{
  "alt": "Dot chart with three panels and interval lines. The top panel shows the rise in the ANES gap from 2012 to 2016 within each mode, {fact:st_a_ftf_1216} points in person and {fact:st_a_web_1216} on the web, and the change from 2022 to 2024 within each GSS mode, {fact:st_g_ftf_2224} points in person and {fact:st_g_web_2224} on the web. The middle panel shows web minus in-person gaps for ANES 2012, 2016 and 2024 and GSS 2022 and 2024. The bottom panel shows web minus in-person Republican shares; the ANES differences are {fact:st_a_moder_2012}, {fact:st_a_moder_2016} and {fact:st_a_moder_2024} points and the GSS differences are {fact:st_g_moder_2022} and {fact:st_g_moder_2024}.",
  "caption": "In the top panel the ANES gap rose by {fact:st_a_ftf_1216} points in person and {fact:st_a_web_1216} on the web between 2012 and 2016. In the lower panels the ANES web estimates are above the in-person estimates in every year, although one gap interval includes zero, while the GSS web and in-person Republican shares differ by {fact:st_g_moder_2024} points in 2024. Dots are estimates and lines are 95% confidence intervals.",
  "data": [
    {
      "comparison": "ANES in person, 2012 to 2016",
      "conf_high": 31.3784,
      "conf_low": 11.4151,
      "estimate": 21.3967,
      "n": 2562,
      "panel": "Change in the gap within one mode"
    },
    {
      "comparison": "ANES web, 2012 to 2016",
      "conf_high": 24.5508,
      "conf_low": 11.4036,
      "estimate": 17.9772,
      "n": 5242,
      "panel": "Change in the gap within one mode"
    },
    {
      "comparison": "GSS in person, 2022 to 2024",
      "conf_high": 18.596,
      "conf_low": -5.4572,
      "estimate": 6.5694,
      "n": 1909,
      "panel": "Change in the gap within one mode"
    },
    {
      "comparison": "GSS web, 2022 to 2024",
      "conf_high": 10.9227,
      "conf_low": -9.60202,
      "estimate": 0.660358,
      "n": 2389,
      "panel": "Change in the gap within one mode"
    },
    {
      "comparison": "ANES 2012",
      "conf_high": 16.3992,
      "conf_low": 0.025528,
      "estimate": 8.21235,
      "n": 4664,
      "panel": "Web minus in-person gap, same year"
    },
    {
      "comparison": "ANES 2016",
      "conf_high": 13.5004,
      "conf_low": -3.91478,
      "estimate": 4.79281,
      "n": 3140,
      "panel": "Web minus in-person gap, same year"
    },
    {
      "comparison": "ANES 2024",
      "conf_high": 24.9606,
      "conf_low": 3.4105,
      "estimate": 14.1855,
      "n": 2432,
      "panel": "Web minus in-person gap, same year"
    },
    {
      "comparison": "GSS 2022",
      "conf_high": 25.0308,
      "conf_low": 1.97419,
      "estimate": 13.5025,
      "n": 2266,
      "panel": "Web minus in-person gap, same year"
    },
    {
      "comparison": "GSS 2024",
      "conf_high": 18.4126,
      "conf_low": -3.22571,
      "estimate": 7.59344,
      "n": 2032,
      "panel": "Web minus in-person gap, same year"
    },
    {
      "comparison": "ANES 2012",
      "conf_high": 15.7022,
      "conf_low": 2.95449,
      "estimate": 9.32833,
      "n": 1827,
      "panel": "Web minus in-person Republican share, same year"
    },
    {
      "comparison": "ANES 2016",
      "conf_high": 18.7253,
      "conf_low": 5.46862,
      "estimate": 12.0969,
      "n": 1476,
      "panel": "Web minus in-person Republican share, same year"
    },
    {
      "comparison": "ANES 2024",
      "conf_high": 27.5404,
      "conf_low": 11.1253,
      "estimate": 19.3329,
      "n": 1125,
      "panel": "Web minus in-person Republican share, same year"
    },
    {
      "comparison": "GSS 2022",
      "conf_high": 22.2475,
      "conf_low": 3.00329,
      "estimate": 12.6254,
      "n": 996,
      "panel": "Web minus in-person Republican share, same year"
    },
    {
      "comparison": "GSS 2024",
      "conf_high": 6.85297,
      "conf_low": -7.68309,
      "estimate": -0.415062,
      "n": 972,
      "panel": "Web minus in-person Republican share, same year"
    }
  ],
  "id": "mode_check",
  "notes": [
    "The 2020 ANES study has no separate weights by mode, so it has no row. The 2024 ANES mode groups are the fresh in-person and fresh web samples; the panel and paper respondents are in neither.",
    "The GSS rows compare 2022 or 2024 in-person with web respondents; the 2021 survey was mostly web and has no in-person comparison."
  ],
  "source": "ANES Cumulative Data File, 2012, 2016 and 2024, and General Social Survey, 2022 and 2024; weighted (weight_ftf and weight_web, the ANES weights for each mode; weight_use)",
  "spec": {
    "facet": "panel",
    "hi": "conf_high",
    "lo": "conf_low",
    "n": "n",
    "reference": 0,
    "sort": "data",
    "x": "estimate",
    "xDomain": [
      -15,
      40
    ],
    "xFormat": {
      "digits": 1
    },
    "xLabel": "Change or difference (percentage points)",
    "y": "comparison"
  },
  "subtitle": "Changes in the Republican minus Democratic gap within one mode, and web minus in-person differences in the gap and in the Republican share; leaners included; 95% intervals",
  "table": {
    "columns": [
      {
        "key": "panel",
        "label": "Panel"
      },
      {
        "key": "comparison",
        "label": "Comparison"
      },
      {
        "format": {
          "digits": 1
        },
        "key": "estimate",
        "label": "Estimate (points)"
      },
      {
        "format": {
          "digits": 1
        },
        "key": "conf_low",
        "label": "Lower"
      },
      {
        "format": {
          "digits": 1
        },
        "key": "conf_high",
        "label": "Upper"
      },
      {
        "format": {
          "thousands": true
        },
        "key": "n",
        "label": "Respondents"
      }
    ]
  },
  "title": "The ANES gap opened within each interview mode, and ANES web respondents were more restrictive than in-person respondents",
  "type": "dot",
  "width": "body"
}