Education Quality Improvement Programme Impact Evaluation Endline Survey 2018

This file contains data at the school level and corresponds to the EQUIP-T IE endline head teacher (HT), headcount (HC), school records (SR) and in-service training coordinator (INCO) questionnaires. It also contains some variables from the front cover of the questionnaires, strata, survey weights and treatment status variables, and some constructed indicators prefixed by n_.

Schools were sampled and for analysis of schools in the treatment group the school weights included in this dataset should be used. Note that there are no survey weights for schools in the control group (see Weighting and Representativeness sub-sections in the Sampling section).

The same sample of schools were interviewed at baseline, midline and endline (panel sample).

Names of districts, schools, teachers and pupils have been removed in the anonymisation process.

This file contains data at the teacher level and corresponds to the EQUIP-T IE endline teacher interview (TI), head teacher interview (HT), headcount (HC) and in-service training coordinator interview (INCO). It also contains some variables from the front cover of the questionnaires, strata, survey weights and treatment status variables, and constructed indicators prefixed by n_.

The teacher roster is a census of all teachers at the school and all teachers eligible for the teacher interview were selected (those who teach maths or Kiswahili to Standards 1 to 3) for the interview. For analysis of teachers in the treatment group, the teacher weights included in this dataset should be used (there are two sets of weight variables: one for teacher interview and another for teacher roster). Note that there are no survey weights for teachers in the control group (see Weighting and Representativeness sub-sections in the Sampling section).

At each round (baseline, midline and endline), a different cross section of teachers were interviewed. Therefore, the interviewed teachers are not a panel sample. However, teachers can be matched to teachers in the baseline and midline datasets using the following ID variable in the dataset: pp_sr_uniqtch, to determine teacher turnover rates for example.

Names of districts, schools and teachers have been removed in the anonymisation process.

This file contains data at the pupil level and corresponds to the EQUIP-T IE endline pupil background and learning assessment questionnaire (PB) and poverty scorecard questionnaire (SC). It also contains some variables from the front cover of the questionnaires, strata, survey weights and treatment status variables. Constructed indicators, prefixed by n_, for pupil learning assessment scores and poverty level of pupils' households are also included in the datasets. For a description of how these indicators were constructed, users should refer to the survey reports included in the meta-data associated with this dataset as well as with the technical document 'Tanzania poverty scorecard instructions (Schreiner 2016)' also included in the meta-data.

Pupils were sampled and for analysis of pupils in the treatment group the pupil weights included in this dataset should be used. Note that there are no survey weights for pupils in the control group (see Weighting and Representativeness sub-sections in the Sampling section).

Note that the pupil IDs cannot be matched to the baseline or midline datasets as a different cross section of Standard 3 pupils was sampled and interviewed at each round (these are not the same pupils).

Names of districts, schools, pupils and parents have been removed in the anonymisation process.

This file contains data at the lesson level and corresponds to the EQUIP-T IE endline lesson observation questionnaire (LO). It also contains some variables from the front cover of the questionnaire, strata, survey weights and treatment status variables.

Lessons were not sampled and for analysis of lessons in the treatment group the school weights included in this dataset should be used. Note that the lesson observations were only conducted in treatment schools at endline.

At each round (baseline, midline and endline), a different cross section of lessons and teachers were observed. Therefore, the lesson observations are not a panel sample and the IDs cannot be matched to baseline or midline datasets.

Names of districts, schools and teachers have been removed in the anonymisation process.

This file contains data at the EQUIP-T residential training session level and corresponds to the EQUIP-T IE endline in-service training coordinator interview (INCO). It also contains some variables from the front cover of the questionnaires, strata, survey weights and treatment status variables.

The in-service training coordinator interview was administered to all treatment schools at endline. For analysis of this data in the treatment group, the school weights included in this dataset should be used. The INCO interview was not administered to control schools.

Names of districts, schools and teachers have been removed in the anonymisation process.

This file contains data at the EQUIP-T school-based training session level and corresponds to the EQUIP-T IE endline in-service training coordinator interview (INCO). It also contains some variables from the front cover of the questionnaires, strata, survey weights and treatment status variables.

The in-service training coordinator interview was administered to all treatment schools at endline. For analysis of this data in the treatment group, the school weights included in this dataset should be used. The INCO interview was not administered to control schools.

Names of districts, schools and teachers have been removed in the anonymisation process.