Thursday, December 15, 2016

Americans don't really want to spend more on education

...we surveyed representative samples of the adult populations in Germany and the U.S. and implemented three randomized experiments on how the provision of information affects support for education spending. ...

We find that a vast majority of the public in both countries underestimates current levels of school spending and teacher salaries. Absent the provision of information, an absolute majority in both countries supports increased government spending on education, with somewhat higher levels of support among Germans than Americans (71 percent vs. 60 percent).

Our first survey experiment shows that citizens of both countries also react similarly to two information treatments, with treatment effects (relative to the control mean) hardly differing. Informing respondents about the current level of annual public education spending per student reduces support for increased spending by more than one quarter (to 50 percent in Germany and 43 percent in the U.S.). Additionally stating that the spending increase would be financed through higher taxation reduces support by more than half compared to the control group (to 30 percent in Germany and 26 percent in the U.S.), with the shares in support no longer differing significantly between the two countries.

When respondents are informed about current salary levels, the share who support increases in teacher salaries declines sharply by about 40 percent (relative to the control mean) in both countries, although baseline support is much lower in Germany. ...

Further analysis confirms that these treatment effects reflect actual information effects, rather than simply the effect of being primed to think about monetary values as opposed to, say, observable conditions in local schools before reporting support for spending increases (Iyengar et al. (1984); Krosnick and Kinder (1990)). In both countries, treatment effects are substantially larger for respondents who underestimated actual levels and are almost zero for respondents who had already been well informed prior to the information treatment.
--Martin West, Ludger Woessmann, Philipp Lergetporer, and Katharina Werner, "How information affects support for education spending: Evidence from survey experiments in Germany and the United States," on why we don't spend more on education