{
  "alt": "A table with one row for each dated item in each of three groups. The best-fitting first new survey is 2016 in every row for white respondents and for white Democrats, with steps of several points for white respondents and larger steps for white Democrats. The white Republican rows show dates of 2004, 2016 or 2018 and smaller steps at the 2016 survey. The kinked-trend error is lower than the step error for discrimination and will power among white respondents and higher for government help and spending, and the kink years fall between about 2011 and 2014.",
  "caption": "Every row for white respondents and for white Democrats has 2016 as the first new survey. For white Republicans the best dates run from 2004 to 2018, so their steps are also given at a fixed 2016 date, and {fact:n_rep_zero} of the six dated items have best-step intervals that include zero. The last columns compare the error of the best step with the error of the best trend with one kink: among white respondents the kink fits the discrimination and will-power shares better, and the step fits the government-help and spending shares better.",
  "data": [
    {
      "best_year": "2016",
      "fix16": 8.9,
      "fix16_high": 11.6,
      "fix16_low": 5.8,
      "group": "White respondents",
      "item_label": "Discrimination",
      "kink": 2014,
      "kink_high": 2015.5,
      "kink_low": 2012,
      "kink_sse": 2.4,
      "p_best_pct": 74,
      "p_earlier_pct": 1,
      "step": 8.9,
      "step_sse": 8
    },
    {
      "best_year": "2016",
      "fix16": -10.6,
      "fix16_high": -7.6,
      "fix16_low": -13.3,
      "group": "White respondents",
      "item_label": "Lack of motivation or will power",
      "kink": 2011.4,
      "kink_high": 2014.8,
      "kink_low": 2008,
      "kink_sse": 3,
      "p_best_pct": 51,
      "p_earlier_pct": 38,
      "step": -10.6,
      "step_sse": 15.2
    },
    {
      "best_year": "2016",
      "fix16": 9.2,
      "fix16_high": 11.6,
      "fix16_low": 7,
      "group": "White respondents",
      "item_label": "At the help end of the scale",
      "kink": 2013.9,
      "kink_high": 2014,
      "kink_low": 2012.9,
      "kink_sse": 12.9,
      "p_best_pct": 100,
      "p_earlier_pct": 0,
      "step": 9.2,
      "step_sse": 9.2
    },
    {
      "best_year": "2016",
      "fix16": 20.3,
      "fix16_high": 23.7,
      "fix16_low": 17.3,
      "group": "White respondents",
      "item_label": "Too little spent",
      "kink": 2013.4,
      "kink_high": 2014,
      "kink_low": 2010.9,
      "kink_sse": 24.1,
      "p_best_pct": 100,
      "p_earlier_pct": 0,
      "step": 20.3,
      "step_sse": 16.1
    },
    {
      "best_year": "2016",
      "fix16": 7.2,
      "fix16_high": 9.8,
      "fix16_low": 4.9,
      "group": "White respondents",
      "item_label": "Favors preferential hiring",
      "kink": 2013.6,
      "kink_high": 2014.7,
      "kink_low": 2006,
      "kink_sse": 9.1,
      "p_best_pct": 87,
      "p_earlier_pct": 10,
      "step": 7.2,
      "step_sse": 9.6
    },
    {
      "best_year": "2016",
      "fix16": 9.3,
      "fix16_high": 11.8,
      "fix16_low": 7.2,
      "group": "White respondents",
      "item_label": "Rejects the special-favors view",
      "kink": 2012.8,
      "kink_high": 2013.9,
      "kink_low": 2011.4,
      "kink_sse": 4.6,
      "p_best_pct": 92,
      "p_earlier_pct": 7,
      "step": 9.3,
      "step_sse": 11.2
    },
    {
      "best_year": "2016",
      "fix16": 18.5,
      "fix16_high": 23.2,
      "fix16_low": 14.1,
      "group": "White Democrats",
      "item_label": "Discrimination",
      "kink": 2012.4,
      "kink_high": 2013.8,
      "kink_low": 2007.3,
      "kink_sse": 5.5,
      "p_best_pct": 78,
      "p_earlier_pct": 20,
      "step": 18.5,
      "step_sse": 13.1
    },
    {
      "best_year": "2016",
      "fix16": -17.6,
      "fix16_high": -13.3,
      "fix16_low": -21.4,
      "group": "White Democrats",
      "item_label": "Lack of motivation or will power",
      "kink": 2008.6,
      "kink_high": 2013.4,
      "kink_low": 2004.1,
      "kink_sse": 6,
      "p_best_pct": 60,
      "p_earlier_pct": 39,
      "step": -17.6,
      "step_sse": 15.5
    },
    {
      "best_year": "2016",
      "fix16": 19.1,
      "fix16_high": 23.4,
      "fix16_low": 15.1,
      "group": "White Democrats",
      "item_label": "At the help end of the scale",
      "kink": 2013.5,
      "kink_high": 2014,
      "kink_low": 2012,
      "kink_sse": 14.5,
      "p_best_pct": 100,
      "p_earlier_pct": 0,
      "step": 19.1,
      "step_sse": 9.2
    },
    {
      "best_year": "2016",
      "fix16": 29.1,
      "fix16_high": 34.3,
      "fix16_low": 24,
      "group": "White Democrats",
      "item_label": "Too little spent",
      "kink": 2013.5,
      "kink_high": 2014,
      "kink_low": 2009.6,
      "kink_sse": 20.1,
      "p_best_pct": 100,
      "p_earlier_pct": 0,
      "step": 29.1,
      "step_sse": 15.3
    },
    {
      "best_year": "2016",
      "fix16": 13.8,
      "fix16_high": 18.1,
      "fix16_low": 9.4,
      "group": "White Democrats",
      "item_label": "Favors preferential hiring",
      "kink": 2011.1,
      "kink_high": 2014.7,
      "kink_low": 2006,
      "kink_sse": 6.4,
      "p_best_pct": 65,
      "p_earlier_pct": 31,
      "step": 13.8,
      "step_sse": 12.6
    },
    {
      "best_year": "2016",
      "fix16": 19.6,
      "fix16_high": 24.6,
      "fix16_low": 15.4,
      "group": "White Democrats",
      "item_label": "Rejects the special-favors view",
      "kink": 2011.6,
      "kink_high": 2013.4,
      "kink_low": 2008,
      "kink_sse": 7.1,
      "p_best_pct": 85,
      "p_earlier_pct": 15,
      "step": 19.6,
      "step_sse": 11.2
    },
    {
      "best_year": "2004",
      "fix16": 0.4,
      "fix16_high": 4.3,
      "fix16_low": -3.3,
      "group": "White Republicans",
      "item_label": "Discrimination",
      "kink": 2015.9,
      "kink_high": 2015.9,
      "kink_low": 2004,
      "kink_sse": 8.3,
      "p_best_pct": 41,
      "p_earlier_pct": 0,
      "step": -3.3,
      "step_sse": 6.4
    },
    {
      "best_year": "2018",
      "fix16": -6.8,
      "fix16_high": -2.5,
      "fix16_low": -11.2,
      "group": "White Republicans",
      "item_label": "Lack of motivation or will power",
      "kink": 2012.6,
      "kink_high": 2015.9,
      "kink_low": 2009.3,
      "kink_sse": 3.8,
      "p_best_pct": 39,
      "p_earlier_pct": 61,
      "step": -10.4,
      "step_sse": 6.5
    },
    {
      "best_year": "2018",
      "fix16": 2.9,
      "fix16_high": 6.1,
      "fix16_low": 0.5,
      "group": "White Republicans",
      "item_label": "At the help end of the scale",
      "kink": 2015.4,
      "kink_high": 2015.9,
      "kink_low": 2004,
      "kink_sse": 10.9,
      "p_best_pct": 50,
      "p_earlier_pct": 50,
      "step": 6.3,
      "step_sse": 11.9
    },
    {
      "best_year": "2016",
      "fix16": 15,
      "fix16_high": 20.2,
      "fix16_low": 10.6,
      "group": "White Republicans",
      "item_label": "Too little spent",
      "kink": 2013.5,
      "kink_high": 2014,
      "kink_low": 2010.2,
      "kink_sse": 14.8,
      "p_best_pct": 98,
      "p_earlier_pct": 2,
      "step": 15,
      "step_sse": 10.9
    },
    {
      "best_year": "2016",
      "fix16": 1.4,
      "fix16_high": 4.2,
      "fix16_low": -1,
      "group": "White Republicans",
      "item_label": "Favors preferential hiring",
      "kink": 2014.2,
      "kink_high": 2015.9,
      "kink_low": 2004,
      "kink_sse": 3.9,
      "p_best_pct": 28,
      "p_earlier_pct": 50,
      "step": 1.4,
      "step_sse": 4.1
    },
    {
      "best_year": "2018",
      "fix16": 2.1,
      "fix16_high": 5.2,
      "fix16_low": -0.4,
      "group": "White Republicans",
      "item_label": "Rejects the special-favors view",
      "kink": 2015.9,
      "kink_high": 2015.9,
      "kink_low": 2004,
      "kink_sse": 13.5,
      "p_best_pct": 62,
      "p_earlier_pct": 38,
      "step": 6.7,
      "step_sse": 13.6
    }
  ],
  "id": "step_dates",
  "notes": [
    "The fit uses the surveys of 2000, 2002, 2004, 2006, 2008, 2010, 2012, 2014, 2016 and 2018, all conducted in person, so the dates cannot come from the move to the web. A step cannot begin before 2004 or leave fewer than one survey after it. The kinked trend is constant before the kink and linear after it, with the kink anywhere from 2004 to just before 2016. A step and a kink are assumptions about the form of change. The bootstrap resamples respondents within each survey year with replacement (seed 12345) and refits; the interval for a step size or a kink year is the 2.5th to 97.5th percentile. Survey spacing is two years, so a date is a survey and not a month."
  ],
  "source": "General Social Survey, 2000-2018; weighted (weight_use); in-person interviews; white respondents (race code white, Hispanic whites included), leaners included for party; seeded bootstrap of respondents within each survey year",
  "spec": {
    "columns": [
      {
        "key": "group",
        "label": "Group"
      },
      {
        "key": "item_label",
        "label": "Item"
      },
      {
        "key": "best_year",
        "label": "Best-fitting first new survey"
      },
      {
        "format": {
          "digits": 1,
          "signed": true
        },
        "key": "step",
        "label": "Step at that survey (points)"
      },
      {
        "format": {
          "digits": 1,
          "signed": true
        },
        "key": "fix16",
        "label": "Step at the 2016 survey (points)"
      },
      {
        "format": {
          "digits": 1,
          "signed": true
        },
        "key": "fix16_low",
        "label": "2016 step, lower end"
      },
      {
        "format": {
          "digits": 1,
          "signed": true
        },
        "key": "fix16_high",
        "label": "2016 step, upper end"
      },
      {
        "format": {
          "digits": 0
        },
        "key": "p_best_pct",
        "label": "Replicates at that survey (%)"
      },
      {
        "format": {
          "digits": 0
        },
        "key": "p_earlier_pct",
        "label": "Replicates at an earlier survey (%)"
      },
      {
        "format": {
          "digits": 1
        },
        "key": "step_sse",
        "label": "Error of the best step"
      },
      {
        "format": {
          "digits": 1
        },
        "key": "kink_sse",
        "label": "Error of the kinked trend"
      },
      {
        "format": {
          "digits": 1
        },
        "key": "kink",
        "label": "Kink year"
      },
      {
        "format": {
          "digits": 1
        },
        "key": "kink_low",
        "label": "Kink year, lower end"
      },
      {
        "format": {
          "digits": 1
        },
        "key": "kink_high",
        "label": "Kink year, upper end"
      }
    ]
  },
  "subtitle": "Best-fitting single step in the yearly share, by item and group, in-person surveys 2000 to 2018, among white respondents (GSS race code, Hispanic whites included); step in percentage points, share of bootstrap replicates choosing each date, and the error of a trend with one kink",
  "title": "The best single step falls at the 2016 survey for each of the six dated items among white respondents and among white Democrats, and a kinked trend fits discrimination and will power better",
  "type": "table",
  "width": "wide"
}