Social network and social normative characteristics of married female adolescents in Dosso, Niger: Associations with modern contraceptive use

Figures & data

Figure 1. Primary respondents were couples interviewed individually (Level 1), during which time they were also asked the questions from the Social Network Module. One alter per participant was then interviewed (Level 2), including a social network survey. Alter’s alters (Level 3) were not interviewed, however, data were provided regarding them from the alter.

Table 1. Respondent and alter attributes, as reported by the respondent.

	Respondent attributes N = 322
	Mean	SD	%
Respondent age	17.4	1.5
Respondent number of children	1.6	1.1
Respondent percent nulliparous			13%
Respondent attended modern school			38%
Respondent only attended Quranic			14%
Respondent ever use of FP			49%
Respondent in the intervention group			75%
Respondent average number of alters	1.4	0.8
Respondent percent with no alters			12%
	Alters nominated (respondent report) N = 403
Alter age (216 missing)	24.5	8.44
Alter number of children	2.61	1.98
Alter was surveyed			57%
Alter relationship to respondent
Sister			11%
Mother			5%
In-law			10%
Friend			37%
Natal family			29%
Other			6%
Believes alter supports longer first birth timing			31%
Believes alter would support respondent FP use			85%
Believes alter approves of men listening to wives fertility wishes			82%
Alter was part of the study			15%

Figure 2. Difference between the types of relationships identified by parous versus nulliparous girls.

Figure 3. Compared to the control group, respondents in the RMA treatment group are more likely to report that their alters would be supportive of their family planning use, agree with a later amount of time between marriage and first birth, and would be supportive of men who listen to their wives fertility preferences.

Table 2. Association between being in treatment group and reporting a belief that alter supports her potential use of modern family planning, dyadic analysis N = 402.

	Beta	SE	P
Intervention group vs control	1.34	0.33	0.00
Respondent age	−0.11	0.12	0.34
Education: Quranic only: ref modern	−0.84	0.48	0.08
Education: none: ref modern	−0.05	0.34	0.87
Age difference husband-wife	0.00	0.04	0.95
Parity	0.27	0.16	0.09
Assets: median vs low	−0.52	0.42	0.21
Assets: high vs low	−0.36	0.44	0.42
Husband migrates	−0.15	0.37	0.69
Number of alters	0.31	0.21	0.15

Figure 4. The association between the likelihood that a respondent believes that their alter supports FP and the respondents report of ever having used FP is significant for new users (those who did not report use at baseline), but not for previous users.

Table 3. Association between the respondents belief that alter supports her FP and respondents reported ever use of FP N = 250.

	Beta	SE	P
Reports belief that alter supports her use of FP	0.99	0.35	0.00
Intervention group vs control	0.88	0.41	0.03
Respondent age	0.09	0.10	0.37
Education: Quranic only: ref modern	−0.49	0.43	0.25
Education: none: ref modern	−0.40	0.33	0.23
Age difference husband-wife	0.03	0.04	0.43
Parity	0.49	0.16	0.00
Assets: median vs low	−0.38	0.37	0.29
Assets: high vs low	−0.48	0.38	0.20
Husband migrates	−0.32	0.32	0.31
Number of alters	−0.17	0.22	0.43

Table 4. Alter attributes from Alter Survey, alters self-report.

	Alters interviewed (N = 251)
	Mean	SD	%
Alter age	26.0	10.04
Alter relationship to respondent
Sister			8%
Mother			6%
In-law			10%
Friend			39%
Natal family			30%
Other			6%
Ever used family planning			50%
Modern Schooling			25%
Quranic schooling			22%
Parity	3.8	2.5
Heard of RMA (treatment village)			75%
Have talked to others about RMA (those who report knowing about it)			50%
Believes community supports FP use			84%
Participated in the RMA program			14%

Figure 5. There is evidence of possible mediation between the association of a respondent’s treatment group, and the use of FP by an alter through the pathway of the alter’s knowledge of the RMA program. Longitudinal data will be necessary to understand this dynamic more clearly.

Table 5. Association between respondent being in the treatment condition and alter reporting ever use of FP N = 250.

	Beta	SE	P	Beta	SE	P
Respondent treatment vs control	0.93	0.47	0.05	0.51	0.52	0.33
Alter knows RMA				0.88	0.38	0.02
Alter is other vs friend	−1.35	0.72	0.06	−1.56	0.74	0.04
Alter is in-law vs friend	0.20	0.52	0.70	0.08	0.53	0.87
Alter is family vs friend	−0.57	0.38	0.13	−0.71	0.39	0.07
Alter is mother vs friend	−0.93	0.84	0.27	−1.03	0.85	0.23
Alter is sister vs friend	0.07	0.63	0.91	0.21	0.65	0.73
Alter age	−0.03	0.03	0.25	−0.03	0.03	0.26
Alter modern education vs Quranic	−0.29	0.45	0.52	−0.28	0.46	0.55
Alter no education vs Quranic	−0.74	0.39	0.06	−0.63	0.40	0.12
Alter parity	0.13	0.10	0.17	0.14	0.10	0.15

Figure 6. The association between the respondents use of FP (y axis), and the alters use of FP differs by relationship of the alter to the respondent (x axis). Respondents use of FP is strongly correlated with that of Inlaws and sisters, but not at all related to that of friends or extended family.

SA Table A1. Association between respondent characteristics with having nominated no alters, individual analysis (N = 322).

	Beta	SE	P
Parity >0	−1.01	0.47	0.03
Intervention group vs control	−0.21	0.43	0.62
Respondent age	−0.10	0.12	0.40
Education: Quranic only: ref modern	−0.13	0.60	0.83
Education: none: ref modern	−0.29	0.40	0.47
Age difference husband-wife > 5	0.90	0.43	0.03
Assets: median vs low	−0.82	0.46	0.08
Assets: high vs low	−0.21	0.43	0.62
Husband migrates	0.63	0.44	0.15

SA Table A2. Association between type of alter nominated and being parous, dyadic analysis N = 402.

	Beta	SE	P
Alter is ‘other’ vs friend	0.27	1.21	0.83
Alter is in-law vs friend	−1.15	0.60	0.05
Alter is family vs friend	0.02	0.47	0.96
Alter is mother vs friend	−1.03	0.59	0.08
Alter is sister vs friend	−0.86	0.49	0.08
Intervention group vs control	−0.47	0.60	0.43
Respondent age	0.70	0.12	0.00
Education: Quranic only: ref modern	−0.90	0.88	0.30
Education: none: ref modern	−0.80	0.56	0.15
Age difference husband-wife	0.08	0.06	0.17
Assets: median vs low	0.01	0.59	0.99
Assets: high vs low	−0.32	0.65	0.62
Husband migrates	0.62	0.49	0.21
Number of alters	0.93	0.38	0.01

SA Table A3. Association between being in the treatment group and having more than 1 alter, individual analysis N = 322.

	Beta	SE	P
Intervention group vs control	0.69	0.33	0.04
Respondent age	−0.04	0.10	0.67
Age difference husband-wife	−0.09	0.40	0.81
Education: Quranic only: ref modern	0.29	0.28	0.29
Education: none: ref modern	−0.08	0.03	0.01
Parity	0.33	0.14	0.02
Assets: median vs low	0.50	0.31	0.11
Assets: high vs low	0.77	0.32	0.02
Husband migrates	−0.43	0.28	0.12

SA Table A4. Association between being in treatment group and reporting a belief that alter supports a longer time until first birth, dyadic analysis N = 402.

	Beta	SE	P
Intervention group vs control	0.77	0.35	0.03
Respondent age	−0.01	0.10	0.93
Education: Quranic only: ref modern	−0.05	0.42	0.91
Education: none: ref modern	−0.14	0.30	0.64
Age difference husband-wife	0.02	0.03	0.44
Parity	−0.09	0.13	0.48
Assets: median vs low	−0.61	0.33	0.06
Assets: high vs low	−0.77	0.34	0.02
Husband migrates	−0.19	0.29	0.52
Number of alters	0.55	0.20	0.01

SA Table A5. Association between being in treatment group and reporting a belief that alter supports men who listen to their wives fertility preferences, dyadic analysis N = 402.

	Beta	SE	P
Intervention group vs control	1.15	0.31	0.00
Respondent age	0.00	0.10	0.97
Education: Quranic only: ref modern	−0.84	0.43	0.05
Education: none: ref modern	−0.23	0.35	0.51
Age difference husband-wife	0.07	0.04	0.07
Parity	0.08	0.14	0.59
Assets: median vs low	−0.16	0.39	0.68
Assets: high vs low	−0.05	0.41	0.90
Husband migrates	−0.23	0.39	0.55
Number of alters	0.49	0.22	0.03

SA Table A6. Association between the respondents belief that alter supports her FP and respondents reported current use of FP, dyadic analysis N = 402.

	Beta	SE	P
Reports belief that alter supports her use of FP	0.73	0.36	0.04
Intervention group vs control	0.94	0.42	0.03
Respondent age	0.07	0.11	0.50
Education: Quranic only: ref modern	−0.58	0.48	0.23
Education: none: ref modern	−0.22	0.33	0.51
Age difference husband-wife	0.01	0.04	0.72
Parity	0.44	0.15	0.00
Assets: median vs low	−0.77	0.37	0.04
Assets: high vs low	−0.61	0.36	0.10
Husband migrates	−0.39	0.33	0.23
Number of alters	−0.01	0.21	0.97

SA Table A7. Association between respondent having reported no alters and respondents reported current use of FP, individual analyses N = 322.

	Model 1: Ego has no alters			Model 2: Controlling for parity
	Beta	SE	P	Beta	SE	P
Intervention group vs control	0.86	0.43	0.04	0.90	0.39	0.02
Respondent age	0.16	0.09	0.09	0.03	0.10	0.79
Education: Quranic only: ref modern	−0.56	0.43	0.19	−0.63	0.44	0.15
Education: none: ref modern	−0.32	0.28	0.25	−0.31	0.28	0.26
Age difference husband-wife	0.06	0.03	0.04	0.55	0.14	0.00
Parity				0.04	0.03	0.17
Assets: median vs low	−0.46	0.32	0.14	−0.48	0.33	0.15
Assets: high vs low	−0.18	0.35	0.61	−0.25	0.35	0.48
Husband migrates	−0.24	0.29	0.42	−0.41	0.30	0.17
Number of alters equal to 0	−0.74	0.40	0.06	−0.63	0.41	0.13

SA Table A8. Association between respondent nominated her mother and ever using FP, individual analysis N = 322.

	Beta	SE	P
Respondent nominated her mother	−1.36	0.63	0.03
Intervention group vs control	0.94	0.38	0.01
Respondent age	0.03	0.10	0.79
Education: Quranic only: ref modern	−0.56	0.44	0.21
Education: none: ref modern	−0.32	0.28	0.26
parity	0.53	0.13	0.00
Age difference husband-wife	0.04	0.03	0.17
Assets: median vs low	−0.51	0.33	0.12
Assets: high vs low	−0.26	0.35	0.46
Husband migrates	−0.47	0.30	0.12
Number of alters equal to 0	−0.71	0.41	0.09

SA Table A9. The association of alter having ever used FP and respondent report of ever having used FP, alter level analysis N = 251.

	Beta	SE	P
Alter use of FP	0.45	0.30	0.14
Respondent number of alters	−0.05	0.21	0.80
Respondent age	0.08	0.11	0.45
Assets: median vs low	−0.50	0.36	0.16
Assets: high vs low	0.02	0.38	0.96
Education: Quranic only: ref modern	−0.71	0.48	0.13
Education: none: ref modern	−0.43	0.31	0.17
Parity	0.53	0.15	0.00

SA Table A10. The association of alter reporting a community norm in support of FP use, and egos use of FP, alter level analysis N = 251.

	Beta	SE	P
Alter norm in support of FP	1.03	0.42	0.01
Respondent number of alters	−0.04	0.21	0.84
Respondent age	0.12	0.11	0.30
Assets: median vs low	−0.50	0.36	0.16
Assets: high vs low	−0.09	0.39	0.82
Education: Quranic only: ref modern	−0.62	0.49	0.21
Education: none: ref modern	−0.41	0.32	0.20
Parity	0.54	0.15	0.00