PAPUA NEW GUINEA: PUBLIC EXPENDITURE AND SERVICE DELIVERY (PESD) Annexes 30 June 2004 TABLE OF CONTENTS ANNEX 1: NOTES ON SAMPLING WEIGHTS ......................................................... 1 ANNEX 2: CONSTRUCTION OF POVERTY MEASURES ...................................... 4 ANNEX 3: CONSTRUCTION OF THE REMOTENESS INDEX.............................. 6 ANNEX 4: ADDITIONAL TABLES ON SCHOOL FACILITIES AND ENVIRONMENT.............................................................................................................. 9 ANNEX 5: SUMMARY MEASURES OF SCHOOL AUTONOMY AND PARENTAL/COMMUNITY PARTICIPATION........................................................ 16 ANNEX 6: ADDITIONAL TABLES ON SCHOOL FINANCES.............................. 21 ANNEX 7: ADDITIONAL TABLES ON TEACHERS AND STUDENTS .............. 25 ii ANNEX 1: NOTES ON SAMPLING WEIGHTS Sample Design Schools were selected through a combination of purposive sampling at the province level, and random sampling at the district and school levels. In the first stage, two provinces were chosen from each of the four main regions in PNG, with one less disadvantaged and one more disadvantaged province selected from each region. The included provinces are: Southern (Papuan) region: Gulf; National Capital District (NCD) Highlands region: Enga; Eastern Highlands Momase region: West Sepik (Sandaun); Morobe Islands region: West New Britain; East New Britain At the second stage, three districts were randomly selected within provinces using a Probability Proportional to Size (PPS) method, where size was defined by the number of schools in the district. Gulf and West New Britain provinces had only two districts each, so in their case both districts were selected. For NCD, which does not have districts but is organized by wards/census enumeration areas, there was no second stage selection. At the final stage, ten schools were then selected at random from each district. For NCD 30 schools were randomly selected. This yielded a total sample of 220 schools, comprising 20 schools each from Gulf and West New Britain, and 30 schools from each of the remaining six provinces (including NCD). However, some of the schools in this original sample could not be surveyed due to school closures and logistical difficulties. The final sample consisted of 214 schools. Its distribution across provinces and districts is shown in Table A1.1. Calculation of weights The sampling weights reflect the probability of a school being selected from all the schools in a given province. The results of the calculations described here are presented in Table A1.1. In order for a given school to be selected into the sample, two random events must transpire. Its district must first be selected, and then the school itself must be chosen from all of the schools in the district. So the overall probability of selection is simply the product of the probabilities of each event occurring. Defining a school Si, in district Di and province Pi, we can write: 1 P( S i selected ) = P( S i selected | Di selected) ⋅ P( Di selected) . Probability of a district being selected Districts in Gulf, West New Britain and NCD were automatically selected, and so have a selection probability of one. Three districts were selected from each of the remaining provinces using PPS sampling. This procedure defines the probability of a district being selected in any draw as the number of schools in the district divided by the number of schools in the province, so the overall probability of selection is three times this ratio:  number of schools in Di  P ( Di selected) = 3 ⋅   number of schools in P  .  i  The calculated probabilities of selection for each district are listed in column (c). In East New Britain, two districts (Gazelle and Pomio) were large enough to be selected twice, so the calculated probabilities for these districts were greater than one. We set these probabilities equal to one, and redistribute the excess probability equally between the other two districts. A Monte Carlo simulation produced empirical estimates of the probabilities which are extremely close to the theoretical results. These estimates are reported in column (d). Probability of a school being selected Each school in a selected district has a probability of selection equal to the number of schools selected from the district, divided by the total number of schools in the district: number of schools selected from Di P ( S i selected | Di selected) = . number of schools in Di The probabilities of each school being selected are reported in column (e). Overall probability of selection The overall probability of selection, reported in column (f), is the product of columns (c) and (e). Column (g) reports expansion factors for each school, which are simply the inverse of the overall probabilities. These give the number of schools in the province represented by each selected school.1 The estimated weights are on average greater than one, so the sum of the weights across schools exceeds the number of schools in the survey. To correct for this, the expansion factors were scaled down by a common factor. This also forces the average normalized weight across all schools to be one. The normalized weights and expansion factors are given in columns (i) and (j). 1 The sum of expansion factors for all selected schools in a province should, by definition, equal the total number of schools in that province. Because of the adjustment to the weights for ENB schools described earlier, the expansion factors for ENB schools sum to slightly more than the total 146 schools in the province. We therefore scale the expansion factors for ENB down slightly so they sum to 146. 2 Table A1.1:Calculation of School Sampling Weights Province / District Number Number Probability of district Probability Overall Expansion Number Normalized Normalized in in survey selection of school probability factor represented weight number district Calculation Monte Carlo selection of selection represented Simulation (c*e) (a) (b) (c) (d) (e) (f) (g) (h) (i) (j) NCD 39 30 39 NCD 39 30 1.000 -- 0.769 0.769 1.300 39 0.218 7 West New Britain 152 16 152 Kandrian - Gloucester 79 8 1.000 -- 0.101 0.101 9.875 79 1.659 13 Talasea 73 8 1.000 -- 0.110 0.110 9.125 73 1.533 12 East New Britain 146 30 146 Gazelle 54 10 1.000 1.000 0.185 0.185 5.338 53 0.897 9 Kokopo 32 10 0.716 0.716 0.313 0.224 4.419 44 0.742 7 Pomio 49 10 1.000 0.996 0.204 0.204 4.843 48 0.814 8 Rabaul 11 -- 0.284 0.281 -- -- -- -- -- -- Enga 138 30 138 Kandep 17 10 0.370 0.362 0.588 0.217 4.600 46 0.773 8 Kompiam - Ambum 23 -- 0.500 0.504 -- -- -- -- Lagaip - Porgera 39 10 0.848 0.847 0.256 0.217 4.600 46 0.773 8 Wabag 31 10 0.674 0.672 0.323 0.217 4.600 46 0.773 8 Wapenamanda 28 -- 0.609 0.608 -- -- -- -- -- -- Sandaun 182 30 182 Aitape - Lumi 50 10 0.824 0.827 0.200 0.165 6.067 61 1.019 10 Nuku 42 10 0.692 0.691 0.238 0.165 6.067 61 1.019 10 Telefomin 36 10 0.593 0.589 0.278 0.165 6.067 61 1.019 10 Vanimo Green River 54 -- 0.890 0.887 -- -- -- -- -- -- Gulf 108 19 108 Kerema 55 9 1.000 -- 0.164 0.164 6.111 55 1.027 9 Kikori 53 10 1.000 -- 0.189 0.189 5.300 53 0.890 9 Eastern Highlands 218 29 218 Daulo 16 -- 0.220 0.219 -- -- -- -- -- -- Goroka 19 -- 0.261 0.265 -- -- -- -- -- -- Henganofi 24 -- 0.330 0.331 -- -- -- -- -- -- Kainantu 27 10 0.372 0.371 0.370 0.138 7.267 73 1.221 12 Lufa 26 -- 0.358 0.359 -- -- -- -- -- -- Obura - Wonenara 34 9 0.468 0.468 0.265 0.124 8.074 73 1.356 12 Okapa 44 -- 0.606 0.598 -- -- -- -- -- -- Unggai - Bena 28 10 0.385 0.384 0.357 0.138 7.267 73 1.221 12 Morobe 291 30 291 Wau - Bulolo 43 -- 0.443 0.441 -- -- -- -- -- -- Finschafen 36 10 0.371 0.374 0.278 0.103 9.700 97 1.629 16 Huon 42 10 0.433 0.437 0.238 0.103 9.700 97 1.629 16 Kabwum 26 -- 0.268 0.267 -- -- -- -- -- -- Lae 19 -- 0.196 0.195 -- -- -- -- -- -- Markham 34 -- 0.351 0.349 -- -- -- -- -- -- Menyamya 26 -- 0.268 0.265 -- -- -- -- -- -- Nawaeb 27 -- 0.278 0.277 -- -- -- -- -- -- Tewae - Siassi 38 10 0.392 0.391 0.263 0.103 9.700 97 1.629 16 Overall 1274 214 1274 214 3 ANNEX 2: CONSTRUCTION OF POVERTY MEASURES Disaggregated maps of poverty in Papua New Guinea are created by combining information from the 1996 National Household Survey (NHS) with data from the 2000 National Census, and from resource and agricultural mapping databases with national coverage. The basic approach involves estimating a model of consumption (per adult equivalent) based on the 1996 NHS and then using the 2000 Census data to predict poverty measures at higher level of spatial disaggregation – up to the LLG-level. For constructing these maps, we have followed the methodology of Elbers et al. (2002), which pays more attention to heteroscedasticity, spatial autocorrelation and other location effects, and which uses simulation methods to calculate the predicted poverty indices and standard errors. The basic consumption model used for the poverty mapping exercise is reported in Table A2.1. For some of the analysis, we classify poverty levels into four categories, using the following bounds on the estimated headcount indices: • Well-off (0 to 0.15 inclusive); • Not poor (0.15 to 0.25 inclusive); • Poor (0.25 to 0.4 inclusive); • Very poor (index greater than 0.4). 4 Table A2.1:Log consumption model used for poverty mapping No. of rooms in dwelling 0.078 (3.63)** Household size (log) -0.478 (6.39)** % of HH age 7-14 years -0.358 (2.42)* School years of HH head 0.033 (4.51)** Wages main income HH head 0.35 (3.24)** Income from running a store 0.272 (2.22)* Income from running a PMV 0.656 (4.33)** Dummy: Altitude 1200-1800 m 0.307 (2.57)* Dummy: Altitude > 1800 m 0.443 (2.83)** Annual rainfall (‘000 mm) -1.183 (3.82)** Annual rainfall squared 0.167 (3.08)** Dummy: Slope > 10 degrees -0.382 (4.28)** Dummy: land inundation occurs -0.173 (1.63) Dummy: Rainfall deficit is rare 0.266 (2.62)* Ag. Syst. Remote from services -0.377 (2.87)** LLG % HH head main income wages 0.908 (2.07)* LLG % HH earning from betel nut 0.553 (2.02)* Constant 2.465 (4.80)** R-squared 0.34 Note: The dependent variable is the log of nominal consumption deflated by a region-specified poverty line. The sample is 830 rural households from the 1996 PNG National Household Survey. Absolute value of t-statistics in parentheses corrected for clustering, stratification and weighting. * significant at 5%; ** significant at 1%. 5 ANNEX 3: CONSTRUCTION OF THE REMOTENESS INDEX Overview Questions 64 to 79 in Section 1C of the S1 (head teacher) questionnaire give measures of each school’s access to 16 commonly-used ‘facilities’ (such as a police station and a bank). The survey asks the respondent for estimates of three measures of each facility’s remoteness from the school: the distance from the school to the facility, the travel time to the facility, and the mode of transport required. This annex describes the construction of a single index of remoteness for each school, which takes all of these measures into account. Construction of the index One could use many methods to combine the three measures of remoteness, and could give different weights to each facility (for instance, by down-weighting a vocational center relative to a police station). We take an agnostic approach, giving equal weight to each measure of remoteness and each facility. Two of the survey’s remoteness measures, distance and mode of transport, are recorded categorically. Distance is recorded as falling into one of six categories, while four modes of transport are listed.2 We take the survey’s categorical codes as the numerical values for these measures.3 These measures, along with the number of hours’ travel to the facility, are normalized according to the following rule: ~ X i − X min Xi = X max − X min where Xi is the measure of remoteness of school i from a given facility, and Xmax and Xmin are the maximum and minimum values of the measure (across schools) for the facility in 2 The distance categories (with codes in brackets) are (0) in school; (1) within 1km; (2) within 5km; (3) within 20km; (4) within 100km; and (5) more than 100km. The mode of transport question has four categories: (1) walk; (2) vehicle; (3) boat; (4) plane. For both questions, we treat ‘don’t know’ responses as missing values. 3 For mode of transport, we give boat and vehicle the same value, so the variable is recoded to (0) walk; (1) vehicle/boat; (2) plane. 6 question. The normalization turns each remoteness measure into a variable between 0 and 1, where 1 is the most remote, and 0 the least remote. A ‘score’ of the remoteness of each facility is then obtained by averaging the three normalized measures for each school. The 16 scores for each school are then averaged again, yielding an aggregate index of remoteness between 0 and 1. Missing responses for any question were excluded from the average. Figure A3.1 shows the distribution of index values for all 214 schools. Figure A3.1: Distribution of remoteness index Table A3.1: Proportion of schools in each remoteness category Num Num 40 40 35 35 Unweighte Weighted 30 30 25 25 Readily Accessible 20.6 12.4 20 20 Accessible 33.0 35.7 15 15 Remote 28.2 30.9 10 10 Extremely Remote 18.2 21.0 100 100 5 5 0 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Cat egory lower bound We classify remoteness into four categories, using arbitrary bounds on the constructed remoteness index.4 The four categories, and their bounds, are: • readily accessible (0 to 0.194 inclusive); • accessible (0.194 to 0.29 inclusive); • remote (0.29 to 0.4 inclusive); • extremely remote (index greater than 0.4). Using this classification, a reasonably large proportion of schools fall into each category. Furthermore, the classification is quite symmetrical. The unweighted and weighted proportions are summarized in Table A3.1. Viewing the distribution of schools across categories on its own can be deceptive, however, since the number of schools in each category varies quite widely across provinces (Table A3.2). NCD, for instance, has only five schools rated less than ‘readily accessible’, and dominates the ‘readily accessible’ category’. Conversely, in the Gulf province nearly all schools are considered ‘remote’ or ‘extremely remote’. 4 There were seven schools for which all of the remoteness questions were missing. These schools have a missing value for the remoteness measure. We were able to allocate remoteness categories to two of the schools based on the interviewers’ field notes. This leaves five of the 214 schools missing both index and category. 7 Table A3.2: Distribution of Remoteness Measures by Province Readily Extremely Accessible Accessible Remote Remote Total NCD 25 4 1 0 30 Enga 6 10 10 2 28 Eastern Highlands 3 15 11 0 29 Morobe 3 7 7 12 29 East New Britain 2 12 7 9 30 Sandaun 2 10 9 9 30 West New Britain 2 8 4 1 15 Gulf 0 3 10 5 18 Total 43 69 59 38 209 The poverty rate and the remoteness index are significantly correlated across the PESD sample. The weighted correlation coefficient is 0.15 while the unweighted correlation is 0.27, both statistically significant at the 5% level or better. 8 ANNEX 4: ADDITIONAL TABLES ON SCHOOL FACILITIES AND ENVIRONMENT Table A4.1: School background, by agency and type Agency Type Government Church Primary Community Mean (SE) Mean (SE) Mean (SE) Mean (SE) Government school (0/1) 0.00 0.00 0.00 0.00 0.57 0.07 0.49 0.06 Primary school (0/1) 0.39 0.06 0.32 0.06 0.00 0.00 0.00 0.00 Year established 1977 1.69 1971 1.42 1967 1.83 1979 1.23 Year became primary (among primary) 1999 0.53 1999 0.39 1999 0.38 0.00 0.00 Number of students 264 24.61 200 16.75 372 22.83 156 7.78 Percent of students girls 0.45 0.01 0.43 0.01 0.43 0.01 0.44 0.01 School land owned by …(0/1) … customary 0.60 0.08 0.48 0.06 0.24 0.07 0.72 0.05 … state 0.29 0.06 0.07 0.02 0.35 0.05 0.10 0.04 … church 0.02 0.02 0.42 0.06 0.28 0.05 0.15 0.04 … school 0.03 0.02 0.03 0.02 0.06 0.03 0.01 0.01 Source: PESD 2002. Means of valid responses. Table A4.2: School Background, by non-grant revenue per student Non-grant per student revenue quintiles Bottom 40% Middle 40% Top 20% Missing non-grant Mean (SE) Mean (SE) Mean (SE) Mean (SE) Government school (0/1) 0.46 0.09 0.45 0.08 0.28 0.16 0.58 0.06 Primary school (0/1) 0.27 0.06 0.35 0.08 0.79 0.15 0.35 0.06 Year established 1977 1.66 1972 2.56 1959 3.29 1976 1.00 Year became primary (among primary) 1999 0.70 2000 0.53 1997 1.06 1999 0.30 Number of students 211 23.78 233 16.94 431 56.72 223 22.66 Percent of students girls 0.44 0.03 0.44 0.01 0.43 0.01 0.44 0.00 School land owned by …(0/1) … customary 0.63 0.09 0.49 0.09 0.10 0.09 0.59 0.08 … state 0.28 0.09 0.14 0.04 0.32 0.12 0.16 0.05 … church 0.09 0.04 0.27 0.07 0.58 0.15 0.18 0.04 … school 0.00 0.00 0.08 0.05 0.00 0.00 0.03 0.02 Source: PESD 2002. Means of valid responses. 9 Table A4.3: Physical infrastructure at schools, by agency and type Agency Type Government Church Primary Community Mean (SE) Mean (SE) Mean (SE) Mean (SE) Classrooms Number of classrooms per 100 students 3.42 0.23 3.52 0.17 3.22 0.16 3.60 0.23 Proportion of classrooms ... made of permanent materials? 0.65 0.04 0.63 0.05 0.80 0.04 0.55 0.04 ... made of semi-permanent materials? 0.18 0.04 0.22 0.05 0.15 0.04 0.23 0.05 ... made of bush materials? 0.17 0.02 0.15 0.03 0.05 0.01 0.22 0.02 … which need to be completely rebuilt? 0.34 0.05 0.31 0.04 0.29 0.03 0.35 0.04 ... with a roof that leaks when it rains? 0.38 0.03 0.36 0.04 0.29 0.02 0.42 0.02 ... a chair and table for the teacher? 0.44 0.06 0.40 0.08 0.47 0.06 0.39 0.08 ... storage space that can be locked ? 0.30 0.04 0.24 0.04 0.33 0.05 0.24 0.02 ... with electricity that works? 0.09 0.02 0.04 0.01 0.13 0.03 0.03 0.01 Other infrastructure Adequate or good provision (0/1) of … … administration block 0.21 0.04 0.12 0.03 0.28 0.07 0.11 0.02 … clear radio reception 0.33 0.05 0.23 0.06 0.32 0.04 0.26 0.05 … school vehicle 0.04 0.02 0.01 0.01 0.02 0.01 0.03 0.02 … agriculture area for student use 0.53 0.05 0.66 0.04 0.55 0.05 0.62 0.05 … agriculture area for teacher use 0.43 0.05 0.49 0.05 0.39 0.05 0.50 0.04 … land for expansion 0.55 0.05 0.65 0.03 0.53 0.04 0.64 0.05 … sports area 0.65 0.04 0.73 0.06 0.71 0.04 0.67 0.05 … sports equipment 0.39 0.06 0.46 0.06 0.49 0.07 0.38 0.05 … specialist science classroom 0.02 0.01 0.01 0.01 0.04 0.01 0.00 0.00 … specialist technology classroom 0.05 0.02 0.02 0.02 0.11 0.04 0.00 0.00 … specialist home economics classroom 0.08 0.03 0.02 0.02 0.13 0.05 0.01 0.01 Source: PESD 2002. Means of valid responses. Table A4.4: Physical infrastructure at schools, by non-grant revenue per student Non-grant per student revenue quintiles Bottom 40% Middle 40% Top 20% Missing non-grant Mean (SE) Mean (SE) Mean (SE) Mean (SE) Classrooms Number of classrooms per 100 students 3.46 0.40 3.67 0.45 2.96 0.44 3.43 0.18 Proportion of classrooms ... made of permanent materials? 0.55 0.05 0.59 0.07 0.92 0.04 0.66 0.04 ... made of semi-permanent materials? 0.25 0.04 0.20 0.07 0.08 0.04 0.20 0.04 ... made of bush materials? 0.20 0.04 0.22 0.07 0.00 0.00 0.14 0.02 … which need to be completely rebuilt? 0.38 0.05 0.28 0.04 0.25 0.10 0.33 0.03 ... with a roof that leaks when it rains? 0.35 0.04 0.31 0.05 0.29 0.09 0.41 0.03 ... a chair and table for the teacher? 0.39 0.10 0.38 0.07 0.45 0.17 0.45 0.09 ... storage space that can be locked ? 0.31 0.07 0.24 0.06 0.37 0.10 0.26 0.03 ... with electricity that works? 0.05 0.04 0.07 0.02 0.23 0.12 0.06 0.01 Other infrastructure Adequate or good provision (0/1) of … … administration block 0.12 0.04 0.13 0.07 0.33 0.16 0.18 0.04 … clear radio reception 0.28 0.09 0.34 0.08 0.30 0.11 0.26 0.05 … school vehicle 0.04 0.04 0.05 0.04 0.02 0.02 0.01 0.01 … agriculture area for student use 0.59 0.07 0.64 0.08 0.71 0.10 0.56 0.06 … agriculture area for teacher use 0.47 0.05 0.59 0.08 0.55 0.10 0.40 0.06 … land for expansion 0.63 0.08 0.60 0.07 0.60 0.15 0.59 0.04 … sports area 0.60 0.11 0.81 0.07 0.96 0.02 0.64 0.03 … sports equipment 0.30 0.05 0.45 0.09 0.73 0.17 0.43 0.07 … specialist science classroom 0.00 0.00 0.04 0.02 0.00 0.00 0.01 0.01 … specialist technology classroom 0.02 0.02 0.04 0.02 0.12 0.12 0.04 0.02 … specialist home economics classroom 0.02 0.02 0.04 0.02 0.12 0.12 0.06 0.03 Source: PESD 2002. Means of valid responses. 10 Table A4.5: Electricity, water and sanitation, by agency and type Agency Type Government Church Primary Community Mean (SE) Mean (SE) Mean (SE) Mean (SE) Public grid/Elkom electricity (0/1) 0.19 0.03 0.11 0.03 0.28 0.04 0.08 0.03 Usable water tank (0/1) 0.51 0.06 0.59 0.09 0.69 0.06 0.47 0.06 Main source of drinking water supply … … none 0.03 0.02 0.01 0.01 0.00 0.00 0.03 0.02 … rain water tank 0.40 0.07 0.49 0.09 0.61 0.08 0.35 0.05 … spring/lake river 0.34 0.08 0.32 0.09 0.11 0.05 0.46 0.07 … well/bore hole 0.06 0.02 0.06 0.02 0.04 0.03 0.07 0.02 … piped water 0.16 0.03 0.12 0.03 0.23 0.04 0.08 0.02 Able to drink from that source today (0/1) 0.90 0.05 0.87 0.03 0.90 0.04 0.88 0.03 Available all year round 2001 (0/1) 0.58 0.04 0.58 0.05 0.46 0.06 0.64 0.05 Toilet facilities … … none available for teachers(0/1) 0.01 0.01 0.00 0.00 0.01 0.00 0.01 0.01 … none available for boys(0/1) 0.03 0.01 0.02 0.02 0.03 0.02 0.02 0.01 … need at least 1 for boys(0/1) 0.42 0.05 0.42 0.05 0.50 0.05 0.37 0.03 … none available for girls(0/1) 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 … need at least 1 for girls(0/1) 0.50 0.04 0.47 0.05 0.54 0.05 0.45 0.06 Source: PESD 2002. Means of valid responses. Table A4.6: Electricity, water and sanitation, by non-grant revenue per student Non-grant per student revenue quintiles Bottom 40% Middle 40% Top 20% Missing non-grant Mean (SE) Mean (SE) Mean (SE) Mean (SE) Public grid/Elkom electricity (0/1) 0.11 0.06 0.17 0.05 0.59 0.19 0.12 0.03 Usable water tank (0/1) 0.51 0.11 0.62 0.10 0.64 0.13 0.53 0.06 Main source of drinking water supply … … none 0.00 0.00 0.04 0.03 0.00 0.00 0.02 0.02 … rain water tank 0.39 0.11 0.49 0.10 0.53 0.14 0.44 0.07 … spring/lake river 0.33 0.11 0.21 0.09 0.15 0.14 0.39 0.07 … well/bore hole 0.06 0.04 0.09 0.05 0.00 0.00 0.05 0.03 … piped water 0.22 0.08 0.17 0.04 0.32 0.13 0.08 0.01 Able to drink from that source today (0/1) 0.96 0.04 0.90 0.04 0.00 0.00 0.84 0.05 Available all year round 2001 (0/1) 0.67 0.08 0.49 0.07 0.53 0.18 0.58 0.05 Toilet facilities … … none available for teachers(0/1) 0.00 0.00 0.01 0.01 0.00 0.00 0.01 0.01 … none available for boys(0/1) 0.02 0.02 0.00 0.00 0.00 0.00 0.04 0.03 … need at least 1 for boys(0/1) 0.44 0.07 0.49 0.08 0.32 0.21 0.40 0.04 … none available for girls(0/1) 0.00 0.00 0.00 0.00 0.00 0.00 0.04 0.03 … need at least 1 for girls(0/1) 0.49 0.09 0.55 0.07 0.55 0.18 0.46 0.05 Source: PESD 2002. Means of valid responses. 11 Table A4.7: Access to facilities, by agency and type Agency Type Government Church Primary Community Mean (SE) Mean (SE) Mean (SE) Mean (SE) Time to nearest …(hours) … High school or Secondary school 3.91 0.89 3.95 0.81 2.39 0.66 4.82 0.97 … Health center/aid post 1.00 0.22 0.81 0.48 0.35 0.17 1.22 0.37 … Vocational center 3.34 0.70 3.38 1.04 2.04 0.48 4.19 0.78 … Nearest store that sells basic materials 4.06 0.91 5.16 1.21 2.15 0.75 6.05 1.23 … Trade store 0.61 0.18 0.91 0.34 0.46 0.30 0.92 0.26 … Postal service 3.93 1.00 3.98 0.84 2.24 0.81 4.97 0.88 … Bank 4.46 1.07 4.56 0.95 2.16 0.72 5.94 0.91 … Police station 6.21 2.84 3.48 0.83 1.27 0.39 7.23 2.10 … Bitumen road 3.81 1.30 4.23 1.13 1.74 0.43 5.59 1.03 … PMV pickup point 3.29 1.17 2.86 0.87 1.60 0.36 4.08 0.91 … Town / station 3.02 0.67 3.29 0.81 1.48 0.41 4.07 0.69 … Provincial capital (of this province) 5.11 0.95 8.33 2.38 2.62 0.55 8.91 2.29 … Air strip 4.21 1.10 4.17 1.01 2.27 0.68 5.33 1.01 … Telephone that is working 3.59 0.83 3.61 0.85 1.58 0.41 4.90 0.93 … VHF radio that is working 1.78 0.46 2.30 0.76 0.77 0.16 2.65 0.64 Source: PESD 2002. Means of valid responses. Table A4.8: Access to facilities, by non-grant revenue per student Non-grant per student revenue quintiles Bottom 40% Middle 40% Top 20% Missing non-grant Mean (SE) Mean (SE) Mean (SE) Mean (SE) Time to nearest …(hours) … High school or Secondary school 4.59 1.48 2.68 0.68 0.28 0.07 4.48 1.12 … Health center/aid post 2.06 1.22 0.88 0.26 0.14 0.07 0.58 0.12 … Vocational center 4.07 1.51 2.82 0.68 0.57 0.09 3.55 0.86 … Nearest store that sells basic materials 7.31 2.30 2.36 0.60 0.38 0.06 4.76 1.30 … Trade store 1.18 0.51 0.47 0.22 0.00 0.00 0.77 0.33 … Postal service 4.95 1.46 3.04 0.89 0.24 0.06 4.29 1.12 … Bank 6.15 1.91 2.65 0.69 0.45 0.06 4.98 1.31 … Police station 4.83 1.49 2.91 0.70 0.22 0.06 6.23 2.61 … Bitumen road 5.57 2.09 2.51 1.28 0.03 0.03 4.48 1.76 … PMV pickup point 3.51 1.74 2.21 1.37 0.01 0.01 3.64 1.60 … Town / station 4.72 1.41 2.35 0.68 0.24 0.06 3.16 0.76 … Provincial capital (of this province) 6.59 1.57 7.33 3.77 4.30 3.40 6.50 1.41 … Air strip 5.62 1.83 3.14 0.80 0.83 0.17 4.34 1.12 … Telephone that is working 4.84 1.79 2.98 0.73 0.33 0.15 3.79 0.87 … VHF radio that is working 3.23 1.56 1.25 0.39 0.08 0.05 1.95 0.46 Source: PESD 2002. Means of valid responses. 12 Table A4.9: School closure and security, by agency and type Agency Type Government Church Primary Community Mean (SE) Mean (SE) Mean (SE) Mean (SE) School schedule Proportion of schools reporting closure ... … 2001 0.58 0.05 0.46 0.06 0.57 0.06 0.50 0.06 Total number of days the school closed in … … 2001 16.31 5.53 12.82 4.76 9.82 4.20 17.31 5.01 … 2002 (through to survey date) 3.01 0.72 3.15 1.00 1.30 0.38 4.06 0.85 Reasons of school closure in 2001 ( 0/1) … lack of water 0.42 0.06 0.31 0.08 0.50 0.10 0.29 0.05 … sewage/toilet problems 0.11 0.06 0.06 0.04 0.13 0.06 0.06 0.04 … poor facilities and maintenance 0.00 0.00 0.06 0.05 0.01 0.01 0.04 0.03 … shortage of teachers 0.03 0.03 0.12 0.05 0.03 0.03 0.09 0.04 … teacher pay problems 0.00 0.00 0.09 0.04 0.04 0.02 0.03 0.02 … school break-ins 0.05 0.03 0.02 0.02 0.03 0.02 0.04 0.03 … death in local community 0.10 0.04 0.09 0.05 0.06 0.04 0.12 0.04 … disputes between communities 0.06 0.03 0.04 0.02 0.07 0.03 0.04 0.03 … dispute between community and school 0.05 0.05 0.03 0.03 0.03 0.02 0.05 0.05 … special events/ poor weather 0.07 0.04 0.07 0.04 0.07 0.04 0.07 0.04 … other 0.02 0.02 0.04 0.04 0.00 0.00 0.04 0.03 Security With effective security fencing in 2002 … around school 0.19 0.06 0.10 0.03 0.23 0.05 0.10 0.03 … around teacher houses 0.10 0.04 0.05 0.03 0.12 0.03 0.06 0.03 Employed security guards in 2002 0.19 0.04 0.11 0.04 0.33 0.05 0.05 0.03 Number of times broken into in … … 2000 0.86 0.16 0.59 0.15 0.88 0.13 0.64 0.15 … 2001 0.78 0.09 0.79 0.14 0.91 0.14 0.71 0.13 … 2002 0.42 0.07 0.36 0.08 0.46 0.07 0.36 0.08 Source: PESD 2002. Means of valid responses. 13 Table A4.10: School closure and security, by non-grant revenue per student Non-grant per student revenue quintiles Bottom 40% Middle 40% Top 20% Missing non-grant Mean (SE) Mean (SE) Mean (SE) Mean (SE) School schedule Proportion of schools reporting closure ... … 2001 0.47 0.09 0.55 0.08 0.68 0.10 0.52 0.06 Total number of days the school closed in … … 2001 17.24 9.78 18.42 10.17 2.51 0.95 13.56 4.45 … 2002 (through to survey date) 2.92 1.22 2.14 1.33 0.84 0.50 3.66 0.76 Reasons of school closure in 2001 ( 0/1) … lack of water 0.43 0.14 0.45 0.09 0.64 0.20 0.29 0.07 … sewage/toilet problems 0.05 0.05 0.04 0.04 0.21 0.18 0.11 0.07 … poor facilities and maintenance 0.05 0.05 0.00 0.00 0.00 0.00 0.03 0.02 … shortage of teachers 0.09 0.08 0.00 0.00 0.00 0.00 0.09 0.05 … teacher pay problems 0.04 0.04 0.07 0.04 0.00 0.00 0.02 0.02 … school break-ins 0.00 0.00 0.10 0.08 0.00 0.00 0.02 0.02 … death in local community 0.14 0.07 0.11 0.08 0.00 0.00 0.09 0.04 … disputes between communities 0.08 0.07 0.03 0.03 0.03 0.04 0.05 0.03 … dispute between community and school 0.00 0.00 0.12 0.11 0.00 0.00 0.02 0.02 … special events/ poor weather 0.00 0.00 0.07 0.07 0.12 0.12 0.09 0.04 … other 0.00 0.00 0.00 0.00 0.00 0.00 0.05 0.04 Security With effective security fencing in 2002 … around school 0.09 0.05 0.18 0.05 0.40 0.15 0.14 0.05 … around teacher houses 0.06 0.04 0.08 0.04 0.24 0.16 0.07 0.03 Employed security guards in 2002 0.13 0.05 0.15 0.08 0.14 0.10 0.17 0.04 Number of times broken into in … … 2000 0.75 0.27 0.96 0.34 0.46 0.19 0.66 0.15 … 2001 0.91 0.24 0.99 0.29 0.49 0.22 0.69 0.13 … 2002 0.50 0.18 0.47 0.18 0.47 0.19 0.32 0.06 Source: PESD 2002. Means of valid responses. Table A4.11: Teaching resources, by agency and type Agency Type Government Church Primary Community Mean (SE) Mean (SE) Mean (SE) Mean (SE) Resource availability … (0/1) … sufficient textbooks for student use 0.20 0.06 0.26 0.05 0.15 0.03 0.28 0.07 … adequate or good provision of library 0.11 0.04 0.16 0.04 0.21 0.05 0.08 0.02 … adequate or good provision (0/1) of staff-rooms 0.15 0.05 0.17 0.04 0.30 0.07 0.08 0.02 … able to produce teaching aids 0.79 0.09 0.77 0.06 0.79 0.06 0.77 0.08 … enough desks for all students 0.49 0.08 0.55 0.06 0.52 0.06 0.52 0.06 … money allocated for classroom use 0.13 0.04 0.11 0.03 0.14 0.04 0.10 0.03 Source: PESD 2002. Means of valid responses. 14 Table A4.12: Teaching resources, by non-grant revenue per student Non-grant per student revenue quintiles Bottom 40% Middle 40% Top 20% Missing non-grant Mean (SE) Mean (SE) Mean (SE) Mean (SE) Resource availability … (0/1) … sufficient textbooks for student use 0.30 0.06 0.23 0.08 0.08 0.08 0.21 0.06 … adequate or good provision of library 0.06 0.05 0.19 0.06 0.42 0.10 0.11 0.03 … adequate or good provision (0/1) of staff-rooms 0.08 0.05 0.13 0.06 0.33 0.16 0.19 0.04 … able to produce teaching aids 0.72 0.09 0.74 0.10 0.88 0.08 0.81 0.06 … enough desks for all students 0.61 0.09 0.51 0.06 0.69 0.11 0.46 0.08 … money allocated for classroom use 0.05 0.03 0.12 0.05 0.02 0.03 0.15 0.03 Source: PESD 2002. Means of valid responses. 15 ANNEX 5: SUMMARY MEASURES OF SCHOOL AUTONOMY AND PARENTAL/COMMUNITY PARTICIPATION There are multiple indicators of the extent to which a school makes its own decisions about things that affect it, as there are multiple indicators of the potential extent to which parents affect those decisions. In order to analyze the patterns in the data, and to assess the relationship between these factors and other school characteristics such as financial status, teacher absenteeism, or test scores, it is useful to derive summary aggregate variables. This section describes how these summary measures, used elsewhere in the report, are calculated. In addition it provides summary statistics of these measures. School “autonomy” The summary measure of school “autonomy” is derived from the variety of questions about who has the most say in various decisions that affect schools. The measure is constructed as a simple average of whether or not the person with the most say in each of the 21 areas is at the “school level” or not. For example, if the response to “who has the most say in determining what teaching methods to use” is the head teacher this counts as school level, but if the response is Provincial Education Advisor the response is counted as non-school level. Formally the responses counted as school level are: board of management, head teacher, teachers, parents/PNC; while those counted as non-school level are national government agencies, provincial government agencies, district/LLG government agencies, inspectors, church agencies, politicians, donors, private businesses, and other. The responses from the head teacher, grade 5 teacher, or the BOM representative could be used to construct the measure. Because of the more limited coverage of the grade 5 and BOM representative surveys, the average response for head teachers is the main variable used. On occasion the average across all three is used, as is the average for grade 5 teachers alone (since they provide a view “from the classroom,” that is a view closely related to the way decision-making is perceived by those who actually interact with students). In those cases, however, the number of schools included in the analysis falls. Table A5.1 shows the percent of respondents who report that the person with the most say is at the school level. The average across all questions for a given respondent gives a “raw” average. This is then divided by the standard deviation in order to yield a 16 “normalized” measure of autonomy. Its interpretation is therefore in terms of standard deviations, that is, a one unit increase in the normalized measure of autonomy is equivalent to a one standard deviation increase in the raw measure of autonomy. Since the measure of autonomy has no intrinsic metric the normalized measure is what is ultimately used in this report. Similarly, the average across all respondents can be normalized (as reported in the last column of Table A5.1) for all valid observations. Table A5.1: Proportion of respondents who say that the person who has the most say is at the “school level” BOM Head Grade 5 represent Teacher Teacher ative All Appointing teachers 0.067 0.119 0.207 0.130 Policy for assessing teachers 0.306 0.307 0.349 0.321 Assessing teachers 0.461 0.474 0.423 0.452 Teacher promotion 0.232 0.161 0.291 0.230 Discipline action against teachers 0.296 0.454 0.487 0.409 Dismissing a teacher 0.155 0.228 0.409 0.264 Selection for inservice 0.193 0.358 0.389 0.310 Who enrols 0.871 0.857 0.879 0.870 Class size 0.752 0.666 0.729 0.718 Teaching methods 0.517 0.515 0.568 0.534 Assessing students 0.908 0.907 0.882 0.899 Policy for assessing students 0.793 0.789 0.640 0.739 Spending school subsidy 0.941 0.944 0.921 0.935 Level of project fees 0.685 0.720 0.816 0.740 How to spend project fees 0.966 0.973 0.958 0.965 Maintenance provision 0.956 0.954 0.950 0.953 Constructing classrooms 0.930 0.909 0.883 0.907 Upgrading school 0.344 0.363 0.390 0.366 Organising pnc activities 0.991 0.971 0.958 0.973 Organising community activities 0.965 0.937 0.926 0.944 Organising fundraising activities 0.975 0.970 0.970 0.972 Average 0.634 0.647 0.668 0.658 Standard deviation 0.154 0.161 0.178 0.120 Minimum 0.000 0.000 0.000 0.238 Maximum 1.000 1.000 1.000 0.937 Normalized average 4.110 4.020 3.763 5.486 Normalized standard deviation 1.000 1.000 1.000 1.000 Normalized minimum 0.000 0.000 0.000 1.984 Normalized maximum 6.486 6.211 5.633 7.802 Source: PESD 2002. Percent of valid responses. How does the summary measure of autonomy vary across the sample of schools? Respondents in NCD clearly feel more autonomous than those in other parts of the country (Table A5.2). Beyond that, however, there is not much significant variation across provinces (As is clear from Figure A5.1). Perhaps surprisingly, respondents in extremely remote locations tend to feel like they have less autonomy—despite the lack of 17 direct and continuous supervision. Neither poverty nor agency type are systematically associated with the summary measure of autonomy. Table A5.2: Average autonomy – for each respondent type and overall. Head teachers Grade 5 teachers BOM representative All respondents Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Province EHP 4.36 0.72 4.11 0.56 3.76 1.05 5.68 0.70 ENBP 4.28 0.75 4.35 0.91 3.69 0.84 5.57 0.74 Enga 4.08 0.79 4.01 0.72 3.97 0.62 5.45 0.52 Gulf 4.30 0.94 3.22 1.72 3.73 1.20 5.32 1.14 Morobe 3.76 1.43 3.88 1.21 3.36 1.09 5.10 1.43 NCD 4.91 0.83 4.70 0.92 4.56 0.89 6.42 0.85 Sandaun 4.03 0.86 3.82 0.76 3.91 0.82 5.35 0.86 WNBP 4.00 0.89 4.28 0.93 4.13 1.15 5.97 0.90 Remoteness Easy access 4.05 1.23 3.96 1.25 3.55 1.14 5.12 1.53 Accessible 4.20 0.88 4.19 0.76 3.98 1.03 5.72 0.81 Remote 4.10 1.11 4.01 1.17 3.69 0.98 5.51 0.95 Extremely remote 4.02 0.90 3.71 0.87 3.63 0.86 5.19 0.93 Poverty Better off 4.26 0.96 4.53 0.86 3.89 0.89 5.98 0.86 NotPoor 4.21 0.77 4.17 0.68 3.73 0.83 5.45 0.72 Poor 4.02 1.17 3.79 1.10 3.65 1.20 5.39 1.07 VeryPoor 4.03 0.93 3.73 1.02 3.86 0.89 5.23 1.13 Agency Government 4.15 0.85 4.01 0.97 3.71 1.07 5.50 0.86 Catholic 4.10 0.92 4.16 0.79 3.89 0.91 5.57 0.86 Others 4.03 1.36 3.91 1.21 3.74 0.96 5.37 1.35 Total 4.11 1.00 4.02 1.00 3.76 1.00 5.49 1.00 Source: PESD 2002. Percent of valid responses. Figure A5.1: Average autonomy as reported by head teachers 6 5 4 3 2 1 0 Sandaun Government NCD Extremely EHP ENBP WNBP Enga Remote Catholic Others Morobe Accessible NotPoor access Gulf Poor VeryPoor Better off Easy remote Remoteness Poverty Agency Total Source: PESD 2002. Percent of valid responses. 18 Parent participation and community partnership Like the summary measure of school autonomy, a single average capturing the extent to which parents and communities participate in school affairs can be constructed. Table A5.3 reports summary statistics on the individual variables that make up the two sub- measures, which are then averaged and normalized to form an overall aggregate measure. Details on the individual variables are discussed above. The aggregate measure has a mean of 2.687 (with a standard normalized deviation to 1). Table A5.3: Proportion of respondents who say that the person who has the most say is at the “school level” Mean Std. Dev. Parent participation PNC exists 0.910 0.287 PNC met more than once in 2001 0.621 0.486 Proportion of parents who attend pnc meetings 0.475 0.314 Proportion of parents who collect assessments 0.346 0.393 Proportion of parents who attend school meetings when these are called 0.503 0.327 At least one parent is identified as being on the BOM 0.925 0.264 Community partnership Community is used as a learning site 0.349 0.478 Community members help develop school programs acnd activities 0.600 0.491 Teachers organize community activities such as adult classes, sporting competitions 0.497 0.501 Community members teach cultural activities in the school? 0.271 0.445 School uses village land for agricultural classes 0.450 0.499 School is used for community meetings 0.634 0.483 School is used for adult classes 0.173 0.379 School is used for community sports events 0.710 0.455 Aggregate measures Average parent participation 0.630 0.225 Normalized parent participation 2.801 1.000 Average community partnership 0.461 0.257 Normalized average community partnership 1.791 1.000 Average normalized parent participation and community partnership 2.296 0.855 Average normalized parent participation and community partnership, normalized 2.687 1.000 Source: PESD 2002. Percent of valid responses. Table A5.4 and Figure A5.2 reports the variation of the aggregate measure of parent participation and community partnership. There is not substantial variation in the amount of involvement. Gulf and WNBP stand out with relatively low levels, and extremely remote areas stand out with relatively high levels. However these are not large variations: the difference between the average in the province with the highest and lowest values ( EHP versus Gulf) is .59, or slightly more than half a standard deviation. 19 Table A5.4: Average parent participation and community partnership Mean Std. Dev. Province EHP 2.70 0.79 ENBP 2.67 0.97 Enga 2.93 0.96 Gulf 2.34 0.82 Morobe 2.75 1.33 NCD 2.61 0.75 Sandaun 2.75 0.74 WNBP 2.53 1.11 Remoteness Easy access 2.83 0.96 Accessible 2.59 0.84 Remote 2.45 1.20 Extremely remote 3.15 0.76 Poverty Better off 2.86 1.10 Not Poor 2.56 0.93 Poor 2.62 1.07 VeryPoor 2.75 0.83 Agency Government 2.72 0.87 Catholic 2.65 0.86 Others 2.65 1.34 Total 2.69 1.00 Source: PESD 2002. Percent of valid responses. Figure A5.2: Average parent participation and community partnership 3.50 3.00 2.50 2.00 1.50 1.00 0.50 0.00 Government Sandaun Extremely NCD EHP ENBP WNBP Remote Enga Catholic Others Not Poor Gulf Morobe Accessible access Poor VeryPoor Better off Easy remote Remoteness Poverty Agency Total Source: PESD 2002. 20 ANNEX 6: ADDITIONAL TABLES ON SCHOOL FINANCES 21 Table A6.1: Correlates of complete school-level financial information: probit regressions Complete information on both revenues & spending (0/1) a/ Poverty rate (0 to 1) 0.037 0.053 0.016 -0.057 (0.13) (0.17) (0.09) (0.32) Remoteness index (0 to 1) -0.035 0.05 0.037 0.076 (0.09) (0.11) (0.10) (0.18) Primary school (0/1) -0.023 -0.096 -0.104 -0.133 (0.21) (0.65) (0.92) (1.02) Church operated school (0/1) 0.112 0.15 0.056 0.082 (1.41) (1.77) (0.73) (0.99) EHP 0.582 0.509 (4.38)*** (2.85)*** ENBP 0.422 0.427 (2.34)** (1.85) Enga 0.416 0.198 (2.46)** (0.87) Gulf 0.359 0.285 (1.50) (1.04) Morobe 0.117 -0.011 (0.67) (0.05) Sandaun 0.243 0.146 (0.97) (0.50) WNBP 0.062 0.076 (0.38) (0.37) Parent and community involvement (0 to 1) 0.110 0.073 0.100 0.07 (1.02) (0.64) (0.92) (0.65) School Autonomy (0 to 1) 0.06 0.052 0.085 0.097 (1.00) (0.91) (1.15) (1.57) School Autonomy* Parent & community (0 to 1) -0.025 -0.018 -0.031 -0.03 (1.16) (0.78) (1.33) (1.35) Head teacher absent (0/1) -0.375 -0.3 -0.344 -0.239 (0.76) (0.44) (0.60) (0.29) Male head teacher (0/1) -0.348 -0.439 -0.286 -0.395 (2.08)** (2.31)** (1.92)* (2.52)** Less than 2yrs as headteacher at this school (0/1) -0.183 -0.227 -0.158 -0.208 (2.56)** (2.54)* (2.13)** (2.31)** Head teacher age -0.014 0.004 -0.004 0.014 (0.31) (0.09) (0.08) (0.29) Head teacher age-squared 0.000 0.000 0.000 0.000 (0.26) (0.19) 0.00 (0.40) Head teacher wants to stay at this school (0/1) 0.233 0.315 0.188 0.237 (2.77)*** (3.09)*** (2.28)** (2.45)** MP from local area (0/1) 0.105 0.127 0.095 0.146 (0.92) (1.05) (0.87) (1.23) Log of total enrolment 2001 -0.002 0.05 0.079 0.111 (0.02) (0.55) (1.44) (1.50) # of inspector visits in 2001 0.023 0.006 0.005 -0.019 (0.77) (0.22) (0.16) (0.64) # of BOM meetings in 2001 -0.007 -0.021 0.005 -0.008 (0.42) (1.05) (0.29) (0.44) One account only (0/1) 0.127 0.138 (1.23) (1.51) At least one joint account (0/1) 0.194 0.196 (2.12)** (2.22)** Observations 204 187 204 187 Pseudo R-squared 0.23 0.23 0.15 0.20 Robust z statistics in parentheses. * significant at 10%; ** significant at 5%; *** significant at 1% 0/1 indicates a binary variable. a/ reports marginal effects of variables rather than the probit coefficients. Source: 2002 PESD. 22 Table A6.2: Regressions: grant revenue per student in 2001 Government grant Non-government grant Donors grant (II) (II) (III) (IV) (V) (VI) (VII) Probability Probability Log of amount Log of amount Log of amount Log of amount Log of amount grant > K50 grant > K50 Poverty rate (0 to 1) 0.291 1.562 0.181 1.389 0.416 -0.780 0.516 (1.34) (1.12) (0.75) (1.58) (0.31) (0.93) (0.44) Remoteness index (0 to 1) 0.016 0.837 -0.109 -0.011 -1.272 0.943 -0.601 (0.06) (0.56) (0.37) (0.01) (0.90) (0.71) (0.48) Church operated school (0/1) -0.002 -0.276 -0.022 0.117 0.197 -0.262 -0.098 (0.05) (0.91) (0.42) (0.39) (0.68) (0.92) (0.39) Primary school (0/1) 0.066 0.559 0.259 1.184 1.257 1.025 1.081 (1.24) (1.77)* (3.91)*** (3.8)*** (4.16)*** (3.44)*** (4.08)*** EHP 0.423 -0.760 0.356 0.105 0.240 (2.64)*** (0.70) (3.26)*** (0.10) (0.27) ENBP 0.359 -0.460 -0.401 -1.396 (2.23)** (0.40) (0.36) (1.44) Enga 0.770 0.943 0.307 -0.780 -0.858 (4.64)*** (0.85) (2.94)*** (0.74) (0.93) Morobe 0.576 -0.132 0.542 1.050 1.747 (3.44)*** (0.12) (4.41)*** (0.98) (1.77)* WNBP 0.544 0.407 -0.346 1.092 (2.64)*** (0.37) (0.33) (1.16) Gulf -0.486 0.601 1.415 -0.927 (0.41) (3.95)*** (1.25) (0.86) Sandaun -1.430 0.459 0.589 -0.568 (1.11) (2.66)*** (0.48) (0.62) Constant 0.954 1.027 1.408 0.962 1.014 (0.88) ( 1.95 )* (1.37) (1.91)* (1.12) Observations 168 168 168 168 168 168 168 Note: Based on 174 schools . Robust z statistics in parentheses * significant at 10%; ** significant at 5%; *** significant at 1% Source: 2002 PESD. 23 Table A6.3: Correlates of revenue per student: OLS regressions Log of per stundent total Log of per student non- Log of per student grant revenue 2001 grant revenue 2001 revenue 2001 Poverty rate (0 to 1) 0.596 -0.276 0.444 -0.666 1.917 0.959 (0.43) (0.39) (0.68) (1.14) (1.67) (1.88)* Remoteness index (0 to 1) 0.634 0.891 -0.769 -1.052 -2.432 -2.332 (0.39) (0.61) (0.55) (0.70) (1.24) (1.29) Primary school (0/1) 0.66 0.916 0.857 0.714 1.144 1.312 (1.19) (2.12)** (2.16)** (2.21)** (2.21)** (2.72)** Church operated school (0/1) 0.053 -0.006 0.206 0.074 -0.373 -0.374 (0.14) (0.01) (1.04) (0.31) (0.96) (0.98) EHP -0.947 -1.227 0.584 (1.58) (3.70)*** (0.66) ENBP -0.435 -0.804 0.153 (0.74) (1.70) (0.22) Enga -1.042 -1.453 0.761 (2.48)** (4.39)*** (1.04) Gulf -0.629 -2.246 1.592 (0.98) (5.40)*** (2.11)** Morobe -0.783 -1.638 0.862 (1.31) (3.54)*** (1.13) Sandaun -1.391 -2.076 0.271 (2.03)* (4.85)*** (0.30) WNBP -0.672 -1.258 0.851 (1.24) (1.68) (1.23) Parent and community involvement (0 to 1) 0.122 0.065 1.854 1.924 -0.005 0.034 (0.19) (0.11) (1.82)* (2.18)** (0.01) (0.07) School Autonomy (0 to 1) -0.203 -0.27 1.038 1.035 -0.783 -0.782 (0.40) (0.61) (1.52) (1.77) (2.04)* (2.05)* School Autonomy* Parent & community (0 to 1) -0.018 0.008 -0.412 -0.417 0.179 0.166 (0.11) (0.05) (1.78)* (2.01)* (1.28) (1.19) Head teacher absent (0/1) -7.709 -7.797 -2.893 -4.004 -12.838 -11.457 (1.52) (1.63) (0.60) (0.90) (2.65)** (2.32)** Male head teacher (0/1) 0.038 -0.018 0.179 -0.083 -0.346 -0.031 (0.06) (0.03) (0.51) (0.18) (0.41) (0.04) Less than 2yrs as headteacher at this school (0/1) 0.012 0.006 -0.101 0.018 -0.024 -0.11 (0.03) (0.02) (0.30) (0.05) (0.05) (0.25) Head teacher age -0.42 -0.408 -0.145 -0.168 -0.663 -0.625 (1.63) (1.72) (0.64) (0.79) (2.92)*** (2.73)** Head teacher age-squared 0.005 0.005 0.002 0.002 0.008 0.008 (1.63) (1.72) (0.60) (0.74) (3.05)*** (2.89)*** Head teacher wants to stay at this school (0/1) 0.408 0.209 -0.241 -0.447 0.249 0.412 (0.79) (0.43) (0.61) (1.19) (0.45) (0.79) MP from local area (0/1) 0.103 0.063 0.104 0.06 -0.266 -0.31 (0.33) (0.23) (0.56) (0.31) (0.82) (1.03) Log of total enrolment 2001 -0.156 -0.305 -0.369 -0.145 -0.548 -0.796 (0.32) (0.85) (1.35) (0.70) (1.11) (2.08)** # of inspector visits in 2001 0.177 0.188 -0.078 -0.055 0.267 0.247 (1.28) (1.67) (0.66) (0.54) (2.27)** (2.18)** # of BOM meetings in 2001 0.121 0.109 0.179 0.196 0.263 0.262 (1.59) (1.48) (2.35)** (2.28)** (2.77)** (2.97)*** Observations 85 85 90 90 166 166 R-squared 0.38 0.35 0.50 0.39 0.32 0.29 Note: Robust z statistics in parentheses* significant at 10%; ** significant at 5%; *** significant at 1% 0/1 indicates a binary variable Source: 2002 PESD. 24 ANNEX 7: ADDITIONAL TABLES ON TEACHERS AND STUDENTS 25 Table A7.1 : Probit regressions : marginal effects (z-stats in parenthesis) of variables on teacher absence (I) (II) (III) A ll te a c h e r M a le te a c h e rs F e m a le te a c h e rs F e m a le (0 /1 ) -0 .0 1 5 (0 .7 5 ) A g e : 3 1 -4 0 yrs 0 .0 0 5 0 .0 1 5 0 .0 0 1 (0 .1 8 ) (0 .3 6 ) (0 .0 3 ) A g e : 4 1 -5 0 yrs -0 .0 3 3 0 .0 0 3 -0 .0 6 8 (1 .1 3 ) (0 .0 8 ) (2 .2 0 )* * A g e > 5 0 yrs 0 .0 2 5 0 .0 0 7 0 .2 1 6 (0 .5 8 ) (0 .1 5 ) (2 .6 7 )* * * H e a d te a c h e r (0 /1 ) -0 .0 5 7 -0 .0 6 9 -0 .0 3 2 (1 .7 1 ) (1 .7 0 ) (0 .5 3 ) T e a c h in g le v e l 3 -0 .0 0 9 -0 .0 0 8 -0 .0 3 0 (0 .3 4 ) (0 .2 5 ) (1 .1 3 ) T e a c h in g le v e l 4 & a b o v e 0 .0 3 3 0 .0 3 8 -0 .0 7 1 (0 .7 9 ) (0 .7 5 ) (2 .0 7 )* * % w ith sc h o o l h o u sin g 0 .0 7 7 0 .1 3 7 -0 .0 2 6 (1 .6 2 ) (2 .6 0 )* * * (0 .4 8 ) P o v e rty ra te (0 to 1 ) 0 .1 5 0 0 .1 9 0 -0 .0 0 5 (1 .5 9 ) (1 .6 2 ) (0 .0 6 ) R e m o te n e ss in d e x (0 to 1 ) 0 .1 4 9 -0 .0 3 4 0 .4 1 0 (1 .1 2 ) (0 .2 3 ) (2 .5 7 )* * * P rim a ry sc h o o l (0 /1 ) -0 .0 4 4 -0 .0 1 7 -0 .1 1 2 (1 .4 2 ) (0 .5 1 ) (2 .6 0 )* * * C h u rc h o p e ra te d sc h o o l (0 /1 ) -0 .0 1 2 -0 .0 3 3 0 .0 1 9 (0 .5 3 ) (1 .1 7 ) (0 .7 1 ) EHP -0 .0 7 9 -0 .0 8 4 -0 .0 6 7 (2 .0 4 )* (2 .0 5 )* * (1 .5 3 ) ENBP -0 .1 0 4 -0 .1 3 6 -0 .0 7 1 (2 .5 7 )* (3 .4 2 )* * * (1 .4 3 ) E nga -0 .0 4 8 -0 .0 4 7 -0 .0 4 0 (1 .0 8 ) (0 .9 7 ) (0 .8 3 ) G u lf -0 .0 5 3 -0 .0 6 1 -0 .0 3 7 (1 .0 1 ) (1 .2 1 ) (0 .5 3 ) M o ro b e -0 .1 1 6 -0 .1 3 3 -0 .0 8 5 (2 .7 2 )* * (2 .7 8 )* * * (2 .0 6 )* * Sandaun -0 .1 3 2 -0 .1 3 8 -0 .0 9 7 (3 .2 6 )* * (2 .6 9 )* * * (2 .5 7 )* * * W NBP -0 .1 0 9 -0 .1 0 8 -0 .1 0 0 (2 .5 6 )* (2 .0 5 )* * (2 .3 5 )* * A d v a n c e n o tic e o f sc h o o l v isit … o n e w e e k o r le ss (0 /1 ) -0 .0 0 2 0 .0 1 7 -0 .0 4 2 (0 .0 7 ) (0 .5 7 ) (1 .2 0 ) … m o re th a n o n e w e e k (0 /1 ) 0 .0 4 6 0 .0 3 4 0 .0 7 2 (1 .1 1 ) (0 .7 7 ) (1 .2 9 ) P a ym e n t d e la y (d a ys) 0 .0 0 2 0 .0 0 4 -0 .0 0 3 (1 .8 7 ) (3 .2 6 )* * * (1 .0 2 ) P a re n ts & c o m m u n ity in v o lv e m e n t -0 .0 3 6 -0 .0 2 2 -0 .0 4 5 (2 .3 4 )* (1 .3 1 ) (2 .8 8 )* * * S c h o o l a u to n o m y 0 .0 0 7 -0 .0 0 5 0 .0 2 3 (0 .5 8 ) (0 .3 0 ) (1 .9 6 )* O b se rv a tio n s 1742 909 833 Note: The regressions also includes dummy variables to control for missing data on teachers' age, teaching level, notice of school visit, payment delay, per student textbooks. Robust z statistics in parentheses. * significant at 10%; ** significant at 5%; *** significant at 1% Source: PESD 2002. 26 Table A7.2 : Teachers’ salary payment and monetary allowances in 2002 Salary payment Monetary allowance Average Average % of teachers additional additional fornight fornight received monetary monetary gross reduction additional allowance allowance salary from monetary average teacher received (kina) salary allowance received (per year, kina) a/ (kina) (per year, kina) c/ a/ b/ 198 schools 198 schools By province Eastern Highlands 427 135 57 90 158 East New Britain 440 153 21 96 454 Enga 454 125 49 407 824 Gulf 426 180 37 83 224 Morobe 443 168 62 109 175 NCD 457 182 54 332 618 Sandaun 413 133 28 75 268 West New Britain 422 98 24 323 1,323 By remoteness Easy access 440 152 49 176 357 Accessible 435 134 39 166 421 Remote 436 154 44 305 686 Extremely remote 432 132 47 107 231 By poverty Well off 438 161 46 223 481 Not poor 454 149 43 258 606 Poor 428 129 47 165 350 Very poor 419 134 35 102 292 By school type Community 431 140 52 128 246 Primary 439 145 39 230 589 By agency Government 435 141 46 190 416 Church 438 146 41 200 490 Total 436 143 44 194 443 Note: a/ 3 teachers were selected from each sample school for an in-depth study (survey data) b/ include zero amount. (allowance>=0) c/ exclude zero amount. (allowance> 0) Source: 2002 PESD and NDOE payroll data. 27 Table A7.3 : Teachers’ salary payment, access and delays Salary payment Salary payment Average salary Average Salary % of teachers delay % of teachers payment delay access delay paid by direct (incl. Zeros, paid by cheques (days) a/ period (weeks) deposits days) By province Eastern Highlands 1 14 1.0 6 94 East New Britain 1 29 1.2 9 91 Enga 1 24 0.6 13 87 Gulf 11 52 0.7 25 75 Morobe 7 29 1.1 9 91 NCD 0 2 0.5 5 95 Sandaun 12 14 1.0 97 3 West New Britain 1 11 1.0 11 89 By remoteness Easy access 1 9 0.6 8 92 Accessible 3 19 0.9 17 83 Remote 5 29 1.1 28 72 Extremely remote 9 22 1.2 33 67 By poverty Well off 4 31 0.8 12 88 Not poor 1 17 0.8 10 90 Poor 2 22 0.9 10 90 Very poor 9 15 1.4 65 35 By type Community 6 22 1.2 28 72 Primary 2 20 0.7 15 85 By agency Government 3 23 0.8 15 85 Church 4 19 1.1 27 73 Total 3 21 0.9 19 81 Note: a/ includes only teachers who answered there is some delay in salary payments. Source: 2002 PESD. 28 Table A7.4 : Allowances the teachers were eligible for but did not receive in 2001 %of teachers did not Type of allowances teachers did not receive receive responsibility or/and Higher Mining or allowances disadvantaged school or/and duty leave fare or which they multigrade or/and domestic market were eligible housing for By province Eastern Highlands 53.1 51.8 0.0 1.4 East New Britain 23.9 23.9 0.0 0.0 Enga 54.7 41.1 8.3 5.3 Gulf 43.3 36.7 6.6 0.0 Morobe 37.9 29.8 7.1 0.0 NCD 23.8 10.9 5.1 7.9 Sandaun 72.1 59.8 9.1 3.2 West New Britain 35.1 32.1 3.0 0.0 By remoteness Easy access 25.8 14.7 7.5 3.7 Accessible 49.8 44.8 2.8 2.2 Remote 49.5 42.3 4.5 2.1 Extremely remote 47.5 44.6 2.9 0.0 By poverty Well off 35.0 24.6 5.9 4.0 Not poor 56.3 49.4 4.9 2.0 Poor 28.3 25.1 1.6 1.5 Very poor 65.3 56.3 6.9 2.1 By type Community 55.8 50.0 4.7 0.8 Primary 36.9 29.6 4.1 3.1 By agency Government 41.5 34.5 5.1 1.9 Church 47.0 40.6 3.1 3.0 Source: 2002 PESD. Table A7.5 : Correlation matrix : Ghost teacher rate, teacher absent rate, teacher shortage rate, teacher turnover rate and student teacher ratio, 2002 Ghost teacher rate Teacher absent rate Teacher shortage rate Teacher turnover rate Student teacher ratio Mean level 14.74 15.06 62.24 39.65 37.80 Teacher absent rate -0.14 1.00 (0.05) Teacher shortage rate -0.12 0.25 1.00 (0.08) (0.00) Teacher turnover rate 0.06 -0.04 0.17 1.00 (0.37) (0.56) (0.01) Student teacher ratio 0.26 -0.03 0.07 0.03 1.00 (0.00) (0.69) (0.31) (0.64) Source: 2002 PESD. 29