Getting tables and pictures out of SPSS isn't as easy as getting them out of Excel was with the chartpic add-in, especially on a Mac Book Pro running bootcamp, but for those of you with the same dilemmas here was my work around:
I got an external USB keyboard (a really cool rubber foldable extended numeric keyboard with Arabic letters from a shop in Bab El Louk near AUC for only around 25 LE, or 5 bucks! I can roll it up and carry it in my jacket!), and with this I can use the "alt-print scrn" commands (not found on the MBP keyboard) to screen capture the active window. I put that into the GIMP using "paste as new image", crop, and save out as a GIF file (uses much less memory).
Here are my preliminary attempts to look at Predictors of Hot Water White Goods:Statistics
We start with the whole sample, once again looking at the frequency of different heater types:
Then we look at the frequency of having ceramic tiles in the bathroom:
Ceramic tile presence is used as a proxy for investment in a bathroom. Anecdotal observations and focus group conversations suggested that families are unlikely to invest in a wall mounted electric or gas appliance water heater if they haven't first "finished" the bathroom. Often they prefer to wait until they can afford to finish the walls with ceramic tiles before installing a wall mounted heater and they will continue to use the stove to heat water until they can afford to do this. Thus, ceramic tiles become a complementary good to hot water heaters. My hypothesis here is that ceramic tiles are the sin qua non of hot water white good presence, and can be used as a useful predictor.
Let's see if the data bear this hypothesis out:
This pie chart shows the 251 households (out of 456 cases surveyed in both communities, i.e. 55 % ) that have water heating white goods (electric or gas wall mounted heaters) with respect to whether or not they also have ceramic tiles in the bathroom. As you can see, 75 % of these white good possessing households have ceramic tiles and 25% do not.This is consistent with observations on the ground because some families do in fact value hot convenient hot water more than a pretty finished bathroom. But they are in the minority. Thus one cannot say that the presence or absence of ceramic tiles is a perfect predictor for hot water heater preference, but it does seem to have some predictive strength.
Now let's look at users of electric heaters only, and then users of gas heaters only.
(BTW: We use the "IF" statement in SPSS (Data, Select Cases) and use "ANY(V10.2TYPEHEATERCODE,1)" to select all electric heater users. If we want gas heater users we use ""ANY(V10.2TYPEHEATERCODE,2)" and if we want to select both we use "ANY(V10.2TYPEHEATERCODE,1,2)" . The nice thing about the "ANY(test, value, value,...)" function is that you can use it to select out, well, ANY cases you like based on their data code value. I put this note in here so that I don't forget when the Sirens of Titan wipe my memory -- Unk)
First, electric.
Doesn't look all that different from the combined case of Electric and Gas, does it? Here 78% of electric heater users have ceramic tiles in their bathrooms.
Now let's look at gas alone:
Here we see something similar. Of the gas users, 79% have ceramic tiles.Now let's look at all those people who don't have white goods and use either the kitchen stove, a one-eye portable stove, a hamil or a babur:
Here we begin to see some difference, and the possibility of the predictive strength of ceramic tiles as an indicator of hot water system preference. Here only 46.6 per cent of the households who have no white goods have ceramic in their bathrooms.The complication here is that we are dealing with aggregate data for both the Zabaleen and the Darb El Ahmar communities, and on the ground observation has revealed that in Darb El Ahmar there are many "finished" bathrooms (such as the one in the flat we rent on Abu Hureyba street) that have no tiles. The old type of buildings in Darb El Ahmar, often made of stone and wood, often favors finishing bathrooms with stucco and white wash rather than tiles. In the Zabaleen informal community almost all buildings are cement and brick, and the unfinished walls, with bare brick, are not suitable for hanging heaters on. So we might expect different results if we look at each community separately.
(Methodologically we considered changing the question to "is your bathroom finished or unfinished", but since the poor live in an "incremental housing" situation we were afraid that many people would answer "unfinished" regardless of what was in the bathroom, and that others might answer "finished" meaning "I can't afford to do anything more" even if the bathroom lacked many amenities. For this reason we chose "presence of ceramic tiles" as a proxy).
The Zabaleen:
To run the same analysis on just the Zabaleen cases where people lack white goods, SPSS lets us simply go to "Data, Select, If" and type "ANY(V10.1.2TYPEHEATERCODE,3,4,5,6,7,8)&ANY(V0.1REGIONCODE,1)"
We start with the case of people who have no hot water white goods:
We see that 46.9% of them have no ceramic tiles in their bathrooms while 51. 7 % do.
We will try to explain this finding below.
Here is the table and chart of Zabaleen who DO have white good water heaters:
This begins to get interesting. 100% of the 79 households that had water heating white goods (of those: 79.7% electric, 20.3 % gas) had ceramic tiles in their bathrooms.
In other words you could use presence of a commercial hot water system as a reliable predictor of whether or not the bathroom was "finished" if you were in the Zabaleen community of Zurayib.
The converse is not true however, since 51.7 % of those who had no water heating white goods still had ceramic tiles in their bathrooms. This is consistent, however, with ground observations: Many households I interviewed told me anecdotally that they were "preparing their bathroom" for a water heater, but hadn't chosen one yet, and often hadn't finished putting in the pipes or the toilet and sink. So while most people who have ceramic in their bathrooms say they are "intending" to get a commercial hot water heater, we can't use ceramic tiles as definite proof that they will (in fact, some of the homes that had ceramic tiles and finished bathrooms had used electric heaters or gas heaters previously and then had abandoned them or failed to replace them when they broke because of the rising costs of gas and electricity. These people had returned to using the stove.)
Let's see how Darb El Ahmar compares:
Darb El Ahmar
While certainly the majority (69.2%) of Darb El Ahmar residents who use white goods for water heating have ceramic, one cannot use the presence of a water heater as a high confidence predictor of how water systems, in contrast to the Zabaleen case. As mentioned, this is probably because many homes in Darb El Ahmar "finish" their bathrooms without tiles.
In contrast to the Zabaleen case, of those families who do not use white goods, the majority (61 %) have no ceramic. Again this may be a cultural/structural difference since ceramic tiles are not seen as being necessary for "finishing" a bathroom.These preliminary frequency statistics do help bolster the argument, however, that there is much more going on than income and the cost of hot water heaters as determinants of hot water system choice.
Unlike the situation in most "developed" countries, where housing stock is usually "finished" before a family moves in, people in developing countries tend to live in an "incremental" housing situation where preferences and choices are guided by the presence or lack of many complentary goods, not just the presence of market substitutes. As Turner is famous for saying "housing is a verb, not a noun."
No comments:
Post a Comment