ANES Time Series

ANES time series studies are administered around each presidential election. They are all two-wave studies, corresponding to a pre-election interview and a post-election interview.

Contents

ANES Time Series
1. Weights

Weights

Since the web mode was added in 2012, separate weights have been made available for the modal samples.

Vintage	Web interviews	Face-to-face interviews	Combined
2012	`weight_web`	`weight_ftf`	`weight_full`
2016	`V160102w`	`V160102f`	`V160102`

For the 2020 vintage:

Sample	Pre-election	Post-election
Full sample	`V200010a`	`V200010b`
Panel sample	`V200011a`	`V200011b`
Web interviews	`V200012a`	`V200012b`
Mixed web/telephony interviews	`V200013a`	`V200013b`
Mixed web/video interviews	`V200014a`	`V200014b`
Fresh sample	`V200015a`	`V200015b`
All excluding mixed web/video interviews	`V200016a`	`V200016b`

For the 2024 vintage:

Sample	Pre-election	Post-election
Face-to-face interviews	`V240101a`	`V240101b`
Web interviews	`V240102a`	`V240102b`
Face-to-face or web interviews	`V240103a`	`V240103b`
Web interviews or PAPI	`V240104a`	`V240104b`
Fresh sample	`V240105a`	`V240105b`
Panel sample	`V240106a`	`V240106b`
Full sample	`V240107a`	`V240107b`
Full sample excluding PAPI	`V240108a`	`V240108b`

Declare the complex survey design like:

proc surveyfreq;
  stratum STRATUM;
  cluster CLUSTER;
  weight WEIGHT;
  table VOTE PARTY VOTE*PARTY ;
run;

Or:

svyset CLUSTER [pweight=WGT], strata(STRATUM) vce(linearize) singleunit(centered)

Until the 2006 pilot, clusters and strata were published as a single variable. This could be broken apart like:

compute STRATUM = trunc(v040103/10).
compute CLUSTER = v040103 ‐ (10*STRATUM)

CategoryRicottone

AmericanNationalElectionStudies/TimeSeries (last edited 2025-12-18 16:05:26 by DominicRicottone)