Sekolah versus madrasah: navigating the varied effects of multilevel factors on student English reading performance.

Figures & data

Figure 1. Conceptual framework.

Note: This framework is applied for sekolah (code = 0) and madrasah (code = 1) groups.

Table 1. Variables used in this study.

Hierarchical Level	Theoretical Dimensions	Latent Variables	Description	Code
School level	School Characteristics	SCSECTOR	School Sector	0 = Public; 1 = Private
(Level 3)		SCSERV	School Service	0 = Non-full-day,
				1 = Full-day
		SCLOC1	School Location (Village)	0 = No, 1 = Yes
		SCLOC2	School Location (District)	0 = No, 1 = Yes
		SCLOC3	School Location (City)	0 = No, 1 = Yes
	School Autonomy	AUTB	Autonomy in Textbook	0 = School
		AUCC	Autonomy in Course Content	1 = Regional Government
		AUAS	Autonomy in Assessment	2 = Central Government
	School Resources	RSCAV	Availability of School Resources	WLE Score
	School Climates	MORALE	Teacher Morale	WLE Score
		DSCPLN	Disciplinary	WLE Score
		PRESS	Achievement Press	WLE Score
		SUPPORT	Teacher Support	WLE Score
Teacher Level	Teacher	EDULV	Education Level	0 = Bachelor; 1 = Master
(Level 2)	Characteristics	TCEXP	Teaching Experience	Scale (Years)
		CERT	Certification	0 = No; 1 = Yes
		JBSTAT	Job Status	0 = Non permanent; 1 = Permanent
	Professional Development	PD	Pedagogical knowledge	<25%–100%
	Job-Related Attitudes	EFF	Teacher Efficacy	WLE Score
		JOBS	Job Satisfaction	WLE Score
	Teaching Effectiveness	ICTUS	ICT use	WLE Score
		READP	Reading Practice	WLE Score
Student Level	Student Demographics	GENDER	Student Gender	0 = Female, 1 = Male
(Level 1)		AGE	Student Age	(14–20 years old)
	Wellbeing	HAPPY	Happiness	WLE Score
		ANX	Anxiety	WLE Score
		PEER	Peer Belonging	WLE Score
		BULLY	Bullying	WLE Score
	Motivation, Anxiety, and Difficulty	MOTREAD	Reading Motivation	WLE Score
		ANXREAD	Reading Anxiety	WLE Score
		DIFREAD	Reading Difficulty	WLE Score
	Reading Achievement	READING	Reading Achievement	WLE Score

Note: WLE = weighted likelihood estimate.

Figure 2. The final three-level model of students’ reading achievement for the sekolah group.

Note: No arrow = not statistically significant (p > .05).

Figure 3. The final three-level model of students’ reading achievement for the madrasah group.

Note: No arrow = not statistically significant (p > .05).

Table 2. Final-model results: fixed effects for English reading achievement.

HLM Models			Group
Student Level	Teacher Level	School Level	Sekolah	Madrasah
For INTRCPT1, π0
For INTRCPT2, β00
INTRCPT3, γ000			−0.54 (0.05)	−0.26 (0.02)
		SCSERV, γ001	−0.17 (0.05)	ns
		PRESS, γ002/ γ001	−0.04 (0.02)	0.10 (0.02)
		DSCPLN, γ003/ γ002	0.04 (0.01)	−0.11 (0.03)
		RSCAV, γ003	ns	0.13 (0.05)
	For PD, β01
	INTRCPT3, γ010		ns	0.11 (0.02)
For GENDER slope, π1
For INTRCPT2, β10
INTRCPT3, γ100			ns	0.09 (0.09)
For DIFREAD slope, π1/ π2
For INTRCPT2, β10/ β20
INTRCPT3, γ100/ γ200			−0.26 (0.06)	−0.38 (0.04)
		AUTB, γ101	−0.10 (0.02)	ns
		AUAS, γ102	0.17 (0.04)	ns
		RSCAV, γ201	ns	−0.07 (0.02)
	For JBSTAT, β11/ β21
	INTRCPT3, γ110/ γ210		−0.09 (0.04)	0.08 (0.03)
	For JOBS, β12
	INTRCPT3, γ120		−0.03 (0.01)	ns
For ANXREAD slope, π2/ π3
For INTRCPT2, β20/ β30
INTRCPT3, γ200/ γ300			−0.05 (0.01)	−0.11 (0.03)
		AUCC, γ301	ns	−0.02 (0.01)
	For ICTUS, β21
	INTRCPT3, γ210		0.01 (0.01)	ns
		SCLOC1, γ211	0.10 (0.04)	ns
For ANX slope, π3
For INTRCPT2, β30
INTRCPT3, γ300			−0.04 (0.01)	ns
For PEER slope, π4
For INTRCPT2, β40
INTRCPT3, γ400			0.20 (0.02)	ns
		MORALE, γ401	−0.05 (0.01)	ns
		SUPPORT, γ402	0.05 (0.01)	ns

Note: HLM = hierarchical linear modelling; INTRCPT1 = intercept at Level 1 represents average test score for a student; INTRCPT2 = intercept at Level 2 represents average classroom-level test score; INTRCPT3 = intercept at Level 3 represents average school-level test score across all schools; SCSERV = school service; ns = not statistically significant; PRESS = achievement press; DSCPLN = disciplinary; RSCAV = availability of school resources; PD = professional development (pedagogical); GENDER = student gender; DIFREAD = reading difficulty; AUTB = autonomy in textbook; AUAS = autonomy in assessment; JBSTAT = job status; JOBS = job satisfaction; ANXREAD = reading anxiety; AUCC = autonomy in course content; ICTUS = information and communication technology use; SCLOC1 = school location (village); ANX = anxiety; PEER = peer belonging; MORALE = teacher morale; SUPPORT = teacher support.

Figure 4. The interaction effects of the predictors on the slope of student-level factors and reading achievement across the groups.

Table 3. Estimations of variance components for English reading achievement across the groups.

Model	Estimation of Variance components
	Sekolah			Madrasah
	Between students	Between teachers	Between schools	Between students	Between teachers	Between schools
	(n = 726)	(n = 34)	(n = 16)	(n = 593)	(n = 30)	(n = 14)
Fully unconditional model	0.87	0.03	0.09	0.57	0.09	0.04
Final Model	0.13	0.01	0.03	0.1	0.01	0.01
Variance at each level
Between students (Level 1)	0.88		0.88	0.81		0.81
Between teachers (Level 2)	0.03		0.03	0.13		0.13
Between schools (Level 3)	0.09		0.09	0.06		0.06
The proportion of variance is explained by the final model.
Between students (Level 1)	0.85		0.85	0.82		0.82
Between teachers (Level 2)	0.67		0.67	0.88		0.88
Between schools (Level 3)	0.67		0.67	0.75		0.75
	The proportion of total variance explained by the final model, .83 = 83%			The proportion of total variance explained by the final model, .81 = 81%

Table

Variable				Sekolah				Madrasah				Total
		M	SD	Skewness	Kurtosis	M	SD	Skewness	Kurtosis	M	SD	Skewness	Kurtosis
School-level factors				n = 16				n = 14				n = 30
School Resource Availability	RSCAV	1.39	1.59	−0.03	0.90	−1.24	0.74	−1.22	0.67	1.58	1.83	0.56	−0.11
Teacher Morale	MORALE	0.32	2.18	−0.28	−1.08	0.13	1.93	0.27	−0.86	0.23	2.03	−0.05	−1.06
Achievement Pressure	PRESS	−0.13	1.66	0.17	0.34	0.96	1.63	0.28	−0.84	0.38	1.71	0.16	−0.21
Disciplinary	DCPLN	−0.05	1.84	0.23	−0.08	1.25	2.82	−0.90	1.64	0.56	1.92	−0.22	−0.38
Teacher Support	SUPPORT	−0.02	1.94	0.17	−1.14	0.21	2.44	−0.19	−0.93	0.09	2.15	−0.17	−0.96
Teacher-level factors				n = 34				n = 30				n = 1,319
Teacher Efficacy	EFF	1.71	2.38	−0.28	−0.82	1.17	3.13	−0.33	−0.39	1.45	2.74	−0.39	−0.31
Job Satisfactions	JOBS	1.33	1.57	0.47	−0.69	1.09	1.78	0.47	−0.70	1.22	1.66	0.43	−0.72
Reading Practice	READP	1.19	3.32	−0.46	−0.57	1.54	3.02	−0.36	−0.59	1.35	3.16	−0.43	−0.57
ICT Use	ICTUS	−0.68	1.75	−0.23	2.16	−0.65	2.03	−0.52	0.91	−0.66	1.87	−0.39	1.27
Student-level factors				n = 726				n = 593			n = 1,319
Happiness	HAPPY	2.00	1.44	−0.25	−0.54	1.89	1.47	−0.17	−0.42	1.95	1.45	−0.21	−0.49
Optimism	OPT	2.39	1.54	−0.01	−0.86	2.19	1.53	0.12	−0.71	2.30	1.54	0.05	−0.80
Anxiety	ANX	1.21	1.63	0.37	0.22	1.31	1.53	0.24	0.49	1.25	1.59	0.31	0.32
Peer Belonging	PEER	1.71	1.69	−0.16	0.15	1.62	1.63	0.09	−0.25	1.67	1.66	−0.05	−0.03
Bullying	BULLY	−1.33	1.54	0.37	0.04	−0.69	1.42	0.13	0.32	−1.04	1.52	0.21	0.02
Reading Motivation	MOTREAD	3.07	2.20	−0.06	−0.09	2.97	2.23	−0.11	−0.25	3.02	2.22	−0.09	−0.17
Reading Anxiety	ANXREAD	1.66	2.09	0.25	1.25	1.87	2.02	−0.21	1.20	1.76	2.06	0.05	1.18
Reading Difficulty	DIFREAD	1.26	1.59	0.44	1.63	1.40	1.54	0.42	1.25	1.32	1.57	0.43	1.46
Reading Achievement	READING	−0.54	0.99	0.42	1.50	−0.24	0.81	−0.03	1.11	−0.41	0.93	0.18	1.30

Table

		School Resource Availability (RSCAV)	Teacher Morale (MORALE)	Achievement Pressure (PRESS)	Disciplinary (DSCPLN)	Teacher Support (SUPPORT)
Sekolah (n = 16)	School Resource Availability (RSCAV)	1	.674**	−.564*	−.024	.614*
	Teacher Morale (MORALE)	.674**	1	−.763**	.063	.894**
	Achievement Pressure (PRESS)	−.564*	−.763**	1	.179	−.673**
	Disciplinary (DSCPLN)	−.024	.063	.179	1	.162
	Teacher Support (SUPPORT)	.614*	.894**	−.673**	.162	1
Madrasah (n = 14)	School Resource Availability (RSCAV)	1	.093	.579*	−.092	.049
	Teacher Morale (MORALE)	.093	1	−.065	−.054	.617*
	Achievement Pressure (PRESS)	.579*	−.065	1	−.121	.046
	Disciplinary (DSCPLN)	−.092	−.054	−.121	1	−.258
	Teacher Support (SUPPORT)	.846**	.15	.839**	−.062	.189

*p < .05 (2-tailed). **p < .01 (2-tailed).

Table

		Reading Practice (READP)	ICT Use (ICTUS)	Teacher Efficacy (EFF)	Job Satisfaction (JOBS)
Sekolah (n = 16)	Reading Practice (READP)	1	.689**	.927**	.599**
	ICT Use (ICTUS)	.689**	1	.724**	.663**
	Teacher Efficacy (EFF)	.927**	.724**	1	.599**
	Job Satisfaction (JOBS)	.599**	.663**	.599**	1
Madrasah (n = 14)	Reading Practice (READP)	1	.093	.121	.800**
	ICT Use (ICTUS)	.093	1	.611**	.163
	Teacher Efficacy (EFF)	.121	.611**	1	.116
	Job Satisfaction (JOBS)	.800**	.163	.116	1

*p < .05 (2-tailed). **p < .01 (2-tailed).

Table

		MOTREAD	ANXREAD	DIFREAD	HAPPY	OPT	ANX	PEER	BULLY	READING
Sekolah (n = 16)	MOTREAD	1	−.023	−.103**	.074*	.035	−.043	.074*	−.065	.128**
	ANXREAD	−.023	1	.288**	−.144**	−.123**	.131**	−.471**	.136**	−.503**
	DIFREAD	−.103**	.288**	1	−.242**	−.258**	.220**	−.381**	.252**	−.693**
	HAPPY	.074*	−.144**	−.242**	1	.218**	−.763**	.238**	−.897**	.278**
	OPT	.035	−.123**	−.258**	.218**	1	−.181**	.368**	−.190**	.276**
	ANX	−.043	.131**	.220**	−.763**	−.181**	1	−.236**	.781**	−.285**
	PEER	.074*	−.471**	−.381**	.238**	.368**	−.236**	1	−.215**	.725**
	BULLY	−.065	.136**	.252**	−.897**	−.190**	.781**	−.215**	1	−.263**
	READING	.128**	−.503**	−.693**	.278**	.276**	−.285**	.725**	−.263**	1
Madrasah (n = 14))	MOTREAD	1	−.054	−.039	−.088*	.058	−.014	−.006	.038	.077
	ANXREAD	−.054	1	.502**	.177**	.021	.062	−.029	−.008	−.701**
	DIFREAD	−.039	.502**	1	.061	.004	.079	−.036	.004	−.750**
	HAPPY	−.088*	.177**	.061	1	.091*	−.493**	.039	−.063	−.115**
	OPT	.058	.021	.004	.091*	1	−.073	.016	.014	−.010
	ANX	−.014	.062	.079	−.493**	−.073	1	−.125**	.146**	−.061
	PEER	−.006	−.029	−.036	.039	.016	−.125**	1	−.848**	.019
	BULLY	.038	−.008	.004	−.063	.014	.146**	−.848**	1	.022
	READING	.077	−.701**	−.750**	−.115**	−.01	−.061	.019	.022	1

Note: MOTREAD = reading motivation; ANXREAD = reading anxiety; DIFREAD = reading difficulty; HAPPY = happiness; OPT = optimism; ANX = anxiety; PEER = peer belonging; BULLY = bullying; READING = reading achievement.

*p < .05 (2-tailed). **p < .01 (2-tailed).

Table

	Sekolah		Madrasah
Level 1 Model		Level 1 Model
READINGijk =	π0jk + eijk	READINGijk =	π0jk + eijk
Level 2 Model		Level 2 Model
π0jk =	γ000 + u00k	π0jk =	γ000 + u00k
Level 3 Model		Level 3 Model
β00k =	γ000 + u00k	β00k =	γ000 + u00k
Mixed Model		Mixed Model
READINGijk=	= γ000+ r0jk + u00k + eijk	READINGijk=	= γ000+ r0jk + u00k + eijk

Table

Sekolah	Madrasah
Level 1 Model	Level 1 Model
READINGijk = π0jk + π1jk*(DIFREADijk) + π2jk*(ANXREADijk) + π3jk*(ANXijk) + π4jk*(PEERijk) + eijk	READINGijk = π0jk + π1jk*(GENDERijk)+ π2jk*(DIFREADijk) + π3jk*(ANXREADijk) + eijk
Level 2 Model	Level 2 Model
π0jk = β00k	π0jk = β00k + β01k*(PDjk)
π1jk = β10k + β11k*(JBSTATjk) + β12k*(JOBSjk) + r1jk	π1jk = β10k
π2jk = β20k + β21k*(ICTUSjk) + r2jk	π2jk = β20k + β21k*(JBSTATjk) + r2jk
π3jk = β30k + r3jk	π3jk = β30k + r3jk
π4jk = β40k + r4jk\
Level 3 Model	Level 3 Model
β00k = γ000 + γ001(SCSERVk) + γ002(PRESSk) + γ003(DSCPLNk) + u00k	β00k = γ000 + γ001(PRESSk) + γ002(RDSCPLNk) + γ003(RSCAVk)
β10k = γ100 + γ101(AUTBk) + γ102(AUASk) + u10k	β01k = γ010 + u01k
β11k = γ110 + u11k	β10k = γ100 + u10k
β12k = γ120 + u12k	β20k = γ200 + γ201(RSCAVk) + u20k
β20k = γ200 + u20k	β21k = γ210 + u21k
β21k = γ210 + γ211(SCLOC1k) + u21k	β30k = γ300 + γ301(AUCCk) + u30k
β30k = γ300 + u30k
β40k = β40k = γ400 + γ401(MORALEk) + γ402(SUPPORTk) + u40k
Mixed Model	Mixed Model
READINGijk= γ000 + γ001*SCSERVk + γ002*PRESSk + γ003*DSCPLNk+  γ100*DIFREADijk + γ101*DIFREDijk*AUTBk +  γ102*DIFREADijk*AUASk + γ110*DIFREADijk*JBSTATjk+  γ120*DIFREADijk*JOBSjk + γ200*ANXREADijk +  γ210*ANXREADijk*ICTUSjk + γ211*ANXREADijk*ICTUSjk*SCLOC1k+  γ300*ANXijk + γ400*PEERijk + γ401*PEERijk*MORALEk +  γ402*PEERijk*SUPPORTk+ r1jk *DIFREADijk+ r2jk *ANXREADijk+  r3jk *ANXijk+ r4jk*PEERijk+ u00k+ u10k *DIFREADijk  + u11k *DIFREADijk*JBSTATjk + u12k *DIFREADijk*JOBSjk  + u20k*ANXREADijk + u21k *ANXREADijk*ICTUSjk +  u30k *ANXijk + u40k *PEERijk + eijk	READINGijk= γ000 + γ001*PRESSk + γ002*RDSCPLNk + γ003*RSCAVk+  γ010*PDjk + γ100*GENDERijk + γ200*DIFREADijk +  γ201*DIFREADijk*RSCAVk+ γ210*DIFREADijk*JBSTATjk +  γ300*ANXREADijk + γ301*ANXREADijk*AUCCk+  r2jk *DIFREADijk+ r3jk *ANXREADijk+ u01k *PDjk +  u10k *GENDERijk + u20k *DIFREADijk +  u21k*DIFREADijk*JBSTATjk + u30k *ANXREADijk + eijk

Note: only significant direct and indirect effects are shown in the final-model equation in each group. READING = reading achievement; DIFREAD = reading difficulty; ANXREAD = reading anxiety; ANX = anxiety; PEER = peer belonging; GENDER = student gender; PD = professional development; JBSTAT = job status; JOBS = job satisfaction; ICTUS = ICT use; SCSERV = school service; PRESS = achievement press; DSCPLN = disciplinary; RSCAV = school resources availability; AUTB = school autonomy in textbook selection; AUAS = school autonomy in assessment; SCLOC1 = school location/village; AUCC = school autonomy in course content; MORALE = teacher morale; SUPPORT = teacher support.

Table

Final Estimation of fixed effects:
Fixed effects	Sekolah			Madrasah
	Coefficient (SE)	t ratio	df	Coefficient (SE)	t ratio	df
For INTRCPT1, π0
For INTRCPT2, β00
INTRCPT3, γ000	−0.55 (0.09)	−6.05	15	−0.23 (0.06)	−3.77	13
Final estimation of Level 1 and Level 2 variance components:
Random Effect	Reliability	Variance Component (SD)	X^2(df)	Reliability	Variance Component (SD)	X^2(df)
INTRCPT1, r0		.03 (0.19)			.09 (0.29)	59.47 (16)
Level 1, e	.47	.87 (0.93)	34.32 (18)	.75	.57 (0.75)
Final estimation of Level 3 variance components:
Random Effect	Reliability	Variance Component (SD)	X^2(df)	Reliability	Variance Component (SD)	X^2 (df)
INTRCPT1/
INTRCPT2, u00	.71	.09 (0.31)	54.30 (15)	.01	.04 (0.01)	14.59 (13)

INTRCPT1 = intercept at Level 1 represents average test score for a student; INTRCPT2 = intercept at Level 2 represents average classroom-level test score; INTRCPT3 = intercept at Level 3 represents average school-level test score across all schools.

Table

Final Estimation of fixed effects:
Random Effect	Sekolah			Madrasah
	Reliability	Variance Component (SD)	X^2(df)	Reliability	Variance Component (SD)	X^2(df)
Final estimation of Level 1 and Level 2 variance components:
INTRCPT1, r0	.05	.01	45.80 (5)	.05	.01 (0.01)	#
DIFREAD slope, r1/ r2	.40	.01	#	.75	.02 (0.13)	#
ANXREAD slope, r2/ r3	.24	.00	179.54 (2)	.73	.01 (0.09)	55.87 (6)
ANX slope, r3	.28	.00	235.95 (5)	–	–	#
PEER slope, r4	.19	.00	188.68 (5)	–	–	#
Level 1, e		.13			.10 (0.32)
Final estimation of Level 3 variance components:
Random Effect	Reliability	Variance Component (SD)	X^2(df)	Reliability	Variance Component (SD)	X^2 (df)
INTRCPT1/INTRCPT2, u00	.87	.03 (0.19)	#	.05	.01 (0.03)	4.07 (2)
INTRCPT1/PD, u01	–	–	–	.05	.00 (0.02)	5.79 (5)
GENDER/INTRCPT2, u10	–	–	–	.14	.00 (0.05)	11.44 (5)
DIFREAD/INTRCPT2, u10	.25	.03 (0.16)	#	.49	.00 (0.11)	19.87 (4)
DIFREAD/ JBSTAT, u11	.06	.00 (0.05)	13.21 (2)	–	–	–
DIFREAD/JOBS, u12	.06	.00 (0.02)	14.22 (2)	.05	.00 (0.02)	3.13 (5)
ANXREAD/INTRCPT2, u20	.18	.00 (0.19)	13.00 (2)	.44	.00 (0.07)	12.11 (4)
ANXREAD/ICTUS, u21	.18	.00 (0.01)	36.34 (1)	–	–	–
ANX/INTRCPT2, u30	.29	.00 (0.03)	56.48 (1)	–	–	–
PEER/INTRCPT2, u40	.48	.00 (0.07)	#	–	–	–

INTRCPT 1 = intercept at Level 1 represents the average test score for a student; # = very small; DIFREAD = reading difficulty; ANXREAD = reading anxiety; ANX = anxiety; PEER = peer belonging; INTRCPT2 = intercept at Level 2 represents average classroom-level test score; PD = professional development; GENDER = student gender; JBSTAT = job status; JOBS = job satisfaction; ICTUS = ICT use.