{
  "alt": "A dot plot of five agreement correlations with intervals, all well above zero. Age and religiosity are highest, the obedience measure is in the middle and education and sex are lowest, and the intervals overlap for the obedience measure and each of the other rows.",
  "caption": "Age and religiosity agree more closely across surveys than the obedience measure does, and education and sex agree less, but all intervals overlap. Each dot compares the country-by-country correlations of left-right with the predictor in the two surveys.",
  "data": [
    {
      "conf_high": 0.900668,
      "conf_low": 0.660316,
      "definition": "WVS and EVS full index against ESS conservation values",
      "estimate": 0.812574,
      "group": "Obedience measure",
      "n_units": 36,
      "predictor": "Obedience measure",
      "spread_wvs": 0.0753776
    },
    {
      "conf_high": 0.934001,
      "conf_low": 0.764304,
      "definition": "Importance of God in the WVS and EVS against attendance and prayer in the ESS",
      "estimate": 0.873502,
      "group": "Reference predictor",
      "n_units": 36,
      "predictor": "Religiosity",
      "spread_wvs": 0.102802
    },
    {
      "conf_high": 0.837495,
      "conf_low": 0.468921,
      "definition": "Three levels",
      "estimate": 0.696621,
      "group": "Reference predictor",
      "n_units": 34,
      "predictor": "Education",
      "spread_wvs": 0.0820707
    },
    {
      "conf_high": 0.950208,
      "conf_low": 0.818265,
      "definition": "Years",
      "estimate": 0.903826,
      "group": "Reference predictor",
      "n_units": 36,
      "predictor": "Age",
      "spread_wvs": 0.100843
    },
    {
      "conf_high": 0.848943,
      "conf_low": 0.515346,
      "definition": "Female against male",
      "estimate": 0.721693,
      "group": "Reference predictor",
      "n_units": 36,
      "predictor": "Sex (female)",
      "spread_wvs": 0.04528
    }
  ],
  "id": "agreement_reference",
  "notes": [
    "Every row uses the countries paired in the obedience comparison, and a country enters a row only if its effective sample is 300 or more for that predictor in both surveys, so the number of countries differs by row.",
    "Intervals are model based (Fisher z on the Kish effective sample size) and ignore clustering in the sample design."
  ],
  "source": "World Values Survey and European Values Study, fieldwork 2005-2023, and European Social Survey, rounds 3 to 10 (2006-2022); weighted (weight within each survey); countries with an effective sample of 300 or more in both",
  "spec": {
    "colors": {
      "Obedience measure": "cat1",
      "Reference predictor": "muted"
    },
    "group": "group",
    "hi": "conf_high",
    "lo": "conf_low",
    "n": "n_units",
    "sort": "desc",
    "x": "estimate",
    "xDomain": [
      0,
      1
    ],
    "xFormat": {
      "digits": 1
    },
    "xLabel": "Correlation across countries between the two surveys",
    "y": "predictor"
  },
  "subtitle": "Correlation across countries between the WVS and EVS and the ESS correlations of left-right with each predictor; 95% Fisher z intervals",
  "table": {
    "columns": [
      {
        "key": "predictor",
        "label": "Predictor"
      },
      {
        "key": "definition",
        "label": "Definition"
      },
      {
        "format": {
          "digits": 2
        },
        "key": "estimate",
        "label": "Correlation between surveys"
      },
      {
        "format": {
          "digits": 2
        },
        "key": "conf_low",
        "label": "Interval low"
      },
      {
        "format": {
          "digits": 2
        },
        "key": "conf_high",
        "label": "Interval high"
      },
      {
        "key": "n_units",
        "label": "Countries"
      },
      {
        "format": {
          "digits": 2
        },
        "key": "spread_wvs",
        "label": "Spread of WVS correlations (standard deviation)"
      }
    ]
  },
  "title": "The two surveys agree as closely on age and religiosity as on obedience, so agreement is not specific to the measure",
  "type": "dot",
  "width": "wide"
}