Threshold BRIGHT I 2007-2008

The data file was constructed by merging several datasets. The household dataset was taken as the master dataset. The household dataset contains three types of data: village-level data, household-level data and child-level data. Observations are at the level of individual children, thus analyses based on child level variables do not require manipulation of the dataset (N=21,773). Analyses based on village-level data require the use of an indicator variable: village_level, where 1=one observation from a village and 0=duplicate observations. The indicator variable village_level was constructed by tagging one observation for each unique village identification number (hc1) in the dataset (N=287). Analyses based on household-level data require the use of an indicator variable: household_level, where 1=one observation from a household and 0=duplicate observations. The indicator variable household_level was constructed by tagging one observation for each unique household identification number (hc2) in the dataset (N=8,467).

The school wave 1 data were merged onto the household data using school ID numbers (matching sch2 from the school wave 1 data and ed6eco from the household data). Merges were successful for 7,675 individual children in the household dataset. School-level analyses on wave 1 school data require the indicator variable school_levelw1. This indicator variable was constructed by tagging one observation from each unique school identification number (sch2) in the school wave 1 dataset (N=278).

The school wave 2 data were merged onto the household data using household and child ID numbers (matching num_na and nuelev from the school wave 2 data and hc2 and hl1 from the household data). Merges were successful for 7,316 individual children in the household dataset, which included 282 children who were reported as not currently attending school. These cases are missing data for variables ed2niv through ed20. School-level analyses on wave 2 school data require the indicator variable school_levelw2. This indicator variable was constructed by tagging one observation from each unique school identification number (ecoleid) in the school wave 2 dataset (N=284).

Applicant data were also merged onto the dataset by matching village ID numbers. These variables came from the applications villages submitted to be part of the BRIGHT school program. These variables are region, province and department. Region is unique data in the dataset, while province and department are text variables that should mirror hc6 and hc7 respectively.
Additionally, these eight variables were merged from other datasets. All are village level variables.
selected
proj_selected
rel_score
hadschool_1
hadschool_2
hadschool_3
hadschool_type

All variables in the dataset can be found in the codebook. Entries for each variable include the variable name, variable label, question text, universe, and total non-missing responses. Some variable listings contain descriptions, construct specifications, ranges, frequencies, means, and/or standard deviations, depending on the type of variable.

To help users, variables are listed here based on the level at which the data were collected, along with the indicator variable that allows use of these variables.

Village-level variables: hc1 hc6 hc7 region province department selected proj_selected rel_score hadschool_1 hadschool_2 hadschool_3 hadschool_type (indicator variable village_level)

Household-level variables: hc2 hc5 hc9 hc10 hc11 hc12niv hc12cla hc14 hc15 hc16a hc16b hc16c hc17a hc17b hc18rad hc18telm hc18mon hc18velo hc18mob hc18veh hc18boe hc19 hc20 hc21ann hc21fre hc22 hc23 hc24 hc25a hc25b hc26 hc27 hc29 (indicator variable household_level)

Child-level variables : hl1 hl3 hl4 hl5 hl7niv hl7cla hl8 hl9 ed2niv ed2cla ed3 ed4 ed5 ed6eco ed6vil ed7 ed8 ed9 ed10 ed11 ed12 ed13 ed14 ed15 ed16 ed17 ed18a ed18b ed18c ed18d ed18e ed18f ed19 ed20 cl3 cl4 cl5 cl6 cl7 cl8 cl9 cl10 cl11 cl12 cl13 ma2_3 ma2_9 ma3chi ma3poi ma4_78 ma4_45 ma4_92 ma5_42 ma5_71 ma6_31 ma6_85 fa1 fa2c fa2t fa3pap fa3v_l fa4eco fa4tom fa5 fa6 ligne num_na nuelev sexe claselev presaj pr_s3jr freqpre absoct absnov absd_c absjan pr_s7jr (no indicator variable needed, as the dataset is at the child level)

School Wave 1 level variables: sch1 sch2 sch5 sch6 sch7 sch8 sch10 sch11 sc1 sc2 sc3 sc4_1gi sc4_1fi sc4_1gr sc4_1fr sc4_2fi sc4_2gi sc4_2gr sc4_2fr sc4_3fi sc4_3gi sc4_3gr sc4_3fr sc4_4gi sc4_4fi sc4_4gr sc4_4fr sc4_5gi sc4_5fi sc4_5gr sc4_5fr sc4_6gi sc4_6fi sc4_6gr sc4_6fr sc5 sc6_c sc6_l sc6_g sc7 sc8 sc9 sc10 sc11 sc12 sp1 sp2 sp3 sp4 sp5_tit sp5_sup sp5_adj sp5_ia sp5_iac sp5_ic sp5_ip sp6_0_5 sp6_5_10 sp6_10 sp7 sp8 ss1 ss2 ss3 ss4 ss5 ss6 ss7 ss8 ss9 ss10 ss11 ss12 ss13 ss14 ss15 ss16 (indicator variable school_levelw1)

School Wave 2 level variables: dateec ouvoct ouvnov ouvd_c ouvjan (indicator variable school_levelw2)

The Burkina Faso Girls' Education Impact Evaluation Survey data contains 21,773 records and 214 variables. Variables are positioned in the file in the following order:
Variables from the Household Survey. Variables are ordered by related questionnaire item number.
Variables from the School Survey Wave 1. Variables are ordered by related questionnaire item number.
Variables from the School Survey Wave 2. Variables are ordered by related questionnaire item number.
Constructed Variables. Constructed variables created from source variables.
Variables from Village Applications and Other Sources. Variables from village applications and other sources are found at the end of the dataset.