99 wave, a random sample of central China was done, which contains about two-thirds of the population. Sampling
weights are not provided for any of the waves. These two surveys are quite different from each other: in the first
wave only 1% of sample was from a town with less than 50,000 people, while in the second wave 63% were. In the
first wave, 60% of respondents were men, the proportion falls to 53% in the second wave. In the 1999-2004 wave the
sampling frame was drawn from a previous nationally representative survey and was conducted throughout the entire
country, with the exception of six remote provinces which comprise 5.1% of the total population. The sample was
also limited to those ages 18-65. Despite some limitations, we believe that the last wave is approximately
representative and include observations from this last wave in our analysis. Observations from the earlier waves are
excluded. If we were to also exclude the final wave, there is no notable impact on our analysis.
The Dominican Republic was included only in the 1995-99 wave. The sample included only 18-49 year olds and
only 4 communities were chosen to be surveyed. We exclude observations from the Dominican Republic from our
analysis.
Egypt was included only in the 1999-2004 wave. The survey notes that a disproportionately high percentage of
housewives were included; examining the survey we find that women are also disproportionately from large urban
areas, in particular Cairo. Since no sample weights are provided, we exclude Egypt from our analysis.
India was included in three waves. In the 1989-93 wave the sample was designed such that 90 percent of respondents
were literate (compared to a population average of less than 50 percent). Interviews were carried out in the eight most
widely-spoken languages of India, but the rural 10 percent of the sample was confined to the five (out of 14) Hindi-
speaking states in the sample. In the 1994-99 wave, the survey was only conducted in Hindi (which is the language of
less than half of the general population), the sample was stratified to allocate 90% of the interviews to urban areas and
10% to rural areas. In 1999-2004, the survey was designed to be representative of 97% of the population and was
conducted in 10 languages. Sample weights were not provided for any of the waves. We include only this last wave
in our analysis.
Nigeria was included in three waves. The 1989-93 and 1994-99 waves focused on the literate and urban portion of
the population. As such, over 40% of the respondents in the first wave had attended university. This proportion falls
to 23% in the second wave, which included a larger rural sample, and to 12% in the 1999-2004 wave which was
designed to be representative of the population. We include only this last wave in our analysis.
Northern Ireland was included in the 1999-2004 wave, but was excluded due to missing GDP data.
Pakistan was included in two waves; however in the 1994-99 wave, sampling was only done in Punjab (a little over
half of Pakistan). In 1999-2004, the sampling frame included the entire country. We include only the 1999-2004
observations in our analysis.
South Africa was included in three waves; however the first wave, 1989-93, over-represents minority races and
blacks were only sampled in certain areas. Sampling weights were not provided. The next two waves, 1994-99 and
1999-2004 were designed to be representative of the population. We exclude observations from the first wave only.
Appendix—4