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 Preloan Year* The preloan year of t he two groups as presented in Table 32 is quite homogenous, which is to be expected since the preloan year of the non-FHA was the same as that for the family that Table 32* Forty FHA families and 84 non-FHA families they named compared b y p re loan year* Year 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 Total Mean Probability selected them* FHA Number Percent 2 5 2 3 4 7 7 2 5 2 1 40 5 13 5 8 10 16*5 16*5 5 13 5 3 100 Non-FHA Number Percent 3 11 4 6 12 13 12 4 12 4 3 4 13 5 7 14 15 14 5 14 5 4 84 100 Dec* 1938 Jan* 1939 ** P>*90 The statistical test substantiated this observation* -101 A comparison of* “the 50 FHA and 84 non-FHA in "the preloan year on the five control variables has just been presented* The four variables of present status will be presented next in order to determine whether or not possible changes occurring between this preloan year and the time of interview were in the direction of the hypothesis* The hypo­ thesis advanced was that the FHA supervised loans would cause partici­ pating families to increase more than ^ control families in organi­ zational participation, level of living, informal participation and inc ome * Present Organizational Participation* Table 33 indicates that present organizational participation scores for both groups are nearly the same, especially if means are compared* Table 33* Score 0 10 20 30 40 50 60 - 9 19 29 39 49 59 69 Total Experimental families had Forty FHA families and 84 non-FHA families they named compared by present organizational partici­ pation* FHA Number Percent Non--FHA Percent Number 13 12 6 5 2 2 0 32 30 15 13 5 5 0 23 30 12 9 6 3 1 27 36 14 11 7 4 1 40 100 84 100 19.50 — 18.75 Mean .75 .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 © O Pi 1 CO CO C•k o> LO •» O o CO o © O O •» LO «>rH to d •rl CM O LO CM 05 * LO LO •» 'cH CO O P CO O 00 <*> OJC 05 ■t P. © Pi *+• CO t- to •* © LO to © to © © to O d CO ■t (0 d p © 1 © p TO biO p. •rl bO 0 1 -P CO O a, • co © © P © d TO © © © © -d g -P Pi bQ d Pi O •rH •rl TO JSi o 1 i d P>H © © cd P=l tH to 05 CM • CM o O ■sjt o oo LO to CM LO o t> LO fc- O LO • o •o LO fc- LO to to LO • to to LO LO • © O -+ LO rH • rH CO to 10 • 0 CO to CD • 05 CM P- P P O O * r—1 CO CO co 00 * 05 CM CO rH • CM CO O CO • CM CO 10 CO * rH CO o © Pi © Pi © 0 3 © Pi © © PM PM -d >» d © p p © o P P © Pi , Pi d -d © bO S Pi •« H 0 TO © rd 0 O rd P £ © TO © TO tO O -d P © w> pi , © N P •h !>> © rH TO © -Pi TO Pi Cm 0 d 0 Pi Pi TO d © ■3 ® H rd P cit. P. 104 findings of this dissertation was supported by an earlier writer* Charles Peters, in a note on Frank A* Ross's criticism of a statis­ tical procedure, wrote in 1933: "Dr. Ross seems to think that, in a problem in which a series of interrelated differences is involved, the statis­ tical significance of the finding hinges upon that of the differences taken individually. This... is a common error. But such an assumption violates a fundamental principle of probability. .. .The differences obtained.•£in a studyj were prevailingly in the same direction, and consequently the total reliability of the showing. ..is to be determined by multiplying together the separate chances through the whole series. This makes the statistical significance of^.^thej] findings overwhelming.n^-20 The fact that most of the differences were not great enough to be statistically significant at the five percent level does not necessarily mean that they do not furnish evidence which can be used in determining whether or not the FHA program benefited participant families according to an article of George W. Snedecor. He stated: "...I think there is little merit in setting arbitrary values to probability, then decreeing that smaller must lead to rejection. The 5 per cent and 1 per cent points are convenient milestones which the investigator will note in passing, but any probability turned up consti­ tutes evidence pertinent to decisions. ...” ”1 am coming to believe that the term, ”test of signifi­ cance” is creating more confusion than it resolves. The phrase is not descriptive of the logical and experimental concepts involved. Partly for this reason, the dictionary definitions of the words blind many to their meaning in terms of probability. Perhaps the time has come to state the probability of the testing function as the end-point of the statistical investigation leaving the researcher to combine this evidence with that already accumulated, then to rest his decision on the whole of it."^l 120 Charles C. Peters. "Hote on a Misconception of Statistical Sig­ nificance •" American Journal of Sociology. Vol. 39. 1933. P. 234. 121 George W. Snedecor. "The Use of Tests of Significance in an Agr cultural Experiment Station." Joumal of the American Statisti Association. Vol. 37. September 1942. P. 386. -136 If the conclusions of authors such as these are accepted, it could "be said that the quite consistent findings in favor of the FHA families provides evidence that the supervised loan program did help these participants increase their social and economic standings in their communities. -136- Chapter VII Summary and Conclusions Part A. Introduction The developing realization that environment could be changed and that the future could be influenced by man lead to attempts to intro­ duce factors which would bring more desirable results. This social planning, the conscious effort to predict what present situations will lead to and then to develop programs to alter future results, has raised the critical problem of how the relationship between a variable introduced in a situation and the resultant change can be measured. This problem is more difficult if the stimulus variable was introduced at some time in the past so that any resulting change would have already taken place. This was the type of problem attacked in this dissertation whose two principal objectives were first, to investigate the effectiveness of the Farmers Home Administration, a social action agency organized to carry out a social plan, and secondly, to develop and experiment with ex post facto research methods combined with matching techniques in measuring change that had already occurred. It was hoped this project would overcome the weakness of past research, as revealed through the literature review, which was the failure to isolate effects of such an action agency from other causative influences. Data available for this analysis were secured by interviewing an experimental group of 97 families who had received loans from the FHA to buy or operate farms and by interviewing a control group of 84 families who had not received such loans. All of these families were living on farms in or near Hillsdale County in southern Michigan, ■137- The 97 FHA families were randomly selected from the 636 receiving loans from 1936 through 1946 in Hillsdale County. The 84 non-FHA families were those named by FHA families as being equal to them before they received FHA loans* Both groups met the requirements of farming at the time of interviewing in 1950 and at the time they received their loans, having had no catastrophe such as a fire, and being un­ broken families. Since this project was planned to be completed in a relatively short time, an ex post facto type of research was utilized in which a base line was established by investigating the past before the experimental variable had been introduced which was the FHA loan. Data obtained in the carefully constructed interviews consisted of scores on scales measuring level of living, participation in organ­ izations, and informal social participation as well as husband's schooling and age, whether or not families had lived in a city, when they started faming, and present family income. FHA families supplied information about their level of living and organizational participation for the year just before they were granted their first Fanners Home Administration supervised loan as well as for the date they were interviewed. Since the comparison of change in these FHA families derived from the before and after situation has limitations in a scientific inquiry, the additional method of controlling unknown causation variables by matching them with the control group of non-FHA was applied. Thus, the control group also furnished information similar to that of the FHA for a preloan equivalent year and for the interview time. The equivalent pre-program year for each non-FHA family was the same as the pre- -133- program year of the FHA. family who had named them as being similar to themselves. Not all of the characteristics obtained in the interview with the families could be used in the matching. used was primarily by testing. Selection of those to be These five characteristics had the highest degree of association and were employed in the matchings preloan organizational participation, husband* s schooling, year family first went into fanning, preloan level of living, and preloan year. The experimental situation for this research, then, consisted of one group of families having participated in a program which was supposed to improve their social and economic status and another group not having taken part in it. The working hypothesis tested was that this Farmers Home Administration supervised loaning increased the social interaction, level of living, and income of participants. The techniques employed to test it were those of exploring the past with ex post facto methods, and those of matching participant families with a control group in a pre-program situation and then comparing them later in a post—program situation. Part B. Generalizations about the Fanners Home Administration Six experiments in the method of comparing the FHA families with the non-FHA control families were employed, each of which involved different degrees of matching the two groups. The first of these methods compared the total groups of 97 FHA and 34- non-FHA families for which data were available before any matching was accomplished, except that due to the fact that FHA families named non-FHA families similar to them. The FHA group had — 139— not participated as much as had the others in community organizational activities during the preloan year* There were other variances be­ tween them which were not statistically significant. In the non-FHA group husbands had slightly more schooling, they had started farming about two years earlier, and their level of living in the preloan year was somewhat higher than that of the FHA. The preloan year for the FHA was slightly later than the preloan year equivalent of the non-FHA but it, also, could have been due to chance. The comparison after the experimental group had FHA loans for a minimum of four years showed them to be somewhat higher on formal organizational participation and informal participation. The controls were a little higher on gross family income while they both were about equal on level of living. Hone of these differences were statis­ tically significant. A significant change, however, had occurred while families had loans when a comparison was made of relative changes between the two time periods. These FHA families had Increased their activity in organizations at a greater rate than the controls so that the statis­ tically significant differences between them in the preloan year was reduced until at the time of interview it was no longer significant* Other significant differences may have been inherent in the data but undiscovered at this time due to the type of analysis. For example, the control families were secured from a list of those named by FHA families as being in about the same social and economic conditions as they were in the pre-program year. It is possible that the FHA selected families "better" or higher in social and economic status than they were. This chosing upward could have influenced the -140 accuracy of their indication of the preloan situation as well as of the present. The second experiment in method tried to reduce the possibility of this influence by matching the total group of 97 with the 84 by their frequency distributions on the five control variables • Through * this procedure they were quite successfully equated in the preloan year which resulted in no significant differences and nearly equal means. During the time they had loans, the FHA families had made larger increases than had the non-FHA so that their post-program organi­ zational participation, level of living, and informal participation means were higher* The non-FHA, on the other hand, increased their income more than the other group. All differences were so slight, however, that they might have occurred by chance. Thus it is not possible to say, on the basis of these data, whether or not the FHA program had any influence on the social and economic status of par­ ticipants . The amount of confidence that could be placed In any generaliza­ tion about an experiment is related to the homogeneity of the two groups in the base period. Even though in this instance they were matched on five characteristics, there may have been other relevant factors uncontrolled. It was known that some of the FHA families were over-represented and some were not represented among the non-FHA families in the control group due to the procedure in securing names of families to be used as control. It was discovered that only 40 out of the 97 FHA families had named all the 84 control families. The prospective control families furnished by the other 57 FHA - 141- families could not be used because they did not meet basic require­ ments such as being in farming and not having had an FHA loan# In the thir'd experiment, the 84 non-FHA families were compared with just the 40 FHA who had named one or more of them# The p re loan analysis somewhat confirmed the suspicion that FHA families had selected upward since the non-FHA were higher than them on formal organizational participation, although there were no statistically significant differences on the other four preloan characteristics# A greater increase in organizational activity of the FHA during the time they had the loan made them quite similar in the post-program year to the controls# Both groups were also similar on the other three present measures of social and economic status, although the FHA were slightly higher in each instance# The fourth experimental method matched the 40 FHA families with the 84 non-FHA they had named by their distribution frequencies on the five control variables* Forty-six non-FHA remained and none of the 40 FHA were removed in the matching process* There was little variance in the five pre-program, variables means upon completion of matching# Analysis of data in the post-program year showed the FHA to be slightly higher on organizational participation, level of liv­ ing, and informal participation, while the non-FKA were a little high­ er on family income, but none of the differences were significant. Matching by distribution frequencies which was done in this com­ parison, might have allowed other factors to exert causative influ­ ences independent of or combined with the supervised loans# A finer type of matching might bring out more definitive differences# The next method of analysis, the fifth, was to match more —142— strictly the total groups of 97 FHA and 84 non—FEA families in a close Pairing procedure# An. FHA family was considered to he closely paired with a non-FHA family if both were in the same class or relatively close to each other’s class in each frequency distribution of the five control variables. The 25 pairs which qualified thus had almost equal means on those preloan variables# Post-program comparisons of members of these 25 pairs indicated that the non-FHA exceed the FHA in organization participation while the FHA T^ere higher on level of living, informal participation and total family income# The sixth approach to the problem of determining effects of the FHA program, one of precise matching, considered only those five pairs of families within which there were no variances on control variables# Post-program results Indicated that the five FHA families were some­ what higher on organizational participation, level of living and income, and lower on informal participation, than were the five nonFHA. The difference in favor of the FHA on level of living was statistically significant and that in favor of the non-FHA on infor­ mal participation epproached the five percent level of significance. It should be recognized that the amount of confidence placed in these results is influenced by the degree to which the small samples rep­ resent the universe from which they were drawn. If pure relevant characteristics of the original universes were maintained within them, generalisations would be valid. The four measures of present status in each of ohe sxx metnods made possible 84 comparisons. In 18 of these the relation of the difference between the experimental and control families was ir. favor 143** Oj. the hypotheses that the FHA would have been influenced by the supervised loan program* There was no apparent difference in one comparison and in the other five it was in favor of the non-FHA thus casting doubt on the hypothesis* However, in only one of the 24 comparisons was the difference between the FHA and non-FHA large enough to be statistically signifi— cant at the five percent level. Even though they were so small, the fact that 18 out of the 24 were consistent seemed to indicate that the participating families did change toward having higher social and economic status in their communities as measured by organizational participation, informal social participation, level of living, and inc ome * The application of research techniques in an evaluation of one agency involved in social planning suggested that it did cause changes. Any attempt to generalize from this experience to other types of plan­ ned social action should be made with caution. Variations in ends desired and means utilized by social planning and action groups might influence the experimental procedures necessary to measure possible charges occurring as a result of their programs • The type of agency itself as well as the characteristics of the subject under observation and the people participating also necessitates extensive consideration of how methods presented here can be utilized elsewhere. Part C. Generalizations About Ex Post Facto Research Methods Three somewhat new techniques in ex post facto and matching re­ search have been applied in this dissertation. The first was a method of securing a list of family names from which controls could be selected# In this research the usual pro­ cedures for obtaining the names of control families were considered not practicable# For this reason a method was worked out whereby experimental families would furnish names to be used in the control group# The question used in the interview was: "Would you mind telling me the names of two or three families that were in about the same condition as you were, in 19__? That is, who were about the same age as you, with about the same number of children, and who ran about the same size farm as you did." The year in which the control families were supposed to be similar was the preloan year for the FHA, although they were not told why that specific year was selected# Previous questions in the interview indicated where the family had been living and what they were doing at that time, enabling the inter­ viewer to assist the interviewee in pinpointing his memory on that specific year# This method of securing control families is especially valuable in scientific evaluation of past experiences when the total group of possible control families is extremely large or unknown. If it is utilized, precautions must be taken with reference to these influences experimental families tend to choose others who are higher socially than themselves; many experimental-type families may be named and thus cannot be used; and the universe may be unknown which might prevent statistical proof of what the control group represents. The second technique was a method of securing data for matching experimental families with control families. Most research utilizing matching procedures obtained data for matching in the pre-program situation from records kept on the families since that time. The lack 145 of adequate records in this project necessitated the development of another source. A schedule of questions was constructed so that the experimental and control families would furnish such information in an interview. Pre-testing in which families furnished data about themselves in the past which could be compared to actual records and the discovery of no irregularities in the data collected seemed to indicate that such an attempt to recreate the past in an interview was successful. If an interview can be used in this manner, many more evaluations can be accomplished of social change that has al­ ready occurred and for which no prior records are available. A third technique attempted to refine comparisons between experi­ mental and control families by six types of matching differing in degree of intensity. An effort was made to increase the homogeneity between the two groups in the pre-pro gram year so that differences between them in the post-program year would more likely be due to the experimental variable. This was done by initially comparing the total groups interviewed, then matching thorn by frequency distributions and later pairing one individual in the experimental group with one in the control, first somewhat loosely and finally very precisely. These six variations in rigor of matching may be useful in other research in determining ways of bringing out significant differences between an experimental and a control group. This dissertation then, has indicated a tendency for the Farmers Home Administration's supervised loaning program to have benefited families in Hillsdale County, Michigan. In investigating this social action organization it has suggested the possibility of securing infor­ mation about the past in interviews so that this type of evaluation can -146 be accomplished where records are unavailable. It has shown how in an interview, names of control families can be obtained from an ex­ perimental group of families. It has also described several refine­ ments in controlling variables possibly influencing change, along with the program under study, by matching experimental families with a control group of families. -end- - 147- Appendix A Bibliography Angell, Robert C. "Difficulties of Experimental Sociology." Forces. Vol. XI. December 1932. Pp. 207-210. Beecher, John. "Save Farm Security." April 26, 1943. Pp. 561-563. Belcher, John C. Economic Scale." Pp. 246-255. Chapin, F. New York. New Republic. Social Vol. 108. "Evaluation and. Restandardization of Sewell's SocioRural Sociology. Vol. 16. No. 3. September 1951. Stuart. Experimental Designs in Sociological Research. Harper and Brothers. 1947. Chapin, F. Stuart. "Matching in Ex Post Facto Studies.” "Letters the Editor." American J o u m a l of Sociology. Vol. LVI. No. 2. September 1950. Pp. 187-190. to Chapin, F. Stuart. "Social Participation and Social Intelligence." American Sociological Review. Vol. 4. April 1939. Pp. 157-166. "College-Aid Land Appropriation. Agricultural-Mechanical Colleges." United States Code Anno ta.ted. Title _[£. Agriculture. St. Paul, Minnesota. ¥est Publishing Company. 1939. P. 157. Cottam, Howard R. Methods of Measuring Level of Living. Social Parti­ cipation and Adjustment of Ohio Farm People. Methodological Supple­ ment to Ohio Agricultural Experiment Station Bulletin 624. Columbus, Ohio. Ohio State University and Ohio Agricultural Experiment Station. Department of Rural Economics and Rural Sociology. Mimeographed Bul­ letin 139. July 1941. Eaton, Joseph V. Exploring Tomorrow's Agriculture. Harper and Brothers. 1943. New York. "Establishment of the Resettlement Administration." 7027. April 30, 1935. Executive Order. "Farmers' Hone 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, B.C. United States Government Printing Office. 1947. P. 1062. Farm Tenancy. Special Commxttee on Farm Tenancy. Washington, D.C* Government Printing Office. Superintendent of Documents. 1937. Freedman, Ronald, and others. Principles of Sociology: A Text with Readings. New York. Henry Holt and Company. 1952. P. 659Freedman, Ronald. "Incomplete Matching in Ex Post Facto Studies." American J o u m a l of Sociology. Vol. LV. No. 5* March 1950. Pp. 435—437* -143- Gaer, Joseph. Toward Farm Security. Department of Agriculture. 1941. Washington, D.C. United States Gibson, Duane L. The Clientele of the Agricultural Extension Service. East Lansing, Michigan. Michigan Agricultural Experiment Station Quarterly Bulletin 26. No. 4. May 1944. Greenwood, Ernest. Experimental Sociology. A. Study in Method. New York. Kings' Crown Press. 1945* Griswold, A. Whitney. Farming and Democracy. New York. Brace and Company, Inc. Hagood, Margaret J. and Hitchcock, Inc. Harding, T. S. Harcourt, 1943. Statistics for Sociologists. New York. 1941. Pp. 46I-465. "Our Underemployed Farm Labor." Tomorrow. Reynal July 1943* Pp. 25-23. Harness, G. Farm Security Administration Rehabilitation Loan Experience in Five Missouri Counties. Missouri Agricultural Experiment Bulletin. Vol. 376. June 1942* Pp. 1-34* Heilman, R. "Farmers Try Group Medicine." December 1940. Pp* 72-30. Harpers. Vol. 132. Henderson, H. and Shaw, S. "Million-Dollar Sharecroppers." Vol. 117. June 12, 1946. Pp. 13-I9f. Collier's. Hill, E. B. Types of Farming in Michigan. East Lansing, Michigan. Michigan Agricultural Experiment Station Special Bulletin 206. (re­ vised June 1939)* House Committee on Agriculture to Investigate Farm Security Adminis­ tration. Washington, D.C. Superintendent of Documents. June 7, 1943 - May 3, 1944Kimball, Solon T. A, Cas® Study in Township Zoning. East Lansing, Michigan. Michigan State College Experiment Station Bulletin. 1947. Kirkpatrick, E. L. Analysis of 70.000 Rural Rehabilitation Families. Washington, D.C. United States Farm Security Administration. Social Research Report No. IX. August 1933. Kolb, John H. and Brunner, Edmund deS. A Study of Rural Society. Houghton Mifflin Company. 1952. P. 92. New York. Larson, Olaf F. "Rural Rehabilitation — Theory and Practice." Rural Sociology. Vol. 12. September 1947. Pp. 223-237. Larson, Olaf F. and others. Ten Years of Rural Rehabilitation in the United States. Washington, D.C. United States Department of Agri­ culture. Bureau of Agricultural Economics. July 1947. - 149- Larzelere, H. E. Farm Debt A_d.iustment in Michigan Through Farm Security Administration. Michigan Agricultural Experiment Station Quarterly Bulletin. Vol. 25. August 1942. Pp. 47-65. Larzelere, H. E. 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. Leighton, Alexander H. Human Relations in a. Changing World. E. P. Dutton and Company, Inc. 1949* P.7. New York. Lindquist, E. F. Statistical Analysis in Educational Research. New York. Houghton Mifflin Co. 1940. Pp. 51-54* Loomis, Alice. Social Forces. "Observation of Social Behavior in Industrial Work." Vol. XI. December 1932. Pp. 211-213. Loomis, Charles P. and Beegle, J. Allan. New York. Prentice-Hall, Inc. 1950. Rural Social Systems. Loomis, Charles P., Beegle, J. Allan, and Longmore, T. Wilson. "Critique of Class as Related to Social Stratification." Sociometry Monograph. No. 19. 1948. Pp. 3-21. Loomis, Charles P. Studies of Rural Social Organization in the United States, Latin America and Germany. East Lansing, Michigan. State College Book Store. 1945* Loomis, Charles P. and Gresham, Glen. "The New Mexican Experiment in Village Rehabilitation." Applied Anthropology. Vol. II. No. 3* June 1943* Pp. 13-37. Mangus, A. R. and Cottam, Howard R. Level of Living. Social Partici­ pation and Adjustment of Ohio Farm People. Wooster, Ohio. Ohio Agricultural Experiment Station Bulletin 624* September 1941. Ogburn, William F. and Nimkoff, Meyer F. Sociology. Houghton Mifflin Company. 1950. P. 585* New York. Peters, Charles C. "Note on a Misconception of Statistical Signifi­ cance." American Journal of Sociology. Vol. XXXIX. 1933. P. 234* Raper, Arthur F. and Reid, Ira DeA. Sharecroppers All* Chapel Hill, North Carolina. University of North Carolina Press. 1941. Sewell, william H. The Construction and Sts.ndardization of a Scale for the Measurement of the Socio-Economic Status of Oklahoma Farm Families. Stillwater, Oklahoma. Oklahoma Agricultural and Mechanical College. Agricultural Experiment Station Technical Bulletin No. 9. April 1940. -150- Snedecor, George W. "The Use of Tests of Significance in an Agricultural Experiment Station."■ Journal of the American Statisti­ cal Association. Vol. 37. September 1942. P. 336. Solem, M. "The Effect of Land Purchase by Farm Security Adminis­ tration Borrowers on Agricultural Production in the Northern Great Plains." Journal of Land and Public Utility Economics. Vol. 19. May 1943. Pp. 231-233. Taylor, Carl C. and others. Alfred A. Knopf. 1949* Rural Life in the United States. New York. Taylor, Carl C., Wheeler, Helen W., and Kirkpatrick, E. L. Dis­ advantaged Classes in American Agriculture. Washington, D.C. United States Farm Security Administration. Social Research Report. No. VIII. April 1938. The Code of Federal Regulations of the United States of America Having General Applicability and Legal Effect in Force June 1< 1933. Title Jo. Agricultural Credit. Washington* D.C. United States Government Print­ ing Office. 1939. The Statutes at Large of the United States of America from March 1933 to June 1934. Vol. 58. Part I. (Under Title II) "Emergency Appropria­ tions, 73rd Congress." Session II. June 19, 1934. United States Census of Agriculture: 1945. Vol. 1. Part 6 . "Michigan Statistics for Counties." Washington, D.C. United States Department of Commerce. Bureau of the Census. 1946. United States Census of Population: 1950. Michigan, Washington, D.C. United States Department of Commerce. Bureau of the Census. Pp. 16—22. Ward, Lester F. Applied Sociology. New York. Wehrwein, G. S. "Appraisal of Resettlement." Vol. 19. February 1937* Pp. 190-205. Ginn and Company. 1906. Journal of Farm Economics. Witt, Lawrence W. Economic Problems of Low Income Farmers in Iowa. Ames, Iowa. Research Bulletin 307. October 1942. — end — -151- Appendix B Rate of Change Analysis This dissertation has been concerned with an analysis of differ­ ences between an experimental and control group of families measured after they had been made equivalent statistically in a pre—program year, ^his Appendix contains an additional method of determining the effect of the FHA supervised loaning program by comparing the rate of changes through the years between preloan year and time of interview­ ing, A comparison of rate of change can be accomplished with organ­ izational participation and level of living scores since comparable data are available for both the preloan and after-loan dates• Rather than involving matching of FHA with non-FHA, this analysis determines how much each FHA family has gained or lost between the pre- and postyears before comparing average amount of gain for all FHA with average gain for all non-FEA families. Organizational Participation, The experimental group with a mean organizational participation score of 11,21 in the preloan year and 15,36 at the time of the interview increased 5,15 points. The control group averaging 16*75 in the preloan year and 19,51 when interviewed had sin increase of 2,76, The t-score test yielded 1,7783 when applied to determine how often this difference between the two changes might occur by chance. This has a probability of between ,05 and ,10, which is above the satisfactory level of ,05 accepted in this analysis. Level of Living. The mean preloan level of living score for FHA families was 17,29 and at the time of interview it was 30,09, making - an average advance of 12 ,8 . 152- The non-FHA had a mean of 18,75 in the preloan year and 31.04 later which was a positive change of 12.3. The mean FHA family made an increase of only .5 points more than the other average family, hut this could be due to chance, since the t-score produced a probability of between .60 and .70. The rate of change for the FHA on organizational participation and level of living was slightly greater than for non-FHA but neither difference was large enough to be statistically significant. It should be pointed out that the starting point already achieved might have had an effect on the amount of change* The FHA had lower begin­ ning scores on both the organizational participation and level of living scales. If families with higher scores increased faster, the non-FHA would have had the advantage. FHA would have had the advantage» If the reverse were true, the The data might indicate that the loan received by the FHA made up for a lower starting score and help­ ed them maintain approximately the same rate as the others. -end-