Onwuegbuzie

Standardized Achievement Within Florida Title I Schools:
Longitudinal Analysis of Third-Grade Performance

Wendy B. Dickinson, Ph.D.
Liberal Arts Program
Ringling School of Art and Design
E-mail: wdickins@rsad.edu

Anthony J. Onwuegbuzie, Ph.D., P.G.C.E., F.S.S.
Professor
College of Education
Sam Houston State University
E-mail: tonyonwuegbuzie@aol.com

The Florida public school system encompasses 67 school districts statewide, providing educational services to more than two million students (n = 2,365,356) at 4,001 discrete school sites (Florida Department of Education, 2004b). As noted by the Florida Department of Education (2004a), all students in Florida in Grades 3-10 take the Reading and Mathematics portions of the FCAT in the spring of each year; all students in Grades 4, 8, and 10 take the Writing components of the FCAT; and all students in Grades 5, 8, and 10 take Science portion of the FCAT. Thus, different grades of students take varying combinations of the FCAT subject examinations.

Despite its importance, scant independent research has been conducted on the FCAT. In particular, little is known about factors that predict FCAT performance. Further, the long-term trends in FCAT achievement have not been the subject of investigations. The few studies that have been undertaken in this area have tended to involve between-group analyses. Yet, conducting within-group analyses has been found to represent an effective way of studying student performance on standardized tests (Onwuegbuzie, 1997). One group that has received little or no attention with regard to FCAT scores involves students who are enrolled in Title I schools. The Elementary and Secondary Education Act (1965), amended in 1994, recognized the need for providing equal opportunities for education for all students (Haladyna, 2002). Consequently, schools are designated as Title I schools if a large proportion of their students are economically disadvantaged. These schools then receive additional federal funding.

Thus, the present study examined the FCAT Reading and Mathematics scores among third graders since inception in 1991 of state testing requirements at each Title I school. Third graders were the focus of this investigation because students in this grade who have not demonstrated proficiency in FCAT Reading and Mathematics are retained and not promoted to Grade 4. Specifically, the current inquiry addressed the following two research questions:

With respect to the second research question, the following school-level variables were examined: socioeconomic status, school size, and school status (i.e., schoolwide vs. targeted assistance). Socioeconomic status (SES) and school size were chosen because they had been found previously to be predictors of standardized test performance using other measures (e.g., Alwin & Thornton, 1984; Borland & Howsen, 2003; Campbell, Voelkl, & Donahue, 1999; Diamond & Onwuegbuzie, 2001; Donahue, Voelkl, Campbell, & Mazzeo, 1999; Coleman, 1999; Okpala, Okpala, & Smith, 2001; Pungello, Kupersmidt, Burchinal, & Patterson, 1996; Snow, Burns, & Griffin, 1998; Valencia, Hiebert, & Kapinus, 1992; Walberg & Tsai, 1984; Walker, Greenwood, Hart, & Carta, 1994; White, 1982). In addition, school status (i.e., schoolwide vs. targeted assistance) was selected as a variable because as a result of 1994 legislation, high-poverty schools (i.e., those with school poverty levels greater than 50%) were allowed to use Title I money, in combination with other federal, state, and local funds, for the purpose of improving the entire educational program for all their students (rather than just Title I students). That is, these schools were permitted to operate schoolwide programs, leading to these schools being referred to as “schoolwide” schools (U.S. Department of Education, Office of the Under Secretary, Policy and Program studies service, 2003, p. 3). Conversely, "targeted-assistance programs” use Title I funds to provide services identified as failing or most at risk of failing to meet a state's content and student performance standards. Schools operating targeted assistance programs are referred to as "targeted assistance schools" (U.S. Department of Education, Office of the Under Secretary, Policy and Program studies service, 2003, p. 3). The 2001 legislation expanded eligibility from schools with 50% or more students eligible for free and reduced price lunch to schools with 40% or more students eligible (U.S. Department of Education, Office of the Under Secretary, Policy and Program studies service, 2003).

Standardized testing continues to play a key role as an indicator of achievement and as a political issue in American education (Haladyna, 2002).

Florida, as one of America’s largest public school arenas, continues to administer standardized student assessment despite dismal past performance and equally dismal projected performance. In fact, according to The Sarasota-Herald Tribune (2005), “nine out of ten Florida schools are going to fail to meet the federal No Child Left Behind law this year unless the state delays tougher measures set to go into effect this year” (p. A1).

Currently, Title I student FCAT scores are not reported separately for Florida students. This study focuses on Title I student performance on standardized testing instruments to “call attention to their needs and provide accurate and fair reporting of how schools are doing with respect to helping this group of students” (Haladyna, 2002, p. 184). Indeed, the present study appears to be the first formal investigation on the performance of Title I schools on the FCAT.

Although the current investigation focuses on standardized testing in the state of Florida, it has national implications. In fact, as stated by Greene et al. (2004):

perhaps the nation’s most aggressive test-based accountability measure is Florida’s A+ program. Florida uses results on the Florida Comprehensive Assessment Test (FCAT) to hold students accountable by requiring all students to pass the third grade administration of the exam before moving to the fourth grade, and by withholding diplomas from students who have not passed all sections of the tenth grade administration of the exam. It also holds schools and districts accountable by using FCAT results to grade schools from A to F on school report cards that are very widely publicized and scrutinized. (p. 1124).

As noted by Dickinson and Hall (2002), “one of the most striking observations about the student population in the state of Florida is the huge disparity in size between districts” (p. 32). Student membership in Florida public schools, for the 67 districts comprising the research population, ranges from Lafeyette (n = 1,048) to Dade (n = 364,554), for a total enrollment of 2,598,231 students. Figure 1 (attached at end of article) presents a map that displays all 67 public school districts in Florida.

Over the past four years, that is, since inception of the third-grade FCAT testing requirements, the percentage of elementary school students receiving free and reduced-price lunch has consistently exceeded 50% of all students in Florida. During this same time period, the number of Title I schools also has increased. For the school year 2003-2004, a total of 1,425 schools were classified as Title I schools statewide. Thus, the sampling frame used in the present study consisted of the 1,425 Title I schools across Florida. Participants comprised 618,203 third-grade students in all Title I schools throughout the state of Florida that had complete data pertaining to both the FCAT Reading (SSS) and FCAT Mathematics (SSS) test scores for all school years since its inception: 2000-2001, 2001-2002, 2002-2003, and 2003-2004. This yielded a total of 1,092 Title I schools being selected for the present investigation. These Title I schools represented all 67 school districts in Florida, with the number of students enrolled in these Title I districts ranging from 19 to 5,065 (M = 618.20, SD = 349.95). The proportion of students on free and reduced lunch in these schools ranged from 20.1% to 100% (M = 72.37%, SD = 16.44%).

Instruments

The Florida Comprehensive Assessment Test (FCAT) arose out of Florida's attempt to improve teaching and learning in elementary and secondary schools throughout the state of Florida. The primary purpose of the FCAT is to assess student achievement of the high-order cognitive skills represented in the Sunshine State Standards (SSS) in Reading, Writing, Mathematics, and Science. The SSS portion of FCAT represents a criterion-referenced test. A secondary purpose of the FCAT is to compare the performance of Florida students to the Reading and Mathematics performance of students across the nation using a norm-referenced test (NRT). Accordingly, each student is compared to other Florida students by indicating into which third of the state score range the student's scores fell.

All students in Grades 3-10 take the FCAT Reading and Mathematics; all students in Grades 4, 8, and 10 take FCAT Writing; finally, all students in Grades 5, 8, and 10 take FCAT Science. FCAT scores are reported numerically, ranging from 100-500. These scores are further categorized into five levels such that these categories range from achievement Level 1 (lowest) to achievement Level 5 (highest). Achievement Level 3 is considered to represent grade-level achievement for the academic subject being tested. Based on Department of Education data, these preliminary results will examine the school differences which resulted in a reported 82% of standard curriculum Caucasian students scoring at level 3 or higher (FCAT Mathematics, 2004), with 66% of Hispanics scoring level 3 or higher, and 48% of African-American students scoring level 3 or higher. Third-grade FCAT reading results report similar student performance, with 84% of standard curriculum Caucasian students scoring level 3 or higher, 68% of Hispanic students scoring level 3 or higher, and 55% of African-American students scoring level 3 or higher.

In 1998, 1999, and 2000, FCAT Reading SSS content scores were reported for the type of passage read (literary or informational). However, since 2001, FCAT Reading SSS content scores have been reported for the following four areas: (a) Words and Phrases in Context; (b) Main Ideas, Plot, and Purpose; (c) Comparisons and Cause/Effect; and (d) Reference and Research. Two types of items are included on the SSS part of FCAT Reading: multiple-choice items and performance task items (only for Grades 4, 8, and 10). Multiple-choice items are machine scored. Each answer to a performance task is scored holistically by at least two trained readers.

FCAT Mathematics SSS content scores are reported for the following five areas: (a) Number Sense and Operations; (b) Measurement; (c) Geometry and Spatial Sense; (d) Algebraic Thinking; and (e) Data Analysis and Probability.

For the 2002 FCAT third-grade administration, cut-scores by achievement level category (i.e., 1-5) and standard error of measurement (SEM) from the FCAT Technical Report (p. 49) are reported in Table 1. Table 2 presents the score reliability estimates from the FCAT 2002 technical report (p. 50), and the number of items for each of the third-grade FCAT Reading reporting categories. Table 3 presents the score reliability estimates from the FCAT 2002 technical report (p. 51) and the number of items for each of the third-grade FCAT Mathematics reporting categories.

Third Grade Student Cut-Scores by Achievement Level
_________________________________________________________
Between Achievement Levels Cut-scores SEM
_________________________________________________________

1-2                                                       253                                    22
2-3                                                       294                                    16
3-4                                                       346                                    16
4-5                                                       398                                    21
________________________________________________________

Cronbach’s Alpha Coefficients and Number of Items per Third-Grade Reading Reporting Category
___________________________________________________________
Reading Reporting Category Cronbach’s Alpha No. of Items
___________________________________________________________

Words and Phrases                                0.67                            6
Main Ideas                                             0.78                            16
Comparisons                                          0.78                            15
Reference Research                               0.55                             3
___________________________________________________________

Cronbach’s Alpha Coefficient and Number of Items per Third-Grade Mathematics Reporting Category
________________________________________________________________
Mathematics Reporting Category Cronbach’s Alpha No. of Items
________________________________________________________________

Number Sense, Concepts, Operations               0.76                                12
Measurement                                                    0.54                                08
Geometry and Spatial Sense                              0.53                                07
Algebraic Thinking                                            0.61                                06
Data analysis/Probability                                   0.67                                07
_______________________________________________________________

Since its inception, the Florida Comprehensive Assessment Test (FCAT) has been administered to all students in Grades 3-10 in the spring of each year. FCAT Scores were available via the Florida Department of Education (2004c) website utilizing two parallel searches. The first search originated through the FCAT homepage, which was linked to the Florida Department of Education main page. From the FCAT home page, the researchers selected the searchable FCAT scores via district and school. By selecting the district, scores were obtained for each school level. After selecting school level data, researchers chose the indicators for inclusion. The functional flaw in the Florida Department of Education database is the absence of linkage between school level indicators and FCAT scores by year. Therefore, FCAT scores by year had to be cross-matched manually by school identification number to provide a complete data picture for each third-grade performance. Once all data were collated, identifying information was removed. This procedure is displayed in Figure 2 (attached at end of article).

Results

Table 4 presents the means and standard deviations pertaining to the FCAT reading and mathematics scores by school year among third-grade Students in all Title I in Schools in Florida. This table reveals that reading scores have increased each year. In fact, the repeated measures analysis of variance (ANOVA) revealed that these four years were statistically significantly different from each other (F[3, 910] = 242.00, p < .0001). The effect size associated with this difference, as measured by ω2, was .55. Using Cohen’s (1988) criteria, this suggested a large effect size. Moreover, a trend analysis revealed that the four sets of FCAT reading scores represented a linear trend that was both statistically significant and practically significant (F[1, 912] = 693.41, p < .0001; ω2 = .43). No quadratic trend (F[1, 912] = 2.62, p > .05) nor cubic trend (F[1, 912] = 0.19, p > .05) emerged. The linear trend can be seen in Figure 3 (attached at end of article).

FCAT Reading and Mathematics Achievement by School Year Among Third-Grade Students in all Title I in Schools in Florida

Reading Mathematics

School Year

2000 - 2001
2001 - 2002
2002 - 2003
2003 - 2004

276.08
279.59
284.11
288.92

22.53
21.05
20.08
21.11

    932
    944
    967
    975

277.33
286.16
290.43
295.15

24.36
24.38
23.26
22.82

   932
   946
   967
   975

A similar pattern of scores emerged with respect to FCAT mathematics achievement, with mathematics scores increasing each year. The repeated measures analysis of variance (ANOVA) revealed that these four years were statistically significantly different from each other (F[3, 912] = 280.24, p < .0001). The effect size associated with this difference, as measured by ω2, was .48. Using Cohen’s (1988) criteria, this suggested a large effect size. More specifically, a trend analysis revealed that the four sets of FCAT mathematics scores indicated a statistically significant quadratic trend (F[1, 914] = 33.27, p < .0001; ω2 = .04). However, the effect size associated with this trend was small. The quadratic trend is illustrated in Figure 4 (attached at end of article). It can be seen from this figure that although there was a statistically significant increase from one year to the next, the largest increase was from school years 2000-2001 to 2001-2002 (M difference = 9.59, SD = 17.77, t = 16.44, p < .001). The difference from 2001-2002 to 2002-2003, although both statistically and practically significant, was smaller than the previous year (Md = 4.42, SD = 17.66, t = 7.67, p < .001). Finally, the difference in mathematics achievement from 2002-2003 to 2003-2004 was similar to the increase associated with the previous year (Md = 4.92, SD = 16.31, t = 9.36, p < .001).

A series of dependent samples t-tests, using the Bonferroni adjustment to control for the inflation of Type I error, revealed that the third-grade students from Title I schools attained statistically significantly higher levels of achievement in mathematics than they did in reading across all four years: 2000-2001 (Md = 1.25, SD = 14.91, t = 2.56, df = 931, p < .01), 2001-2002 (Md = 6.47, SD = 12.88, t = 15.43, df = 943, p < .001), 2002-2003 (Md = 6.32, SD = 12.67, t = 15.49, df = 966, p < .001), and 2003-2004 (Md = 6.22, SD = 12.77, t = 15.21, df = 974, p < .001). The Cohen’s (1988) d effect sizes associated with these differences were .05, .29, .29, and .28, respectively.

The Shapiro-Wilk test (Shapiro & Wilk, 1965; Shapiro, Wilk, & Chen, 1968) indicated that the distribution of the FCAT reading scores for third-grade students in the 2003-2004 school year departed slightly from normality (W = .99, p < .001). However, it should be noted that this test, like all other null hypothesis significance tests, is extremely sensitive to large sample sizes (Kirk, 1996; Thompson, 1993, 1996), such as was the case in the present study. Indeed, an examination of the standardized skewness coefficient (skewness divided by its standard error) and standardized kurtosis coefficient (kurtosis divided by its standard error) indicated that the skewness coefficient was within the range of normality, although there was slight evidence of a leptokurtic distribution that was characterized by a distributional shape that was more peaked (Onwuegbuzie & Daniel, 2002b, 2003). However, this peak was not considered severe enough, thereby justifying use of multiple regression. In addition, evaluation of assumptions of linearity and homogeneity revealed no threat to multiple regression analysis. This multiple regression analysis was conducted using customized SAS 8.0 programming (Cody & Smith, 1997). Every elementary school (i.e., n = 975) having a complete set of data since third-grade FCAT inception (2001-2004) was included in the analysis.

Using school as the unit of analysis, an all possible subsets (APS) multiple regression (Onwuegbuzie, 2003; Thompson 1995) was used to identify an optimal combination of independent variables (i.e., socioeconomic status, school size, and school type [schoolwide vs. targeted assistance]) that predicted 2003-2004 FCAT reading scores among third-grade students in Title I schools throughout the state of Florida. Using this technique, all possible models involving some or all of the independent variables were examined. This method of analysis has been recommended by many researchers (e.g., Thompson, 1995). In APS regression, separate regressions are computed for all independent variables singly, all possible pairs of independent variables, all possible trios of independent variables, and so forth, until the best subset of independent variables is identified according to some criterion. For this study, the criterion used was the maximum proportion of variance explained (R2), which provides an important measure of effect size (Cohen, 1988). An additional index used was Mallow’s Cp (Myers, 1986; Sen & Srivastava, 1990).

Squared semi-partial correlation coefficients, also known as part correlations, represent the amount by which R2 is reduced if a particular independent variable is removed from the regression equation. That is, squared semi-partial correlation coefficients express the unique contribution of the independent variable as a proportion of the total variance of the dependent variable (Cohen, 1988). Also, squared partial correlation coefficients indicate the unique contribution of the independent variable as a proportion of R2. In the present study, R2 was used directly as an effect size estimate, as recommended by Cohen (1988). According to Cohen (1988), for multiple regression models in the field of social and behavioral science, squared partial correlation values between 2% and 12.99% suggest small effect sizes, values between 13% and 25.99% indicate medium effect sizes, and values of 26% and greater suggest large effect sizes. These same criteria were used to assess whether the proportion of variance explained by the independent variables, R2, was suggestive of a small, medium, or large effect.

The APS multiple regression analysis revealed that a model containing one variable provided the best fit to these data. In fact, the best two-variable model, in which school size was added to the model, only increased the proportion of variance explained by 0.08%. In addition, Mallow’s Cp was closer in value to the number of regressor variables (Myers, 1986; Sen & Srivastava, 1990) with the one-variable solution than with any two-variable solution.

The selected model indicated that socioeconomic status contributed statistically significantly (F[1, 973] = 872.15, p < .0001) to the prediction of 2003-2004 FCAT reading scores (Table 4). This variable explained 47.27% of the variation in FCAT reading scores (adjusted R2 = 47.21%). Using Cohen’s (1988) criteria for assessing the predictive power of a set of independent variables in a multiple regression model, the proportion of variance explained indicated an extremely large effect size, because it well exceeded 26%.

An inspection of the studentized residuals generated from the model (Myers, 1986) suggested that the assumptions of normality, linearity, and homoscedasticity were met. Using the Bonferroni adjustment, none of the studentized residuals suggested that outliers were present. Also, the following additional influence diagnostics were examined: (a) the number of estimated standard errors (pertaining to the regression coefficient for socioeconomic status) that the coefficient changes if the ith observation were set aside (i.e., DFBETAS); (2) the number of estimated standard errors that the predicted value changes if the ith point is removed from the data set (i.e., DFFITS); and (3) the reduction in the estimated generalized variance of the coefficient over what would have been produced without the ith data point (i.e., COVRATIO). Using criteria recommended in the literature (e.g., Myers, 1986; Sen & Srivastava, 1990), no participant generated DFBETAS, DFFITS, or COVRATIO values that were large enough to indicate an outlying observation--again suggesting sample invariance.

The equation for the final regression model was as follows:

FCAT reading achievement    =                      352.84       –    0.88 (% FARL)
                                                                     (SE = 2.22)         (SE = 0.03)                  (1)

FARL indicates free and reduced lunch status and SE represents the standard error of the corresponding coefficient. The standardized beta weight was –0.69. This regression equation indicates that a 1% increase in the number of third-grade students who are eligible for free and reduced lunch at Title I schools is associated with a decrease of 0.88 in FCAT reading scores. Alternatively stated, a 10% increase in a Title I school’s eligibility rate is associated with a decrease of 8.8 in FCAT reading scores.

As was the case for the FCAT reading scores, the Shapiro-Wilk test (Shapiro & Wilk, 1965; Shapiro, Wilk, & Chen, 1968) indicated that the distribution of the FCAT mathematics scores for third-grade students in the 2003-2004 school year departed slightly from normality (W = .99, p < .001). However, an examination of the standardized skewness and kurtosis coefficients again indicated that the FCAT mathematics scores was characterized by a skewness coefficient that was within the range of normality coupled with a slight leptokurtic distribution. In addition, the assumptions of linearity and homogeneity appeared to be met. Thus, use of multiple regression was considered justified. This multiple regression analysis was conducted using every elementary school (i.e., n = 975) having a complete set of FCAT mathematics scores since third-grade FCAT inception (2001-2004).

Using school as the unit of analysis, as before, the APS multiple regression was used to identify an optimal combination of independent variables (i.e., socioeconomic status, school size, and school type) that predicted 2003-2004 FCAT mathematics scores among third-grade students in Title I schools throughout the state of Florida. The APS multiple regression analysis revealed that a model containing one variable provided the best fit to these data. In fact, the best two-variable model, in which school size was added to the model, only increased the proportion of variance explained by 0.64%. In addition, Mallow’s Cp was closer in value to the number of regressor variables (Myers, 1986; Sen & Srivastava, 1990) with the one-variable solution than with any two-variable solution.

The selected model indicated that socioeconomic status contributed statistically significantly (F[1, 973] = 586.40, p < .0001) to the prediction of 2003-2004 FCAT reading scores (Table 4). This variable explained 37.60% of the variation in FCAT reading scores (adjusted R2 = 37.5%). Using Cohen’s (1988) criteria, the proportion of variance explained indicated an extremely large effect size.

An inspection of the studentized residuals generated from the model (Myers, 1986) suggested that the assumptions of normality, linearity, and homoscedasticity were met. Using the Bonferroni adjustment, none of the studentized residuals suggested that outliers were present. Also, examination of additional influence diagnostics indicated that no participant generated DFBETAS, DFFITS, or COVRATIO values that were large enough to indicate an outlying observation--again suggesting sample invariance.

FARL indicates free and reduced lunch status and SE represents the standard error of the corresponding coefficient. The standardized beta weight was –0.61. This regression equation indicates that a 1% increase in the number of third-grade students who are eligible for free and reduced lunch at a Title I school is associated with a decrease of 0.85 in FCAT mathematics scores. In other words, a 10% increase in a Title I school’s eligibility rate is associated with a decrease of 8.5% in FCAT mathematics scores.

The present study examined FCAT Reading and Mathematics scores among third graders over a four-year period at every Title I school in Florida with four years of performance data. The Florida Department of Education (2004d) provides summary scores on all sections of the FCAT for schools, districts, and the state. However, despite the prevalence of Title I schools in the state of Florida, FCAT performance (by Title I school status), has not received the attention it deserves. As such, the current investigation represents the first attempt to examine the longitudinal FCAT achievement among this under-researched population.

The following two research questions were posed in this study: (1) What historical trends have been exhibited by the Title I schools’ student performance since the inception of third-grade FCAT testing? and (2) What role do school-level factors play on third-grade student academic performance in Title I schools, as measured by FCAT Reading (SSS) and FCAT Mathematics (SSS) scores? With regard to the first research question, findings revealed that FCAT reading scores increased monotonically over these years, revealing a linear trend. FCAT mathematics scores also increased; however, the trend was quadratic. These results are encouraging inasmuch as they indicate that FCAT reading and mathematics performance has been increasing each year since FCAT inception in Title I schools.

However, these scores are well below the overall state average for FCAT reading and mathematics performance. For example, in the 2003-2004 school year, the state average FCAT performance was 303 for reading and 310 for mathematics (Florida Department of Education, 2004d). As reported earlier, FCAT performance in this same year was 288.93 (SD = 21.11) 295.15 (SD = 22.82) for reading and mathematics, respectively. Using the Title I standard deviations indicates that the performance of Title I schools for the 2003-2004 school year represents 0.67 and 0.65 standard deviations below the state performance as a whole. Thus, clearly, there has been a large gap in student performance between Title I schools and non-Title I schools for the past years.

With regard to the second research question, the two regression models illustrated that third-grade students from Title I schools belonging to families characterized as having the lowest SES (i.e., percent eligible for free and reduced lunch) tended to attain the lowest FCAT reading scores and mathematics scores. This finding supports the myriad of studies in this area in which a large relation between SES and achievement in the areas of reading and mathematics has been documented (e.g., Alwin & Thornton, 1984; Campbell et al., 1999; Diamond & Onwuegbuzie, 2001; Donahue et al., 1999; Coleman, 1999; Okpala et al., 2001; Pungello et al., 1996; Snow et al., 1998; Valencia et al., 1992; Walberg & Tsai, 1984; Walker et al., 1994; White, 1982). Because the effect sizes associated with the relationship between SES and achievement was so large, an important implication of this study is that poor FCAT performance cannot be attributed only to teacher-related variables (e.g., competency) and school-related variables (e.g., organizational structure). Level of poverty appears also to play an important function.

It is therefore crucial to investigate further the influence of poverty on FCAT performance because important educational decisions for students are made based on FCAT scores. For third-grade students, the stakes are especially high: per the Florida educational statute, 1008.25(5)(b),
“Beginning with the 2002-2003 school year, if the student's reading deficiency, as identified in paragraph (a), is not remedied by the end of grade 3, as demonstrated by scoring at Level 2 or higher on the statewide assessment test in reading for grade 3, the student must be retained”. Because the decision to retain a student is made using a single measure of performance (i.e., FCAT Reading), this performance measure must accurately measure reading skills and not merely reflect the level of poverty (or non-poverty) of the children themselves.

By definition, Title I schools in Florida provide educational services to a higher proportion of students with lower incomes and minority students than do non-Title I schools.
As Diamond and Spillane (2004) noted, “While education is viewed by many as an important mechanism for social mobility, many scholars argue that schools reproduce rather than challenge social inequality” (p. 1145). Accordingly, the exact role of poverty and FCAT performance remains undefined, and provides direction for further research efforts. As many states have enacted mandatory standardized achievement testing (similar to the Florida Comprehensive Achievement Test) to meet federal accountability requirements, this research has implications not only for Florida’s Title I schools, but for every Title I school in all states within our nation
which employ high-stakes testing instruments.

Alwin, D., & Thornton, A. (1984). Family origins and the schooling process: Early versus late influences of parental characteristics. American Sociological Review, 49, 784-802.

Bishop, J. H. (2002, Fourth Quarter). Do students’ achievement standards raise performance? New England Economic Review, 14-16.

Borland, M. V., & Howsen, R. M. (2003). An examination of the effect of elementary school size on student academic achievement. International Review of Education, 49, 463-474.

Campbell, J., Voelkl, K., & Donahue, P. (1999, February 28). Report in brief: NAEP 1996 trends in academic progress. Retrieved February 5, 2005, from http://nces.ed.gov/nationsreportcard/96report/

Cody, R., & Smith, J. (1997). Applied statistics and the programming language. Upper Saddle River, N.J.: Prentice-Hall.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, N.J.: Erlbaum.

Darling-Hammond, L. (2004). Standards, accountability, and school reform. Teachers College Record, 106, 1047-1085.

Diamond, P. J., & Onwuegbuzie, A. J. (2001). Factors associated with reading achievement and attitudes among elementary school-aged students. Research in the Schools, 8(1), 1-11.

Diamond, J. B., & Spillane, J. P. (2004). High-stakes accountability in urban elementary schools: Challenging or reproducing inequality? Teachers College Record. 106, 1145-1176.

Dickinson, W. B., & Hall, B. W. (2003). Students behaving badly: Policies on weapons violations in Florida schools. Journal of School Violence, 2(3), 27-52.

Donahue, P., Voelkl, K., Campbell, J., & Mazzeo, J. (1999). NAEP 1998 reading report card for the nation and the states. Washington, D. C.: U.S. Department of Education Office of Educational Research and Improvement.

Florida Department of Education. (2004a). Florida Assessment Comprehension Test. Retrieved September 30, 2004, from http://www.firn.edu/doe/sas/fcat.htm

Florida Department of Education. (2004b). Florida School Indicators Report. Retrieved September 30, 2004, from http://info.doe.state.fl.us/fsir

Florida Department of Education. (2004c). 2003-2004 final list of Title I schools. Retrieved September 30, 2004, from http://fldoe.org

Florida Department of Education. (2004d). Florida Comprehension Assessment Test: State
Report of District Results: Grade 03 Reading. Retrieved September 30, 2004, from http://fldoe.org

Florida Department of Education. (2002). Technical report for test administrations of the FCAT 2002. San Antonio, TX: Harcourt Educational Measurement.

Greene, J., Winters, M., & Forster, G. (2004).Testing high-stakes tests: Can we believe the results of accountability tests? Teachers College Record, 106, 1124-1144.

Haladyna, T. M. (2002). Essentials of standardized testing. Boston, MA: Allyn & Bacon.
Kirk, R. E. (1996). Practical significance. A concept whose time has come. Education and Psychological Measurement, 56, 746-759.

Myers, R. H. (1986). Classical and modern regression with applications. Boston, MA: Duxbury Press.

Okpala, C. O., Okpala, A. O., & Smith, F. E. (2001). Parental involvement, instructional
expenditures, family socioeconomic attributes, and students achievement. Journal of Educational Research, 95, 110-115.

Onwuegbuzie, A. J. (1997, April). The impact of magnet schools on within-race academic
achievement. Paper presented at the annual meeting of the National Association of School Psychologists (NASP), Anaheim, CA.

Onwuegbuzie, A. J., & Daniel, L. G. (2002a). A framework for reporting and interpreting
internal consistency reliability estimates. Measurement and Evaluation in Counseling and Development, 35(2), 89-103.

Onwuegbuzie, A. J., & Daniel, L. G. (2002b). Uses and misuses of the correlation coefficient. Research in the Schools, 9(1), 73-90.

Onwuegbuzie, A. J., & Daniel, L. G. (2003, February 12). Typology of analytical and
interpretational errors in quantitative and qualitative educational research. Current Issues in Education [On-line], 6(2). Retrieved February 8, 2005, from http://cie.ed.asu.edu/volume6/number2/

Pungello, E., Kupersmidt, J., Burchinal, M., & Patterson, C. (1996). Environmental risk factors
and children’s achievement from middle childhood to early adolescence. Developmental Psychology, 32, 755-767.

Sen, A. K., & Srivastava, M. (1990). Regression analysis: Theory, methods, and applications.
New York: Springer-Verlag.

Shapiro, S. S., & Wilk, M. B. (1965). An analysis of variance test for normality and complete samples. Biometrika, 52, 592-611.

Shapiro, S. S., Wilk, M. B., & Chen, H. J. (1968). A comparative study of various tests for normality. Journal of the American Statistical Association, 63, 1343-1372.

Snow, C., Burns, S., & Griffin, P (Eds.). (1998). Preventing reading difficulties in young children. Washington, D.C.: National Academy Press.

Thompson, B. (1993). The use of statistical significance research: Bootstrap and other alternatives. The Journal of Experimental Education, 61, 361-377.

Thompson, B. (1995). Stepwise regression and stepwise discriminant analysis need not apply here: A guidelines editorial. Educational and Psychological Measurement, 55, 525-534.

Thompson, B. (1996). AERA editorial policies regarding statistical significance testing: Three suggested reforms. Educational Researcher, 25(2), 26-30.

Understanding FCAT Reports. (2004). Assessment and school performance. Tallahassee, FL: Florida Department of Education

U.S. Department of Education, Office of the Under Secretary, Policy and Program Studies
Service. (2003). A snapshot of Title I schools, 2000-2001 (pp. 1-11). Washington, DC: Author.

Valencia, S., Hiebert, E., & Kapinus, B. (1992). National Assessment of Educational Progress: What do we know and what lies ahead? The Reading Teacher, 45, 730-734.

Walberg, H., & Tsai, S. (1984). Reading achievement and diminishing returns to time. Journal of Educational Psychology, 76, 442-451.

Walker, D., Greenwood, C., Hart, B., & Carta, J. (1994). Prediction of school outcomes based on early language production and socioeconomic factors. Child Development, 65, 606-621.

White K. (1982). The relation between socioeconomic status and academic
achievement. Psychological Bulletin, 91, 461-481.

You are invited to join AE Extra staff!
Send your ideas and/or writing sample to the Editor-in-chief:
Elizabeth Haller
Kent State University (e-mail: editoraee@hotmail.com)

Academic Exchange Extra invites reader response to any writings in this issue--especially articles advancing the scholarly debate of issues raised.