The list from which the control group of non-FHA families was selected was compiled by asking FHA families to give names of two or three families similar to them during the preloan year in sociological and economic characteristics using this questions "Would you mind giv ing me the names of two or three families who were in about the same condition as you were in ? (The preloan year), ‘ ^hat is, who were about the same age, had about the same number of children, and who -55- had about the same size farm as you *”93 A total of 201 families were named by the 97 FHA families in this manner* The 201 families were then investigated to determine the possi bility of using them in the study. This process is illustrated in the sample flow chart in Figure 7* Forty-five could not be included for reasons similar to those eliminating FHA families; they had moved away from the county, either the husband or wife were deceased, they were no longer farming, or they could not be located* Sixty-seven were eliminated because they had FHA loans, which was one indication that experimental families had suggested people on the same social and economic level as themselves. After those not satisfying the criteria were eliminated, 89 were left to be interviewed* ^f these, two refused and three were either not farming or could not be included for other reasons. Interviewing was accomplished between June 1 and September 15, 1950* Completion of the long interview sometimes required two visits* When a family was not at home or when the male head was absent, additional trips were made to obtain all the information* The same questions were used for this group as for the FHA fami lies except for certain questions applying specifically either to the FHA experimental sample or to the control families. Major emphasis in the interviewing was put upon the collection of evidence that could be employed in determining whether or not level of living and social par ticipation of families receiving supervised FHA loans had increased as S3 The fact that the year forwhich information was desired was the year before they received their loan was not mentioned in the interview* -56- a result of the program. The questions asked to provide this infor mation had to meet several criteria. First, they had to be of a type that could be remembered by the families at some time in the past. Second, they were required to be similar for the past and the present in order to reveal any indication of change occurring through time. Finally, they needed to be the same for both FHA, and non-FHA families in order to determine differential changes between the two groups. Figure 7. Non-FHA sample flow chart* All families who might have been named as similar by the FHA informants Named as Had FHA loans 67 by FHA families Refused 2 Eligible for interview 89 Successfully interviewed 84 Left area Deceased 45 Not farming Broken family Couldn’t locate Not farming or other reasons -57- To meet these requirements, scales developed in studies for that purpose were selected for recording level of living and community organization participation*®^ Other questions were asked concerning informal social participation and income* Numerical values were designated for different types of informal activities* No extensive testing of the validity and reliability of these scales was carried out for two reasons: (l) interest was primarily in broad group com parisons and not in individual predictions and ( 2 ) in the case of two of the scales the items and scoring followed very closely the items and weighting of standardized scales — the Sewell scale in the case of level of living and the Chapin scale in the case of form al social participation* In addition to the statistical evaluation of the scales done in Ohio and Oklahoma, there is a precedent for their use in Michigan* Gibson used these scales of socio-economic 94 William H. Sewell * The Construction and Standardization of & Scale for the Measurement of the 3 ocio-Bc onomic Status of Oklahoma Farm Families * Stillwater, Oklahoma. Oklahoma Agricultural and Mech anical College* Agricultural Experiment Station* Technical Bull etin No. 9. April, 1940* A. R. Mangus and Howard R. Cottam. Level of Living, Social Part icipation, and Adjustment of Ohio Farm People. Wooster, Ohio. Ohio Agricultural Experiment Station. Bulletin 624. September 1941. Howard R. Cottam. Methods of Measuring Level of Living, Social Participation and Adjustment of Ohio Farm People. Methodological Supplement to Ohio Agricultural Experiment Station. Bulletin 624. Columbus, Ohio. Ohio State University and Ohio Agricultural Ex periment Station. Department of Rural Economics and Rural Soci ology. Mimeographed Bulletin 139. July 1941. Sewell, Mangus and Cottam utilized Chapin’s contributions on social participation as a background for their work. See Chapin, F. Stuart. "Social Participation and Social Intelligence." American Sociological Review. Vol. 4f No. 2. April 1939. Pp. 157-16C. John C. Belcher. "Evaluation and Restandardization of Sewell’s Socio-economic Scale.” Rural Sociology. Vol. 16, No. 3. September 1951. Pp. 246-2BF7 -58- status in studying the clientele of the Agricultural Extension Service which included most of the same level of living and organizational participation items utilized here.95 The items for both the preloan and present level of living scores were the same* They are presented with their respective scoring values in Table 3. Table 3* Items included in the determination of living, with values• Item given in response to questions Do you have «•••? Response to question Value of each response Telephone Yes No 2 0 Re f ri ge rat ion Deep freeze Refrigerator Ice 4 3 Electric Gasoline Hand 3 2 Mangle Electric Stove 4 2 Electric Carpet-hand 3 Piped hot Piped cold Hand pump inside 3 2 Yes No 1 Indoor Outdoor 4 0 Washing machine Iron Sweeper Water Sink Toilet 1 1 1 1 1 0 95 Duane L. Gibson* The Clientele of the Agricultural Extension Service. East Lansing, Michigan. Michigan Agricultural Experiment Station* Quarterly Bulletin 26, No* 4* May 1944* -59- Table 3. (Cont'd) Item given in response to questions Do you have....? Response to question Value of each response Bathroom Tub in room Tub not in room 4 2 Heating Furnace Heaters 4 1 Living room Painted floors Rugs in good condition Linoleum or worn, rugs 1 3 2 Magazines 12 or more 7 through 11 4 through 6 2 through 3 4 3 2 1 Books 100 or more 50 through 99 10 through 49 3 2 1 Bookcase Yes Ho 1 0 Sewing machine Electric Foot 4 1 Cooking stove Gas Electric Oil 2 3 1 Organizational participation was secured by asking whether or n ot the husband or wife or both were members, attended, served on committees, or were officers of organizations indicated in Table 4, They were also asked what public offices were held such as school board member, village officer, or political party worker. Table 4. Organizations for which participation status was obtained. Organization Farm Bureau Grange Organization Fraternal lodge: Oddfellows, Masons, Rebeccas, or others. 60- Table 4* (Cont'd) Organization Organization Gleaners Farmers Union Farmers Cooperatives American Legion, Veterans of Foreign Wars, and auxiliaries Parent Teachers Association Agricultural Adjustment Administra. Hillsdale Soil Conservation District Dairy Herd Improvement Assn., Fruit Assn., and other similar groups Home Extension Club Civil Club: Lions, Kiwanis, Rotary, Chamber of Commerce Church Sunday School, or other relig ious organizations; Ladies Aid, Missionary Community Club or other local formal organizations Other The complete organizational participation score for husband and wife was determined in this manners membership of either spouse in any of the organizations listed in Table 4 received one point, attendance two points, committee memberships three points, and officerships three points. Each public office or political party work was given four points • Informal participation was obtained by asking nine questions The questions and the values for different responses are in Table 5. In the first eight of these questions both husband and wife were granted the values while in the last one only one score was given. The family score was secured by adding both spouses' scores to the value given in the last question* 96 See Uangus and Cottam, o£. cit., P. 37, for * discussion of infor mal participation. Also Cottam, ££. cit., P. Z7. Table 5* Questions, responses and values used in determining infor mal participation status. Question asked Classification of replies Value given to each reply 1. How often do you or your wife go out to dinner at other peo ples homes or have other people over for dinner? Regularly, once a week either way occasionally, once a month either way seldom, once a year 10 3 1 2# How often do either of you visit away from home? Regularly, twice a week occasionally, once a month seldom, twice a year 10 4 1 3. How often do either you or your wife go to a restaurant or bar to visit with other people? Regularly, twice a week occasionally, once a month seldom, twice a year 10 3 1 4. Most of them Do either you or once in a while your wife attend parties, potluck suppers, picnics, meetings or gatherings put on by organizations or clubs such as the school, 4-H Club or neighborhood groups? 4 2 5* Do either of you go to auctions? Most of them some of them once in a while 4 3 2 6. Do you or your wife hunt or fish with other people? Once a week, in season once a month, in season once a year 4 2 1 7* Do you or your wife participate in any sports or athletic programs such as bowling, baseball, or skating? Once a week once a month once a year 5 2 1 8. How often do you or your wife play cards or other games with other people? Regularly, once a week occasionally, once a month seldom, once a year -62- Table 5, (Cont*d) 9* Question asked Classification of replies How often do people visit you? Regularly, twice a week occasionally, once a month seldom, once a year Value given to each reply 10 3 1 Total family income was secured in the interview and will be used in this analysis* Income and informal participation were recorded only for the year preceding the interview* Level of living and organ izational participation scores were obtained for the year preceding the interview and for the year preceding the time of the loan* For the control group, the ”preloan year equivalent” was the preloan year of the experimental families who had named them as their equals • In addition to the information about level of living, organiza tional participation, income, and informal social participation record ed for both the FHA participants and the non-FHA group, their attitudes toward the FHA program, mobility, medical symptoms, diet and other characteristics were obtained. These data were analyzed in other re ports apart from this dissertation*97 A discussion of how community leaders rated FHA and non-FHA families on social status and prestige items and an analysis of FHA families not now farming was also included in reports prepared by the Social Research Service of Michigan State College and will not be part of this dissertation* The method of analysis used in this dissertation involved compar ing experimental with control families to ascertain differences that 97 Reference to" them may be obtained from the Director of the Social Research Service, Michigan State College, East Lansing, Michigan. 65- might be due to the known factor, the FHA program, on which they differ ed* This type of analysis required that control families be similar to the experimental before the FHA loans were granted. Thus, the prob lem of controlling influences on change in these families was consid ered. Chapter V discusses extensively the factors controlled, and the remainder of the dissertation presents the results; that is, an evaluation of the Farmers Home Administration as an influence in change under varied precision of control. - 64- Chapter V Selection of Variables to be Controlled In Chapter III it was noted that the methodology of this research utilized a combination of ex post facto techniques with the matching of experimental with a control group. data available for analysis. Chapter IV described the It pointed out that the experimental sample was comprised of 97 families who participated in the Farmers Home Administration program while the control sample included 84 fami lies who did not participate. The aim of this analysis is to determine changes that have occur red as a result of the FHA program among those families receiving the supervised loaning service. Changes studied involved scores on level of living and organizational participation for which similar data were available both in the preprogram situation and at the time of the interview and income and informal participation for which data existed only at the time of the interview. It has been shown in Chapter III that in order to realize the objective of deducing a real causal re lationship, it is necessary to eliminate the effect of variables whose influence we do not wish to study. One effective way of doing this is to match the experimental group with a control before comparisons are made • Critical to the process is the selection of variables to be con trolled in the matching process. Greenwood says that "the first step in experimental control is to identify those factors which are known definitely to be relevant to the specific phenomenon being observed* ...control in an experiment need not be absolute but only selective. - 65- For the validity of a scientific result very careful control is nece ssary with respect to variables which might affect the result, while very little control is necessary with respect to those variables that would not affect the result* Selective control is specifically di rected toward the objectives in view.”9® The problem of importance is how the relevant factors can be determined, since a factor is relevant if it contributes to the effect being studied, according to this same author." the social Angell stated that in field, scientists can scarcely hope to identify all sig nificant variables, because there are so many to deal w i t h . ^ ^ The Yale Institute observational studies of industrial employees bears this out. Loomis, in describing the Yale research, stated that the biggest trouble was to discover the significant situational variables. As the work progressed, the number of such variables grew. Variables were often identified, but their presumable effect overlooked, only to have it appear weeks later that an unsuspected variable was exert- ing a strong influence. 101 Max Weber felt that an investigator familiar with his data could easily spot the relevant situational factors on the basis of his ex perience, with the result that many factors could be shown to be causally irrelevant on the basis of factual knowledge. If one knew the usual function of factors, he might dispose of them mentally in 98 Greenwood, Experimental Sociology. Op. cit♦ P. 73. 99 Loc. cit. 100 Robert C. Angell. "Difficulties of Experimental Sociology." Social Forces. Vol. XI. December 1932. Pp. 207—210. 101 Alice Loomis. "Observation of Social Behavior in Industrial Work.” Social Forces. Vol. XI. December 1932. Pp. 211—213. 66- order "to ascertain whether or not their absence could have an effect on the actual course of events. The factors which could be thought away in this manner were causally irr©levant• There seem to be three general ways in which variables can be chosen, One is selection by intuition, that is, picking those that "everyone knows" to have a causal relationship. those that others have found relevant. A second is to use The third is to find and use those variables having effects in this problem* Very little treatment is given as to how variables were selected in most research reports where matching was utilized. to consider the intuitive method to be satisfactory. Chapin seemed In his discuss ion of Christiansen*s education study in St. Paul, Minnesota, he stated that the two variables to be observed in the experiment were school progress and economic adjustment. For him, it was perfectly obvious that he was using rather crude measures of these variables and that factors of age difference, sex, home status, nationality, neighborhood, and mental ability as well as the interrelationship of the network of factors might affect the economic adjustment variable. He maintained that "since every one of these variable factors is re cognized by sensible p e o p l e as influencing the course of indi vidual economic progress, the way to obviate their disturbing influ ence is to control them"*-1*0^ St. Paul study. All six factors were controlled in the One wonders whether or not the six factors were arrived at by lining up a group of "sensible people" and asking them to list them. 102 Greenwood^ Experimental Sociology. Op. oit♦ P. 73. 103 Italics inserted by the author of this dissertation* 104 Chapin. Experimental Designs. P. 100. 67- - The second method of selecting variables for matching, that of using what others have found relevant, requires a search for applica ble studies in literature* This was indicated, for example, in rChapinfs statement that in the Hew York City T* B* and housing study, some factors for matching were selected because other studies had shown positive relationships between those factors and T. B* rates Continuity in research would contribute to this technique since the researcher could utilize his own past findings as well as results of others about causal relationships for selecting variables* This bas ing of relevant variable selection on other findings is probably the most valid way for selecting such variables, but quite often other related research does not exist or Is not sufficiently comparable* The third method, that of discovering variables having effects in the problem under study, is accomplished by analyzing all varia bles for which there are available data to determine their associa tion with the factor being investigated. The total number of cases in both the experimental and control groups are utilized in this comparison. Those having a high association with the dependent vari ables (in this case; level of living, organizational participation and informal participation, etc.) are then used in achieving the matching. This method is actually validating data with itself, but it is the most feasible course when the second method is not possible. It is recognized that some variables even though important in the experiment are difficult to control, posing a problem of select ing those that can be controlled and on which, measurable data can be 105 Ibid. P. 129. 68- obtained as well as selecting those that are relevant• For example, in his review of the Christiansen experiment, which utilized the in tuitive method, Chapin noted that persistence, physical health, num ber of broken homes, and exact money income should have been controll ed in addition to high school education#106 In the present work, a combination of the three methods for selecting relevant variables was used with the main emphasis on the last procedure# Intuitive feelings and knowledge along with results of other research were utilized in developing the schedule of questions used in securing the original data# This placed limitations on the variables that could eventually be used in matching, since only data collected in the interviews would be available# The knowledge and memory of the experimental group was relied on to some extent when they were asked to name other families which were to serve as the basis for the control group* A type of control was exerted when cer tain characteristics were accepted as a significant reason for ex cluding families from both the experimental and control groups. These characteristics, broken homes through separation or death, severe fires, and other catastrophic incidents, were selected on the basis of knowledge of their relation to the variable under consideration and on intuitive feelings# Once the data had been collected, the third method, that of test ing variables to determine whether or not they are significantly re lated to the dependent variables under consideration, was vised* three primary measures of change which were of interest here were 106 IbidU P. 112# The -69 level of living, organizational participation, and informal partici pation# It should be reiterated that the major hypothesis under con sideration here is: Has the FHA program influenced those three fac tors in the families who have participated in the program? Since there are a number of variables that may influence these factors of level of living, organizational participation, and informal partici pation independent of the effect of the FHA, they must be controlled# Seven such factors having a possible effect on those three character istics were delineated. All seven were available from the preloan year or preloan year equivalent which was the year before the family became involved in the F H A # ^ ^ Preloan level of living, preloan organizational participation, husbands* schooling, year the family first went into farming, year they received the loan, whether or not husbands had lived in a city, and the age of husbands were analyzed to determine the degree of association with the three primary measures of change. Coefficient of contingency was used to determine the amount and kind of association. in Table 1# The values of C for these tests are presented The higher the value of C the more effect that factor would have on the results and, therefore, the more important it was to match the experimental with the control group on it, according to the process of variable selection used here# If the groups were not matched on a factor having a high C value, associated with level of 107 For non-FHA families a preloan year equivalent was assumed which was the same as that of the FHA families who had named them as being equal to them as was described in Chapter IV. - 70- living, for example, differences in level of living after the fami lies had received a supervised loan might be due to that factor rather than to the loaning program. Table 6 shows that the C values were relatively low, although all of them were positive as determined by inspection. The average values and the range of C values were utilized in rating the importance of factors. On the basis of these findings, the following variables were selected to be used in matching the experimental group and the control families: Preloan organizational participation, husbands* schooling, year first went into farming, preloan level of living, and preloan year. The remaining two, whether or not husbands had lived in a city and age of husband were dropped because they demonstrated the lowest association with the present factors. It would have been extremely difficult to utilize all seven in the matching and if carr ied out the number of control and experimental families left after matching would have been extremely small. The nature of several experimental situations in which these variables are used in matching will be set forth in detail in the next chapter. - Table 6# 71- Coefficient of contingency values for the association of factors in the past with factors in the present* C values for factors in the present^ Factors in the past Preloan organi zational participation Level Organi of zational living partici pation Informal partici pation Summary Average Range of C C value values .47 •6 1 ^ .22 .43 .22 - .61 .24 .44 .40 .36 .24 - .44 Year first went into farming .36 .24 .29 .30 .24 - .36 Preloan level of living .43 .40 .19 .34 .19 - •43 .30 *24 .21 •25 .21 - .30 .11 .23 .34 .23 .11 - •34 .14 .18 .25 .19 .14 - .25 Husbands * schooling Preloan year If husbands lived in city Age of husbands * The C values have been corrected for the understatement of corre lation due to computing the coefficient of contingency. For a dis cussion sees McCormick, Thomas C* E 1ement ary Social Statistics * New York* McGraw-Hill Book Co*, Inc* 1941* Pp« 207—208* ♦♦Some small cells were present in this table* - 72- Chapter VI Findings Part A* Introduction This chapter is a discussion of six attempts to determine the influence of participation in the FHA supervised loan program on selected characteristics of families. The characteristics employed as indices for studying possible change occurring during the partici pation were level of living, organizational participation, informal participation, and income. the purposes of the FHA. Advancement in these seemed to epitomize The working hypothesis was that FHA influ ence would produce for participating families an increase over their former status in each of these indices. Measurement of such change was attempted through use of ex post facto methods combined with matching experimental FHA families with control families before they had received loans on the five variables indicated in the last chapter. The nature of various experimental situations will be set forth in detail in each of the next six parts of this chapter. Part B. Comparison of Total Groups To begin with, the method of securing the 84 non-FKA families achieved a degree of matching in two ways • First, participating FHA families were asked to name other families similar to them in the preloan year with respect to age of parents, size of farm, number of children, and tenure. Second, families having unusual characteristics such as broken homes, deaths, and severe fires were eliminated from both groups before the interview was completed. Thus the 97 FHA and - 73- the 84 non-FHA comprising the control group were matched in a rough manner. In order to show whether the two groups thus roughly matched were similar with respect to the five variables which might need to be controlled, (See Chapter V), data regarding them are presented and analyzed in the pages that follow* Preloan Organi zational Participation* Preloan organizational participation, the variable showing the highest positive correlation with the indices in the present period, is presented in Table 7, A study of the observed numbers of FHA and non—FHA families, distributed by scores received on an organizational participation scale in the preloan year, indicates a tendency for the FHA to be lower with 23 families scoring zero compared to only nine of the non-FHA who did* Table 7* Total groups of 97 FHA and 84 non-FHA families distributed b y preloan organ izational participation scores. Score 0 - FHA Uon-FHA 23 17 21 9 7 5 3 3 5 1 1 1 1 0 9 12 13 9 8 10 7 5 5 2 2 1 0 1 Total 97 84 Mean Probability 13.6 1 5 10 15 20 25 30 35 40 45 50 55 60 4 9 14 19 24 29 34 39 44 49 54 59 64 19*6 F < .01 74- ■ Similarly, 61 FHA families scored nine or lower while only 34 of the other group*s scores were as low* The mean preloan organizational participation score of the ex perimental group (FHA) was 13.6 and for the control, 19.6. The t-test of significance showed a probability of less than once out of 100 for the difference to occur by chance if successive random samples were drawn from the same u n i v e r s e . I t could be concluded, therefore, that the two groups of families, when compared by their scores on preloan organizational participation, were not very similar since the FHA group was lower than the other despite the rough matching which took place at the time of interview. Schooling of Husband. Husband*s schooling, the second most important variable having an influence on the present status of families, showed that the mean years of schooling for the FHA was 9.01 and for the non-FHA 9.67, as presented in Table 8. gave a probability slightly above the five percent level. The t-test If one adheres to the five percent level as the "critical*' level of signifi cance, the conclusion would be that schooling of FHA and non-FHA husbands is not significantly different. 108 For a discussion of this statistical test see Margaret J. Hagood. Statistics for Sociologists. Hew York. Reynal and Hitchcock, Inc. 1941. Pp. 461-465. Alsq see E. F. Lindquist. Statistical Analysis in Educational Research. New York. Houghton Mifflin Co. 1940. Pp. 51-54 75- Table 8. Total groups of 97 FHA and 84 non-FHA families distributed by schooling of husbands# Schooling Non-FHA FHA One through four years grammar Five through seven years grammar Completed eight grades Hine through 11 grades Completed 12 grades One through three years college Completed college Graduate work 3 7 32 23 21 6 2 0 0 4 23 24 26 5 0 1 Total 94* 83** Mean Probability 9.67 9.01 ,05
.95 112' TKe seTec^Ton of these variables was discussed in Chapter V. especially Table 6. See 87- "the non-FHA has none in 55 - 59 and one in 60 — 64. This means that they are nearly identical on this variable, an improvement in homo geneity from the comparison before matching as given in Table 7, page ^ this chapter, which showed the PHA to be statistically sig nificantly lower than the others. Schooling of Husbands. Comparison of the two groups in Table 19 by schooling of husbands shows very little difference between them. Six FHA as compared to four non-FHA had not completed the eighth Table 19* Total groups of FHA and non-FHA families matched by distributions on schooling of husbands• Schooling Non-FHA FHA 1 through 4 years grammar 5 through 7 years Completed 8 grades 9 through 11 grades Completed 12 grades 1 through 5 years college Completed college Graduate work No answer 1 5 20 19 20 4 0 0 1 0 4 20 19 20 5 0 1 1 Total 70* 70 Mean Probability 9.83 10.16 P > .90 ♦Schooling of one person was not known. grade* Two of the FHA had completed college compared to one of the others, but the non-FHA husband had gone beyond college into some graduate work. Observation of Table 8 indicates that the distribu tions here are more homogenous after matching even though differences were not quite statistically significant before. Year First Started Farming* The matched groups are compared by -88~ the year the families first went into farming on their own in Table 20. Differences between the two distributions prior to matching, as pre sented in Table 9, page 4, were not statistically significant* be seen by comparing tables 9 and It can that had the variance been sig nificant, it would have been greatly reduced through the matching process, even though there are now more non-FHA than FHA in classes Table 20. Total groups of FHA and non-FHA families matched by distributions on the year they first started farming# Year FHA Non-FHA 1905 - 1908 1909 - 1912 1913 - 1916 1917 - 1920 1921 - 1924 1925 - 1928 1929 - 1932 1933 - 1936 1937 - 1940 1941 - 1944 1945 - 1948 No answer 0 4 2 4 6 3 6 14 15 9 6 2 1 2 4 2 4 7 10 12 17 6 4 1 Total 71 70 Mean Probability Jan. 1932 Oct. 1932 >.90 1925 - 1928 and 1929 - 1932, 17 compared to nine, and a few more FHA than non-FHA in 1941 - 1944 and 1945 - 1948, 15 compared to 10. In general, such differences are reduced when a class is combined with the one above or below it • Preloan Level of Living. Table 21 presents a comparison of preloan level of living scores for the two family groups. Again variance between distributions is slight. The non-FHA have one more than the others in the lowest range. The non-FHA have one more - 89« Table 21 • Total groups of FHA and non-FHA families matched by distributions on preloan level of living. Score Non-FHA FHA 5 - 9 10 - 14 15 - 19 20 - 24 25 - 29 30 - 34 35 - 39 40-44 6 20 13 15 8 4 4 1 7 20 14 14 9 3 3 0 Total 71 70 18.35 19.25 Mean Probability P>.90 than the others in the lowest range. * The non-FHA have one more in the class of 15 - 19, but this is compensated for by their having one less in the class above. Similarly, the FHA have a deficit of one in the class of 25 - 29 and an excess of one in the next class of 30 - 34. There is one more FHA family scoring 35 - 39 and one more scoring 40 — 44. It can be recalled that preloan level of living comparison in Table 10 also showed no statistical difference even before this matching was done. Preloan Year. The distribution of both groups by preloan year, shown in Table 22, indicates few differences between them.113 It should be remembered that preloan year had the lowest association of all five variables with the present measures of family status. TlFT~pr'eloan-year for the FHA is one year before they received a loan. For the n o n - F H A the preloan year equivalent is the same as the preloan year of the FHA family naming them. 90- ■ Table 22 0 Total groups of FHA and non-FHA families matched by distributions and compared b y preloan year. Year Non-FHA FHA 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 4 7 3 4 6 13 9 7 4 6 8 2 8 4 5 11 12 10 3 8 4 3 Total 71 70 Mean Probability Dec. 1938 May 1939 P>*90 This means that least emphasis was placed on homogeneity of dis tributions in preloan year, and for this reason greater variance be tween them would be expected* The table indicates that eight FHA compared to three non-FHA were in the 1944 class. This difference was reduced when three classes are combined since 18 FHA were in the range of 1942 through 1944 compared to 15 of the others. Five more of the non-FHA, 11 compared to six, were in the 1938 class. differences are slight* Other Before matching, as can be seen in Table 11, the loan year of the FHA was later than the non-FHA. Matching has reduced these differences so that they could now be due to chance. The five comparisons just presented of the groups after matching show very few differences between FHA and non-FHA matched groups in the pre-program situation. For this reason, changes between then and the time they were interviewed, the postprogram situation, might be due to a variable such as the FHA supervised loan, since it is known ~ that they differ on this* 91» The comparison of the experimental and control groups as they emerged after this matching by the four meas ures of success will now be presented. Present 0rgani 2ational Participation* Scores on this variable for FHA and non-FHA after they were matched by distributions, are shown in Table 23* The appearance of similarity is borne out in statistical comparisons* The FHA mean is 19*4 and for the control group 18*0 or a difference of only 1.4 in favor of the hypothesis that the FHA program increased participants1 scores* A variance as wide as this could happen by chance between 40 and 60 times out of 100 if successive random samples were drawn from the same universe* Thus both groups increased their scores over the preloan year about the same amount* Table 23. Total groups of FHA and non-FHA families matched by distribution and compared by present organizational participation scores* Score Non-FHA FHA 0 - 4 5 - 9 10 - 14 15 - 19 20 - 24 25 - 29 30 - 34 35 - 39 40-44 45 - 49 50 - 54 55 - 59 60 - 64 65 - 69 70 - 74 13 8 9 12 11 4 4 5 0 2 2 0 0 0 1 11 10 11 16 7 3 4 3 2 1 1 1 0 1 0 Total 71 70 Mean Probability Relation to hypothesis 18,0 19,4 ,40
*■*
17.2
14.5
Mean
Probability___________________ .30 .90
-108in Table 39-
The mean for the experimental families is 9.6 and 10.0 for
■feeothers indicating that the two groups are quite similar.
^ear ffirst Started Farming#
The third control variable, the year
the families started fanning on their own, is in Table 40.
The means
are equal with both being April, 1932.
Table 40.
FHA families matched with non-FHA families they
named compared by year they first started farm
ing.
Year
1909
1920
1930
1934
1936
1938
1940
1942
1944
1946
-
FHA
1919
1929
1933
1935
1937
1939
1941
1943
1945
1948
Total
6
4
7
4
5
2
5
4
2
0
7
6
7
4
4
5
4
3
3
2
39*
45*
Mean
April 1932
Probability
♦One response was not secured.
Preloan Level of Living.
Non-FHA
April 1932
P>.95
The 46 non-FHA families and the 40 FHA
are compared by preloan level of living in Table 41 after matching was
completed.
The means for the experimental and control groups are
almost identical being 18.5 and 18.4 respectively*
Table 41.
FHA families matched with non-FHA families they
named compared by preloan level of living.
Score
1
5
10
15
20
25
30
35
-
4
9
14
19
24
29
34
39
Total
Mean
Probability
Non-FHA
FHA
1
4
10
9
7
3
4
2
0
6
11
10
9
5
3
2
40
18.5
46
18.4
Pj> .90
-109-
Preloan Year*
This was the last of the five variables on which
the FHA and Non-FHA were matched*
Table 42 indicates that the mean
preloan year for experimental form was December 1938 and that the
preloan year equivalent for the control was January 1939*
Thus,
after matching they differed by only one month.
Table 42*
FHA families matched with non-FHA families they
named compared by preloan year*
Year
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
Total
Mean
Probability
Non-FHA
FHA
2
5
2
3
4
7
6
3
5
2
1
2
4
3
5
7
6
6
3
4
3
3
40
46
January 1939
December 1938
P > .90
Present measures of change for the 46 non-FHA families matched
with the 40 FHA families who named them as similar in the pre-program
year will now be discussed to determine whether or not any changes
taken ple.ce support the working hypothesis that the supervised loan
aided those that received it.
Present 0rganizational Participation.
Table 43. presents the
two groups by present organizational participation scores.
The mean
for the experimental group was 19.20 while for the others it was 16.48.
Since the difference of 2.72 which was in the direction of supporting
the hypothesis, was not statistically signiiicant at the five percent
level, it might have been due to chance.
-
Present Level of Living*
110-
The 40 FHA have a mean level of living
score at the time of the interview of 32*13 compared to 29*83 for the
others, (Table 44)*
Table 43*
Since this difference of 2*3 is not a statis-
FHA families matched with non-FHA families they
named compared by present organizational par
ticipation scores*
Score
1
5
10
15
20
25
30
35
40
45
50
55
60
65
70
0
-
FHA
4
9
14
19
24
29
34
39
44
49
54
59
64
69
74
Total
Mean
Probability
Relation to hypothesis
Non-FHA
0
6
8
4
8
3
2
2
3
0
2
1
0
0
0
1
2
0
0
1
0
0
0
40
46
19.20
16*48
3
4
8
8
9
4
3
3
I
Support
tieally significant one, it can be concluded that chance factors
might have operated rather than the participation in the supervised
loaning program*
Table 44*
S c ore
15
20
25
30
35
40
45
_
_
_
mm
19
24
29
34
39
44
49
FHA families matched with non-FHA families they
named compared by present level of living scores.
FHA
2
5
10
7
7
7
2
ITon—FHA
1
12
11
10
6
6
0
-
111-
Table 44. (Cont'd)
Score
Non-FHA
46
FHA
Total
40
Mean
Probability
Relation to hypothesis
29.83
52.13
.10 < P <*15
Support
Present Informal Participation.
Comparison on this factor pro
duced a mean score of 55*50 for the experimental group and 52*54 for
the control (Table 45)*
The variance of 2*96 again in favor of the
hypothesis, might have been due to chance.
Table 45.
FHA families matched with non-FHA families they
named compared by present informal participation
scores•
10
20
30
40
50
60
70
80
90
~
-
Non-FHA
FHA
Score
19
29
39
49
59
69
79
89
99
Total
Mean
Probability
Relation to hypothesis
Present Gross Income*
0
5
6
4
6
8
7
3
1
2
1
8
8
14
3
8
1
1
40
46
52.54
55*5
* 4 0 < P < .45
Support
Family gross in come shown in Table 4i
indicates that the mean of the FHA is only #75 lower than the nonFHA*
This difference was not enough for statistical significance*
The 84 non-FHA families were reduced to 46 when matched by
frequency distributions on the five control variables.
None of the
40 FHA families who named either one or more of them were removed
in matching.
By matching, it was hoped to make the groups homogenous
-112
Table 46#
FHA families matched with non-FHA families they
named compared by present gross income#
Income in. Dollars
1
1,001
2,001
3,001
4,001
5,001
6,001
7,001
8,001
9,001
10,001
11,001
12,001
13,001
14,001
15,001
16,001
—
-
-
-
-
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
10,000
11,000
12,000
13,000
14,000
15,000
16,000
17,000
Total
Mean
Probability
Relation to hypothesis
FHA
Uon-FHA
1
6
5
7
7
4
0
4
2
0
2
0
1
0
0
0
1
0
2
10
8
7
7
6
1
3
0
0
0
0
0
0
1
1
40
46
#5000
#4925
P > .90
Hot support
in the pre-program year in order that differences in the post-program
year might then be due to one group participating in the program de
signed to increase their social and economic status#
Table 47 indicates that it was possible to make the means of the
two nearly identical on all pre-program control variables.
Table 47.
Variable
A summary
Summary of the preloan and present variablesx dis
tribution means of the FHA families after they
were matched with the non-FHA families they had
named •
FHA
Hon—FHA
Relation to
hypothesis*
Preloan Comparisons
Organizational Participation
15.7
16.3
Schooling of Husbands
9.6
10.0
Year First Started Farming April 1932 April 1932
Level of Living
18*5
18*4
Preloan year
Dec# 1938 Jan# 1939
-113-
Table 47 • (Cont’d)
Variable
FHA
Non-FHA
Relation to
hypothesis*
Present Comparison*
Organizational Participation
19*20
16.48
Support
Level of Living
32*13
29.83
Support
Informal Participation
55.5
52.54
Support
Gross Income
$4925.00 $5000.00
Not support
♦That the FHA supervised loan influence would increase the
participating families 1 rating on each of four measures of
present status.
of the post-program measures of family status, also presented in
Table 47, shows that the FHA participating families were slightly
higher on three of the four indicators.
This tendency for three of
the differences to be in the same direction and the fourth to be
negligible might demonstrate that the supervised loaning program did
effect them positively even though individual differences were not
statistically significant.
It should be remembered that the equating of the two groups in
the pre-program year was either because the experimental families
named the controls as being equal to them, or because they were
matched by distribution frequencies.
Distributions made similar
through matching could still allow relevant unknown causation factors
to operate which might have effected changes occurring through the
years when only one group of families had loans with the Farmers Home
Administration.
A more vigorous equating procedure, that of matching
by individual characteristics rather than by distributions, will be
applied in Part F.
Part F.
Comparison of Closely Paired FHA and Non-FHA Families
The utilization of a more stringent method of matching wherein
experimental families are paired by individual preloan characteristics
-114-
with the control group may add to our knowledge of the effects of the
supervised loan program.
This close pairing will concern all five variables selected in
Chapter V for matching in the preloan year.
The definition of close
pairing states that the two families must be in the same frequency
class for the first two variables, preloan organizational partici
pation, and husbands* schooling.
coding.
A class is the range set up for
They may vary one class in the third variable, year first
started farming, providing they do not vary in the fourth, level of
living, or the fifth preloan year.
They may differ by two classes
in the fourth variable if they do not vary in the fifth, or they may
vary one class in the fourth if the difference in the fifth is not
more than one.
The difference on the fifth variable may be two
classes if no other difference exists or one if they also vary by
one class in the fourth variable.
The method used in pairing was to first sort all 181 FHA and
non-FHA families by organizational participation and then to take
all families in each class of this distribution and sort on schooling
of husbands.
After that, all families in each class on schooling was
sorted on level of living with the same being done with the other two
factors.
Any FHA cards meeting the requirements stated above were
considered closely matched with the equivalent non-FHA families.
When
this process was completed, 25 FHA were found to be thus closely paired
with 25 non-FHA.
Preloan Organizational Participation.
The class range for organ
izational participation was 10 points which meant that one member of
the pair might vary that much.
Table 4# presents the 25 pairs with
their actual scores in order to provide a finer measure for comparison.
-115-
Table 48*
Closely paired FHA and Hon-FHA families compared
by preloan organizational participation*
Score
Pair Humber
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
FHA 1
Hon-FHA
1
0
0
8
29
26
28
23
20
14
16
1
8
3
4
6
9
4
0
0
8
8
7
6
7
23
20
27
21
27
19
16
5
5
0
4
1
5
3
0
3
1
0
0
0
0
6
8
25
8.56
25
8.64
0
0
Total
Mean
Probability
P>.90
Six pairs received the same score and the rest varied up to seven
points*
The mean score for the two was almost identical with the FHA
being only *08 higher than the non-FHA indicating the closeness of
matching*
Schooling of Husbands♦
It would be expected that the variance
between paired families would be slight, since the requirement was
for them to be in the same class*
The exact number of years of school
ing for the male head of the household for members of each pair is in
Table 49.
Seventeen pairs had the same amount of formal schooling,
-116-
six varied by one year, and two by two years*
The means were 9.48
and 9.32 years which gives an extremely small difference of .16.
Thus they were well matched on schooling.
Table 49.
Closely paired FHA and non-FHA families compared by
number of years of schooling of husbands.
Pair Humber
Years of Schooling
FHA
1
2
3
4
5
8
8
9
12
8
12
12
10
8
10
8
12
11
11
10
11
10
10
8
8
8
8
8
7
10
8
8
10
12
8
12
12
9
8
9
8
12
12
10
10
9
11
10
8
8
8
8
8
5
10
25
9.48
25
9 .32
6
7
8
9
10
11
12
13
14
15
16
17
13
19
20
21
22
23
24
25
Total
Mean
Probability
Year First Started Farming.
Non-FHA
P > .90
The 25 pairs of experimental and
control families are distributed in Table 50 to show how much members
of pairs differed from one another.
identical.
Both members of four pairs were
All but two others were within seven years of each other.
-117
Table 50*
Closely paired FHA and non-FHA compared by year
first started fanning.
Year started farming
Non-FHA
FHA
Pair Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
1924
39
34
43
17
45
38
36
30
33
23
38
34
45
41
37
33
23
44
44
40
31
11
38
26
1926
38
34
42
13
42
39
38
28
32
30
36
34
43
40
37
35
29
48
29
40
30
28
43
22
25
25
Total
Mean
Probability
March 1934
November 1933
P>.90
The twentieth pair varied fifteen years and the twenty-third pair
varied 17 years, since each member of the pair were actually only one
class apart as set up in the coding.
The mean for the FHA was Nov
ember 1933 and for the non-FHA March 1934 or a difference of only
four months •
This could just as well have been due to chance as to
any significant difference between them.
Preloan Level of Living.
Level of living in the preloan year
was the fourth control variable.
Actual scores are shown in Table 51.
-
Table 51 •
118-
Closely paired FHA. and Non-FHA compared by
preloan level of living.
________ Score____________
FHA
Non-FHA
Pair Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
10
11
12
23
15
27
17
15
27
21
10
10
16
13
10
25
19
12
9
15
6
7
11
6
10
11
9
12
29
14
21
15
18
20
21
12
14
10
15
11
21
17
5
13
12
19
8
9
10
14
Total
25
25
Mean
Probability
14,40
14.28
P > .90
Variance within all pairs was seven or less, except number 21 where
it was 13.
This difference of 13 points was allowed because they
were identical on three other variables and in the same class on the
other one.
The mean for the participation families was 14.28 and
14.40 for those not having the loans.
These differences and the
variance within pairs is so small that it could be due to chance*
Preloan Year.
Table 52 contains the 25 pairs with the pre
loan year of paired families.
Hven though this was the last control
-119-
Table 52*
Closely paired FHA and Non-FHA families compared
by preloan year.
Preloan year
Pair Number
FHA
Non-FHA
1938
38
36
43
35
43
36
39
35
39
36
42
35
44
40
39
41
39
40
43
40
40
34
38
34
1938
38
36
43
35
42
39
40
36
40
36
42
36
44
38
37
39
39
40
42
40
38
34
38
35
25
25
August 1938
Mean
Probability
July 1S38
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Total
P > .90
factor and thus might vary within pairs more than the other four,
there are extremely small differences.
Thirteen pairs are identical,
seven are one year apart, four vary by two years, and one is sepa
rated by three years.
The FHA has a mean of August 1938 and the
other group July 1938.
The close pairing was able to achieve a high degree of homo
geneity between families in pairs on the five control factors.
If
the 2.5 families were equal before the experimental half received loans
and if these loans did influence the present status of the families,
-
120-
such an influence might show up in differences between them at the
time of interviewing#
A discussion of the four measures of present
status should show whether or not such a hypothesis was supported#
Present Organizational Participation#
Table 53 has the dis
tributions of the pairs by organizational participation at the time
Table 53*
Closely paired FHA and non-FHA families compared
by present organizational participation#
Score
Pair Humber
FHA
Non-FHA
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
1
14
12
18
19
8
45
29
10
17
13
22
15
1
20
12
39
12
1
6
8
11
0
3
8
4
5
18
16
7
21
48
21
25
15
32
19
9
0
1
1
9
18
13
22
10
0
40
0
7
Total
25
25
13,76
Mean
Probability
Relation to hypothesis
of the interview.
14,44
P > .90
Hot support
Differences between members of the pairs is 19 or
less except for pair number 17 which varied 30 and pair number 23
varying 40.
The means are nearly equal, with the FHA being 13.76 and
-
121~
the non—FHA being 14*44 or a difference of only *68 in favor of the
non—FHA*
This indicates that both groups changed about the same, re
gardless of the loaning program*
Present Level of Living*
The FHA mean for level of living scores
after they had participation in the loaning program was 29*64 while it
was 27.08, or 2.56 points lower, for the non-FHA*
within pairs was from zero to 23 points*
These are not large enough,
however, to be statistically significant.
Table 54.
The difference,
They are shown in Table 54.
Closely paired FHA and non-FHA compared by
present level of living*
Score
Pair Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Total
Mean
Probability
Relation to hypothesis
FHA
40
28
29
31
29
40
29
41
2G
37
20
31
28
18
29
40
48
21
24
23
23
38
18
16
40
Non-FHA
33
21
23
24
23
29
26
30
38
27
35
25
27
21
34
21
30
21
33
23
25
26
41
16
25
25
25
27.08
29*64
P?
*90
Support
-
122-
Present Informal Participation*
Table 55 contains the scores
on the informal participation scale of the 25 pairs*
The mean for
the FHA is nearly the same as for the control group, 55*40 compared
to 54*80.
There was a great deal of variance within pairs with it
Table 55*
Closely paired FHA and Non-FHA families com
pared by present informal participation*
Score
Pair Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Total
Mean
Probability
Relation to hypothesis
Non-FHA
FHA
25
41
52
52
74
67
87
38
20
38
34
71
85
62
53
67
58
38
59
68
62
20
31
80
94
45
75
68
36
28
65
88
67
15
32
91
32
52
49
63
70
72
38
76
75
38
42
52
32
69
25
25
54.8
55.4
P > .90
Support
being about as much in favor of the experimental families as of the
others.
Thus whether or not the family was involved in the super
vised loan program did not seem to be a major factor in this instance.
Gross Family Income*
Total family income before farm expenses
-123-
were subtracted is presented in Table 56.
There was extensive var
iance, as much as $14,469, within pairs, but when the amounts that
the FHA families exceed the others are compared to the amounts by
Table 56.
Closely paired FHA and Non-FHA compared by
present gross family income.
Pair Number
1
2
3
4
5
6
7
8
9
10
11
12
13
1415
16
17
18
19
20
21
22
23
24
25
Mean
Probability
Relation to hypothesis
Income in dollars
FHA
Non-FHA
3993
1500
4320
2892
3868
3000
3452
2193
8293
4379
3550
6000
3050
3697
8000
1900
6500
5230
2899
3000
1578
4200
5262
7000
3800
5300
800
2200
4260
3200
4171
7358
5960
5000
9000
2000
4000
1600
2122
2000
3061
1610
3151
4084
15789
1320
4000
23 60
4900
11000
4174.52
4377.56
P > .90
Support
which the non-FHA exceed, it can be seen that the totals are nearly
equal.
The mean gross income for FHA families was $4377.56 while it
was #4174.52 for the non-FHA.
Again, the fact that one-half of each
pair had the loan did not seem to have a great effect on their pres©ivfc income even though the income tends to lie in the direction of
the hypothesis.
-
124-
Data in Part F have shown that it was possible to closely pair
25 non FHA families similtaneously on the five control variables*
Later it was discovered that there were no statistically significant
differences between the means of the two groups even though FHA
families has been involved in the supervised loaning program for a
minimum of six years*
This seemed to indicate that the loan did not
cause the families having it to change any more than did those with
out it although three out of the four comparisons were in favor of
the FHA families*
Table 57 summarizes these comparisons of the five
control variables and the four present measures of change by present
ing their means#
Table 57.
Factor
Summary of the closely paired FHA and Non-FHA
families on the control variables and present
measures of status#
FHA
Preloan Comparisons
Organizational Partici
8.56
pation
Husbands* schooling
9.48
Year first started
farming
Nov# 1955
14.28
Level of living
Preloan year
Aug. 1938
Non-FHA
Relation to
hypothesis*
8.64
9.32
March 1934
14.40
July 1938
Present Comparisons
Organizational Partici
Not support
14.44
13.76
pation
27.08
Support
29.64
Level of living
54.80
Support
Informal participation 55.40
Support
Gross income
#4,377*56 #4,174.52
*That the FHA supervised loan influence would increase
the participating families* rating on each of the four
measures of present status#
This lack of any statistically significant difference between
the experimental and control families may be due to factors such as
these fours
126
1*
The techniques and instruments used to discern differences
may not be fine enough*
2*
The program actually may not have had an influence in the
success of the participating family*
3.
It may have had an influence but the state of the business
cycle and other effects may have caused those families
without loans to increase as much as those with them*
4*
The matching of the two groups in the preloan or before the
program year may not have been precise enough for changes
to show up*
This last factor can still be altered in this analysis and for
this reason a more precise method of matching will be attempted in
Part G.
Part G*
Comparison of Precisely Paired FHA and Non-FHA Families
This part compares the experimental and control groups of’families
after they have been precisely paired on the final control variables to
determine whether or not the FHA families were affected by the super
vised loan program.
This precision matching procedure has strict re
quirements as to the homogeneity of the members of the pairs and thus
the number of pairs that can qualify is small*
116
Every FHA experi
mental fWAy had to have one non-FHA control family who was in the
same class as it was in all five preloan variables.
Only five pairs
qualified under this requirement*
Tables comparing the five families in each group on each preloan
scale are not presented because the distribution would be identical.
For each class in preloan organizational participation, husbands*
schooling, year first started farming, preloan level of living, and
preloan year contraining one or more FHA families, there was also the
116 For a discussion of the influence of the closeness of matching on
size of sample see Chapin, Experimental Designs * Op* cit. Pp.
101 - 103.
-125-
same nuiaker of non-FHA families*
Even within classes very little
variance between scores of the members of each pair existed*
Five families from the original 97 FHA, selected because they
could be matched with five from the total group of 84 non-FHA fami
lies, might, even though the total is low, give an indication of
changes emanating from the loan program because of their homogeneity
in the preprogram situation#
The four measures of present status
will be discussed to determine their relationship to the hypothesis
that the loan program aided the participants in achieving economic
and sacial gains#
Present Organ!zational Participation♦
The 10 families are
compared in Table 58 by organizational participation scores at the
Table 58,
Precisely paired FHA and non-FHA families com
pared by present organizational participation
scores•
Scores
Pair Number
4
5
18
16
7
1
14
12
18
19
1
2
3
4
5
Mean
Probability
Relation to hypothesis
time of interviewing*
Non-FHA
FHA
10.0
12*8
.40 < P < # 5 0
Support
The difference between members of pairs ranged
from two to 12 , but the mean difference was only 2.8 when the three
pairs in which the FHA were higher are combined with the two in which
the others had larger scores*
Since the mean difference in favor of
the FHA could have happened by chance, it is not known whether or not
the FHA changed more or less than non-FHA families#
Present Level of Living#
Variance between members of the pre
-127-
cisely paired families in Table 59 was six points for two of them and
seven for three, all in the direction of a higher score for FHA fami
lies*
This is highly significant because the t-test yielded a proba
bility that the difference would occur by chance less than one time
Table 59*
Precisely paired FHA and non-FHA families com
pared by present level of living*
Scores
Pair Humber
FHA
Hon-FHA
40
28
29
31
29
33
21
23
24
23
1
2
3
4
5
Means
out of 100.
24.8
31.4
Probability
Relation to hypothesis
P>.01
Support
Thus it can be said, therefore, that the five FHA fami
lies were somewhat higher than the others on present level of living
s cores•
Present Informal Participation.
This index of present status,
as presented for the five pairs in Table 60, showed a variance
Table 60.
Precisely paired FHA and non-FHA families com
pared by present informal participation*
bcores
Pair Humber
Hon-FHA
FHA
45
75
68
36
28
25
41
52
52
74
Means
Probability
Relation to hypothesis
within pairs to be from 16 to 46.
48.8
50.4
P>.90
Hot Support
Three non-FHA families scored
higher than their partners, while the reverse was true for the other
128-
two.
The mean difference of 1*6 in favor of the non-FHA thus casting
doubt on the hypothesis that the FHA would be higher, could have
happened b y chance*
Present Gross Income*
Table 61 gives the distribution of the two
members of each pair on gross family income which means their total
Table 61.
Precisely paired FHA and non-FHA families compared by present gross income.
Inc ome in dollars
Non-FHA
1500
2892
3000
3452
4379
Pair Number
1
2
3
4
5
FHA
3993
4320
3868
2193
8293
Means
Probability
Relation to hypothesis
4533.40
3044.60
.05 P .10
Support
income before farm expenses are subtracted.
FHA exceeded that of their partners*
than that of the non-FHA.
The income of four of the
Their mean was $1488.80 more
This difference was just above the five per
cent level of probability established as necessary before the two
groups could be co ns idered sxgnificantly different*
Thus, there was a
tendency for the FHA to have higher incomes, but this was not quite
statistically significant.
A summary of the precisely paired comparisons in Table 62 showed
the FHA families to have rated higher on the present organizational
DS.i*tici'C£|'fci<3n level of Irving and Income than non—FHA, while the
informal participation of the non-FHA was larger.
The size of
organizational participation difference in favor of the FHA and the
size of the informal participation in favor of the non-FHA were not
large enough to be statistically significant.
However, the level
of living difference in favor of the FHA was highly significant and
the income difference also in favor of the FHA w as just above the
-129-
five percent level of significance.
It must be remembered that rig
orous matching reduced tshe families to five pairs.
The degree of
confidence which can be placed in a generalization coming from the
data is related to the degree these small samples represent the
universes from which they were extracted and to the pure character
istics which have been maintained within them.
Table 62.
Summary of comparison of means of the five precisely
paired FHA and non-FHA families on present measures
of social and economic status.
Present comparison
FHA
Non-FHA
Relation to
hypothesis*
Organizational partici
Support
12.8
10.0
pation
Support
24.8
Level of living
31.4
Not support
50.4
48.8
Informal participation
3044.60
Support
4533.40
Income
♦That the FHA supervised loan influence would increase the par
ticipating families* rating on each of the four measures of
present status.
One additional method of studying the effect would be to observe
the rate of change between the pre-program year and the time of
interviewing or "after" year.
This would give relative change re
gardless of where either group was at the beginning.
This has been
analyzed in Appendix B.
Part H.
Summary of Six Methods
One of the two purposes of this project was to objectively da
te m i n e whether or not this Governmental agency, which had made loans
for buying machinery, livestock, or farms to low income farm families
unable to obtain credit elsewhere, had benefited the participating
families.
The six combinations of ex post facto techniques with
matching were devised in the attempt to appraise this supervised
loaning program.
A summary of findings in the post-program year may
130-
indicate whether or not any useful generalizations about the Farmers
Home Administration can be made.
Post-program means on the four measures of family status for
both the FHA and non-FHA families were presented in Table 63 for
each of the six methods of approaching this evaluative problem.
The
direction of the mean differences with respect to the working hypo
thesis that taking part in the Farmers Home Administration would
cause an increase in social and economic status was also indicated.
In five out of the six methods FHA families increased more than
non-FHA in their participation in community organizations.
However,
none of the differences between them and the non-FHA were large
enough to be statistically significant at the five percent level.
Similarly in five out of the six comparisons on level of living,
the FHA exceeded the non-FHA, and they were equal the other time.
Only one of these five differences was statistically significant.
Informal social participation comparisons also showed the FHA
to be more active in five out of the six methods with the reverse
being true in one comparison.
tistically significant.
the same pattern.
These differences were not sta
Family income comparisons did not follow
The FHA had higher incomes one-half of the time
with the non-FHA being high the other half.
One of the times when
the FHA had higher incomes the difference approached the five
percent level of significance.
to statistical significance.
None of the others were that close
PI
o
Pi
P
d
O
•H Pi
Pi
•rl
©
i>» rH
rH rH
•rl O
CD
TO
•H
Q
TO
*
.
Pi