Geographical variation and determinants of women unemployment status in Ethiopia; A multilevel and spatial analysis from 2016 Ethiopia Demographic and Health Survey data.

BACKGROUND: Unemployment is a major problem in both developed and developing countries. In Ethiopia, women unemployment is particularly high, and this makes it a grave socio-economic concern. The aim of this study is to assess the spatial distribution and identify the determinant factors of women unemployment in Ethiopia.
METHODS: The data used for the study is the Ethiopian Demographic and Health Surveys of 2016. A total of 15683 women are involved in the study. Global Moran's I statistic and Poisson-based purely spatial scan statistics are employed to explore spatial patterns and detect spatial clusters of women unemployment, respectively. To identify factors associated with women unemployment, multilevel logistic regression model is used.
RESULTS: A spatial analysis showed that there was a major spatial difference in women unemployment in Ethiopia with Global Moran's index value of 0.3 (p<0.001). The spatial distribution of women's unemployment varied significantly across the country. The major areas of unemployment were Afar and Somalia; southwest Tigray; North and west Oromia, and Eastern and southern parts of Amhara. Women with primary level of education(AOR = 0.88, 95%CI: 0.80, 0.98), secondary and above level of education (AOR = 0.71, 95%CI: 0.62, 0.82), women with rich wealth index (AOR = 0.79, 95% CI: 0.70, 0.90), pregnant women (AOR = 1.24, 95% CI: 1.06, 1.5), women with a male household head(AOR = 1.4, 95% CI: 1.28, 1.50), and urban women(AOR = 0.60, 95% CI: 0.50, 0.70) statistically associated with women unemployment.
CONCLUSION: The unemployment rate of women in Ethiopia showed variation across different clusters. Improving entrepreneurship and women's education, sharing business experiences, supporting entrepreneurs are potential tools for reducing the unemployment women. Moreover, creating community-based programs that prioritize participation of poor households and rural women as well as improving their access to mass media and the labor market is crucial.

Introduction

Unemployment is a critical issue the world is facing. Nonetheless, the effect and severity of unemployment varies between men and women. Globally, the rate of unemployment is higher among women than it is among men. The quality of employment is also disparate among men and women [1-3].Women, for instance, account to 38.8 percent of all labor force participants. However, women rather than spend more time on unpaid work such as childcare and housework. Compared to the 41 million (1.5 percent) of men, 606 million (21.7 percent) women offer full-time unpaid care globally, unpaid labor takes women 4 hours and 22 minutes per day, while for males, it takes only 2 hours and 15 minutes [4,5].

Unemployment not only causes economic uncertainty but also numerous social problems such as violence, human capital erosion, poverty and civil conflict [6]. It further causes social problems such as desperation, anger, violence and the eventual drift of unemployed women into all forms of criminal behavior [7-9]. High women unemployment cause HIV/AIDS spread in developing countries [10]. Long-term unemployment contributes more to financial deprivation, hunger, insecurity, violence, discontent, family conflict, social loneliness, lack of trust and self-esteem which eventually results in the deterioration of a stable community [10]. Moreover, unemployment is a waste of scarce resources since it results in a loss of potential national output [11,12].

Although Ethiopia is among the fast-growing countries in Africa, it has not been able to make effective use of the workforce necessary for sustaining economic development. The government didn’t create adequate job opportunities that accommodate future labor force through adopting a viable workforce policy. Consequently, women in Ethiopia not only lack the economic opportunities that allow them engage in alternative income-generating activities but they also lack alternative income sources that left them dependent on their spouse, and non participant party in decision-making of their household [3,13]. As a result of the numerous barriers they face in different aspects of their life, many Ethiopian women migrate to Arab and European countries [14,15]. The migration causes them several physical and mental health disorders. Various reports indicate that migrants are victims of fraud, forced labor, physical, sexual and psychological harassment by their employers, or smugglers. Many among the respondents experience psychological barrier [16,17].

Unemployment is hence a critical and pressing agenda for the Ethiopian Government. Although the Government commits itself to achieving the objectives of reducing unemployment, no significant change is noticed with respect to the reducing unemployment, and the aggregate job growth performance has remained slow.

Prior studies investigated the determinants of women’s unemployment in small urban areas [1,10,18], though only on individual-level factors [1,10,18]. At the community-level, unemployment among women is affected by factors, such as residence, media exposure, region and cluster (enumeration area) [19,20]. In addition, recent studies that employed conventional logistic regression that disregarded clustering effects investigated the rate of women’s unemployment and predicted that women’s individual characteristics associated with their unemployment. Yet, observations within a cluster tend to be more similar to observations, and an analysis that neglected this remains inadequate. Disregarding clustering in analysis may lead to overstating or understating the accuracy of the results, and determinant variables may be incorrectly reported as important. Thus, considering the shortcomings of existing studies, this study identified factors associated with women’s unemployment using a multi-level logistic regression model that considers the association between responses of interest to respondents from within the same cluster [21,22]. Spatial analyses helped locate hotspots and information to decision makers on strategic planning. The aim of this study is, therefore, to assess the spatial distribution and identify the determinant factors of women unemployment by adopting a multilevel model.

Materials and methods

Source of data

Data from the Ethiopian Demographic and Health Surveys 2016 were used for this analysis. The EDHS 2016 survey was organized to allow estimates of key indicators for the country as a whole, for urban and rural areas separately, and for each of the nine regions and the two administrative cities. Each region was stratified into urban and rural areas, yielding 21 sample strata. Samples of Enumeration Areas (EAs) were selected independently in two stages in each stratum. In the first stage, based on the 2007 PHC, an independent selection was implemented in each sampling stratum involving a total of 645 EAs (202 in urban and 443in rural) areas with probability proportional to EA size. In the second stage, a fixed number of 28 households per cluster were selected through an equal probability systematic selection from the newly created household listings.

Outcome variable

Unemployment status of women was the dependent variable containing two categories: unemployed women and employed women. According to ILO’s definition, people who are simultaneously “without work”, “currently available for work” and “seeking work” are considered as unemployed. Thus, the outcome for the i woman is represented by a random variable Y with two possible values coded as 1 and 0. In view of this, the outcome of the i woman Y was measured as a dichotomous variable.

Data management and statistical analysis

Data were extracted and re-categorized using SPSS version 21 software.Spatial analyses were performed using Geoda V.1.8.10 (geode center.github.ib), QGIS V.2.18.0 (qgis.org) and Arch GIS software V.10.1 (arcgis.com), and base files of the administrative regions for Ethiopia were obtained from DIVA (diva-gis.org). Multilevel logistic regression model was made using R software version 3.5.3.

Spatial analysis

Spatial analysis was conducted by joining the occurrence of unemployment (as proportions) with each cluster to the corresponding geospatial location (survey cluster values). The values of Demographic Health Survey data were merged with the geographic positioning system (GPS) dataset in Geoda software, and the values were imported into the QGIS software. Proportions of unemployment were then computed at lower (cluster), zonal, and regional levels using QGIS.” [23]. The spatial patterns of the rate of unemployment among women were visualized, and a spatially smooth proportion was obtained through empirical Bayes estimation method. The smoothed proportions presented clearer patterns in contexts where the problem was most severe. The spatial empirical Bayes ‘smooth’ estimates technique was applied for the spatial heterogeneity. The estimation technique guarantees that estimates in the neighboring states are more alike than estimates in states that are further away [24].Getis-OrdGi* statistics was used for this spatial analysis. Local Getis-OrdGi* statistics helped identify the hot and cold spotted areas for unemployment in women using GPS latitude and longitude coordinate readings that were taken at the nearest community center for EAs or EDHS 2016 clusters [25].Using Kuldorff’sSaTScan version 9.6 software, we used a Bernoulli model spatial Scan statistics to know the locations of statistically significant clusters for unemployment rate [23]. This model is used when unemployment are in the study area, as is the case for employed as controls, because of the scanning window which moves across the area. It was possible to detect small and large clusters which contained more than themaximum limit with the circular shape of the window, using the default maximum spatial cluster size of less than half the population. P- Values and log-likelihood ratios

Multilevel logistic regression analysis

The 2016 EDHS data is hierarchical in its structure andwomen are nested within enumeration area (communities). Hence, considering the hierarchical nature of the data, multilevel logistic regression models were applied to identify factors associated with women unemployment [26].The log of the probability of women unemployment was modeled using a two-level multilevel model as follows:

Where, i and j are the level 1 (individual) and level 2 (community) units, respectively; X and Z refer to individual and community-level variables, respectively; π is the probability of unemployment for the i women in the j community; and the β indicates the fixed coefficients. β0 is the intercept-the effect on the probability of women unemployment in the absence of influence of predictors; u showed the random effect (effect of the community on women unemployment for the j community, and e showed random errors at the individual level. Assuming each community had different intercepts (β0) and fixed coefficient (β), the hierarchical (clustered) nature of the data and the within and between community variations were taken into account. Four models were fitted to identify community and individual level factors associated with women unemployment. The first model (Model 1 or empty model) contained no explanatory variables. Instead, it was fitted to decompose the total variance into its individual- and community-level components. Individual-level factors were incorporated in the second model. In the third model, house hold level factors were included. In the fourth model, community-level factors were included. Finally, individual and community-level factors were included in the fourth model. Model comparison was made using deviance information criteria (DIC), Akaike’s Information Criterion (AIC), and Bayesian’s Information Criterion (BIC). The model with the smallest value of the information criterion was selected as the final model of the analysis [27]. For the result of fixed effect, odds ratio (ORs) with 95% confidence intervals (CIs) determined the statistical significance. The P-value of ≤ 0.05 is considered statistically significant.The measures of variation (random-effects) were summarized using ICC, Median Odds Ratio (MOR) and proportional change in variance (PCV) to measure the variation between enumeration areas (clusters). ICC, which is a measure of within-cluster and variation between individuals within the same cluster was calculated using the formula: , where V is the estimated variance in each model described elsewhere [28]. The total variation attributed to individual or/and community level factors at each model was measured with a proportional change in variance (PCV)calculated as: , where VA = variance of the initial model, and VB = variance of the model with more terms [28]. The MOR is the median odds ratio between a person with a higher propensity and a person with a lower propensity that compares two people from two different randomly chosen clusters and measures unexplained cluster heterogeneity as well as variation between clusters by comparing two people from two different randomly chosen clusters. It was calculated with the following formula:

, where V is the cluster level variance [28,29]. The MOR measure is always greater than or equal to 1. If the MOR is 1, there is no variation between clusters [30].

Ethical consideration

Publicly available EDHS 2016 data were used for this study. Informed consent was taken from each participant, and all identifiers were removed.

Results

Socio-demographic characteristics of study participants

A total of 15,683 women were included in the study. An unemployment rate of about 29.8% was observed among uneducated women. 48.9% of unemployment occurred in women with secondary education and above. Less than half (45.8%) of the unemployment were attributed to rich women. The unemployment of urban and rural women was 50.7% and 28.7%, respectively. Married women (32.5%) had the lowest unemployment. While the unemployment rate of uneducated husbands was 24.4%, husbands with educational level of secondary and above had an unemployment rate of 42.7%. The rich women had an unemployment rate of 45.8% ().

Incremental spatial autocorrelation

A peak in the graph denotes the distance at which the clustering is most pronounced. The color of each point on the graph corresponds to the statistical significance of the z-score values. The incremental spatial autocorrelation demonstrated that with 10 distance bands beginning at 2 km, women unemployment clustering was detected at 2.94 km distance. Statistically, a significant z-score (10.55) indicates that spatial clustering ofwomen unemployment was most pronounced at 2.94 Km distance (Fig 1).

Fig 1

Incremental spatial autocorrelationof women unemployment status in Ethiopia, 2016 EDHS.

Spatial pattern of women unemployment

In Ethiopia, the spatial distribution of women’s unemploymentwas spatially clustered with Global Moran’s I = 0.064298 (z-score = 16.375386, P-value <0.0001). This demonstrated that spatial hotspot and coldspot clustering was identified in Ethiopian regions. Given the z-score of 16.375386, there was less than a 1% chance that this high-clustered pattern was the result of randomchance. The bright red and blue colors on the tails indicate a higher level of significance (Fig 2).

Fig 2

Spatial autocorrelation analysis of women unemployment status in Ethiopia, 2016 EDHS.

Hot spot (Getis-Ord Gi*) analysis

Fig 3 summarizes a hot and cold spot analysis of the risk locations for women’s unemployment in Ethiopia. Beneshagul and Gambela, Oromia, WesternAmhara, and northeren SNNPR are the regions with the highest unemployment rates. Tigray, Afar, Dire Dawa, Hariri, northern Amhara, and Somalia on the other hand, have been designated as cold-spot (low risk) regions (Fig 3).

Fig 3

Hot spot and cold spot identification of unemployment women in Ethiopia, 2016 EDHS (Source of shapefiles: https://africaopendata.org/dataset/ethiopia-shapefiles).

Spatial SaTScan analysis of unemployment women across region

Most likely (primary clusters) and secondary clusters of unemployment women were identified. In 2016 EDHS, spatial scan statistics identified a total of high and modest performing spatial clusters of unemployment women. A total of 164 significant clusters were identified. Of these, among the significant clusters, 120 were most likely (primary cluster) and the other 44 were secondary.The spatial window of the primary cluster was located at Tigray, Amhara, southern Gambela,eastern Oromia and Afar which was centered at (12.376936 N,38.357984 E) / 318.07 km, RR = 1.43, and log likelihood ratio (LLR) of 200.78 at p-value < 0.0001. It showed that house holds in the spatial window had 1.43 times higher unemployment women ratethan those outside the window (S1 Table and Fig 4).

Fig 4

SaTScanAnalysisof unemployment women in Ethiopia, 2016 EDHS (Source of shapefiles: https://africaopendata.org/dataset/ethiopia-shapefiles).

Spatial interpolation

The spatial kriging interpolation analysis predicted high risk regions for women Unemployment in Ethiopia. Predication of high risk areas were indicated by Blue color. Beneshagul, western and centeral Oromia regional state were predicted as more risky areas compared to other regions. women in this areas were endangered to unemployment in Ethiopia. In other hand; women in Somalia, Afar, Eastern Amhara, Dire Dawa, central Tigray and eastern Oromia were identified as vulnerable to poor unemployment in Ethiopia (Fig 5).

Fig 5

Kriging interpolation of unemployment women in Ethiopia,2016 EDHS (Source of shapefiles: https://africaopendata.org/dataset/ethiopia-shapefiles).

Factors associated with women unemployment in Ethiopia

Multivariable-multilevel logistic analysis revealed that residence, age of women, marital status, women education level, husband/partners level of education, husband/partners occupation, sex of the household’s head, wealth index, family size, region and pregnancy were statistically significant factors for women unemployment. Women with a primary level of education were 0.88 (AOR = 0.88, 95%CI: 0.80, 0.98) times less likely to be unemployed than women who were not educated. Women whose level of education was secondary and above were 0.71 (AOR = 0.71, 95%CI: 0.62, 0.82) times less likely to be unemployed than women who have no education. Women whose husbands had a primary education level were 0.82 (AOR = 0.82, 95%CI: 0.73, 0.92) times less likely to be unemployed than women whose husbands were not educated. Rich women were 0.79 (AOR = 0.79, 95% CI: 0.70, 0.90) times less likely to be unemployed than poor women. Compared with women aged less than 25 years, the likelihood of unemployment among women aged 25–34 years was 0.50 (AOR = 0.50, 95% CI: 0.44, 0.54) times lower. The likelihood of unemployment for a family whose size was 6–10 was 1.2 (AOR = 1.2, 95% CI: 1.11,1.31) times higher than those whose family size was less than 6. Women who are already pregnant were1.24 (AOR = 1.24, 95% CI: 1.06, 1.5) times more likely to be unemployed than women who were not pregnant. Women living in households headed by men were 1.40 (AOR = 1.4, 95% CI: 1.28,1.50) times more likely to be unemployed than women living in households headed by women. Women with non-agricultural husbands were 0.47(AOR = 0.47, 95% CI: 0.30, 0.70) times less likely to be unemployed than women whose husbands are unemployed. The odds of unemployment for women living in Afar (AOR = 2.2, 95% CI: 1.62, 2.99), Amhara (AOR = 1.70, 95% CI: 1.30, 2.20), Somali(AOR = 1.90, 95% CI: 1.45, 2.60), and Dire Dawa(AOR = 1.44, 95% CI: 1.06, 1.96) was higher than women living in Tigray. Women living in Benishangule-Gumuz were 0.43 (AOR = 0.43, 95% CI: 0.32, 0.60) times less likely to be unemployed than women living in Tigray. Women living in urban areas were 0.60 (AOR = 0.60, 95% CI: 0.50, 0.70) times less likely to be unemployed than women living in rural areas (Table 2)

Table 2

Factors associated with women unemployment in Ethiopia, EDHS 2016.

Variable	Null	Model 1	Model 2	Model 3
	OR(95%CI)	OR(95%CI)	OR(95%CI)	OR(95%CI)
Marital status
Living alone		1		1
Married		1.09(0.96,1.25)		1.08(0.94,1.24)
Other		0.53(0.45,0.62)*		0.51(0.43,0.613)*
Mother education level
No Education		1		1
Primary		0.823(0.74,0.91)*		0.88(0.80,0.98)*
Secondary and above		0.65(0.57,0.74) *		0.71(0.62,0.82) *
Husband education level
No Education		1		1
Primary		0.78(0.691,0.88)*		0.82(0.73,0.92)*
Secondary and above		0.83(0.72,0.96)*		0.89(0.77,1.03)
Wealth index
Poor		1		1
Middle		0.88(0.77,1.01)		0.95(0.83,1.09)
Rich		0.64(0.57,0.72)*		0.79(0.70,0.91)*
Drug addiction
No		1		1
Yes		0.96(0.87,1.05)		0.97(0.88,1.06)
Age of women in year
15–24		1		1
25–34		0.459(0.414,0.5)*		0.5(0.44,0.54)*
34–49		0.457(0.41,0.51)*		0.49(0.44,0.62)*
Family size
2–5		1		1
6–10		1.24(1.14,1.34)*		1.20(1.11,1.31)*
>10		1.57(1.21,2.02)*		1.51(1.13,1.9)**
Child under age of 5 years
No		1		1
Yes		1.51(0.82,8.20)		1.62(0.34,88)
Sex of household head
Female		1		1
Male		1.39(1.27,1.27)*		1.40(1.28,1.51)*
Pregnancy
No		1		1
Yes		1.27(1.08,1.48)**		1.24(1.06,1.50)*
Migration status
Visitors		1		1
Usual Residence		0.95(0.75,1.22)		0.97(0.76,1.23)
Husbands occupation
Not Working		1		1
Agric Employee		0.45(0.3,0.7)*		0.47(0.31,0.70) *
Non-Agric Employee		0.6(0.5,0.72) *		0.63(0.52,0.71) *
Region
Tigray			1	1
Afar			2.70(2.00,3.71) *	2.21(1.62,2.99) *
Amhara			1.67(1.26,2.23) *	1.70(1.31,2.22) *
Oromia			1.06(0.81,1.38)	0.95(0.74,1.23)
Somali			2.52(1.9,3.36) *	1.92(1.45,2.61) *
Benishangul-Gumuz			0.49(0.37,0.66)*	0.43(0.32,0.6) *
SNNP			1.02(0.78,1.33)	0.921(0.71,1.19)
Gambela			0.964(0.72,1.3)	0.91(0.72,1.21)
Harari			1.18(0.86,1.61)	1.10(0.81,1.53)
Dire Dawa			1.56 (1.14,2.14)*	1.44(1.06,1.96) *
Addis Ababa			0.81(0.6,1.094)	0.80(0.62,1.04)
Residence
Rural			1	1
Urban			0.37(0.32,0.44)*	0.63(0.52,0.71) *
Exposed to Mass media
No			1	1
Yes			1.14(0.99,1.30)	1.12(0.95,1.27)

1 reference category for categorical variable and * reference P-value < 0.0001.

Random effect analysis (Measures of variation)

The findings of the random logistic regression analysis are summarized in Table 3. The empty model (Model I) shows that there are discrepancies in the community’s unemployment rates for women. Women’s unemployment rates varied by roughly 21.48 percent due to community-level factors (ICC = 22.48 percent). In the null model, women’s unemployment had the highest MOR value (2.46), showing that there was variation between communities clustered as MOR was 2.46 times more than the comparison (MOR = 1). Furthermore, the PCV of the whole model (model IV) revealed that individual and community factors accounted for roughly 59 percent of the difference in women’s unemployment among populations. The unexplained community variance in women unemployment was decreased to MOR of 1.78 when all factors were incorporated to the null model. When all factors are taken into account, the effect of clustering remains statistically significant in the overall model (Table 3).

Table 3

Measures of variation and model fit statistics on women unemployment in Ethiopia.

Measures of variations	Model 1	Model 2	Model 3	Model 4
Variance	0.90(0.77, 1.05)*	0.63(0.51, 0.70)*	0.42(0.35,0.50)*	0.37(0.30,0.45)*
ICC (%)	21.48	16.07	11.32	10.11
PCV (%)	Reference	30.00	53.33	58.89
MOR	2.46	2.13	1.85	1.78
Model fit statistics
DIC (-2log likelihood)	19127.6	18245.9	18778.122	18033.86
AIC	19131.59	18287.9	18806.12	18099.86
BIC	19146.91	18448.77	18913.37	18352.65

*reference P-value < 0.0001.

Discussion

In Ethiopia, approximately 63.8 percent of women are unemployed, with rates varying by location. The distribution of women’s jobless status is clustered, showing that it is not random in Ethiopia, according to the spatial analysis. The western, north-west, central and south-western regions of Ethiopia are identified as high-risk (hot spot) areas for women’s unemployment rates, whereas the northern, south eastern, and north easternares are identified as low-risk areas. In the regions of Tigray, Amhara, southern Gambela,eastern Oromia and Afar, SaTscan identified significant primary clusters (most likely clusters).This could be because the bulk of people living in the this area are pastoralists. A similar result was found in Ethiopia [19] and Turkey [20]. Ethiopia’s unemployment rate depends on geographic location and gender, as well as public policies and directions to combat unemployment and its’ harmful consequences. Although Ethiopia’s economy has shown impressive declines in unemployment, women have not benefited as much. In fact, women have higher unemployment rates [31-33].

Individual and community-level factors accounted for around 58.89 percent of the variance in women’s unemployment, according to the results of the random effects model. In contrast to illiterate women, educated women were less likely to be jobless. Women with educated spouses had a higher rate of unemployment than women with uneducated husbands. Previous findings [7,19,31,32] are supported by this. People with a higher level of education have a better chance of landing a good job. Furthermore, it empowers women to make their own decisions, to be accepted by their family and society, and to have more job prospects. Unemployment was lower among pregnant women than among non-pregnant women. This finding is consistent with earlier research [1,19,33], which found that pregnant women were less likely than non-pregnant women to work in a given year.

Women who lived alone were more likely than widowed or divorced women to be unemployed. A study in Ethiopia’s Harari Region [1], Ethiopia [19,32], and Pakistan’s Sahiwal District [2] supports this conclusion. Female-headed families had a lower unemployment rate than male-headed households. A research report from Halaba town, in Southern Ethiopia (SNNPR) [32] supports this conclusion. Women’s increased participation in household development activities may be to blame for these phenomena.

Women from wealthy families were less likely than those from poor families to be unemployed. This finding is in line with earlier research from Ethiopia [18,19], South Africa [34], and Ghana [35], all of which found that women from low-income households had the greatest unemployment rate. This could be because women from higher-income families may have better job-searching resources or access to the first capital needed to start their own business.

Women in urban were less likely than those in rural areas to be unemployed. This conclusion is supported by research from Ethiopia [19] and South Africa [36]. This could be due to the high shadow value of home production activities among women in rural areas. Unemployment was higher among women under the age of 25 than among older women. This finding is consistent with a study done in Urban Districts, Harari Region, Ethiopia [1], Halaba town, SNNPR, Southern Ethiopia [34], Ethiopia [19] and Sahiwal District, Pakistan [2] which found that women in the youngest age group studied are most influenced by unemployment.

Women’s unemployment rates were greater in households with bigger family sizes than in homes with smaller families. This finding is consistent with findings from studies conducted in Harari Region, Ethiopia [1], Halaba Town, SNNPR, Southern Ethiopia [32], Bahir Dar City, Northwest, Ethiopia [18], and Sahiwal District, Pakistan [2], which found that women with families of five or more were more likely to be unemployed than women with families of less than five. The unemployment rate of women whose husband has a non-agricultural employee was less than women whose husband was unemployed. Women whose husband has an agricultural employment were less likely to be unemployed than women whose husband was unemployed. This is supported by findings from other studies in Ethiopia [1] and Spain [35].

Furthermore, women’s unemployment was significantly influenced by their geographic location. Women in the Afar, Amhara, Somalia, and Dire Dawa areas had greater unemployment rates than women in the Tigray region. The findings of this study are consistent with those of earlier Ethiopian studies [19]. Women who have access to the media have a lower chance of becoming unemployed than women who do not. This finding is backed up by findings from a research report conducted in Halaba, Southern Ethiopia [32].

Conclusion

The study revealed that about 63.8% of Ethiopian women are unemployed and the unemployment rate varied across the regions. The spatial analysis indicated that the distribution of women’s unemployment status is clustered indicating that it was not random. High-risk (hot spot) areas for women’s unemployment rates were in the eastern, north-eastern and south-eastern parts of Ethiopia, while low-risk areas were in the northern, southern and western Ethiopia. Residence, age, marital status, educational level, husband’s education level, husband occupation, sex of household head, wealth index, family size, region and pregnancy were statistically important factors affecting women’s unemployment. The unemployment rate of women in Ethiopia differed from cluster to cluster. Hence, improving entrepreneurship and women’s education, sharing business experiences, supporting entrepreneurs could be useful measures to reduce women’s unemployment. Furthermore, community-based programs that prioritize the participation of poor women and improve their access to the media and the labor market need to be developed.

Significant SaTScan spatial scan clusters for unemployment women across region in Ethiopia, 2016 EDHS.

(DOCX)

Click here for additional data file.

28 Jul 2021

PONE-D-21-08056

Geographical variation and determinants of women unemployment status in Ethiopia; A multilevel and spatial analysis from 2016 Ethiopia Demographic and Health Survey data

PLOS ONE

Dear Dr. Fenta,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Sep 09 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see:  http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at  https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Ceyhun Elgin

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

You can read the reviewer letters below. As you can see one reviewer is positive and suggests some minor improvements, whereas another reviewer is more skeptical of what you are doing and recommended a rejection. I am giving you a chance to revise your paper (it is a major revision though), however, for this paper to be accepted; you have to convince the reviewer 2.

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please amend either the abstract on the online submission form (via Edit Submission) or the abstract in the manuscript so that they are identical.

3. Your ethics statement should only appear in the Methods section of your manuscript. If your ethics statement is written in any section besides the Methods, please move it to the Methods section and delete it from any other section. Please ensure that your ethics statement is included in your manuscript, as the ethics statement entered into the online submission form will not be published alongside your manuscript.

4. We note that Figures 4.1,4.2 and 4.3 in your submission contain map images which may be copyrighted. All PLOS content is published under the Creative Commons Attribution License (CC BY 4.0), which means that the manuscript, images, and Supporting Information files will be freely available online, and any third party is permitted to access, download, copy, distribute, and use these materials in any way, even commercially, with proper attribution. For these reasons, we cannot publish previously copyrighted maps or satellite images created using proprietary data, such as Google software (Google Maps, Street View, and Earth). For more information, see our copyright guidelines: http://journals.plos.org/plosone/s/licenses-and-copyright.

We require you to either (1) present written permission from the copyright holder to publish these figures specifically under the CC BY 4.0 license, or (2) remove the figures from your submission:

A. You may seek permission from the original copyright holder of Figures 4.1,4.2 and 4.3 to publish the content specifically under the CC BY 4.0 license.

We recommend that you contact the original copyright holder with the Content Permission Form (http://journals.plos.org/plosone/s/file?id=7c09/content-permission-form.pdf) and the following text:

“I request permission for the open-access journal PLOS ONE to publish XXX under the Creative Commons Attribution License (CCAL) CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). Please be aware that this license allows unrestricted use and distribution, even commercially, by third parties. Please reply and provide explicit written permission to publish XXX under a CC BY license and complete the attached form.”

Please upload the completed Content Permission Form or other proof of granted permissions as an "Other" file with your submission.

In the figure caption of the copyrighted figure, please include the following text: “Reprinted from [ref] under a CC BY license, with permission from [name of publisher], original copyright [original copyright year].”

B. If you are unable to obtain permission from the original copyright holder to publish these figures under the CC BY 4.0 license or if the copyright holder’s requirements are incompatible with the CC BY 4.0 license, please either i) remove the figure or ii) supply a replacement figure that complies with the CC BY 4.0 license. Please check copyright information on all replacement figures and update the figure caption with source information. If applicable, please specify in the figure caption text when a figure is similar but not identical to the original image and is therefore for illustrative purposes only.

The following resources for replacing copyrighted map figures may be helpful:

USGS National Map Viewer (public domain): http://viewer.nationalmap.gov/viewer/

The Gateway to Astronaut Photography of Earth (public domain): http://eol.jsc.nasa.gov/sseop/clickmap/

Maps at the CIA (public domain): https://www.cia.gov/library/publications/the-world-factbook/index.html and https://www.cia.gov/library/publications/cia-maps-publications/index.html

NASA Earth Observatory (public domain): http://earthobservatory.nasa.gov/

Landsat: http://landsat.visibleearth.nasa.gov/

USGS EROS (Earth Resources Observatory and Science (EROS) Center) (public domain): http://eros.usgs.gov/#

Natural Earth (public domain): http://www.naturalearthdata.com/

5. Thank you for submitting the above manuscript to PLOS ONE. During our internal evaluation of the manuscript, we found significant text overlap between your submission and the following previously published works.

- https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-019-7529-z

- https://bmjopen.bmj.com/content/9/4/e027276

- https://pubmed.ncbi.nlm.nih.gov/24411023/

We would like to make you aware that copying extracts from previous publications, especially outside the methods section, word-for-word is unacceptable. In addition, the reproduction of text from published reports has implications for the copyright that may apply to the publications.

Please revise the manuscript to rephrase the duplicated text, cite your sources, and provide details as to how the current manuscript advances on previous work. Please note that further consideration is dependent on the submission of a manuscript that addresses these concerns about the overlap in text with published work.

We will carefully review your manuscript upon resubmission, so please ensure that your revision is thorough.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I am glad to review and assess this interesting article, entitled, Geographical variation and determinants of women unemployment status in Ethiopia; A multilevel and spatial analysis from 2016 Ethiopia Demographic and Health Survey data. The organization of this article is good and satisfactory. The Introduction part and methodology portions are adequate. I suggest the authors improve the Introduction section by adding some latest articles' citations to enhance the work quality.

Overall, the manuscript is a good piece of work. I recommend that authors do a little more work and add the latest literature to support the study, as suggested. The English level is not good and smooth, e.g., the language standard, specifically the grammar, of not sufficient quality to meet scientific merit for publication. I accept this manuscript after minor revision, as I have recommended.

Reviewer #2: This work contains two simple empirical exercises about female unemployment in Ethiopia.

In the first one, the authors routinely use standard spatial techniques to detect clusters of areas with low or high unemployment. While technically well carried out (mainly because it is simply implemented in usual software), results are barely commented and it is particularly unclear the contribution of this analysis. The potential contribution should be focused on commenting about spatial correlation. To start with, both Figure 4.2 and 4.3 are barely commented. The terms used in Figure 4.2 are confusing: "High Cluster" can also apply to what the authors call "High outlier", because (1) both are clusters (but in the latter case this is not mentioned), (2) the term outlier is not appropriate because being a high-low area is NOT weird (=outlier), but a simple indication of negative spatial correlation. If the authors could just READ their software output (labels in the picture their software provided), they could find the standard term: high-high cluster and high-low cluster. In any case, authors just provide a purely descriptive minimum comments on the areas with high or low unemployment incidence, but no discussion at all about the implications or how these results can be useful. Commenting that female unemployment is different in different clusters is very vague and can be concluded from any other study. And the general recommendations in the last two sentences of the Conclusions are NOT directly supported by the analysis.

Therefore, it is unclear the value added by the spatial analysis. The authors do not seem to feel that it was useful, because the second part of their work does not use the information about spatial correlation among clusters (enumeration areas). They just apply a multilevel logistic model, but they do not seem to take into account spatial correlation to compute standard errors (and therefore confidence interval). It is, of course, important to account for both individual level and community/clusters level factors to avoid biased results (i.e., as the authors mention, incorrectly concluding that a factor is significant or not). But the spatial correlation analysis is not needed to account for this. If one finds that spatial correlation exists, this should be incorporated into the subsequent analysis. In their current analysis, the authors allow for within-cluster (community) correlation, but assume no between-clusters correlation despite of their previous findings!

The (logistic) regression analysis is just a sophisticated but ultimately simple exercise about which variables are correlated with observed female unemployment. On the one hand, no really new result is found. The authors again present the results and just comment on the sign and significance of different factors, showing that previous papers have already found such results. However, no clear implication of these results are provided. On the other hand, results lack any causal interpretation since they are obviously plagued by endogeneity and self-selection problems (eg., labour market participation), Therefore, these correlation results are barely useful for policy interventions as the authors claim. For example, imagine that job opportunities in urban areas are lower for woman, while in rural areas some opportunities always exist. Then, most women in urban areas decide NOT to join to the labour force (they are not “currently available for work” and “seeking work”), but almost all of the few ones that join find a job (fill those few job opportunities); therefore, unemployment is low ( (“without work” AMONG those in the labour force: “currently available for work” and “seeking work”). The policy recommendation by authors is to focus on rural areas where the problem exists, while job opportunities are equally or less scarce in urban areas.

In summary, this work has a number of important shortcomings and its contribution to the literature is unclear. I do not feel that it reaches the standards of a scientific publication in an international journal.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

21 Jan 2022

First of all thanks for your worth and essential over all comments. On the very beginning I am glad for your nice reviewing and assessing this article, entitled, Geographical variation and determinants of women unemployment status in Ethiopia; A multilevel and spatial analysis from 2016 Ethiopia Demographic and Health Survey data.

Editor Comments :

Comment 1:Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

Response: Thank you for your important comments. We have followed all PLOS ONE's style requirements, including those for file naming.

Comment 2:Please amend either the abstract on the online submission form (via Edit Submission) or the abstract in the manuscript so that they are identical.

Response: Thank you for your suggestion. We correct the abstract in our manuscript to be identical.

Comment 3: Your ethics statement should only appear in the Methods section of your manuscript. If your ethics statement is written in any section besides the Methods, please move it to the Methods section and delete it from any other section. Please ensure that your ethics statement is included in your manuscript, as the ethics statement entered into the online submission form will not be published alongside your manuscript.

Response: Thank you for your suggestion.we have incorporated the points that the ethical statement is only appear in the method section.

Comment 4: We note that Figures 4.1,4.2 and 4.3 in your submission contain map images which may be copyrighted. All PLOS content is published under the Creative Commons Attribution License (CC BY 4.0), which means that the manuscript, images, and Supporting Information files will be freely available online, and any third party is permitted to access, download, copy, distribute, and use these materials in any way, even commercially, with proper attribution. For these reasons, we cannot publish previously copyrighted maps or satellite images created using proprietary data, such as Google software (Google Maps, Street View, and Earth). For more information, see our copyright guidelines: http://journals.plos.org/plosone/s/licenses-and-copyright.

We require you to either (1) present written permission from the copyright holder to publish these figures specifically under the CC BY 4.0 license, or (2) remove the figures from your submission:

A. You may seek permission from the original copyright holder of Figures 4.1,4.2 and 4.3 to publish the content specifically under the CC BY 4.0 license.

We recommend that you contact the original copyright holder with the Content Permission Form (http://journals.plos.org/plosone/s/file?id=7c09/content-permission-form.pdf) and the following text:

“I request permission for the open-access journal PLOS ONE to publish XXX under the Creative Commons Attribution License (CCAL) CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). Please be aware that this license allows unrestricted use and distribution, even commercially, by third parties. Please reply and provide explicit written permission to publish XXX under a CC BY license and complete the attached form.”

Please upload the completed Content Permission Form or other proof of granted permissions as an "Other" file with your submission.

In the figure caption of the copyrighted figure, please include the following text: “Reprinted from [ref] under a CC BY license, with permission from [name of publisher], original copyright [original copyright year].”

B. If you are unable to obtain permission from the original copyright holder to publish these figures under the CC BY 4.0 license or if the copyright holder’s requirements are incompatible with the CC BY 4.0 license, please either i) remove the figure or ii) supply a replacement figure that complies with the CC BY 4.0 license. Please check copyright information on all replacement figures and update the figure caption with source information. If applicable, please specify in the figure caption text when a figure is similar but not identical to the original image and is therefore for illustrative purposes only.

The following resources for replacing copyrighted map figures may be helpful:

USGS National Map Viewer (public domain): http://viewer.nationalmap.gov/viewer/

The Gateway to Astronaut Photography of Earth (public domain): http://eol.jsc.nasa.gov/sseop/clickmap/

Maps at the CIA (public domain): https://www.cia.gov/library/publications/the-world-factbook/index.html and https://www.cia.gov/library/publications/cia-maps-publications/index.html

NASA Earth Observatory (public domain): http://earthobservatory.nasa.gov/

Landsat: http://landsat.visibleearth.nasa.gov/

USGS EROS (Earth Resources Observatory and Science (EROS) Center) (public domain): http://eros.usgs.gov/#

Natural Earth (public domain): http://www.naturalearthdata.com/

Response: Thank you for your suggestion. We incorporate your worth comments and the map images are not copyrighted, which are my work.

Comment 5: Thank you for submitting the above manuscript to PLOS ONE. During our internal evaluation of the manuscript, we found significant text overlap between your submission and the following previously published works.

- https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-019-7529-z

- https://bmjopen.bmj.com/content/9/4/e027276

- https://pubmed.ncbi.nlm.nih.gov/24411023/

We would like to make you aware that copying extracts from previous publications, especially outside the methods section, word-for-word is unacceptable. In addition, the reproduction of text from published reports has implications for the copyright that may apply to the publications.

Please revise the manuscript to rephrase the duplicated text, cite your sources, and provide details as to how the current manuscript advances on previous work. Please note that further consideration is dependent on the submission of a manuscript that addresses these concerns about the overlap in text with published work.

We will carefully review your manuscript upon resubmission, so please ensure that your revision is thorough.

Response: Thank you for your suggestion. We incorporate your worth comments and we revise and paraphrase all the manuscript in detail.

Reviewer #1:

Comment 1:I am glad to review and assess this interesting article, entitled, Geographical variation and determinants of women unemployment status in Ethiopia; A multilevel and spatial analysis from 2016 Ethiopia Demographic and Health Survey data. The organization of this article is good and satisfactory. The Introduction part and methodology portions are adequate. I suggest the authors improve the Introduction section by adding some latest articles' citations to enhance the work quality.

Overall, the manuscript is a good piece of work. I recommend that authors do a little more work and add the latest literature to support the study, as suggested. The English level is not good and smooth, e.g., the language standard, specifically the grammar, of not sufficient quality to meet scientific merit for publication. I accept this manuscript after minor revision, as I have recommended.

Response: Thank you for your suggestion and comments. This is a very essential comment, we have incorporated the points. We repeatedly read the whole document and consider all the corrections (adding the most recent citation and also incorporate the latest literature) and the language is edited by professionals.

Reviewer #2:

Comment1:This work contains two simple empirical exercises about female unemployment in Ethiopia.

In the first one, the authors routinely use standard spatial techniques to detect clusters of areas with low or high unemployment. While technically well carried out (mainly because it is simply implemented in usual software), results are barely commented and it is particularly unclear the contribution of this analysis. The potential contribution should be focused on commenting about spatial correlation. To start with, both Figure 4.2 and 4.3 are barely commented. The terms used in Figure 4.2 are confusing: "High Cluster" can also apply to what the authors call "High outlier", because (1) both are clusters (but in the latter case this is not mentioned), (2) the term outlier is not appropriate because being a high-low area is NOT weird (=outlier), but a simple indication of negative spatial correlation. If the authors could just READ their software output (labels in the picture their software provided), they could find the standard term: high-high cluster and high-low cluster. In any case, authors just provide a purely descriptive minimum comments on the areas with high or low unemployment incidence, but no discussion at all about the implications or how these results can be useful. Commenting that female unemployment is different in different clusters is very vague and can be concluded from any other study. And the general recommendations in the last two sentences of the Conclusions are NOT directly supported by the analysis.

Therefore, it is unclear the value added by the spatial analysis. The authors do not seem to feel that it was useful, because the second part of their work does not use the information about spatial correlation among clusters (enumeration areas). They just apply a multilevel logistic model, but they do not seem to take into account spatial correlation to compute standard errors (and therefore confidence interval). It is, of course, important to account for both individual level and community/clusters level factors to avoid biased results (i.e., as the authors mention, incorrectly concluding that a factor is significant or not). But the spatial correlation analysis is not needed to account for this. If one finds that spatial correlation exists, this should be incorporated into the subsequent analysis. In their current analysis, the authors allow for within-cluster (community) correlation, but assume no between-clusters correlation despite of their previous findings!

The (logistic) regression analysis is just a sophisticated but ultimately simple exercise about which variables are correlated with observed female unemployment. On the one hand, no really new result is found. The authors again present the results and just comment on the sign and significance of different factors, showing that previous papers have already found such results. However, no clear implication of these results are provided. On the other hand, results lack any causal interpretation since they are obviously plagued by endogeneity and self-selection problems (eg., labour market participation), Therefore, these correlation results are barely useful for policy interventions as the authors claim. For example, imagine that job opportunities in urban areas are lower for woman, while in rural areas some opportunities always exist. Then, most women in urban areas decide NOT to join to the labour force (they are not “currently available for work” and “seeking work”), but almost all of the few ones that join find a job (fill those few job opportunities); therefore, unemployment is low ( (“without work” AMONG those in the labour force: “currently available for work” and “seeking work”). The policy recommendation by authors is to focus on rural areas where the problem exists, while job opportunities are equally or less scarce in urban areas.

In summary, this work has a number of important shortcomings and its contribution to the literature is unclear. I do not feel that it reaches the standards of a scientific publication in an international journal.

Response: On the very beginning I am glad for your nice reviewing and assessing this article, entitled, Geographical variation and determinants of women unemployment status in Ethiopia; A multilevel and spatial analysis from 2016 Ethiopia Demographic and Health Survey data.

Then for your question on fig4.2 and fig4.3, I need to clarify that, initially the research routinely used to show the geographical variation/distribution of unemployment rate of women by using spatial auto-correlation. So, the Global Moran’s I Spatial auto-correlation indices measure the spatial dependence between values of the same variable(unemployment rate) in different places in space. Then we say that, Spatial auto-correlation is positive when similar values(high unemployment rate with high unemployment rate or low unemployment rate with low unemployment rate) of the variable to be studied are grouped geographically.Spatial auto-correlation is negative when the dissimilar values of the variable(high low or low high) to be studied come together geographically.

In other words, spatial dependence exists when statistical values are correlated. However, in our case Global Moran’s I values I=0.33 (p-value=0.001) , this can only explain the clustering(non-random) effects in a given enumeration areas without clearly highlighting the clustering regions. This method was most commonly used in testing global spatial auto-correlation. Often, our interest lies not only in determining whether the data as a whole exhibit spatial auto-correlation, but also, in identifying the specific observations that exhibit spatial auto-correlation with their neighbors.So,the study following local measures called local spatial auto-correlation indicators(LISA), which is a set of local indicators for inferring the scope of clustering regions (in our case enumeration areas). Once a significance level is set(if the global spatial auto-correlation is significant) , values can also be plotted on a map to display the specific locations of hot spots and potential outliers(dissimilar values). The results of LISA map are displayed on fig4.2 and fig4.3. Statistical significance tests are mainly performed to test the scope of clustering spatial elements relative to the entire scope, where a higher significance denotes that spatial clustering is more prominent.

However, in fig4.2 result we try to clarify the potential outliers areas on unemployment rate of women as:high outlier(high-low) refers to high proportion of women unemployment surrounded by low proportion of women unemployment.; low outlier(low-high)refers to low proportion of women unemployment surrounded by high proportion of women unemployment. And the rest was similar values.

And then, in fig4.3 ahot spot(high risk areas) and cold spot(low risk areas) analysis of the risk areas for women'sunemployment rate in Ethiopia has been summarized.

After all the spatial analysis result were discussed on the first paragraph in discussion part and concluded from any other study.

After all, concluding that the spatial distribution of women's unemployment status is non-random( I=0.33 (p-value=0.001)) at the national level and associated with neighboring values; and the data set taken for this study was from EDHS 2016. These data are hierarchical structure and surveys are obtained from nested sampling in heterogeneous subgroups or a sampling method was multistage stratified cluster sampling. For multistage clustered samples, the dependence among observations often comes from several levels of the hierarchy. The problem of dependencies between individual observations also occurs in survey research, where the sample is not taken randomly but cluster sampling from geographical areas is used instead. In this case, the use of single-level statistical models is no longer valid and reasonable. Hence, in order to draw appropriate inferences and conclusions from multistage stratified clustered survey data we may require tricky and complicated modeling techniques like multilevel modeling. That is why we use multilevel logistic regression model by take into account for lack of independence(correlation) across levels of nested data (i.e., individuals (women) nested within enumeration areas). This information also supported by the intra-class correlation coefficient (ICC) measures the proportion of variance in the outcome explained by the grouping structure. This ICC is an indication of the correlation of the women unemployment rate belonging to the same enumeration areas, i.e. it is an indication of the dependency of the unemployment rate among women within the enumeration areas. So,from the research analysis, About 21.48% of the variation in women's unemployment rate occurred due to community-level factors (ICC =21.48%).The MOR (2.46) value of women's unemployment was the largest in the null model, which showed that there was variation between communities clustered as MOR was 2.46 times higher than the comparison (MOR=1.78). In addition, the highest (58.89%) PCV in the full model (model IV) revealed that about 59%of the difference in women's unemployment across populations was due to both the individual and community-level factors. The unexplained community variation in women unemployment decreased to MOR of 1.78 when all factors were added to the null model (empty model). This indicates that when all factors are included, the effect of clustering is still statistically significant in the full model (Table 3).

Submitted filename: 2 nd revisoin plos.docx

Click here for additional data file.

25 Mar 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Hugh Cowley

Senior Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I am glad to review this exciting article . I directly recommend this study for publication . Fully satisfied

Reviewer #2: I appreciate that the authors have tried to clarify some of my previous comments. But my comments have not addressed no proper argument has been provided to justify their choices.

I do not see any contribution in finding spatial correlation in unemployment: I doubt that in any country one can find that unemployment is randomly distributed across regions (which are endogenously formed in a way clearly linked to unemployment). The authors' response has ignored my point about the incorrect use of the word "outliers" as the authors do: this is not a standard way to name that situation, a better one exists and it is confusing.

I do not see again that my comments on the second part of the paper (factors associated with women unemployment) have been addressed. As I said, this provides some mere correlations between factors and unemployment. This descriptive evidence can have some interest (this is an editorial choice), but it should be properly carried out AND it must be clear that NO CAUSAL interpretation should be concluded from it. On the one hand, if the authors carefully read my previous previous, I mention that STANDARD ERRORS should be account for clustering; I did not mention at all, as the authors focus on their reply, on using single-level or multilevel statistical models. If standard errors are not properly computed, results are not credible. On the other hand, the authors should be much clearer about the merely descriptive implications of their results. They cannot claim (as they implicitly and explicitly do) that a changing some factors (eg., improving entrepreneurship,etc.) could have an effect on unemployment: this a causal claim that cannot be derived from the analysis and this should be clear in a serious scientific paper.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

8 Apr 2022

First of all thanks for your worth and essential over all comments. On the very beginning I am glad for your nice reviewing and assessing this article, entitled, Geographical variation and determinants of women unemployment status in Ethiopia; A multilevel and spatial analysis from 2016 Ethiopia Demographic and Health Survey data.

Reviewer #1:I am glad to review this exciting article . I directly recommend this study for publication . Fully satisfied

Response: We authors owe you a great debt of gratitude.

Reviewer #2: I appreciate that the authors have tried to clarify some of my previous comments. But my comments have not addressed no proper argument has been provided to justify their choices. I do not see any contribution in finding spatial correlation in unemployment: I doubt that in any country one can find that unemployment is randomly distributed across regions (which are endogenously formed in a way clearly linked to unemployment). The authors' response has ignored my point about the incorrect use of the word "outliers" as the authors do: this is not a standard way to name that situation, a better one exists and it is confusing. I do not see again that my comments on the second part of the paper (factors associated with women unemployment) have been addressed. As I said, this provides some mere correlations between factors and unemployment. This descriptive evidence can have some interest (this is an editorial choice), but it should be properly carried out AND it must be clear that NO CAUSAL interpretation should be concluded from it. On the one hand, if the authors carefully read my previous previous, I mention that STANDARD ERRORS should be account for clustering; I did not mention at all, as the authors focus on their reply, on using single-level or multilevel statistical models. If standard errors are not properly computed, results are not credible. On the other hand, the authors should be much clearer about the merely descriptive implications of their results. They cannot claim (as they implicitly and explicitly do) that a changing some factors (eg., improving entrepreneurship,etc.) could have an effect on unemployment: this a causal claim that cannot be derived from the analysis and this should be clear in a serious scientific paper.

Response: At the outset, we want to thank you for taking the time to review and evaluate this article again.

The research is routinely used to demonstrate the initial goal of investigating the spatial structure of unemployment rate across different clusters to provide implications for policymakers, investigating the hot spots of unemployment rate, and showing a visual picture of unemployment rate. Mapping was used as a preliminary step in conducting a visual inspection for the unemployment rate. In addition, mapping is important in the monitoring of unemployed people. Maps can reveal spatial patterns that were previously unknown or unnoticed when examining a table of statistics, as well as high-risk communities or problem areas. The goal of spatial analysis is to find patterns in geographic data and try to explain them. Spatial autocorrelation is the term used for theinterdependence of the values of a variable over space.As a result, the most important contribution in determining spatial correlation/statistics in unemployment is not only interested in answering the "how much unemployment rate women in Ethiopia" question, but also the "how much is where" question. As stated in the first law of geography, "everything is related to everything else, but closer things are more related than distant things," the application of statistical techniques to spatial data faces significant challenges. Two spatial autocorrelation statistics based on sharing boundary neighbors, known as global and local Moran's I, were used to investigate global clustering and local clusters, respectively. Based on visual inspection of the mapping, global clustering was discovered in unemployment rate, which was confirmed by the significant statistic discovered by global Moran's I(Moran’s I values I=0.33 (p-value=0.001).

Our interest is frequently not only in determining whether the data as a whole exhibits spatial auto-correlation, but also in identifying specific observations that exhibit spatial auto-correlation with their neighbors. As a result, the study relied on local measures known as local spatial auto-correlation indicators (LISA), which are a collection of local indicators used to determine the scope of clustering regions (in our case enumeration areas). Once a significance level has been determined (if the global spatial auto-correlation is significant), values can be plotted on a map to show the precise locations of hot spots and potential outliers (dissimilar values) or A positive local Moran value indicates local stability, such as a cluster with a high/low unemployment rate surrounded by another cluster with a high/low unemployment rate. A negative local Moran value indicates local instability(high/low outlier ), such as a cluster with low unemployment surrounded by a cluster with high unemployment, or vice versa. Figures 2 and 3 show the results of the LISA map.

And ,Thank you for your suggestion and comments. This is a very essential comment, we have incorporated the points. We repeatedly read the whole document and consider all the corrections.

After all, concluding that the spatial distribution of women's unemployment status at the national level is non-random (I=0.33; p-value=0.001) and associated with neighboring values. The result indicates that there is spatial autocorrelation /dependency or the unemployment rate is interdependent across clusters. As a result, we employ a multilevel logistic regression model that accounts for a lack of independence (spatial correlation) across levels of nested data (i.e., individuals (women) nested within clusters). The spatial correlation result was also supported by the intra-class correlation coefficient (ICC), which measures the proportion of variance in the outcome explained by the grouping structure while accounting for variance across clusters. This ICC represents the correlation of the women's unemployment rate within the same enumeration areas, i.e. the dependency of the unemployment rate among women within the clusters.So,from the research analysis, About 21.48% of the variation in women's unemployment rate occurred due to community-level(due to clusters) factors (ICC =21.48%).

Submitted filename: Response_2 for plos one.docx

Click here for additional data file.

29 Apr 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Wen-Wei Sung, M.D., Ph.D.

Academic Editor

PLOS ONE

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #2: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: No

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #2: I do not see any real improvement with respect to the previous versions. None of my comments that the editor mentioned have been properly addressed.

I feel that a crucial point in any serious scientific paper as making clear that the results cannot be interpreted in a causal way CANNOT be addressed just by adding a "could be". Instead it should be explicitly mentioned the true purely descriptive scope of the results.

The authors do not only insist in using the term "outlier" in a unconventional sense, but they have now extended its use. Even more surprisingly they do not even attempt to provide a simple explanation about their choice, although I have mentioned this in my two previous reports. It is really a complete outlier the way in which the author insist in using the term "outlier" (positive correlation between two variables, say, high education and high income is never called a positive outlier). And the way in which the authors completely ignore this point (to accept it or to provide a convincing alternative explanation about their unusual way of using the term) is more outlier in peer review process.

I perfectly know that the interesting question is to show where the unemployment is located and how is related to nearby areas. I never raised this point that the authors insist in explaining while ignoring other more important and explicitly mentioned. However, the paper is more focused precisely in trivial and uninteresting things: to discussed the global Moran statistics that reveals the TRIVIAL point that unemployment is not uniform across regions (randomly distributed) as I already mentioned.

I perfectly know how to interpret the local spatial auto-correlation indicators. If the authors carefully read and understand this and my previous reports, they would see that the literature in this area NEVER calls OUTLIERS to what is simply positive or negative correlation amongst clusters (high/high, high/low, etc.). The term outliers is NOT used in this or any other literature to name dissimilar values (high/low); and, as I already mention, the standard output of statistical software already provides a proper term for this (and it is not outlier). Even more, if the authors were using outlier in the sense of dissimilar values (which is unconventional and not justified or explained by the authors), high-high and low-low cases are also identified as "outliers" but in this case the values are not dissimilar...

Finally the authors have completely ignored (again) my point about standard errors. Statistical modelling (i.e., using a multilevel model or not) is not the only relevant part of empirical results; and in any case it was never my point. In any empirical papers, researcher should be very careful about inference. And in this case it means to use proper standard errors. Otherwise, the results could be meaningless: some factors that the authors claim to be correlated to unemployment might actually be not significantly related to it. Once again the authors do not even mention why this crucial point is not even properly addressed.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

17 May 2022

Response to Reviewers

First of all thanks for your worth and essential over all comments. On the very beginning I am glad for your nice reviewing and assessing this article, entitled, Geographical variation and determinants of women unemployment status in Ethiopia; A multilevel and spatial analysis from 2016 Ethiopia Demographic and Health Survey data.

Reviewer #2: I do not see any real improvement with respect to the previous versions. None of my comments that the editor mentioned have been properly addressed.

I feel that a crucial point in any serious scientific paper as making clear that the results cannot be interpreted in a causal way CANNOT be addressed just by adding a "could be". Instead it should be explicitly mentioned the true purely descriptive scope of the results.

The authors do not only insist in using the term "outlier" in a unconventional sense, but they have now extended its use. Even more surprisingly they do not even attempt to provide a simple explanation about their choice, although I have mentioned this in my two previous reports. It is really a complete outlier the way in which the author insist in using the term "outlier" (positive correlation between two variables, say, high education and high income is never called a positive outlier). And the way in which the authors completely ignore this point (to accept it or to provide a convincing alternative explanation about their unusual way of using the term) is more outlier in peer review process.

I perfectly know that the interesting question is to show where the unemployment is located and how is related to nearby areas. I never raised this point that the authors insist in explaining while ignoring other more important and explicitly mentioned. However, the paper is more focused precisely in trivial and uninteresting things: to discussed the global Moran statistics that reveals the TRIVIAL point that unemployment is not uniform across regions (randomly distributed) as I already mentioned.

I perfectly know how to interpret the local spatial auto-correlation indicators. If the authors carefully read and understand this and my previous reports, they would see that the literature in this area NEVER calls OUTLIERS to what is simply positive or negative correlation amongst clusters (high/high, high/low, etc.). The term outliers is NOT used in this or any other literature to name dissimilar values (high/low); and, as I already mention, the standard output of statistical software already provides a proper term for this (and it is not outlier). Even more, if the authors were using outlier in the sense of dissimilar values (which is unconventional and not justified or explained by the authors), high-high and low-low cases are also identified as "outliers" but in this case the values are not dissimilar...

Finally the authors have completely ignored (again) my point about standard errors. Statistical modelling (i.e., using a multilevel model or not) is not the only relevant part of empirical results; and in any case it was never my point. In any empirical papers, researcher should be very careful about inference. And in this case it means to use proper standard errors. Otherwise, the results could be meaningless: some factors that the authors claim to be correlated to unemployment might actually be not significantly related to it. Once again the authors do not even mention why this crucial point is not even properly addressed..

Response: And ,Thank you for your suggestion and comments. This is a very essential comment, we have incorporated the points. We repeatedly read the whole document and consider all the corrections.

At the outset, we want to thank you for taking the time to review and evaluate this article again. The manuscript was revised (all the spatial out put was updated).

Your comment is clear but please understated the aim of this study.

This research is routinely used to assess the spatial distribution (varation) of unemployment rate across different clusters to provide implications for policymakers, investigating the hot spots of unemployment rate(the highest risk region/areas), and showing a visual picture of unemployment rate in the first section.

In this part we identify that:

Spatial pattern of women unemployment

In Ethiopia, the spatial distribution of women's unemploymentwas spatially clustered with Global Moran's I = 0.064298 (z-score = 16.375386, P-value <0.0001). This demonstrated that spatial hotspot and coldspot clustering was identified in Ethiopian regions. Given the z-score of 16.375386, there was less than a 1% chance that this high-clustered pattern was the result of randomchance. The bright red and blue colors on the tails indicate a higher level of significance (Fig. 2).

Then we find the hot spot also:

Hot spot (Getis-Ord Gi*) analysis

Figure3 summarizes a hot and cold spot analysis of the risk locations for women's unemployment in Ethiopia. Beneshagul and Gambela, Oromia, WesternAmhara, and northeren SNNPR are the regions with the highest unemployment rates. Tigray, Afar, Dire Dawa, Hariri, northern Amhara, and Somalia on the other hand, have been designated as cold-spot (low risk) regions (Fig 3).

And we can identify the most likely cluster by Spatial scan statistics:

Spatial SaTScan analysis of unemployment women across region

Most likely (primary clusters) and secondary clusters of unemployment women were identified. In 2016 EDHS, spatial scan statistics identified a total of high and modest performing spatial clusters of unemployment women. A total of 164 significant clusters were identified. Of these, among the significant clusters, 120 were most likely (primary cluster) and the other 44 were secondary.The spatial window of the primary cluster was located at Tigray, Amhara, southern Gambela,eastern Oromia and Afar which was centered at (12.376936 N,38.357984 E) / 318.07 km, RR = 1.43, and log likelihood ratio (LLR) of 200.78 at p-value < 0.0001. It showed that house holds in the spatial window had 1.43 times higher unemployment women ratethan those outside the window (Table 2 and Fig. 4).

Here we are addressing the spatial distribution and the risk area/cluster of unemployment, which means we speak about the first aim of this study.

In the next section of this study: We need to identify the determinant factors of women unemployment rate in Ethiopia

After all, concluding that the spatial distribution of women's unemployment status at the national level is non-random and associated with neighboring values. The result indicates that there is spatial autocorrelation /dependency or the unemployment rate is interdependent across clusters. As a result, we employ a multilevel logistic regression model that accounts for a lack of independence (spatial correlation) across levels of nested data (i.e., individuals (women) nested within clusters). The spatial correlation result was also supported by the intra-class correlation coefficient (ICC), which measures the proportion of variance(indirectly we know the standard deviation) in the outcome explained by the grouping structure while accounting for variance across clusters. This ICC represents the correlation of the women's unemployment rate within the same enumeration areas, i.e. the dependency of the unemployment rate among women within the clusters.So,from the research analysis, About 21.48% of the variation in women's unemployment rate occurred due to community-level(due to clusters) factors (ICC =21.48%).and MOR is showing that there was variation between communities clustered.So, here we know that the within and between cluster variation.

The result and interpretation was clearly mentioned in the manuscript.

Next, we examined the factors that influenced Ethiopia's unemployment rate using MLM, since this model incorporates the cluster correlation between unemployment and various factors.

Submitted filename: Response_3_ for plos one.docx

Click here for additional data file.

13 Jun 2022

16 Jun 2022

Thank you for your comments

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited

papers that have been retracted, please include the rationale for doing so in the manuscript text,

or remove these references and replace them with relevant current references. Any changes to

the reference list should be mentioned in the rebuttal letter that accompanies your revised

manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the

References list and also include a citation and full reference for the retraction notice.

Response: Please accept our sincere thanks for your comment. We carefully review our reference list and ensure that it is accurate. Additional references are also added to the list.

31. Broussard, N. and T.G. Tekleselassie, Youth unemployment: Ethiopia country study. International Growth Centre. Working Paper, 2012. 12(0592): p. 1-37.

32. Nganwa, P., D. Assefa, and P. Mbaka, The nature and determinants of urban youth unemployment in Ethiopia. Nature, 2015. 5(3): p. 197-203.

33. Batu, M.M., Determinants of youth unemployment in urban areas of Ethiopia. International Journal of Scientific and Research Publications, 2016. 6(5): p. 343-350.

Reviewer #2:

I do not see how my points were at odds with the aim of the study. Identifying hot spots areas is fine, but one cannot use concepts (i.e., outlier) in a completely different way as the literature does or one must provide good arguments for such unconventional choice. The authors have gone much beyond this modest point and now provide a completely new set of results. While I did not ask to follow this approach, I acknowledge that it has improved a lot the text and I prefer these results to the previous ones even if they had included my proposed corrections. The current results are in the line with the previous ones, although they are more complete and provide better insights.

In the previous version, the authors defined/used "outlier" in an incorrect way according to the literature. Now the term "outlier" is only used a couple of times in the Discussion section, but never properly defined. An "outlier" does not mean any value just above or below the mean or a reference value, but a really extreme value. It is never clear in the main text how an outlier is defined; for instance, how many times apart from mean/reference value and WHY, as it is routinely done in any rigorous statistical publication dealing with outliers.

Response: Please accept our sincere thanks for your comment. All your comments have been incorporated into the current version of the manuscript.

Submitted filename: point by point response to women unemployment R4.docx

Click here for additional data file.

22 Jun 2022

Geographical variation and determinants of women unemployment status in Ethiopia; A multilevel and spatial analysis from 2016 Ethiopia Demographic and Health Survey data

PONE-D-21-08056R4

Dear Dr. Setegn Muche Fenta,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Wen-Wei Sung, M.D., Ph.D.

Academic Editor

PLOS ONE

Reviewers' comments:

28 Jun 2022

PONE-D-21-08056R4

Geographical variation and determinants of women unemployment status in Ethiopia; A multilevel and spatial analysis from 2016 Ethiopia Demographic and Health Survey data

Dear Dr. Fenta:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Wen-Wei Sung

Academic Editor

PLOS ONE

Table 1

Socio-demographic characteristics of study participants.

		Women Unemployment rate
Variables	Categories	No (%)	Yes (%)
Women education level	No education	4937(70.2)	2096(29.8)
	Primary	3317(63.6)	1896(36.4)
	Secondary and above	1757(51.1)	1680(48.9)
Husband education level	No education	3352(75.6)	1079(24.4)
	Primary	1961(64.2)	1093(35.8)
	Secondary and above	4698(57.3)	3500(42.7)
Marital status	Living alone	2722(63.6)	1556(36.4)
	Married	6636(67.5)	3188(32.5)
	Other	653(41.3)	928(58.7)
Economic status	Poorest	4398(74)	1542(26)
	Middle	1416(70.7)	586(29.3)
	Richest	4197(54.2)	3544(45.8)
Residence	Rural	7374(71.3)	2961(28.7)
	Urban	2637(49.31)	2711(50.7)
Age of women	<25 years	4500(70.3)	1901(29.7)
	25–34 years	3020(59.4)	2066(40.6)
	>34years	2491(59.4)	1705(40.6)
Household size	0–5	5045(59.2)	3474(40.8)
	6–10	4629(68.9)	2086(31.1)
	>10	337(75.1)	112(24.9)
Sex of household	Female	2600(53.8)	2230(46.2)
	Male	7411(68.3)	3442(31.7)