Importantly,
our null effect estimates from the random experiment differ substantially from those found from an analysis of CPS data, raising concerns about the potential for selection bias in non-experimental estimates of returns.
Not exact matches
If those without phone service have a higher prevalence of obesity and are likely to live in areas where minorities predominate, and thus advertisements are prevalent, our
effect estimates are likely to be biased towards the
null hypothesis.
Estimates from regressions with detailed controls, nearest - neighbor models, and propensity score models all indicate large, positive, and statistically significant relationships between computer ownership and earnings and employment, in sharp contrast to the
null effects of our experiment.
Under the
null hypothesis, we would expect 4 of the 84
effect estimates displayed in Table 3 to be statistically significant at the p = 0.05 level.
This would likely result in conservative
estimates of the relationships between adverse childhood experiences as persons who had potentially been exposed to an experience would always be misclassified as unexposed; this type of misclassification would bias our results toward the
null.32 However, to assess this potential
effect, we repeated our analyses after excluding any respondent with missing information on any one of the adverse childhood experiences.