ANES Time Series 2016: Timing Variables (version: 20191104)

1) The timing data files contain the four digit Case ID variable found in the main 2016 ANES Time
Series data file (V160001).

2) Timing data from Westat was received rounded to the nearest second. More precies increments
of time are not available.

3) Generally speaking, the timing variable names match the corresponding variable names in the 2016
ANES Time Series data file with an "_t" at the end. Exceptions include instances where there was not 
a one for one correspondnace. For example, in some cases, a series of variables from the Time Series 
data might have only one timing data variable (usually because there was only one screen in the 
instrument for more than one question). Ranges of Time Series variables captured in one timing variable
are noted in the variable label.

4) In some cases, the timing data from the face-to-face and web instruments was captured in slightly different
ways. For example, sometimes an intro screen in the face-to-face instrument was a separate screen from the first
question in that module whereas in the web instrument both the intro screen and the first question appeared on 
the same screen. In that case, the web timing variable represents the entire time the respondent spent on the 
one screen containing both the intro and the first question. The face-to-face instrument, on the other hand, 
yielded two separate timing variables. These differences are noted in the variable labels.

5) Missing data values are negative. Please note that these should be set to missing before performing numeric 
computations using the timing data.

6) Problems were encountered in the face-to-face timing data with certain branching questions. If there was a
main question and a follow up questions, the same timing was output for both. The timing data for the follow
up questions, in these cases, was correct. Timing data for the main question was recoded to -4 (error). ANES
is still investigating the issue and will release revised files if the correct timing data can be recovered.

7) There are a small number of variables for which no timing data was recorded. 