IHDS asked a series of questions about what goods the household owned and about the quality of the housing. Similar housing and consumer goods questions are now widely used in developing country surveys as an easily administered scale measuring household economic level. The summary score (hhassets) represents one of three principal variables measuring a household's economic status. The others are consumption and income. The three economic measures differ primarily in their volatility over time: the household assets measure is the least volatile and measures the economic status of the household over the last several years while the income measure generally would show the most year-to-year fluctuations in economic position. Because of the temporal stability of the assets measure, it is often the best correlate of other household behaviors and outcomes. It is also the simplest measure of the three. The hhassets scale sums 30 dichotomous items measuring household possessions and housing quality. It has a range from zero to 30, with a unweighted mean of 12.25 and unweighted standard deviation of 6.25. See the distributions and India map. A simple sum was chosen as the most appropriate measure for comparing to other similar scales and for eventual comparisons over time of the IHDS scale. Other surveys use more complex principal components scales or weighted sums. The user is free to construct such scales from these items.
|wa1d||Piped indoor water||EH||9||5||1||25%|
The hhassets scale has a Cronbach's reliability coefficient alpha of 0.914 (unweighted). Three items that had been candidates for the scale were dropped because they substantially lowered the reliability coefficient suggesting that they did not measure the same dimension as the remaining 30 items; cg5 owns a generator; hq2a any excrement observed around the house; and hq2b any standing water around the house. Four consumer goods items were modified because they were less expensive alternatives for other items in the scale: cg9 air coolers (vs. cg19 air conditioners); cg7 a black and white television (vs. cg8 a colour television); cg6 a motor scooter (vs. cg18 an automobile); and a cg2 bicycle (vs. a cg6 scooter or cg18 automobile). In these cases, if the household owned the more expensive alternative (e.g., an air conditioner), then the less expensive item (e.g., air cooler) was recoded as owned, regardless of whether the household reported owning the less desirable item. Without this modification, the less expensive items did not scale well; a household might not own them either because they were too affluent (and owned the more expensive alternative) or because they were too poor (and owned neither alternative).
Back to Constructed Variables.