Supplementary appendix

This appendix formed part of the original submission and has been peer reviewed.
We post it as supplied by the authors.

Supplement to: Ewerling F, Lynch JW, Victora VG, van Eerdewijk A, Tyszler M, Barros AJD.
The SWPER index for women’s empowerment in Africa: development and validation
of an index based on survey data. Lancet Glob Health 2017; published online July 26.
http://dx.doi.org/10.1016/S2214-109X(17)30292-9.
  The SWPER index for women’s empowerment in Africa: 
development and validation of an index based on survey data 
Ewerling, F et al. The Lancet Global Health, 2017. 

Supplementary tables and figures

Table S1. Latest DHS surveys from African countries included in the analyses and number of
women in the samples.
Country Year N Country Year N
1. Benin 2011 11206 18. Malawi 2010 15067
2. Burkina Faso 2010 13192 19. Mali 2012 8737
3. Burundi 2010 5138 20. Morocco 2003 8619
4. Cameroon 2011 9039 21. Mozambique 2011 8643
5. Comoros 2012 2930 22. Namibia 2013 3520
6. Congo DR 2013 12045 23. Niger 2012 9165
7. Côte d’Ivoire 2011 5817 24. Nigeria 2013 26500
8. Egypt 2014 20406 25. Rwanda 2014 6778
9. Ethiopia 2011 10004 26. S Tomé & Príncipe 2008 1715
10. Gabon 2012 4100 27. Senegal 2014 5251
11. Gambia 2013 6423 28. Sierra Leone 2013 10254
12. Ghana 2014 5343 29. Swaziland 2006 1888
13. Guinea 2012 6641 30. Tanzania 2010 6233
14. Kenya 2014 8833 31. Togo 2013 6161
15. Lesotho 2009 4069 32. Uganda 2011 5150
16. Liberia 2013 5708 33. Zambia 2013 9088
17. Madagascar 2008 11102 34. Zimbabwe 2010 5444

Table S2. DHS variables included in the empowerment index and reasons for exclusion.
Variable Included Reason for exclusion
Beating justified if wife goes out without telling husband Yes
Beating justified if wife neglects the children Yes
Beating justified if wife argues with husband Yes
Beating justified if wife refuses to have sex with husband Yes
Beating justified if wife burns the food Yes
Frequency of reading newspaper or magazine Yes
Respondent worked in last 12 months Yes
Woman’s education Yes
Education difference: woman’s minus husband’s years of Yes
schooling
Age difference: woman’s minus husband’s age Yes
Age at first cohabitation Yes
Age of respondent at 1st birth Yes
Who usually decides on respondent's health care Yes
Who usually decides on large household purchases Yes
Who usually decides on visits to family or relatives Yes
Who usually decides on daily household purchases No Not available in all surveys
Whether she can refuse to have sex with the husband No Not available in all surveys
Ever used anything or tried to delay or avoid getting pregnant No Not available in all surveys
Type of earnings from respondent’s work No Asked only to women employed in last year
Can you go to a health centre or hospital alone or with your No Not available in all surveys
young children?
Decision maker for using contraception No Asked only to women using contraception
Ownership of house or land No Not available in all surveys
Ownership of a personal mobile phone No Usually a household question; personal
information is not available in all surveys.

1 
Table S3. Composition patterns of the SWPER domains.
Variable numbers*
Attitude to
Country N Social independence Decision making
violence
1 2 3 4 5 6 87 9 10 11 12 13 14 11 6 7
Benin 11206 x x x x x x xx x x x x
Burkina Faso 13192 x x x x x x xx x x x x
Burundi 5138 x x x x x x xx x x x x
Cameroon 9039 x x x x x x xx x x x x x
Comoros 2930 x x x x x x xx x x x x
Côte d’Ivoire 12045 x x x x x x xx x x x x
Gambia 5817 x x x x x x xx x x x x
Ghana 20406 x x x x x x xx x x x x
Guinea 10004 x x x x x x xx x x x x
Kenya 4100 x x x x x x xx x x x x x
Madagascar 6423 x x x x x x xx x x x x
Mali 5343 x x x x x x xx x x x x
Niger 6641 x x x x x x xx x x x x
Nigeria 8833 x x x x x x xx x x x x x
Senegal 4069 x x x x x x xx x x x x x
Sierra Leone 5708 x x x x x x xx x x x x
Swaziland 11102 x x x x x x xx x x x x
Togo 15067 x x x x x x xx x x x x x
Uganda 8737 x x x x x x xx x x x x
Zambia 8619 x x x x x x xx x x x x
Zimbabwe 8643 x x x x x x xx x x x x
Congo DR 3520 x x x x x x xx x x x x x
Gabon 9165 x x x x x x xx x x x x x
Liberia 26500 x x x x x x xx x x x x x
S Tomé & Príncipe 6685 x x x x x x xx x x x x x
Egypt 1715 x x x x x xx x x x x
Ethiopia 5251 x x x x x xx x x x x
Lesotho 10254 x x x x x xx x x x x
Malawi 1888 x x x x x xx x x x x
Morocco 6233 x x x x x xx x x x x
Rwanda 6778 x x x x x xx x x x x
Mozambique 5150 x x x x x x x x x x x x
Tanzania 9088 x x x x x x x x x x x x
Namibia 5444 x x x x x x x x x x x x x
Key to variable numbers: Beating not justified: (1) if woman goes out without telling husband; (2) if woman neglects the
children; (3) if woman argues with husband; (4) if woman refuses to have sex with husband; (5) if woman burns the food;
(6) Frequency of reading newspaper; (7) Education; (8) Age at 1st birth; (9) Age at 1st cohabitation; (10) Education
difference (woman’s minus husband’s years of schooling); (11) Work; Who usually decides on: (12) respondent’s
healthcare; (13) large household purchases; (14) visits to family or relatives.

2 
Table S4. Tertiles (green=top; yellow=intermediate; red=bottom) and country rankings of the
average national scores of the three domains. Top tertile represent the countries with higher
average empowerment levels.
Tertiles Ranking
Attitud Social Decisio Attitud Social Decisio
Country Year e to Independenc n e to Independenc n
violence e making violence e making
Namibia 2013 ❸ ❸ ❸ 6 1 1
Lesotho 2009 ❸ ❸ ❸ 10 5 7
Zimbabwe 2010 ❸ ❸ ❸ 9 7 3
Ghana 2014 ❸ ❸ ❸ 8 9 6
Swaziland 2006 ❸ ❸ ❷ 2 2 12
Egypt 2014 ❸ ❸ ❷ 11 3 15
Gabon 2012 ❷ ❸ ❸ 15 4 4
Rwanda 2014 ❷ ❸ ❸ 13 6 9
Kenya 2014 ❷ ❸ ❸ 14 8 5
S Tomé &
2008 ❸ ❷ ❸ 3 12 10
Principe
Madagascar 2008 ❸ ❷ ❸ 7 16 2
Comoros 2012 ❷ ❸ ❷ 16 10 21
Zambia 2013 ❷ ❷ ❸ 22 14 11
Malawi 2010 ❸ ❷ ❷ 1 22 22
Benin 2011 ❸ ❶ ❷ 5 24 20
Mozambique 2011 ❸ ❶ ❷ 4 25 17
Liberia 2013 ❷ ❶ ❸ 18 27 8
Morocco 2003 ❶ ❸ ❶ 31 11 27
Burundi 2010 ❶ ❷ ❷ 26 13 14
Togo 2013 ❷ ❷ ❶ 12 17 24
Congo DR 2013 ❶ ❷ ❷ 29 19 19
Cameroon 2011 ❷ ❷ ❶ 19 20 23
Nigeria 2013 ❷ ❷ ❶ 17 21 30
Tanzania 2010 ❶ ❷ ❶ 24 15 25
Senegal 2014 ❶ ❷ ❶ 28 18 33
Uganda 2011 ❶ ❶ ❷ 23 23 13
Gambia 2013 ❶ ❶ ❷ 25 26 18
Cote d’Ivoire 2011 ❷ ❶ ❶ 21 28 28
Burkina Faso 2010 ❷ ❶ ❶ 20 31 31
Ethiopia 2011 ❶ ❶ ❷ 32 32 16
Sierra Leone 2013 ❶ ❶ ❶ 27 29 26
Mali 2012 ❶ ❶ ❶ 33 30 34
Guinea 2012 ❶ ❶ ❶ 34 33 29
Niger 2012 ❶ ❶ ❶ 30 34 32

3 
 How to calculate the SWPER for a specific survey

The equation used to estimate individual standardized scores for each of the PCA j components is given
by:

̅ ̅ ⋯ ̅
1

where Sij are the individual standardized scores for individual i and component j; x1j,…, x15j are the
individual values for variables x1-x15 included in the PCA analyses; are the standard deviations of
the predicted scores of each component j. The weight given to each of the 15 variables in each
component j is defined as:

Where is the PCA loading for each of the variables in each domain j and is the standard
deviation of each variable in the combined dataset.
By using simple algebra, we can simplify the equation above to:
∑ ̅ ∑
3

Please, follow the next steps to calculate the standardized individual SWPER scores for any African
country of your interest1:

1. Recode variables
The first step is to recode the variables as it is shown in Table S5, below.

1.1. Imputation of woman’s age at first birth

To procced the imputation of age at first birth for nulliparous women, we used single hotdeck
imputation. This method randomly selects the value to be imputed for a missing case from a group of
individuals that are similar to it in terms of a variable or a group of variables. In this case, women were
clustered in groups of age at first cohabitation. This variable was selected because it had the highest
correlation with age at first birth, and other variables did not add much predictive power in a regression
model. Despite the current preference for multiple imputation, we used a single imputation approach
because procedures for principal component analysis with multiple imputation data are not commonly
available, and the percentage of missing information was not so high that overall variance would be
significantly reduced by use of single imputation.

2. Calculate the individual scores

Using the equations below, it is possible to estimate the scores for the three SWPER domains:
0.950 ∑

1.818
5.360 ∑

1.475
0.857 ∑

1.417

                                                            
1
A Stata do-file with all procedures required for the calculation of the SWPER Index scores is available from
the Dropbox link <https://www.dropbox.com/sh/6jw57xtp824qsbh/AAC5kOeVBG2o_aBbzdHPCnlYa?dl=0>.

4 
Table S5. Variables used in the development of the survey-based women’s empowerment index.
Variable Code or unit
Beating justified if:
1. wife goes out without telling husband
2. wife neglects the children
Yes = -1; DK=0; No=1
3. wife argues with husband
4. wife refuses to have sex with husband
5. wife burns the food
6. Frequency of reading newspaper or magazine Not at all=0;
<once a week=1;
≥once a week=2
7. Respondent worked in last 12 months No = 0;
In the past year = 1;
Have a job, but on leave last 7 days = 2;
Currently working = 2
8. Woman education in completed years of schooling Years
9. Education difference: woman’s minus husband’s years of Years
schooling
10. Age difference: woman’s minus husband’s age Years
11. Age at first cohabitation Years
12. Age of woman at first birth* Years
Who usually decides on:
Husband/other alone= -1;
13. Respondent's health care
joint=0;
14. Large household purchases
respondent alone=1
15. Visits to family or relatives
* This variable was imputed for women who had not had a child, please see section 1.1 for details.

Where is the value of variables for each individual and are the variable weights, that
can be found in Table S6.

Table S6. Variable weights used in the equations for estimating individual scores for each
domain of the SWPER Index.
λv1 λv2 λv3
Attitude to Social Decision-
Variable (v) violence independence making
1. Beating not justified if wife goes out without telling
husband 0·489 -0·006 -0·001
2. Beating not justified if wife neglects the children 0·493 -0·020 -0·040
3. Beating not justified if wife argues with husband 0·501 0·000 0·007
4. Beating not justified if wife refuses to have sex with
husband 0·493 0·000 0·026
5. Beating not justified if wife burns the food 0·546 -0·003 -0·014
6. Frequency of reading newspaper or magazine 0·056 0·549 0·150
7. Woman education 0·015 0·090 0·026
8. Age of respondent at first birth -0·008 0·141 -0·019
9. Age at first cohabitation -0·004 0·131 -0·006
10. Age difference: woman’s minus husband’s age 0·002 0·026 0·012
11. Education difference: woman’s minus husband’s
years of schooling -0·004 0·050 -0·009
12. Who usually decides on respondent's health care 0·008 0·004 0·770
13. Who usually decides on large household purchases -0·034 -0·013 0·831
14. Who usually decides on visits to family or relatives 0·008 -0·052 0·768
15. Respondent worked in last 12 months -0·001 -0·060 0·180

5