regression is a type of analysis that uses one or more independent variables to predict one dependent variable. So, for example, children’s test scores (the variable under observation, i.e., the dependent variable) could be influenced by many factors (such as parental reading frequency, household income, and parental education—the independent variables). Using all of the data available, a regression model finds the estimates that best represent the data. So, in our sample, the regression results would tell us whether children whose parents read to them more have higher test scores than other children whose parents have the same level of education and household income but read to them less. A coefficient (i.e., estimated effect) of 10 on the reading variable would mean that each additional time per week that a parent reads to a child, that child’s test score is expected to rise by 10 points. Using a p-value (see above) we can determine whether that coefficient is significant.