Appendix 2

It is impossible to present a single and unified picture of the adolescent girl in Samoa and at the same time to answer most satisfactorily the various kinds of questions which such a study will be expected to answer. For the ethnologist in search of data upon the usages and rites connected with adolescence it is necessary to include descriptions of customs which have fallen into partial decay under the impact of western propaganda and foreign example. Traditional observances and attitudes are also important in the study of the adolescent girl in present-day Samoa because they still form a large part of the thought pattern of her parents, even if they are no longer given concrete expression in the girl’s cultural life. But this double necessity of describing not only the present environment and the girl’s reaction to it, but also of interpolating occasionally some description of the more rigid cultural milieu of her mother’s girlhood, mars to some extent the unity of the study.

It is impossible to give a single, unified picture of the adolescent girl in Samoa while fully answering every expected question. An ethnologist seeking information about adolescent customs and rites needs descriptions of customs that have partly faded. These customs have declined under the influence of Western propaganda and foreign examples. Traditional practices and attitudes also matter when studying the adolescent girl in present-day Samoa. They still shape much of her parents’ thinking, even if they no longer appear in the girl’s cultural life. But the study must describe both the girl’s current surroundings and her response to them. It must also occasionally describe the more rigid cultural setting of her mother’s girlhood. This need to describe both settings makes the study somewhat less unified.

The detailed observations were all made upon a group of girls living in three practically contiguous villages on one coast of the island of Taū. The data upon the ceremonial usages surrounding birth, adolescence and marriage were gathered from all of the seven villages in the Manu’a Archipelago.

All detailed observations focused on a group of girls living in three villages that nearly adjoined one another. The villages were on one coast of the island of Taū. Information on ceremonies surrounding birth, adolescence, and marriage came from all seven villages in the Manu’a Archipelago.

The method of approach is based upon the assumption that a detailed intensive investigation will be of more value than a more diffused and general study based upon a less accurate knowledge of a greater number of individuals. Dr. Van Waters’ study of The Adolescent Girl Among Primitive Peoples has exhausted the possibilities of an investigation based upon the merely external observations of the ethnologist who is giving a standardised description of a primitive culture. We have a huge mass of general descriptive material without the detailed observations and the individual cases in the light of which it would be possible to interpret it.

This approach assumes that a detailed, intensive investigation will be more valuable than a broader, more general study. The broader study would cover more people but rely on less accurate knowledge of each person. Dr. Van Waters’ study of The Adolescent Girl Among Primitive Peoples has exhausted the possibilities of research based only on ethnologists’ external observations. These observations come from ethnologists providing standardized descriptions of a primitive culture. We have a vast amount of general descriptive material. But we lack the detailed observations and individual cases that would allow us to interpret that material.

The writer therefore chose to work in one small locality, in a group numbering only six hundred people, and spend six months accumulating an intimate and detailed knowledge of all the adolescent girls in this community. As there were only sixty-eight girls between the ages of nine and twenty, quantitative statements are practically valueless for obvious reasons: the probable error of the group is too large; the age classes are too small, etc. The only point at which quantitative statements can have any relevance is in regard to the variability within the group, as the smaller the variability within the sample, the greater the general validity of the results.

The writer therefore chose to work in one small community of only six hundred people. She spent six months gaining a close, detailed understanding of all the adolescent girls in the community. There were only sixty-eight girls between the ages of nine and twenty. With so few girls, statistical statements have practically no value, for obvious reasons. The group's probable margin of error is too large, the age groups are too small, and so on. Statistical statements can be relevant only when describing variation within the group. This is because the less variation there is within the sample, the more broadly valid the results are.

Furthermore, the type of data which we needed is not of the sort which lends itself readily to quantitative treatment. The reaction of the girl to her stepmother, to relatives acting as foster parents, to her younger sister, or to her older brother,—these are incommensurable in quantitative terms. As the physician and the psychiatrist have found it necessary to describe each case separately and to use their cases as illumination of a thesis rather than as irrefutable proof such as it is possible to adduce in the physical sciences, so the student of the more intangible and psychological aspects of human behaviour is forced to illuminate rather than demonstrate a thesis. The composition of the background against which the girl acts can be described in accurate and general terms, but her reactions are a function of her own personality and cannot be described without reference to it. The generalisations are based upon a careful and detailed observation of a small group of subjects. These results will be illuminated and illustrated by case histories.

Furthermore, the kind of data we needed does not lend itself readily to analysis in numbers. A girl’s reactions to her stepmother, relatives acting as foster parents, younger sister, or older brother cannot be compared numerically. Doctors and psychiatrists have found it necessary to describe each case separately. They use their cases to shed light on an argument rather than offer indisputable proof, as the physical sciences can. Similarly, researchers studying the less tangible, psychological aspects of human behavior must clarify an argument rather than prove it. The background in which a girl acts can be described accurately and in general terms. But her reactions depend on her own personality and cannot be described without reference to her personality. The generalizations are based on careful, detailed observation of a small group of subjects. Case histories will clarify and illustrate these results.

The conclusions are also all subject to the limitation of the personal equation. They are the judgments of one individual upon a mass of data, many of the most significant aspects of which can, by their very nature, be known only to herself. This was inevitable and it can only be claimed in extenuation that as the personal equation was held absolutely constant, the different parts of the data are strictly commensurable. The judgment on the reaction of Lola to her uncle and of Sona to her cousin are made on exactly the same basis.

All the conclusions are also limited by the researcher's personal perspective. The conclusions are one person's judgments about a large body of data. Many of the data's most significant aspects can, by their very nature, be known only to the researcher herself. This limitation was inevitable. The only defense is that, because the researcher's perspective remained entirely constant, the different parts of the data are directly comparable. Lola's reaction to her uncle and Sona's reaction to her cousin are judged on exactly the same basis.

Another methodological device which possibly needs explanation is the substitution of a cross sectional study for a linear one. Twenty-eight children who as yet showed no signs of puberty, fourteen children who would probably mature within the next year or year and a half, and twenty-five girls who had passed puberty within the last four years but were not yet classed by the community as adults, were studied in detail. Less intensive observations were also made upon the very little children and the young married women. This method of taking cross sections, samples of individuals at different periods of physical development, and arguing that a group in an earlier stage will later show the characteristics which appear in another group at a later stage, is, of course, inferior to a linear study in which the same group is under observation for a number of years. A very large number of cases has usually been the only acceptable defence of such a procedure. The number of cases included in this investigation, while very small in comparison with the numbers mustered by any student of American children, is nevertheless a fair-sized sample in terms of the very small population of Samoa (a rough eight thousand in all four islands of American Samoa) and because the only selection was geographical. It may further be argued that the almost drastic character of the conclusions, the exceedingly few exceptions which need to be made, further validate the size of the sample. The adoption of the cross section method was, of course, a matter of expediency, but the results when carefully derived from a fair sample, may be fairly compared with the results obtained by using the linear method, when the same subjects are under observation over a period of years. This is true when the conclusions to be drawn are general and not individual. For the purposes of psychological theory, it is sufficient to know that children in a certain society walk, on the average, at twelve months, and talk, on the average, at fifteen months. For the purposes of the diagnostician, it is necessary to know that John walked at eighteen months and did not talk until twenty months. So, for general theoretical purposes, it is enough to state that little girls just past puberty develop a shyness and lack of self-possession in the presence of boys, but if we are to understand the delinquency of Mala, it is necessary to know that she prefers the company of boys to that of girls and has done so for several years.

Another research method that may need explanation is using a cross-sectional study instead of a longitudinal study. A cross-sectional study compares different groups at one time. A longitudinal study follows the same group over several years. Twenty-eight children who showed no signs of puberty were studied in detail. So were fourteen children who would probably reach puberty within the next year or year and a half. The detailed study also included twenty-five girls who had passed puberty within the last four years. The community did not yet consider these girls adults. Very young children and young married women were also observed, but less intensively. The cross-sectional method takes samples of people at different stages of physical development. It assumes that an earlier-stage group will later show the characteristics found in a later-stage group. This method is, of course, less reliable than following the same group for several years. Usually, only a very large number of cases has been accepted as justification for using the cross-sectional method. This study included very few cases compared with the numbers studied by any researcher working with American children. Nevertheless, the sample was fairly large given Samoa’s very small population and because participants were selected only by location. The total population of all four islands of American Samoa was roughly eight thousand. The striking nature of the conclusions may offer further support for the sample’s size, given the extremely few exceptions needed. The cross-sectional method was, of course, chosen for practical reasons. But results carefully drawn from a fair sample can reasonably be compared with results from following the same subjects for years. This comparison holds when the conclusions concern groups in general, rather than individuals. For psychological theory, knowing the average ages when children in a particular society begin walking and talking is enough. For example, children may walk at twelve months on average and talk at fifteen months on average. For someone making a diagnosis, the necessary information is that John walked at eighteen months and did not talk until twenty months. So, for general theoretical purposes, a statement about girls just past puberty is enough. Such girls become shy and less at ease around boys. But to understand Mala’s delinquency, we need to know that she prefers boys’ company to girls’ company. We also need to know that she has preferred boys’ company for several years.

The description of the cultural background was obtained in orthodox fashion, first through interviews with carefully chosen informants, followed by checking up their statements with other informants and by the use of many examples and test cases. With a few unimportant exceptions this material was obtained in the Samoan language and not through the medium of interpreters. All of the work with individuals was done in the native language, as there were no young people on the island who spoke English.

The cultural background was described using standard methods, beginning with interviews with carefully chosen informants. Their statements were then checked with other informants and through many examples and test cases. With a few minor exceptions, this information was gathered in Samoan rather than through interpreters. All work with individuals was done in Samoan because no young people on the island spoke English.

Although a knowledge of the entire culture was essential for the accurate evaluation of any particular individual’s behaviour, a detailed description will be given only of those aspects of the culture which are immediately relevant to the problem of the adolescent girl. For example, if I observe Pele refuse point blank to carry a message to the house of a relative, it is important to know whether she is actuated by stubbornness, dislike of the relative, fear of the dark, or fear of the ghost which lives near by and has a habit of jumping on people’s backs. But to the reader a detailed exposition of the names and habits of all the local ghost population would be of little assistance in the appreciation of the main problem. So all descriptions of the culture which are not immediately relevant are omitted from the discussion but were not omitted from the original investigation. Their irrelevancy has, therefore, been definitely ascertained.

Although understanding the entire culture was essential to accurately assess any individual’s behaviour, I will describe only certain aspects in detail. These are the aspects directly relevant to the question of adolescent girls. For example, suppose I observe Pele flatly refuse to carry a message to a relative’s house. It is important to know whether her refusal comes from stubbornness, dislike of the relative, or fear of the dark. Or perhaps she fears the ghost that lives nearby and has a habit of jumping on people’s backs. But a detailed account of all the local ghosts’ names and habits would do little to help readers understand the main question. So descriptions of all cultural aspects not directly relevant are left out of this discussion. But those aspects were not left out of the original investigation. Their lack of relevance has therefore been firmly established.

The knowledge of the general cultural pattern was supplemented by a detailed study of the social structure of the three villages under consideration. Each household was analysed from the standpoint of rank, wealth, location, contiguity to other households, relationship to other households, and the age, sex, relationship, marital status, number of children, former residence, etc., of each individual in the household. This material furnished a general descriptive basis for a further and more careful analysis of the households of the subjects, and also provided a check on the origin of feuds or alliances between individuals, the use of relationship terms, etc. Each child was thus studied against a background which was known in detail.

Knowledge of the general cultural pattern was supplemented by a detailed study of social structure in the three villages studied. Each household was examined for rank, wealth, location, closeness to other households, and relationships with other households. The study also recorded each household member’s age, sex, family relationships, marital status, number of children, previous residence, and other details. This information provided a general description for a further, more careful analysis of the study participants’ households. It also helped check the origins of feuds or alliances between individuals, the use of relationship terms, and similar matters. Each child was therefore studied within a background known in detail.

A further mass of detailed information was obtained about the subjects: approximate age (actual age can never be determined in Samoa), order of birth, numbers of brothers and sisters, who were older and younger than the subject, number of marriages of each parent, patrilocal and matrilocal residence, years spent in the pastor’s school and in the government school and achievement there, whether the child had ever been out of the village or off the island, sex experience, etc. The children were also given a makeshift intelligence test, colour-naming, rote memory, opposites, substitution, ball and field, and picture interpretation. These tests were all given in Samoan; standardisation was, of course, impossible and ages were known only relatively; they were mainly useful in assisting me in placing the child within her group, and have no value for comparative purposes. The results of the tests did indicate, however, a very low variability within the group. The tests were supplemented by a questionnaire which was not administered formally but filled in by random questioning from time to time. This questionnaire gave a measure of their industrial knowledge, the extent to which they participated in the lore of the community, of the degree to which they had absorbed European teaching in matters like telling time, reading the calendar, and also of the extent to which they had participated in or witnessed scenes of death, birth, miscarriage, etc.

I also gathered a large amount of detailed information about the subjects. This included approximate age (actual age can never be determined in Samoa), birth order, and numbers of brothers and sisters. I recorded which siblings were older or younger than the subject and how many times each parent had married. I noted whether families lived with the husband’s or the wife’s family. I recorded years spent in the pastor’s school and the government school, along with achievement at each. Other information included whether the child had ever left the village or the island, sexual experience, and so on. The children also took an improvised intelligence test. The tasks included naming colours, memorising by repetition, identifying opposites, substitution, ball and field, and interpreting pictures. All these tests were given in Samoan. Standardising the tests was, of course, impossible, and ages were known only in relation to one another. The tests mainly helped me place each child within her group and have no value for comparisons with other groups. The results did indicate, however, very little variation within the group. The tests were supplemented by a questionnaire. Rather than administering the questionnaire formally, I filled it in through occasional, unsystematic questioning. The questionnaire measured the children’s knowledge of practical work and how much they shared in the community’s traditional knowledge. It also measured how much European teaching they had absorbed in matters such as telling time and reading the calendar. It also measured how much they had taken part in or witnessed scenes of death, birth, miscarriage, and so on.

But this quantitative data represents the barest skeleton of the material which was gathered through months of observation of the individuals and of groups, alone, in their households, and at play. From these observations, the bulk of the conclusions are drawn concerning the attitudes of the children towards their families and towards each other, their religious interests or the lack of them, and the details of their sex lives. This information cannot be reduced to tables or to statistical statements. Naturally in many cases it was not as full as in others. In some cases it was necessary to pursue a more extensive enquiry in order to understand some baffling aspect of the child’s behaviour. In all cases the investigation was pursued until I felt that I understood the girl’s motivation and the degree to which her family group and affiliation in her age group explained her attitudes.

But these numbers give only the barest outline of the material gathered over months of observation. I observed individuals and groups, alone, in their households, and at play. Most conclusions about the children’s attitudes toward their families and one another come from these observations. So do conclusions about their religious interests, or lack of interest, and the details of their sex lives. This information cannot be reduced to tables or statistical statements. Naturally, the information was fuller in some cases than in others. In some cases, I had to investigate further to understand a puzzling aspect of a child’s behavior. In all cases, I continued until I felt I understood the girl’s motives. I also needed to understand how much her family and her ties within her age group explained her attitudes.

The existence of the pastor’s boarding-school for girls past puberty provided me with a rough control group. These girls were so severely watched that heterosexual activities were impossible; they were grouped together with other girls of the same age regardless of relationship; they lived a more ordered and regular life than the girls who remained in their households. The ways in which they differed from other girls of the same age and more resembled European girls of the same age follow with surprising accuracy the lines suggested by the specific differences in environment. However, as they lived part of the time at home, the environmental break was not complete and their value as a control group is strictly limited.

The pastor’s boarding school for girls past puberty gave me a rough control group. These girls were watched so closely that sexual activity with the opposite sex was impossible. They were grouped with other girls their age, regardless of family relationships. They lived more orderly and regular lives than girls who stayed in their households. The boarding-school girls differed from other girls their age and more closely resembled European girls their age. These differences matched, with surprising accuracy, what the specific differences in their environments would suggest. However, the boarding-school girls spent part of their time at home, so the change in environment was not complete. Their value as a control group is therefore strictly limited.