Jackson, N. & Kearney, J.
Talent Development III, pp. 203-217
Gifted Psychology Press
This book chapter excerpted from Talent Development III. Authors Nancy Jackson and Julie Kearney look at the case histories of two precocious readers. The chapter reports on achievement scores for 64 others and relates them to a model of development that indicates that a) precocious reading predicts positive performance in later years, b) continuity may be heterotypic, and c) similar achievements may come from highly intelligent children who are not precocious readers.
Case histories are reported for two young adults who were gifted performers during their preschool years. We also report the 4th-, 5th-, and 6th-grade achievement test performances of 64 children who had been identified as precocious readers at the end of kindergarten. Taken together with other findings in the literature, the present results are consistent with a model of the development of gifted performances in childhood in which (a) precocious reading predicts good, but not necessarily excellent, reading and language arts performance in later years; (b) continuity may be heterotypic, with precocious readers later showing aptitude for learning in other domains based on closed symbol systems, such as mathematics; and (c) highly intelligent children who are not precocious readers may show intellectual achievements similar to those who do read early.
Achievement of Precocious Readers in Middle Childhood and Young Adulthood
Developmental psychologists have a special reason not to mind growing older; the older we get, the better our opportunities to do long-term longitudinal studies! By 1995, some of the young children that the first author and her colleagues studied in the late 1970s were young adults, and the following discussion of what happens to precocious readers in later years is framed by accounts of what two of these children were doing as their adult lives began to take shape. The first case history suggests that precocious reading may predict a later-developing talent in mathematics. This hypothesis was tested in a large sample study of the elementary school achievement of a later generation of precocious readers who were studied first as post kindergarteners in 1987.
Taken together, case studies, other literature, and the longitudinal findings that follow are consistent with a conceptual model or the development of gifted performance in childhood that we will present. However, answers to questions that parents and teachers often ask about precocious readers, such as, "Are precocious readers gifted?" or "What will they be like later on?" are neither simple or clearcut.
Consider the case of a young man who participated in the Seattle Project longitudinal study in the late 1970's, beginning when he was four years old. Highlights of his case history are summarized in Table 1.
TABLE 1. Jeff: A Precocious Reader (Born 1972).*
As a preschooler, this young man, whom we have called Jeff, was a very precocious reader. By the time he came to the Seattle research group at age 4 years, 9 months, he was already reading quite well, so his record doesn't include much information about his movement into reading. His mother's reports suggest that he learned letters and numbers before age 2 and was reading fairly well at least by age 3. We gave him an informal reading test when he was four, and he was easily able to read sentences such as, "Reading is one of my favorite things to do."
The summary of his scores from the Peabody Individual Achievement Test, or PIAT (Dunn & Markwardt, 1970), shows that Jeff's ability to pronounce words (Reading Recognition score) was at the sixth grade level by the time he was about 5 1/2 years old. Over the next year, his Reading Recognition scores stayed about the same (the apparent up and down reflects a very small difference in items correct), and his ability to comprehend what he was reading caught up with his word identification skill. Although Jeff's tested achievement in mathematics was above average for his age, it did not then come close to his precocious reading achievement.
On two separate occasions a year apart, Jeff earned extremely high Stanford-Binet IQs (Terman & Merrill, 1973) and also performed exceptionally well on two tests that emphasize spatial-perceptual reasoning to some extent: the WISC-R Block Design and Mazes (Wechsler, 1974). These scores suggest that his high intelligence was not just in the verbal realm emphasized by the Stanford-Binet of that era.
Jeff's mother and preschool teacher reported many instances of his enthusiastic, intellectually advanced, and creative work with graphic arts materials, puzzles, games, and music. Reading those descriptions more than fifteen years later, one gets the impression of a child who was delightful and challenging to work with. For example, his teacher reported that at age 4 years, Jeff's imaginary play was complicated and original. "One day he and another boy built a house of hollow blocks complete with bedrooms, 2 stories and a bath. He then urinated in the bathroom. When asked why he did it, he said, I did it for a joke, but now I don't think it's so funny.
Not surprisingly, Jeff did well in an accelerated public school program designed for academically advanced students and went on to a highly selective undergraduate program. At last report, he was working happily in a challenging, sophisticated computer programming job. He also had maintained strong avocational interests in studying a second language that has a writing system different from English and in participating in music and community theater activities. Jeff's adult interest in drama recalls his preschool venture into performance art, but his adult efforts were much more conventional.
Although Jeff's early interests and achievement did not directly suggest that he might develop talent and interest in a mathematical, logical field such as computer programming, the model presented later in this paper suggests that his career choice might be a likely one for precocious readers. His adult interests in a second language and in music also are consistent with the argument we will make that precocious readers often are children with a special affinity for mastering new symbol systems.
Jeff's history may sound familiar to readers who know the case study literature on child prodigies and children with extremely high IQs. He has much in common with the children described by Miraca Gross (1992), David Feldman (1986), and others. In this literature, and in retrospective reports of the early accomplishments of children who later do well in the SAT talent searches (VanTassel-Baska, 1989), one is struck by how often children who have demonstrated advanced talent in mathematics, computer science, and other scientific and technical disciplines are described as having been precocious readers.
Late Elementary School Achievement Levels of Precocious Readers
Why should so many prodigies and high achievers in math and science have histories as precocious readers? Is it just that both kinds of accomplishments are associated with high general or verbal intelligence, or with growing up in an intellectually stimulating environment? Or is there some more specific link? A first step toward answering this question would be to establish the prospective probability that a child identified as a precocious reader would later show very high achievement in mathematics.
We collected 4th-, 5th-, and 6th-grade achievement test data from a sample of children who had all been precocious readers by the time they finished kindergarten-reading at least at the second grade level and typically at the late third grade level (Jackson, Donaldson, & Mills, 1993). We knew from previous research that these children were likely to have remained good, but not necessarily exceptional readers (Mills & Jackson, 1990). Although one purpose of the present study was to replicate earlier findings on the subsequent reading and language arts achievement of precocious readers, our primary interest was in the extent to which precocious readers, as a group, would later turn out to be high achievers in mathematics.
Table 2 summarizes the characteristics of the sample of children who participated in the follow-up study. Although the follow-up sample of 64 was just slightly more than half of the original group, they were similar to the full sample in a number of key characteristics. This is important, because the full sample was probably quite representative of precocious readers in public school kindergartens in the greater Seattle area at the time we conducted the original study (Jackson et al., 1993).
TABLE 2. Comparison of precocious readers with follow-up MAT or CAT scores (n=64) with full original sample (n=116).*
For practical reasons, we had to rely on grade-level test scores routinely administered by the children's schools and forwarded to us either by the parents themselves or, with their permission, by the school district. Using grade-level test scores as our achievement measure meant that there probably was a substantial ceiling effect in our follow-up data. That is, some children in the sample probably could have answered much harder questions than those given on these tests, and small individual differences in scores at the top end probably reflect chance as much as any real difference in ability.
The follow-up achievement test scores that we had available for the largest groups of children are summarized in Table 3. The largest sample is for the Metropolitan Achievement Test (Harcourt Brace Jovanovich, 1985) because that test was given statewide early in grade 4. The California Achievement Test scores (CTB/McGraw-Hill, 1985) from the end of the 4th, 5th, and 6th grades are mostly for children who attended school in the largest participating district, Seattle.
Table 3. Distributions of precocious readers' 4th-, 5th, and 6th-grade Metropolitan Achievement Test (MAT) and California Achievement Test (CAT) normal curve equivalent (NCE) scores.*
Because of the ceiling effect, scores were summarized in terms of whether they were less than one, one to less than two, or two standard deviations above the national mean in normal-curve- equivalent (NCE) scores, which are scaled to an M of 50 and an SD of about 21. This strategy reduced the effect of score compression and also lets one see how many children were scoring in the highest range. Scores two or more standard deviations above the mean often are needed to qualify for a gifted education program or participation in a talent search involving out-of-level testing. Only a few of the children had been accelerated. Most of these performed very well by the more stringent grade norms, and these are the scores reported in Table 3.
The reading score distributions in Table 3 essentially replicate an earlier finding (Mills & Jackson, 1990) that precocious readers are likely to remain good, but not necessarily exceptional, readers. Three of the four sets of scores indicate only about 40% of the sample performing at the "gifted" level in reading achievement, although most did perform at least one standard deviation above the mean. This is solid performance, but far less exceptional than this group's post kindergarten performance in reading.
The language arts scores on tests that primarily reflect mastery of spelling, punctuation, and other mechanics, are roughly similar to those for reading. The present scores tend to be higher than those from an earlier study. The difference may reflect the fact that in the previous study, out-of-level tests were used and performance relative to grade-level norms was estimated using procedures available from the test publisher (Mills & Jackson, 1990). Another likely source of the performance differences across studies is that the present data were collected during the children's school day, under standard conditions for the administration of group achievement tests. In the earlier study, testing was done individually in after school sessions, and the children may have been less alert or less strongly motivated to work carefully on the language arts questions.
To us, the most interesting scores in the present data set are those for mathematics achievement. These weren't quite as high as those for reading and language arts, but they did come quite close to those distributions on three of the four tests. On all tests, more than a third of these children who were identified initially as precocious readers performed at the highest level in grade-level mathematics achievement. Median NCE scores for mathematics achievement ranged from 0 to 8 points below those for reading achievement. This pattern of results suggests that precocious readers, as a group, are a population of young children likely to contain a high proportion of children who later will do well in mathematics. Of course, the design of the present study did not permit comparative evaluation of post kindergarten reading achievement, mathematics achievement, and verbal intelligence, alone or in combination, as alternative screening criteria for identifying children who later will earn high mathematics achievement scores.
Individual Differences in Later Achievement
Because precocious readers vary widely in their later achievement, it is reasonable to ask which precocious readers are most likely to be the highest achievers in various domains in later school years. We did not attempt to predict individual differences in MAT Language Arts scores because these scores are hard to interpret in any theoretically interesting way. However, we did look at what postkindergarten characteristics were associated with individual differences in the children's later reading and mathematics achievement. Table 4 summarizes the results of these analyses.
TABLE 4. Correlations (r) among postkindergarten measures and associations (eta) between these measures and 4th-grade MAT score levels (n = 57).*
The data in the top portion of Table 4 indicate relations among measures obtained when the children were first tested, just after kindergarten. Relations between postkindergarten measures and 4th-grade MAT scores are in the bottom portion of the table. Longitudinal associations are reported in Table 4 only with 4th grade MAT scores because that was the largest set of scores. However, we also tested and considered associations with the other follow-up test scores. These analyses are mentioned whenever their results reflect on the interpretation of the associations reported in Table 4.
A detailed explanation of the tasks from which the predictor variables were derived and the confirmatory factor analysis in which some of their identities were validated is available elsewhere (Jackson, Donaldson, & Mills, 1993). The following brief descriptions are intended to give a general sense of what these scores mean.
Two of the predictor variables were derived from parent questionnaire items. Starting age refers to a composite score that was computed from parents' answers to retrospective questions about when their child passed a sequence of early reading milestones. Math interest was a weak measure--parents' responses to a single question about the extent of their postkindergartener's current interest in mathematics.
The remaining predictors were derived from scores on the test battery administered to the precocious readers during the summer following their kindergarten year. Factor scores used in the current analyses were not the actual latent factor scores but simple unweighted composites of standard scores for measures loading on a particular factor in the confirmatory analysis.
Three scores summarizing aspects of the children's performance on a set of tests of cognitive abilities were included among the predictors of later achievement. Reflectivity refers to a composite score on a test of how slowly and accurately a child makes a series of judgments about matches within sets of line drawings. Previous researchers have found that children who do this task slowly and accurately may be more successful scholars (Kurtz & Borkowski, 1987). Verbal knowledge was a composite score derived from a set of language tests and probably approximated verbal intelligence. Letter-naming speed was used as an index of how rapidly a child can retrieve familiar information from long-term memory. Efficiency on this kind of elementary cognitive task tends to be good in precocious readers and poor in children who are poor readers (Wolf, 1991).
The predictors included two measures of text reading. The first, Oral text-reading accuracy, was a precision measure. Word-by-word accuracy in reading text is not something we particularly want to encourage in children, but we included this measure among the predictors because we thought it might indicate a plodding, code-emphasizing style (Olson, Kleigl, Davidson, & Foltz, 1985). In contrast, Text-reading effectiveness was the most global measure of good reading in the initial battery. This score was a composite of the child's score on the PIAT Reading Comprehension subtest (Dunn & Markwardt, 1970) and measures of oral reading speed. Comprehension accuracy and oral reading speed were combined into a single score because a confirmatory factor analysis indicated that they represented a single construct.
These predictors were chosen from the larger set available in part because they were not strongly related to one another, and the top section of the matrix in Table 4 shows that this was the case. The highest correlation, that between verbal knowledge and text-reading effectiveness, was only .44. This low level is typical of correlations between verbal intelligence and beginning reading achievement in unselected samples. Figuring out the code of an alphabetic language seems to require skills that are only moderately related to verbal intelligence (e.g., Curtis, 1980).
Because the follow-up achievement scores had been aggregated into three categories, we did not use Pearson correlation coefficients to estimate longitudinal associations. Instead, we used a statistic called eta that measures both linear and nonlinear association and that is not marked plus or minus. The direction and linearity or nonlinearity of an eta association are revealed in a pattern of cell means. All of the reported associations were substantially linear and their directions are discussed below in any instance where it is not the obvious association of earlier and later good performances.
The only two significant predictors of individual differences in precocious readers' later reading achievement, as measured by 4th-grade MAT scores, were postkindergarten reading effectiveness and verbal knowledge. This pattern replicates the results of an earlier study (Mills & Jackson, 1990). Verbal knowledge also was associated strongly with 4th-, 5th-, and 6th-grade CAT scores (eta's = .67, .71., .65, all p < .01). Text-reading effectiveness was significantly associated with fourth-and sixth-grade CAT Reading (eta's = .51 and .53) and tended to be associated with fifth-grade CAT Reading (eta = .41, p < .10). A tendency for girls to score better than boys on fourth-grade MAT reading was not evident on any of the subsequent tests.
As might be expected given the set of available predictors, the overall prediction was not as strong for fourth-grade MAT mathematics scores. As hypothesized, postkindergarten verbal knowledge was a significant predictor of later mathematics achievement. This pattern was replicated as a trend in the fourth-grade CAT scores (eta = .39, p <.10) and as a significant association in the 5th-grade scores (eta = .72, p < .0001) but was not evident in the 6th-grade scores.
Although the sample as a whole achieved well in mathematics, those individual precocious readers who had the highest text reading effectiveness scores at initial testing did not consistently earn the highest mathematics achievement scores later on. Text-reading effectiveness was a significant predictor of mathematics achievement only in fifth grade (eta = .51, p < .05).
More surprising than the predictive association between verbal knowledge and later mathematics achievement was the substantial power of reflectivity to predict fourth-grade MAT mathematics scores. This pattern did not reappear in the subsequent sets of test scores and most likely indicated either an individual difference that becomes less important as children get older or something specific to the content of that particular achievement test.
One pattern that is only suggested slightly in the 4th-grade MAT data appeared more strongly and significantly (eta = .42, p < .05) in the 5th-grade CAT data. Those readers who were reported by their parents to be the earliest starters in reading earned the highest 5th-grade mathematics achievement scores. Although this pattern was very tenuous in the present data set, it is consistent with the model of the development of gifted performance to be presented here.
Another pattern emerged only in the three sets of CAT data, but it was substantial and significant in all three. Oral text reading accuracy was associated with mathematics achievement on these later tests (eta's = .42, .46., .53, all p < .05). Although the best, fastest readers were not consistently likely to be the best mathematicians, the most precise readers were, at least when tested on the CAT.
A Conceptual Model of Development
Results like those just presented, case histories of children such as Jeff, and the literature on gifted children and on development in general all suggest that gifted performance in childhood is best thought of in the context of a model of both homotypic and heterotypic continuity, such as that outlined in Figure 1. Continuity in development is homotypic if an underlying trait is expressed in similar ways across age; it is heterotypic if the trait is expressed in different ways (Kagan, 1971).
FIGURE 1. A conceptual model of the development during childhood of gifted performance in reading, mathematics, music, and language.* Mathematics is intended to include related fields such as computer science. Behaviors that have been studied in laboratory research but that cannot be observed easily in everyday life are enclosed in rounded boxes. From Jackson & Klein (in press).
Overall, the model in Figure 1 is meant to reflect ways in which children's early predispositions to process information in particular ways might interact with their subsequent experiences to produce a variety of different kinds of gifted performances at different ages. There is at least some theoretical or empirical basis for each hypothesized connection (Jackson, in press), but the model is necessarily a patchwork of reasoning from many sources. Its advantage over traditional, global conceptions of childhood giftedness (e.g., Terman & Oden, 1947) is that it reflects the domain specificity that characterizes creative-productive giftedness in adulthood (Siegler & Kotovsky, 1986).
This conceptual model emphasizes hypothesized continuities of both types in cognitive performances across age periods. Individual differences in genotype have an initial and lasting impact. Influences of experience, which are expected to be large and complex, are suggested by the boxes enclosing each set of possible age-related demonstrations of gifted performance. Darker arrows suggest relations hypothesized to be stronger. Additional information about the derivation of elements in the model has been given elsewhere (Jackson, in press).
Three ideas are key to how the model portrays the development of gifted performance, and all were inspired in part by Csikszentmihalyi and Robinson (1986), Sternberg (1993), and Tannenbaum (1993). One is that giftedness is likely to take different forms and require different skills at different ages and levels of achievement. In reading, breaking the alphabetic code requires different skills from comprehending complex texts. In mathematics, doing algebraic reasoning requires skills different from those needed to master basic computation.
The second idea is that what "counts" as gifted performance depends on a child's age. Reading gets a lot of attention when a child is four, but not when she is eight. Conversely, a child's rapid advancement in mathematics may get more attention when she reaches school age and is exposed to formal instruction in that area.
A third key idea is that, for at least some precocious readers, the essential element of their accomplishment is that they have a special interest in and aptitude for breaking the alphabetic code of print, which is a closed symbol system. A closed symbol system is one that is explicit, fixed, and limited in scope. Once mastered, such a system can be used productively in a great variety of ways. The alphabet can be used to represent countless different words in oral language; mathematical symbols can be used in an endless variety of problems; the structure and notation of a musical system, such as that of traditional western music, can be used to generate anything from "Chopsticks" to the works of Mozart. Similarly, learning a computer language (or a new natural language) gives a person a code that she or he can use in myriad creative ways.
Note that there is no direct relation between early oral language development and precocious reading. This finding is consistent with literature showing little or no relation between precocity in these two realms (Crain-Thoreson & Dale, 1992). However, language and verbal intelligence become more closely related to reading and other forms of literacy in later years.
The data presented in this paper do not prove a specific relation between precocious reading and high achievement in mathematics, but they do add one more small piece to the prior literature suggesting such a relationship. What accomplishments in these two domains might have in common is that both involve use of a closed symbol system. The same predilections that might make a child learn to sound out words at the age of two or three might help him or her master elementary (and perhaps higher) mathematics with equal zest and efficiency. The same might be said for achievement in areas involving other closed symbol systems, such as music, computer programming, and new languages. Recall that Jeff, who became a highly skilled programmer, started out as a precocious reader and also had longstanding interest in music and a second language. Codes upon codes! Another precocious reader, whose early skills probably were similar to Jeff's, became highly skilled at mathematics, but she eventually chose a career as a scholar and teacher of classical languages. (Jackson, 1992b)
The model portrays precocious reading as a phenomenon with diverse roots and sequelae. This picture fits data suggesting that not all precocious readers of English start out with an interest in and talent for breaking the grapheme-phoneme correspondence code and not all are likely to develop special interests in other closed symbol systems. In other studies, my colleagues and I have demonstrated that reasons for reading early and initial skill patterns vary widely (e.g., Jackson, 1988; 1992a). Some precocious readers may not be especially adept alphabetic code breakers, but may be very good at and interested in comprehending language in larger units and picking out partial correspondences between language they hear and what they see on the printed page.
There is a slight suggestion in the data I have presented earlier in this paper that the earliest-starting readers are the most likely to be the code-breaking sort, and hence the most likely to move on to interest in mathematics. Additional, albeit also tenuous, support for a link between a particular kind of reading precocity and later achievement in mathematics comes from the relation reported earlier between text reading accuracy and mathematics achievement. Perhaps being a precise reader is characteristic of a code-breaker's style.
The model in Figure 1 was derived primarily from studies of children who were precocious readers. What, if any, later demonstrations of gifted performance would one expect from comparably bright children who did not begin reading early? The case history of a second young man from a group of extraordinarily gifted young children who were studied in the late 1970's illustrates one possible answer to this question.
TABLE 5. Philip: Not a Precocious Reader (Born 1973).*
Philip, who is about the same age as Jeff, was not a precocious reader, although otherwise the two had much in common as preschoolers. Both came from similarly advantaged families and both were very high in verbal intelligence, although Philip did not show Jeff's extreme strength in performance on visual-perceptual tests. Like Jeff, Philip gave his mother and preschool teachers many signs of his high intelligence in his everyday behavior. However, the PIAT achievement scores in Table 5 show that his reading development initially was much slower than Jeff's. Philip was not able to pronounce even simple, familiar words until the last test session, when he attempted, systematically but unsuccessfully, to sound them out. The contrast with Jeff's rapid progress in reading is even more striking because Philip, unlike Jeff, spent his preschool and kindergarten years in a program that gently but systematically encouraged the development of academic skills. Philip was approaching seven, when, much to his mother's relief, he finally got the hang of reading and began to make rapid progress, gaining years of achievement in a few months. She later reported that he became a fluent and voracious reader within the next few years.
Does the fact that Philip was not an early reader mean that his subsequent intellectual development and career choice were very different from Jeff's? Not at all. Like Jeff, Philip attended a highly selective college and became a high-achieving journeyman scientist, although in computer hardware design rather than in software. A cognitive psychologist's analysis of their respective work might reveal important differences between the kinds of problems the two young men have worked at and how they have approached them, but, given the broad array of career paths two young men could take in our society, there is certainly some similarity here in type of profession chosen as well as in high level of achievement.
This second case study provides a reminder that whatever the statistical implications of precocious reading might be, precocious reading achievement is not necessary as a start for an excellent career in school or beyond; it is not even necessary for developing the kind of "code breaking" mathematical orientation described earlier. In children's development, nothing is simple. Those of us who do basic research often get excited, and appropriately so, at patterns we find in our data that suggest possible, partial cause-effect relations in development. However, we do need to be cautious about leaping from these statistical associations in groups to making recommendations about the lives of individual children. Precocious reading is a fascinating phenomenon, but there is no reason to be concerned about bright children who do not read early.
Note: This paper was presented at the third biennial Wallace Symposium on Talent Development, University of Iowa, Iowa City, IA, May 19, 1995, with a different title. Portions of the data presented here were generated in a series of collaborations with Wendy C. Roedell, Gary Donaldson, Joseph Mills, and the late Halbert B. Robinson. A grant to our colleague Susan Assouline provided partial support for collection of the longitudinal follow-up data reported here, and National Science Foundation Grant BNS 8509963 to the first author supported the initial study. We thank the participating families and school districts, especially Micahel O'Connell of the Seattle Public Schools, for their assistance. We also thank Nancy M. Robinson for providing access to the files which made the case studies possible.
Correspondence concerning this article should be addressed to Nancy Ewald Jackson, 361 Lindquist Center, University of Iowa, Iowa City, IA, 52242. Electronic mail may be sent via the Internet to firstname.lastname@example.org.
*Please refer to original text for all tables and figures.
Crain-Thoreson, C. & Dale, P. S. (1992). Do early talkers become early readers? Linguistic precocity, preschool language, and emergent literacy. Developmental Psychology, 28, 421- 429.
Csikszentmihalyi, M., & Robinson, R. E. (1986). Culture, time, and the development of talent. In R. J. Sternberg & J. E. Davidson (Eds.), Conceptions of giftedness (pp. 264-284). New York: Cambridge University Press.
CTB/McGraw-Hill. (1985). CAT Technical Report, Form E and F. Monterey, CA: Author.
Curtis, M. E. (1980). Development of components of reading skill. Journal of Educational Psychology, 72, 656-669.
Dunn, L. M. & Markwardt, F. C. (1970). Manual for the Peabody Individual Achievement Test. Circle Pines, MN : American Guidance Service.
Feldman, D. H. (1986). Nature's gambit: Child prodigies and the development of human potential. New York: Basic Books.
Gross, M.U.M. (1992). The early development of three profoundly gifted children of IQ 200. In P. S. Klein & A. J. Tannenbaum (Eds.), To be young and gifted (pp. 94-140). Norwood, NJ: Ablex.
Harcourt Brace Jovanovich (1985). The Metropolitan Achievement Tests: MAT6. Author.
Jackson, N. E. (1988). Precocious reading ability: What does it mean? Gifted Child Quarterly, 32, 200-204.
Jackson, N. E. (1992a). Precocious reading in English: Sources, structure, and predictive significance. In P. Klein & A. J. Tannenbaum (Eds.), To be young and gifted (pp. 171-203). Norwood, NJ: Ablex.
Jackson, N. E. (1992b). Understanding giftedness in young children. Lessons from the study of precocious readers. Talent Development: Proceedings from the 1991 Henry B. and Jocelyn Wallace National Research Symposium on Talent Development. Unionville, NY: Trillium Press.
Jackson, N. E. (in press). Strategies for modeling the development of giftedness in children: Reconciling theory and method. Chapter to appear in R. Horowitz & R. Friedman (Eds.), The gifted and talented: Theories and reviews. Washington, D. C.: American Psychological Association.
Jackson, N. E., Donaldson, G., & Mills, J. R. (1993). Components of reading skill in postkindergarten precocious readers and level-matched second graders. Journal of Reading Behavior, 25, 181-208.
Jackson, N. E. & Klein, E. (in press). Gifted performance in young children. Chapter to appear in N. Colangelo & G. A. Davis (Eds.), Handbook of gifted education (2nd ed.). Boston: Allyn & Bacon.
Kagan, J. (1971). Continuity and change in infancy. New York: Wiley.
Kurtz, B. E. & Borkowski, J. G. (1987). Development of strategic skills in impulsive and reflective children. Journal of Experimental Child Psychology, 43,129-148.
Mills, J. R., & Jackson, N. E. (1990). Predictive significance of early giftedness: The case of precocious reading. Journal of Educational Psychology, 82,410-419.
Olson, R. K., Kleigl, R., Davidson, B. J. & Foltz, G. (1985). Individual and developmental differences in reading disability. In G. E. MacKinnon & T. G. Waller (Eds.). Reading research: Advances in theory and practice (Vol. 4, pp. 1-64). New York: Academic Press.
Siegler, R. S., & Kotovsky, K. (1986). Two levels of giftedness: Shall ever the twain meet? In R. J. Sternberg & J. E. Davidson (Eds.), Conceptions of giftedness (pp. 417-436). New York: Cambridge University Press.
Sternberg, R. J. (1993). Procedures for identifying intellectual potential in the gifted: A perspective on alternative "metaphors of mind." In K. A. Heller, F. J. Monks, & A. H. Passow (Eds.), International hand book for research and development of giftedness and talent (pp. 185-207). Oxford: Pergamon Press.
Tannenbaum, A. J. (1993). History of giftedness and gifted education in world perspective. In K. A. Heller, F. J. Monks, & A. H. Passow (Eds.), International handbook for research and development of giftedness and talent (pp. 3-27). Oxford: Pergamon Press.
Terman, L. M. & Oden, M. H. (1947). Genetic studies of genius. Vol. IV. The gifted child grows up. Stanford, CA: Stanford University Press.
Terman, L. M. & Merrill, M. A. (1973). Stanford-Binet intelligence scale: Manual for the third revision, form L-M. Boston: Houghton-Mifflin.
VanTassel-Baska, J. (1989). Profiles of precocity: A three-year study of talented adolescents. In J. L. VanTassel-Baska & P. Olszewski-Kubilius (Eds.), Patterns of influence on gifted learners (pp. 29-39). New York: Teachers College Press.
Wechsler, D. (1974). Manual for the Wechsler Intelligence Scale for Children-Revised. New York: Psychological Corp.
Wolf, M. (1991). Naming speed and reading. Reading Research Quarterly, 26, 123-141.