Predicting Reading Fluency Growth from Grade 2 to Age 23 with Parental and Child Factors

Figures & data

Figure 1. a) the specification of the main growth curve models 1–5 for reading fluency and the predictors: b) the specification of the additional models 1–5 with two separate growth curves.

Note. Models 1, 2, 3, 4, and 5 were very similar and only differed on their predictors. The exact list of predictors included in each of the four models can be found in Table 3. Asterisks (*) indicate that the loadings were allowed to be estimated freely.

Figure 2. All reading fluency trajectories.

Table 1. Descriptive statistics for all variables.

	N	Minimum	Maximum	Mean	SD	Skewness (Std. Error)	Kurtosis (Std. Error)
Reading fluency
Grade 2	178	1.23	10.22	5.15	2.06	.25 (.18)	−.70 (.36)
Grade 3	200	1.55	11.30	5.83	1.87	.38 (.17)	−.20 (.34)
Grade 8	182	2.96	13.85	8.69	2.01	−.08 (.18)	−.10 (.36)
Age 23	136	4.10	17.48	10.18	2.21	.22 (.21)	.37 (.41)
Children’s phonological awareness
	159	−2.18	1.58	−.18	.72	−.07 (.19)	−.03 (.38)
Children’s RAN
	195	31.50	144.35	59.91	16.57	1.76 (.17)	5.44 (.35)
Children’s Verbal IQ
	198	72	137	99.84	11.60	.25 (.17)	−.19 (.34)
Children’s Performance IQ
	198	64	135	100.83	13.20	−.09 (.17)	−.29 (.34)
Children’s letter knowledge
	197	.00	23.00	11.59	7.44	−.00 (.17)	−1.43 (.34)
Reading motivation (ages 8–9)
	163	1.50	3.67	2.78	.57	−.31 (.19)	−.53 (.38)
Task avoidance (age 8)
	173	.00	3.60	1.52	.80	.25 (18)	−.29 (.37)
Task avoidance (ages 13–15)
	133	1.80	4.70	3.20	.66	.11 (.21)	−.54 (.42)
Leisure reading (ages 13–15)
	112	1.00	5.00	3.36	.92	−.60 (.23)	−.21 (.45)
Parental education
	198	1.00	7.00	4.03	1.09	.66 (.17)	.77 (.34)
Parental income
	150	1.50	14.00	7.48	1.96	.01 (.20)	.96 (.39)
HLE (shared reading)
	156	1.00	4.53	2.89	.58	.09 (.19)	.64 (.39)
HLE (teaching letters)
	190	1.00	5.00	2.26	.91	.89 (.18)	.49 (.35)

Table 2. Pearson correlation coefficients between all variables.

Variable, age in years	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18
1. Reading fluency, 8	1
2. Reading fluency, 9	.89***	1
3. Reading fluency, 14	.74***	.79***	1
4. Reading fluency, 23	.72***	.76***	.81***	1
5. Parental education	.17*	.09	.11	.13	1
6. Parental income	.07	.00	.04	.01	.43**	1
7. Parental dyslexia, y/n	−.33***	−.24***	−.27***	−.25***	−.10	−.04	1
8. HLE, shared read 4–6	.08	.06	−.04	.07	.09	−.04	.06	1
9. HLE, teaching, 4	.21**	.14	.18*	.35***	.09	−.01	−.18*	.13	1
10. Verbal IQ, 8	.37***	.34***	.23 **	.35***	.17*	−.00	−.12	.08	.19*	1
11. Perform. IQ, 8	.27***	.21**	.15*	.17*	.14	.09	−.08	.15	.07	.46***	1
12. RAN, 5–6	−.38***	−.39***	−.43***	−.39***	−.13	.00	.24***	.07	−.07	−.31***	−.13	1
13. Phonol. skills, 3–5	.36***	.31***	.25***	.34***	.25***	.17*	−.31***	.15	.33**	.48***	.31***	−.36***	1
14. Letter know., 5	.55***	.47***	.38***	.34***	.22**	.16*	−.21**	.14	.34***	.47***	.26***	−.43***	.67***	1
15. Reading motiv., 8–9	.49***	.42***	.21**	.27**	.08	−.02	−.07	.34***	.22**	.33***	.32***	−.08	.31***	.28***	1
16. Task avoidance, 8	−.16*	−.23**	−.18*	−.16	−.05	−.07	.11	−.23**	.00	−.20**	−.16*	.13	−.17*	−.18*	−.25**	1
17. Task avoid., 13–15	−.01	−.07	−.11	−.08	−.21*	−.13	.09	−.11	−.20*	−.11	−.07	.07	−.03	−.04	−.06	.19*	1
18. Leisure read., 13–14	.06	.12	.07	.14	.07	−.06	.01	.22*	.03	.09	.14	−.01	.15	.07	.36***	−.17	.20*	1

*p < .05, **p < .01, ***p < .001.

Figure 3. The associations between significant predictors and the reading fluency trajectory.

Note. In Figures a-c, the base trajectory was constructed using the mean scores of reading fluency derived from the whole sample. The other trajectories were constructed using the mean scores of reading fluency derived from the subsamples of those who demonstrated a) RAN, b) letter knowledge, c) reading motivation above or below one standard deviation in the total sample, or d) those who received the amount of literacy-related teaching that was above or below one standard deviation in the total sample.

Table 3. The fully standardized estimates of the models and the model fit statistics.

	Intercept	Slope	Model fit statistics and Chi-square difference tests
	(age 23 level) Estimate (SE)	Estimate (SE)
Model 1: Cognitive predictors
Verbal IQ, age 8	.118 (.100)	.010 (.144)	Model 1 fit: x2 (13) = 25.83, p = .02, CFI = .98, TLI = .96, RMSEA = .07 (90% CI.03–.11), SRMR = .04. Chi-square difference test to compare if FR and control models are different: Δχ2(10) = 16.17, p = .095
Performance IQ, age 8	.013 (.077)	−.087 (.130)
RAN, ages 5–6	−.374 (.075)***	−.316** (.108)
Phonological awareness, age 5	−.073 (.109)	−.014 (.181)
Letter knowledge, age 5	.200* (.094)	−.339* (.169)
R-square	.274 (.059)***	.141 (.069)*
Model 2: Motivational predictors
Reading motivation, ages 8–9	.184 (.101)	−.395 (.123)**	Model 2 fit: x2 (11) = 18.88, p = .06, CFI = .99, TLI = .98, RMSEA = .06 (90% CI.00–.10), SRMR = .04. Chi-square difference test to compare if FR and control models are different: Δχ2(21) = 28.21, p = .134
Task avoidance, age 9	−.130 (.089)	−.035 (.108)
Task avoidance, ages 13–15	−.135 (.095)	−.197 (.126)
Leisure reading, ages 13–15	−.020 (.097)	.039 (.137)
R-square	.084 (.047)	.181 (.096)
Model 3: Parental predictors
Parental education	.114 (.091)	.017 (.112)	Model 3 fit: x2 (17) = 29.54, p = .03, CFI = .98, TLI = .97, RMSEA = .06 (90% CI.02–.09), SRMR = .04 Chi-square difference test to compare if FR and control models are different: Δχ2(8) = 9.99, p = .266 (we compared the models in which parental difficulties were excluded from predictors because the family risk variable was used for grouping)
Parental income	−.058 (.124)	−.025 (.135)
Parental dyslexia	−.208 (.076)**	.047 (.105)
Shared reading, ages 4–6	−.045 (.570)	−.169 (.118)
Teaching letters, 4	.217 (.076)**	.226 (.103)*
R-square	.119 (.048)*	.069 (.054)
Model 4: All significant predictors
Parental dyslexia	−.116 (.077)	.044 (.097)	Model 4 fit: x2 (16) = 30.17, p = .02, CFI = .98, TLI = .97, RMSEA = .06 (90% CI.02–.10), SRMR = .04.
RAN, ages 5–6	−.384 (.072)***	−.288 (.109)**
Letter knowledge, age 5	.103 (.083)	−.371 (.116)**
Reading motivation, ages 8–9	.142 (.086)	−.375 (.104)***
Teaching letters, age 4	.122 (.076)	.315 (.110)**
R-square	.291 (.056)***	.323 (.098)**
Models 5: Types of parental dyslexia predictors (dummy predictors; the controls are the reference group)
Persistent parental difficulties	−.313 (.073)***	−.059 (.105)	Model 5 fit: x2 (7) = 10.91, p = .14, CFI = .99, TLI = .99, RMSEA = .05 (90% CI.00–.11), SRMR = .05
Resolving parental difficulties	−.038 (.099)	.137 (.118)
R-square	.092* (.043)	.027 (.034)

*p < .05, **p < .01, ***p < .001.

Table A1. Patterns of missingness in the predictor variables with largest missingness (>20%).

Predictor with missing values/missing percent	Children’s reading fluency whose values are present on the predictor N, Mean (SD)				Children’s reading fluency whose values are missing on the predictor N, Mean (SD)
	Gr 2	Gr 3	Gr 8	Age 23	Gr 2	Gr 3	Gr 8	Age 23
Parental income, 22.5%	135, 5.32 (1.99)	150, 5.93 (1.74)	137, 8.79 (1.84)	102, 10.45 (2.07)	43, 4.63 (2.22)	50, 5.54 (2.22)	45, 8.36 (2.45)	34, 9.37 (2.45)
Task avoidance, ages 13–15, 34.6%	126, 5.25 (1.95)	133, 5.89 (1.91)	127, 8.60 (1.92)	100, 10.04 (2.22)	52, 4.90 (2.30)	67, 5.72 (1.80)	55, 8.89 (2.20)	36, 10.58 (2.15)
Leisure reading, ages,13–15, 44.5%	104, 5.14 (2.00)	112, 5.81 (1.91)	105, 8.57 (2.00)	85, 10.04 (2.23)	74, 5.17 (2.16)	88, 5.87 (1.84)	77, 8.85 (2.02)	51, 10.41 (2.18)

Table A2. Patterns of missingness in reading fluency at age 23.

Predictor with missing values/missing percent	Children’s reading fluency whose values are present on age 23 fluency N, Mean (SD)			Children’s reading fluency whose values are missing on age 23 fluency N, Mean (SD)
	Gr 2	Gr 3	Gr 8	Gr 2	Gr 3	Gr 8
Age 23 reading fluency, 32.5%	128, 5.31 (2.03)	136, 5.99 (1.88)	132, 8.87 (1.98)	50, 4.75 (2.10)	64, 5.49 (1.82)	50, 8.22 (2.03)

Tables A1 and A2 do not reveal apparent patterns of systematic missingness but they do indicate that parents whose children had lower reading scores were slightly more likely not to report their income. In addition, children with lower reading scores were slightly more likely not to self-report their task avoidance at ages 13 and 15 and not to complete the reading fluency assessment at age 23.

Table A3. The fully standardized estimates of the additional models that included two growth specifications.

Additional model 1 with two growth specifications: Cognitive predictors
Predictors	Intercept for the first growth model (Age 8), Estimate (SE)	Slope for the first growth model (Ages 8–9), Estimate (SE)
Verbal IQ, age 8	.107 (.076)	.009 (.093)
Performance IQ, age 8	.059 (.072)	−.013 (.089)
RAN, ages 5–6	-.173** (.063)	−.065 (.083)
Phonological awareness, age 5	−.013 (.098)	−.119 (.114)
Letter knowledge, age 5	.419*** (.080)	-.219* (.090)
Predictors	Intercept for the second growth model (Age 14), Estimate (SE)	Slope for the second growth model (Ages 14–23), Estimate (SE)
Verbal IQ, age 8	.043 (.084)	.215 (.120)
Performance IQ, age 8	.034 (.066)	−.016 (.105)
RAN, ages 5–6	-.348*** (.070)	.050 (.086)
Phonological awareness, age 5	−.117 (.094)	.203 (.123)
Letter knowledge, age 5	.276** (.082)	−.153 (.116)
Additional model 2 with two growth specifications: Motivational predictors
Predictors	Intercept for the first growth model (Age 8), Estimate (SE)	Slope for the first growth model (Ages 8–9), Estimate (SE)
Reading motivation, ages 8–9	0.468*** (.068)	-.315*** (.088)
Task avoidance, age 9	−.076 (.081)	−.098 (.083)
Task avoidance, ages 13–15	.017 (.074)	−.108 (.079)
Leisure reading, ages 13–15	.093 (.086)	−.161 (.113)
Predictors	Intercept for the second growth model (Age 14), Estimate (SE)	Slope for the second growth model (Ages 14–23), Estimate (SE)
Reading motivation, ages 8–9	.186* (.094)	.176 (.108)
Task avoidance, age 9	−.124 (.085)	.019 (.090)
Task avoidance, ages 13–15	−.078 (.093)	−.063 (.105)
Leisure reading, ages 13–15	.001 (.092)	.099 (.141)
Additional model 3 with two growth specifications: Parental predictors
Predictors	Intercept for the first growth model (Age 8), Estimate (SE)	Slope for the first growth model (Ages 8–9), Estimate (SE)
Parental education	.152 (.087)	-.152* (.075)
Parental income	−.022 (.113)	−.076 (.101)
Parental dyslexia	-.269*** (.066)	.149* (.073)
Shared reading, ages 4–6	.065 (.071)	−.059 (.088)
Teaching letters, 4	.111 (.075)	−.049 (.093)
Predictors	Intercept for the second growth model (Age 14), Estimate (SE)	Slope for the second growth model (Ages 14–23), Estimate (SE)
Parental education	.109 (.087)	.026 (.095)
Parental income	−.055 (.119)	−.001 (.094)
Parental dyslexia	-.215** (.069)	.005 (.082)
Shared reading, ages 4–6	−.042 (.072)	.027 (.077)
Teaching letters, 4	.108 (.073)	.267*** (.071)
Additional model 4 with two growth specifications: All significant predictors
Predictors	Intercept for the first growth model (Age 8), Estimate (SE)	Slope for the first growth model (Ages 8–9), Estimate (SE)
Parental dyslexia	-.169** (.057)	.138* (.070)
RAN, ages 5–6	-.188** (.059)	−.077 (.083)
Letter knowledge, age 5	.347*** (.063)	-.217** (.081)
Reading motivation, ages 8–9	.385*** (.055)	-.178* (.078)
Teaching letters, age 4	−.046 (.062)	.024 (.095)
Predictors	Intercept for the second growth model (Age 14), Estimate (SE)	Slope for the second growth model (Ages 14–23), Estimate (SE)
Parental dyslexia	−.118 (.067)	−.022 (.086)
RAN, ages 5–6	-.346*** (.066)	.020 (.085)
Letter knowledge, age 5	.163* (.075)	−.078 (.095)
Reading motivation, ages 8–9	.163* (.076)	.089 (.100)
Teaching letters, age 4	.003 (.068)	.290*** (.076)
Additional model 5 with two growth specifications: Dummy predictors of parental dyslexia
Predictor	Intercept for the first growth model (Age 8), Estimate (SE)	Slope for the first growth model (Ages 8–9), Estimate (SE)
Persistent parental difficulties	-.323*** (.067)	.166* (.077)
Resolving parental difficulties	-.159*(.078)	.156* (.070)
	Intercept for the second growth model (Age 14), Estimate (SE)	Slope for the second growth model (Ages 14–23), Estimate (SE)
Persistent parental difficulties	-.300***(.067)	−.030 (.097)
Resolving parental difficulties	−.053 (.086)	−.029 (.070)

*p < .05, **p < .01, ***p < .001. All models in this table are saturated and for this reason fit statistics are not reported.