Data from an experimental study evaluating the effectiveness of different reading instruction methods on mathematics achievement gains, drawn from a multilevel education study. The dataset contains 1,190 students nested within 312 classrooms across 107 schools, with measures of student demographics, classroom-level teacher characteristics, and school-level poverty indicators. Used in regression teaching to illustrate ANOVA with categorical predictors, ANCOVA controlling for prior achievement, and nested data structures.
Format
A tibble with 1,190 rows and 12 columns:
- girl
Student gender indicator. Type: numeric. Binary (0/1) where 1 = girl, 0 = boy.
- minority
Minority status indicator. Type: numeric. Binary (0/1) where 1 = minority student, 0 = non-minority.
- mathkind
Mathematics score at kindergarten entry. Type: numeric. Range: (290, 629). Standardized test score.
- mathgain
Gain in mathematics score from kindergarten to first grade. Type: numeric. Range: (-110, 253). Computed as first grade score minus kindergarten score.
- ses
Socioeconomic status of the student's household. Type: numeric. Range: (-1.61, 3.21). Continuous standardized composite measure.
- yearstea
Teacher's years of teaching experience. Type: numeric. Range: (0, 40). Classroom-level variable.
- mathknow
Teacher's mathematical knowledge score. Type: numeric. Range: (-2.50, 2.61). 109 missing values. Standardized measure of teacher content knowledge.
- housepov
Proportion of households in the school neighborhood below the poverty line. Type: numeric. Range: (0.01, 0.56). School-level variable.
- mathprep
Teacher's number of math content and methods courses. Type: numeric. Range: (1, 6). Classroom-level variable.
- classid
Classroom identifier. Type: numeric. Range: (1, 312).
- schoolid
School identifier. Type: numeric. Range: (1, 107).
- childid
Student identifier. Type: numeric. Range: (1, 1190).
Source
Hill, H. C., Rowan, B., & Ball, D. L. (2005). Effects of
teachers' mathematical knowledge for teaching on student achievement.
American Educational Research Journal, 42(2), 371-406.
Original data file: instruction.dta
Details
This dataset is used in Chapter 3 (ANOVA) and Chapter 2 (ANCOVA) to illustrate one-way ANOVA comparing group means, ANCOVA adjusting for covariates, and the use of dummy variables with categorical predictors. Key analyses include: comparing math gains across groups, adjusting for prior math scores (mathkind) and SES, examining teacher and school-level predictors of student math gains, and nested data analysis.
Examples
data(instruction)
head(instruction)
#> # A tibble: 6 × 12
#> girl minority mathkind mathgain ses yearstea mathknow housepov mathprep
#> <int> <int> <int> <int> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 1 1 448 32 0.460 1 NA 0.0820 2
#> 2 0 1 460 109 -0.270 1 NA 0.0820 2
#> 3 1 1 511 56 -0.0300 1 NA 0.0820 2
#> 4 0 1 449 83 -0.380 2 -0.110 0.0820 3.25
#> 5 0 1 425 53 -0.0300 2 -0.110 0.0820 3.25
#> 6 1 1 450 65 0.760 2 -0.110 0.0820 3.25
#> # ℹ 3 more variables: classid <int>, schoolid <int>, childid <int>
# ANCOVA: math gain predicted by minority status, controlling for prior math
lm(mathgain ~ minority + mathkind, data = instruction)
#>
#> Call:
#> lm(formula = mathgain ~ minority + mathkind, data = instruction)
#>
#> Coefficients:
#> (Intercept) minority mathkind
#> 266.4329 -8.7261 -0.4349
#>