Supplementary Material > Volume 113 Issue 490 (October 2003)

Volume 113 Issue 490 (October 2003)

Precautionary saving and social learning across generations: an experiment

T. Parker Ballinger, Michael G. Palumbo, Nathaniel T. Wilcox

 File size: 459KB

Description of data:
This is a description of the data file "EJdATA.txt" found in this folder. That file contains
all of the data used in the paper PRECAUTIONARY SAVING AND SOCIAL LEARNING ACROSS GENERATIONS:
AN EXPERIMENT, by T. Parker Ballinger, Michael G. Palumbo and Nathaniel T. Wilcox, currently
planned for publication in the October 2003 issue of the ECONOMIC JOURNAL.

The data is arranged in five columns.

Column 1 is a numeric treatment indicator.
1 - Low Variance Income with communication (LV-C).
2 - High Variance Income with communication (HV-C).
3 - High Variance Income with no communication (HV-NC).
4 - High Variance Income and subjects wait 30 minutes before starting (HV-30). See the discussion associated with Table 7.
5 - High Variance Income and subjects observe random income sequence before starting (HV-R). See the discussion associated with Table 7.
6 - High Variance Income and subjects receive extra instruction regarding the payoff
function (HV-EI). See the discussion associated with Table 7.
7 - High Variance Income with rational pretest result (HV-Rat). See the discussion associated with Table 8.
8 - High Variance Income with naive pretest result (HV-Naive). See the discussion associated with Table 8.

Column 2 are subject identification numbers (SINs) for each treatment. There are 18 subjects in LV-C, HV-C, HV-NC. There are 6 subjects in HV-30 and HV-R. There are 7 subjects in HV-EI. There are 13 subjects in HV-Rat. There are 14 subjects in HV-Naive.

Only treatments 1, 2 and 3 (LV-C,HV-C and HV-NC) are treatments with "families" of three generations. The convention of SINs within each of these treatments is the same: Subjects 1, 2 and 3 are family one; subjects 4, 5 and 6 are family two; and so on, with subjects 16, 17 and 18 making up family six. Within each family, the lowest SIN is the first generation subject, the middle SIN is the second generation subject and the highest SIN is the third generation subject. For instance, in family four, SIN=10 indicates the first generation subject, SIN=11 indicates the second generation subject and SIN=12 indicates the third generation subject.

Column 3 is the period identifier. 1 to 60 for each subject in each treatment.

Column 4 is the income draw in each period, for each subject, in each treatment.

Column 5 is the spending decision in each period, for each subject, in each treatment.

Note: In LV-C, there is one subject (subject 10) who, due to a computer glitch, has missing observations for periods 17 and 43. In HV-c, there is one subject (subject 5) who, due to a computer glitch, has a missing observation in period 49. Each of these missing observations is clearly annotated with the word "missing."