Presidential Leadership of Public Opinion: An Embedded Survey Experiment

The Promise of Experiments for Studying the Presidency

“…They found support for the priming hypotheses in both the experiment and survey. Since Druckman and Holmes (), more presidency research using experiments has been published, although, compared to other subfields, the number is still relatively small (Braman ; Christenson and Kriner forthcoming; Cohen ; Kriner and Reeves ; Reeves and Rogowski ; Rottinghaus and Tedin ; Sirin and Villalobos ; Streb et al ; Tedin, Rottinghaus, and Rodgers ; Villalobos and Sirin )…”

“…There are several recurrent themes in this research. First, much of it looks at the ability of presidents to influence public opinion (e.g., Cohen ; Rottinghaus and Tedin ; Sirin and Villalobos ; Tedin, Rottinghaus, and Rodgers ; Villalobos and Sirin ), drawing on the interest of the earlier round of experimental research on the presidency. But another set of studies has turned its attention to public support for presidential unilateral behavior (Braman ; Kriner and Christenson ; Reeves and Rogowski ).…”

“…But another set of studies has turned its attention to public support for presidential unilateral behavior (Braman ; Kriner and Christenson ; Reeves and Rogowski ). Second, many of these studies embed their experiments in surveys with nationally representative samples, owing to concerns of generalizability of findings to the real world (e.g., Braman ; Cohen ; Kriner and Christenson ; Reeves and Rogowski ; Streb et al ).…”

“…For instance, Cohen () investigated whether political predispositions, in particular partisanship, moderate the impact of a presidential communication. I hypothesized that same‐party and opposition‐party identifiers would respond differently to a presidential communication depending upon whether the president is viewed as taking a party‐leadership versus national‐leadership role.…”

“…One way to address this concern is to use a weak treatment (Levendusky ). Referring back to Cohen () discussed above, the treatment provides respondents with two conditions, both using George W. Bush's own and identical words from two speeches, on the president's positions on immigration policy. The treatment merely identified briefly the venue and date of the speech, that is was either delivered to the nation or to a Republican audience.…”