LAPOP Multi-Line Time Series Graph Pre-Processing

This function creates a dataframe which can then be input in lapop_mline for to show a time series plot with multiple lines. If one "outcome" variable and an `xvar` variable is supplied, the function produces the values of a single outcome variable, broken down by a secondary variable, across time. If multiple outcome variables (up to four) are supplied, it will show means/percentages of those variables across time (essentially, it allows you to do lpr_ts for multiple variables).

Usage

lpr_mline(
  data,
  outcome,
  rec = c(1, 1),
  rec2 = c(1, 1),
  rec3 = c(1, 1),
  rec4 = c(1, 1),
  xvar,
  use_wave = FALSE,
  use_cat = FALSE,
  ci_level = 0.95,
  mean = FALSE,
  filesave = "",
  cfmt = "",
  ttest = FALSE,
  keep_nr = FALSE
)

Arguments

data: A survey object. The data that should be analyzed.
outcome: Character vector. Outcome variable(s) of interest to be plotted across time. If only one value is provided, the graph will show the outcome variable, over time, broken down by a secondary variable (x-var). If more than one value is supplied, the graph will show each outcome variable across time (no secondary variable).
rec, rec2, rec3, rec4: Numeric. The minimum and maximum values of the outcome variable that should be included in the numerator of the percentage. For example, if the variable is on a 1-7 scale and rec is c(5, 7), the function will show the percentage who chose an answer of 5, 6, 7 out of all valid answers. Can also supply one value only, to produce the percentage that chose that value out of all other values. Default: c(1, 1).
xvar: Character. Variable on which to break down the outcome variable. In other words, the line graph will produce multiple lines for each value of xvar (technically, it is the z-variable, not the x variable, which is year/wave). Ignored if multiple outcome variables are supplied.
use_wave: Logical. If TRUE, will use "wave" for the x-axis; otherwise, will use "year". Default: FALSE.
use_cat: Logical. If TRUE, will show the percentages of category values of a single variable; otherwise will show percentages of the range of values from rec(). Default FALSE.
ci_level: Numeric. Confidence interval level for estimates. Default: 0.95
mean: Logical. If TRUE, will produce the mean of the variable rather than rescaling to percentage. Default: FALSE.
filesave: Character. Path and file name to save the dataframe as csv.
cfmt: Character. changes the format of the numbers displayed above the bars. Uses sprintf string formatting syntax. Default is whole numbers for percentages and tenths place for means.
ttest: Logical. If TRUE, will conduct pairwise t-tests for difference of means between all individual x levels and save them in attr(x, "t_test_results"). Default: FALSE.
keep_nr: Logical. If TRUE, will convert "don't know" (missing code .a) and "no response" (missing code .b) into valid data (value = 99) and use them in the denominator when calculating percentages. The default is to examine valid responses only. Default: FALSE.

Value

Returns a data frame, with data formatted for visualization by lapop_mline

Author

Luke Plutowski, luke.plutowski@vanderbilt.edu & Robert Vidigal, robert.vidigal@vanderbilt.edu

Examples


# Single Variable
if (FALSE) { # \dontrun{
lpr_mline(ym23,
outcome = "ing4",
rec = c(5, 7),
use_wave = FALSE)

# Multiple Variables
lpr_mline(cm23,
outcome = c("b13", "b18", "b21"),
rec = c(5, 7),
rec2 = c(1, 2),
rec3 = c(5, 7),
rec4 = c(1, 1),
use_wave = TRUE)

# Binary Single Variable by Category
lpr_mline(bra23,
outcome = "jc1",
use_cat = TRUE,
use_wave = TRUE)

# Recode Categorical Variable (max 4-categories)
lpr_mline(gm,
outcome = "a4n",
rec = c(1,4),
use_cat = TRUE,
use_wave = TRUE)} # }