CHAPTER V
Conclusions
Chapter III detailed each one of the statistical methods used in the study and the results obtained. Chapter IV discussed, at some length, the major findings resulting from the applied statistics. This chapter will summarise what has been learned from this investigation about language development in bilingual school children of Puerto Rican extraction. In addition, it will delineate an interpretation of the major findings in light of Cummins’ hypothesis. The educational practices of the school system under study as well as the management of these bilingual programs will also serve as background in understanding the linguistic profiles shown by the study.
Linguistic Profiles
The previous chapter has described two clearly distinguishing linguistic profiles in the context of school interlinguistic development in language minority children. The first is consonant with Cummins’ linguistic interdependence hypothesis, that second language development relies strongly on first language proficiency in this specific population of bilingual children.
This was observed to be independent of the action of intervening variables in a short term paradigm. The second involves the assumed action of intervening variables which seem to generate a developmentally regressive interlinguistic profile throughout the span of three years.
Both the linguistic interdependence hypothesis and the regressive interlinguistic development will be treated below.
Linguistic Interdependence Profile
Cummins’ linguistic interdependence hypothesis is sustained by the following statistical analyses. First, two groups of bilingual third graders were compared: high and low achievers on language tests of Spanish proficiency. CTBS Espanol test scores of 1982 served as the grouping variable. The median of these was the cut off point between high and low proficient groups. Scores on second language tests administered to these children on three consecutive years served as the dependent variables: the MAC in 1982 and the LAB (1982 version) in 1983 and 1984. Independent t tests on each one of the dependent variables were performed in order to test for significant mean differences between both groups: the high and the low proficient.
Significantly higher levels of English language competence only on the high proficient Spanish group (group scoring above the median on the 1982 CTBS Espanol) were generated, thus strongly supporting the hypothesis that Spanish language competence is in fact a strong factor accounting for English language proficiency.
Secondly, a significant correlation was observed between first and second language proficiency scores on the sample of bilingual elementary school children of Puerto Rican extraction undergoing bilingual education. The correlations between English proficiency in 1983 and Spanish Proficiency in 1982 and 1983 were .46 and .45 respectively (see Table 15, Chapter III). Both coefficients are significant, meaning in this case, that the probability of these values to occur by chance is smaller than one percent. In addition, these scores share a fairly large common Variance or r square: 21 percent (coefficient of .46) and 20 percent (coefficient of .45). This means that at least 21 percent of the variance of English proficiency scores of 1983 may be accounted for by the Variance of Spanish proficiency scores obtained in 1982 and 1983.
This profile is clearly underscored in the factor analytic study of all the variables of the study. In all, 16 independent variables were used in the Factor Analysis.
These are the following:
The factor analysis indicated which variables could be grouped together into a superordinate construct called a "factor." This included the variables which appeared to be highly correlated among each other indicating whether English scores and Spanish scores belonged to the same (bilingual proficiency), factor generated by the 75 cases and sixteen variables. An inspection of the correlation matrix revealed that the language variables (within and across languages) positively correlated with each other at significant levels. A second important observation was that language variables measured across time, although for only a short term of one year, likewise, positively and significantly correlated with each other (see table 15, Chapter III). The longitudinal dimension of linguistic interdependence, which seems to be at the crux of Cummins’ hypothesis, is shown to exist. The factor loadings which resulted from this analysis clearly showed Spanish and English academic language scores loading on one single factor (the strongest factor) reflecting a pattern of bilingual interlinguistic developmentally related skills. This specific trend shows how these children’s languages relate previous to a long term exposure to formal teaching.
Regressive Interlinguistic Developmental Profile
Descriptive statistics for the Spanish and English tests mean scores showed a downward trend on these values from one year to the next beginning with the scores obtained in 1982 and ending in 1984 (see Table 6, Chapter IV). A decrease in English proficiency levels in both groups, the high and the low Spanish proficient groups, was observed in each consecutive year. A diminishing of the range of the means of the grouping variable year by year was also seen. A similar pattern was observed for the Spanish proficiency scores.
The correlation between Spanish proficiency scores in 1982 and English proficiency scores in 1984 was .26; between Spanish 83 and English 84 was .29. Both coefficients were much lower than would be expected if both languages were to follow a long term interdependence course. Furthermore, these coefficients were unexpectedly smaller than those observed between English proficiency in 1983 and Spanish Proficiency in 1982 and 1983, .46 and .45 respectively.
Such a trend may result when school bilingualism exists under curricular goals not specifically designed to account for differential first language proficiency and its relationship to English proficiency. It may, in fact, occur when Cummins’ paradigm of language interdependence is not given priority in curricular design by allowing L1 skills to be consolidated academically for the effective development of L2 academic skills. This profile may well be descriptive of languages in contact in the context of bilingualschooling when subjected to the strains of a heterogenous mix of language skills, students and an unclear curricular philosophy. Cummins (1979) has indicated that these children’s cognitive deficiencies are not to be attributed to their basic abilities but to the way the school handles bilingual education. Fishman (1980) and Garcia (1983) claim means of the grouping variable year by year was also seen. A similar pattern was observed for the Spanish proficiency scores.
The correlation between Spanish proficiency scores in 1982 and English proficiency scores in 1984 was .26; between Spanish 83 and English 84 was .29. Both coefficients were much lower than would be expected if both languages were to follow a long term interdependence course. Furthermore, these coefficients were unexpectedly smaller than those observed between English proficiency in 1983 and Spanish Proficiency in 1982 and 1983, .46 and .45 respectively.
Such a trend may result when school bilingualism exists under curricular goals not specifically designed to account for differential first language proficiency and its relationship to English proficiency. It may, in fact, occur when Cummins’ paradigm of language interdependence is not given priority in curricular design by allowing L1 skills to be consolidated academically for the effective development of L2 academic skills. This profile may well be descriptive of languages in contact in the context of bilingual schooling when subjected to the strains of a heterogenous mix of language skills, students and an unclear curricular philosophy. Cummins (1979) has indicated that these children’s cognitive deficiencies are not to be attributed to their basic abilities but to the way the school handles bilingual education. Fishman (1980) and Garcia (1983) claim that bilingual education and curriculum should aim at a diglossic arrangement for the teaching of languages so that
these maintain functional significance. This strategy will allow languages to co-exist in an orderly fashion, thus preventing confusion and ultimately, barring the disappearance of the vernacular.
The following comments regarding bilingual education extend beyond Cummins’ linguistic interdependence hypothesis. However, they may relate considerably to the long term educational outcome involving the student population being served. Bilingual education schools are not managed or controlled by the minority language communities they serve,
but by local, state and national boards, agencies and departments controlled and operated by the white majority population. It appears necessary for the wider community of Hispanics to organize and set their own educational goals in conjunction with the Hispanic population directly engaged in the education of Hispanic children.
Furthermore it may not be until parents create their own ethnic-mother-tongue schools that long lasting and authentic support of the vernacular will ever be effected. Fishman (1980), Fishman and Milan (1980), and Garcia (1983) argue for the social organization of Hispanics as an ethnolinguistic group, the Spanish-mother-tongue school as a social institution and community oriented planning of a curriculum to teach both English and Spanish in schools controlled by Hispanics.
Sufficient economic, social and political power must be achieved by language minorities before their own school system can become instrumental in creating competitiveeducational excellence for their children (Fishman 1980;Garcia, 1983). Further evidence supporting this outcome (regressive interdependence) is shown on the results of the factor analysis, specifically, the second factor. The second of the six factors generated by the factor analysis combined two aspects of language proficiency, academic and oral. The subjects’ number of years spent in their country of origin or old country and their academic English proficiency loaded on a single factor.
This combination yielded a negative loading on years lived in the old country and positive loadings on all English proficiency variables, thus allowing for the following inference: children who lived longer on the mainland knew English better than children who lived in a
non English environment. Their having learned more English (oral language skills) possibly influenced their language performance in school, therefore, allowing these two variables, time lived on the mainland (measured indirectly) and academic language proficiency, to load on one factor.
On the other hand, Cummins’ hypothesis would predict a gradual decreasing trend in academic English proficiency for these children, assuming little support for second language
development would come from their diminished command of Spanish. The significant drop in English proficiency scores from 1982 to 1984 for the full sample of children under
study, including those who had lived longer on the mainland, confirms this hypothesis.
It supports Cummins’ assertion about the need for maintaining cognitive and academic skills in the first language in order for second language skills to evidence positive development. Longer residence on the mainland does allow for better communicative skills in English to develop, but it may also explain a lack of development of cognitive and academic skills in Spanish and English.
This specific trend additionally indicates how communicative skills in English assumed to be superior in children who have lived longer in the mainland may interact with curriculum variables to generate a negatively skewed development of English. Teachers may be misled to believe that a Puerto Rican child who speaks English without an "accent" may be as good as a native speaker of English on tasks which require elaborate thinking in that language.
This misconception may lead to decisions which possibly delay his or her linguistic progress. To deemphasize these children's need for progress in Spanish because a successful transition into English has been presumably accomplished may be an erroneous strategy applied to these children.
Paralinguistic Profiles
The ultimate linguistic developmental outcome shown in the study has generated several by-products which are not strictly speaking linguistic in nature. One is related to intelligence.
Verbal Intelligence
Verbal intelligence, as shown by the results of the Spanish WISC’s Verbal intelligence test scores, exhibited an average score of 89 IQ, eleven points below the average normed score of 100. This difference is highly significant indicating that intelligence scores for this group of children are not chance variations from the population mean, but systematic, reflecting the possible influence of an extraneous variable. The Reliability and the Standard Error of Measurement of the Verbal Scale Score of the WISC (10 and a half year olds) were .96 and 3.00 respectively.
This test’s yield concurs with the low yield also exhibited by the English language test scores of English in 1983 and 1984, both relative measures of verbal intelligence. A similar trend, although minor, was observed for the Spanish language tests administered in 1982 and
1983. These show mean scores that although not deviating from the population mean, slightly decreased from one year to the next (see Table 6, Chapter IV). On the other hand the
Performance Intelligence was l03, fourteen points higher than the Verbal intelligence and somewhat above average (100 IQ is average for the normed population).
Therefore, this profile is marked by a below average mean Verbal IQ yield, by decreasing mean language tests’ scores for three consecutive years, and an otherwise slightly above average Performance IQ on the mean scores of the Spanish WISC. As verbal intelligence (or verbal ability) has diminished throughout the first three years of schooling (excluding kindergarten), performance intelligence has seemingly maintained an average yield. This maintenance of performance Intelligence, which may be otherwise presumed to be equal or similar to verbal intelligence, may point to these children's need to compensate for the low verbal intelligence yield. Concrete intelligence, what is in fact measured in performance tests, is observed to have partially replaced abstract intelligence as the dominant coping strategy to handle the academic linguistic shortcomings.
One additional complementary piece of information is to say, pragmatic, requiring the effective use of perceptual and motor skills involving speedy responses to rapidly shifting experiences. The school ambience seems to foster the maintenance of performance skills such as attention to the environment, awareness of the social situation, social intelligence, non-verbal concept formation, and visual-motor coordination.
It would seem reasonable that verbal and related skills mighty be the weakest area of development for these children, considering the large variety of bilingual skills exhibited by the student population and the ambiguous philosophical and curricular goals for the teaching of both languages.
The second para-linguistic profile is not necessarily and by-product the the linguistic problem It relates sot the school-family relationship, and and more specifically with the school grades and the presence of he father in the house. This profile indicates that children whose father lived in stye house did better in school, in linguistic and other related activities. This pacific paradigm may indicate that the presence of the father in the house, and the consequent father-child relationship may provide sufficient security to allow the child to cope successfully with the demands of the classroom and the school, including the linguistic ones, whatever they may be.
Suggestions for Further Research
This research effort has in fact seriously questioned the effectiveness of bilingual education, not in principle, but under the specific conditions which prevailed in theNewark public school system at the time this study was performed. Cummins’ hypothesis clearly suggests the need for a maintenance bilingual education curriculum and the need for supportive attention to the status of the vernacular. Both of these elements seem to be crucial to an effective bilingual education program. However, the scarcity of basic and operational research needed to improve program quality is still the rule. Funds appropriated for this purpose are less than one—half of one percent of the spending in bilingual education (Troike, 1978).
The present investigation has looked at one specific area of bilingual education. This concerns the role of first language development on second language competence. Normal
constraints in using information about tests and other materials of students stored in board of education files does not allow for a thorough review of the abundant quantities of data found there. One important step in furthering this specific area of research is to make more use of relevant archival information which already exists. Following are research areas suggesting what can be done with this information.
Archival Data
Rigorous developmental analysis of first and second language test data on a single sample of subject throughout the span of more than five years is vital in constructing a long term profile of bilingual child development. It has been very difficult at this time to gather longitudinal data due to the nature of the population, which evidences rapid geographical mobility. However, using a larger portion of the population of bilingual students, for example, from a state or a region, may allow for a much larger sample, and therefore, more reliable results. Frequently, the fact that language tests administered to these children might change from one year to the next, or every so many years, increases the error variance of the child's true scores, also weakening the reliability of the results. If language tests change, internal and external validity criteria for item analysis change; so do items and norms.
Along these lines, it would be very useful to compare students who have failed. Although it is known that a large number of factors contribute to either outcome, a factor
analytic study, including, parent and child demographic data, may generate critical information about the importance of the student's language resources for school success.
School success may be measured by using more than one criterion. For example, success and failure might be categorized as post high school leadership, employment or post-secondary education, other than just graduation.
Statistics about the proportion of students who succeed versus those who do not among those with a poorly developed first language and those with a highly developed first language will yield vital information related to the role of the first language on post high school education success.
It may be hypothesized that a large number of students deserting the school system may have had lagging first language academic skills in school from the beginning delaying school progress, therefore, lowering these students’ motivation to continue in school.
The correlation between first and second language development may be tested by analyzing individual items of language tests. How specific errors of syntax, morphology or semantics are carried over from the first to the second language. This way, a correlation of error similarities from one language to the other on one subject and on a group may be computed. The hypothesized construct is that children will correlate semantic proficiency, syntactical ability and morphological facility across languages.
Evaluation Research
A second area of research should necessarily involve program evaluation. The first point is the policy supporting these bilingual education programs, the objectives and goals these programs have set out to accomplish and the methods used to evaluate their effectiveness. Already available information on policy matters related to these programs should be used.
In addition, personnel involved in policy making decisions in each school system may be addressed to provide policy related information; so may program directors, principals,
teachers, students and parents involved in bilingual education. This kind of research, as for most evaluation research, must be initiated by those who have decisions to make and seek for an evaluation to find answers. The program being evaluated may cover the nation, a region, a state, city, neighborhood or be limited to one specific site, a classroom.
Such an endeavor must include an evaluation of current beliefs held by bilingual populations to whom these services are directed (parents of bilingual education children) about the nature of these bilingual education programs. This kind of research may provide vital information about what consumers expect of the program and what values are attached to such services. Program effectiveness may be highly dependent upon such expectations. It is hypothesized that the parents of children undergoing bilingual education may lack sufficient information about the role played by first language on second language development. This may also be true of high level personnel who must make decisions about bilingual education programs (modifying programs, etc.), and need to know more about the effectiveness of the program.
One last issue regarding the bilingual community served by bilingual schools pertains to its views of societal bilingualism. Are Hispanics interested in maintaining Spanish as a functionally active language fostering ethnic unity and traditions? If so, what kind of diglossic arrangement of language codes would Hispanics prefer for their own community? How seriously do Hispanics themselves take Spanish literacy? Once they take their own languages seriously will these languages be utilized also to assist the mastery of school English? Maximal answers to these questions seem to be necessary if bilingual schools belonging truly to the ethnic communities they serve are to seriously develop advanced formal skills in the language of those communities.
Personality
A third area of research is personality and language development. To be specific, it was observed that children reported in the present study to live without their father (in the house) did not fare as well in school (including verbal skills) as children registering data on both parents’ presence in the house. Personality variables, then, seem to play a role in language development, assuming that children need paternal closeness, on one hand, or need to be free from serious parental conflict, on the other hand, to evidence successful progression into school, for at least the first few years of schooling.
Intelligence
A fourth area which this investigation opens for further research is intelligence. The children of this study demonstrated a tendency to balance out performance and verbal intelligence. As verbal intelligence seemed to have weakened throughout the several years of schooling, performance intelligence seemed to have become dominant. These children looked very keen when handling tasks which required concrete intelligence and day to day handling of basic interpersonal skills. They looked and acted street-wise. This seem to be an adaptive and compensatory mechanism. Further evidence should be sought on this area. A longitudinal study of these children’s intelligence should be designed, beginning infirst grade or kindergarten. Verbal and performance intelligence should be sampled in both languages.
This profile is clearly underscored in the factor analytic study of all the variables of the study. In all, 16 independent variables were used in the Factor Analysis.
These are the following:
The factor analysis indicated which variables could be grouped together into a superordinate construct called a "factor." This included the variables which appeared to be highly correlated among each other indicating whether English scores and Spanish scores belonged to the same (bilingual proficiency), factor generated by the 75 cases and sixteen variables. An inspection of the correlation matrix revealed that the language variables (within and across languages) positively correlated with each other at significant levels. A second important observation was that language variables measured across time, although for only a short term of one year, likewise, positively and significantly correlated with each other (see table 15, Chapter III). The longitudinal dimension of linguistic interdependence, which seems to be at the crux of Cummins’ hypothesis, is shown to exist. The factor loadings which resulted from this analysis clearly showed Spanish and English academic language scores loading on one single factor (the strongest factor) reflecting a pattern of bilingual interlinguistic developmentally related skills. This specific trend shows how these children’s languages relate previous to a long term exposure to formal teaching.
Regressive Interlinguistic Developmental Profile
Descriptive statistics for the Spanish and English tests mean scores showed a downward trend on these values from one year to the next beginning with the scores obtained in 1982 and ending in 1984 (see Table 6, Chapter IV). A decrease in English proficiency levels in both groups, the high and the low Spanish proficient groups, was observed in each consecutive year. A diminishing of the range of the means of the grouping variable year by year was also seen. A similar pattern was observed for the Spanish proficiency scores.
The correlation between Spanish proficiency scores in 1982 and English proficiency scores in 1984 was .26; between Spanish 83 and English 84 was .29. Both coefficients were much lower than would be expected if both languages were to follow a long term interdependence course. Furthermore, these coefficients were unexpectedly smaller than those observed between English proficiency in 1983 and Spanish Proficiency in 1982 and 1983, .46 and .45 respectively.
Such a trend may result when school bilingualism exists under curricular goals not specifically designed to account for differential first language proficiency and its relationship to English proficiency. It may, in fact, occur when Cummins’ paradigm of language interdependence is not given priority in curricular design by allowing L1 skills to be consolidated academically for the effective development of L2 academic skills. This profile may well be descriptive of languages in contact in the context of bilingualschooling when subjected to the strains of a heterogenous mix of language skills, students and an unclear curricular philosophy. Cummins (1979) has indicated that these children’s cognitive deficiencies are not to be attributed to their basic abilities but to the way the school handles bilingual education. Fishman (1980) and Garcia (1983) claim means of the grouping variable year by year was also seen. A similar pattern was observed for the Spanish proficiency scores.
The correlation between Spanish proficiency scores in 1982 and English proficiency scores in 1984 was .26; between Spanish 83 and English 84 was .29. Both coefficients were much lower than would be expected if both languages were to follow a long term interdependence course. Furthermore, these coefficients were unexpectedly smaller than those observed between English proficiency in 1983 and Spanish Proficiency in 1982 and 1983, .46 and .45 respectively.
Such a trend may result when school bilingualism exists under curricular goals not specifically designed to account for differential first language proficiency and its relationship to English proficiency. It may, in fact, occur when Cummins’ paradigm of language interdependence is not given priority in curricular design by allowing L1 skills to be consolidated academically for the effective development of L2 academic skills. This profile may well be descriptive of languages in contact in the context of bilingual schooling when subjected to the strains of a heterogenous mix of language skills, students and an unclear curricular philosophy. Cummins (1979) has indicated that these children’s cognitive deficiencies are not to be attributed to their basic abilities but to the way the school handles bilingual education. Fishman (1980) and Garcia (1983) claim that bilingual education and curriculum should aim at a diglossic arrangement for the teaching of languages so that
these maintain functional significance. This strategy will allow languages to co-exist in an orderly fashion, thus preventing confusion and ultimately, barring the disappearance of the vernacular.
but by local, state and national boards, agencies and departments controlled and operated by the white majority population. It appears necessary for the wider community of Hispanics to organize and set their own educational goals in conjunction with the Hispanic population directly engaged in the education of Hispanic children.
Furthermore it may not be until parents create their own ethnic-mother-tongue schools that long lasting and authentic support of the vernacular will ever be effected. Fishman (1980), Fishman and Milan (1980), and Garcia (1983) argue for the social organization of Hispanics as an ethnolinguistic group, the Spanish-mother-tongue school as a social institution and community oriented planning of a curriculum to teach both English and Spanish in schools controlled by Hispanics.
Sufficient economic, social and political power must be achieved by language minorities before their own school system can become instrumental in creating competitiveeducational excellence for their children (Fishman 1980;Garcia, 1983). Further evidence supporting this outcome (regressive interdependence) is shown on the results of the factor analysis, specifically, the second factor. The second of the six factors generated by the factor analysis combined two aspects of language proficiency, academic and oral. The subjects’ number of years spent in their country of origin or old country and their academic English proficiency loaded on a single factor.
non English environment. Their having learned more English (oral language skills) possibly influenced their language performance in school, therefore, allowing these two variables, time lived on the mainland (measured indirectly) and academic language proficiency, to load on one factor.
On the other hand, Cummins’ hypothesis would predict a gradual decreasing trend in academic English proficiency for these children, assuming little support for second language
development would come from their diminished command of Spanish. The significant drop in English proficiency scores from 1982 to 1984 for the full sample of children under
study, including those who had lived longer on the mainland, confirms this hypothesis.
It supports Cummins’ assertion about the need for maintaining cognitive and academic skills in the first language in order for second language skills to evidence positive development. Longer residence on the mainland does allow for better communicative skills in English to develop, but it may also explain a lack of development of cognitive and academic skills in Spanish and English.
This specific trend additionally indicates how communicative skills in English assumed to be superior in children who have lived longer in the mainland may interact with curriculum variables to generate a negatively skewed development of English. Teachers may be misled to believe that a Puerto Rican child who speaks English without an "accent" may be as good as a native speaker of English on tasks which require elaborate thinking in that language.
Paralinguistic Profiles
The ultimate linguistic developmental outcome shown in the study has generated several by-products which are not strictly speaking linguistic in nature. One is related to intelligence.
Verbal Intelligence
Verbal intelligence, as shown by the results of the Spanish WISC’s Verbal intelligence test scores, exhibited an average score of 89 IQ, eleven points below the average normed score of 100. This difference is highly significant indicating that intelligence scores for this group of children are not chance variations from the population mean, but systematic, reflecting the possible influence of an extraneous variable. The Reliability and the Standard Error of Measurement of the Verbal Scale Score of the WISC (10 and a half year olds) were .96 and 3.00 respectively.
This test’s yield concurs with the low yield also exhibited by the English language test scores of English in 1983 and 1984, both relative measures of verbal intelligence. A similar trend, although minor, was observed for the Spanish language tests administered in 1982 and
1983. These show mean scores that although not deviating from the population mean, slightly decreased from one year to the next (see Table 6, Chapter IV). On the other hand the
Performance Intelligence was l03, fourteen points higher than the Verbal intelligence and somewhat above average (100 IQ is average for the normed population).
One additional complementary piece of information is to say, pragmatic, requiring the effective use of perceptual and motor skills involving speedy responses to rapidly shifting experiences. The school ambience seems to foster the maintenance of performance skills such as attention to the environment, awareness of the social situation, social intelligence, non-verbal concept formation, and visual-motor coordination.
It would seem reasonable that verbal and related skills mighty be the weakest area of development for these children, considering the large variety of bilingual skills exhibited by the student population and the ambiguous philosophical and curricular goals for the teaching of both languages.
The second para-linguistic profile is not necessarily and by-product the the linguistic problem It relates sot the school-family relationship, and and more specifically with the school grades and the presence of he father in the house. This profile indicates that children whose father lived in stye house did better in school, in linguistic and other related activities. This pacific paradigm may indicate that the presence of the father in the house, and the consequent father-child relationship may provide sufficient security to allow the child to cope successfully with the demands of the classroom and the school, including the linguistic ones, whatever they may be.
Suggestions for Further Research
This research effort has in fact seriously questioned the effectiveness of bilingual education, not in principle, but under the specific conditions which prevailed in theNewark public school system at the time this study was performed. Cummins’ hypothesis clearly suggests the need for a maintenance bilingual education curriculum and the need for supportive attention to the status of the vernacular. Both of these elements seem to be crucial to an effective bilingual education program. However, the scarcity of basic and operational research needed to improve program quality is still the rule. Funds appropriated for this purpose are less than one—half of one percent of the spending in bilingual education (Troike, 1978).
The present investigation has looked at one specific area of bilingual education. This concerns the role of first language development on second language competence. Normal
constraints in using information about tests and other materials of students stored in board of education files does not allow for a thorough review of the abundant quantities of data found there. One important step in furthering this specific area of research is to make more use of relevant archival information which already exists. Following are research areas suggesting what can be done with this information.
Archival Data
Rigorous developmental analysis of first and second language test data on a single sample of subject throughout the span of more than five years is vital in constructing a long term profile of bilingual child development. It has been very difficult at this time to gather longitudinal data due to the nature of the population, which evidences rapid geographical mobility. However, using a larger portion of the population of bilingual students, for example, from a state or a region, may allow for a much larger sample, and therefore, more reliable results. Frequently, the fact that language tests administered to these children might change from one year to the next, or every so many years, increases the error variance of the child's true scores, also weakening the reliability of the results. If language tests change, internal and external validity criteria for item analysis change; so do items and norms.
Along these lines, it would be very useful to compare students who have failed. Although it is known that a large number of factors contribute to either outcome, a factor
analytic study, including, parent and child demographic data, may generate critical information about the importance of the student's language resources for school success.
School success may be measured by using more than one criterion. For example, success and failure might be categorized as post high school leadership, employment or post-secondary education, other than just graduation.
Statistics about the proportion of students who succeed versus those who do not among those with a poorly developed first language and those with a highly developed first language will yield vital information related to the role of the first language on post high school education success.
It may be hypothesized that a large number of students deserting the school system may have had lagging first language academic skills in school from the beginning delaying school progress, therefore, lowering these students’ motivation to continue in school.
The correlation between first and second language development may be tested by analyzing individual items of language tests. How specific errors of syntax, morphology or semantics are carried over from the first to the second language. This way, a correlation of error similarities from one language to the other on one subject and on a group may be computed. The hypothesized construct is that children will correlate semantic proficiency, syntactical ability and morphological facility across languages.
Evaluation Research
A second area of research should necessarily involve program evaluation. The first point is the policy supporting these bilingual education programs, the objectives and goals these programs have set out to accomplish and the methods used to evaluate their effectiveness. Already available information on policy matters related to these programs should be used.
In addition, personnel involved in policy making decisions in each school system may be addressed to provide policy related information; so may program directors, principals,
teachers, students and parents involved in bilingual education. This kind of research, as for most evaluation research, must be initiated by those who have decisions to make and seek for an evaluation to find answers. The program being evaluated may cover the nation, a region, a state, city, neighborhood or be limited to one specific site, a classroom.
Such an endeavor must include an evaluation of current beliefs held by bilingual populations to whom these services are directed (parents of bilingual education children) about the nature of these bilingual education programs. This kind of research may provide vital information about what consumers expect of the program and what values are attached to such services. Program effectiveness may be highly dependent upon such expectations. It is hypothesized that the parents of children undergoing bilingual education may lack sufficient information about the role played by first language on second language development. This may also be true of high level personnel who must make decisions about bilingual education programs (modifying programs, etc.), and need to know more about the effectiveness of the program.
One last issue regarding the bilingual community served by bilingual schools pertains to its views of societal bilingualism. Are Hispanics interested in maintaining Spanish as a functionally active language fostering ethnic unity and traditions? If so, what kind of diglossic arrangement of language codes would Hispanics prefer for their own community? How seriously do Hispanics themselves take Spanish literacy? Once they take their own languages seriously will these languages be utilized also to assist the mastery of school English? Maximal answers to these questions seem to be necessary if bilingual schools belonging truly to the ethnic communities they serve are to seriously develop advanced formal skills in the language of those communities.
Personality
A third area of research is personality and language development. To be specific, it was observed that children reported in the present study to live without their father (in the house) did not fare as well in school (including verbal skills) as children registering data on both parents’ presence in the house. Personality variables, then, seem to play a role in language development, assuming that children need paternal closeness, on one hand, or need to be free from serious parental conflict, on the other hand, to evidence successful progression into school, for at least the first few years of schooling.
Intelligence
A fourth area which this investigation opens for further research is intelligence. The children of this study demonstrated a tendency to balance out performance and verbal intelligence. As verbal intelligence seemed to have weakened throughout the several years of schooling, performance intelligence seemed to have become dominant. These children looked very keen when handling tasks which required concrete intelligence and day to day handling of basic interpersonal skills. They looked and acted street-wise. This seem to be an adaptive and compensatory mechanism. Further evidence should be sought on this area. A longitudinal study of these children’s intelligence should be designed, beginning infirst grade or kindergarten. Verbal and performance intelligence should be sampled in both languages.
