get_cumulative_estimates {nrba}

Calculate cumulative estimates of a mean/proportion

Description

Calculates estimates of a mean/proportion which are cumulative with respect to a predictor variable, such as week of data collection or number of contact attempts. This can be useful for examining whether estimates are affected by decisions such as whether to extend the data collection period or make additional contact attempts.

Usage

get_cumulative_estimates(
  survey_design,
  y_var,
  y_var_type = NULL,
  predictor_variable
)

Arguments

survey_design

A survey design object created with the survey package.

y_var

Name of a variable whose mean or proportion is to be estimated.

y_var_type

Either NULL, "categorical" or "numeric". For "categorical", proportions are estimated. For "numeric", means are estimated. For NULL (the default), then proportions are estimated if y_var is a factor or character variable. Otherwise, means are estimated. The data will be subset to remove any missing values in this variable.

predictor_variable

Name of a variable for which cumulative estimates of y_var will be calculated. This variable should either be numeric or have categories which when sorted by their label are arranged in ascending order. The data will be subset to remove any missing values of the predictor variable.

Value

A dataframe of cumulative estimates. The first column–whose name matches predictor_variable–gives describes the values of predictor_variable for which a given estimate was computed. The other columns of the result include the following:

outcome

The name of the variable for which estimates are computed

outcome_category

For a categorical variable, the category of that variable

estimate

The estimated mean or proportion.

std_error

The estimated standard error

respondent_sample_size

The number of cases used to produce the estimate (excluding missing values)

References

See Maitland et al. (2017) for an example of a level-of-effort analysis based on this method.

• Maitland, A. et al. (2017). A Nonresponse Bias Analysis of the Health Information National Trends Survey (HINTS). Journal of Health Communication 22, 545-553. doi:10.1080/10810730.2017.1324539

Examples

# Create an example survey design
# with a variable representing number of contact attempts
library(survey)
data(involvement_survey_srs, package = "nrba")


survey_design <- svydesign(
  weights = ~BASE_WEIGHT,
  id = ~UNIQUE_ID,
  fpc = ~N_STUDENTS,
  data = involvement_survey_srs
)

# Cumulative estimates from respondents for average student age ----
get_cumulative_estimates(
  survey_design = survey_design |>
    subset(RESPONSE_STATUS == "Respondent"),
  y_var = "STUDENT_AGE",
  y_var_type = "numeric",
  predictor_variable = "CONTACT_ATTEMPTS"
)

# Cumulative estimates from respondents for proportions of categorical variable ----
get_cumulative_estimates(
  survey_design = survey_design |>
    subset(RESPONSE_STATUS == "Respondent"),
  y_var = "WHETHER_PARENT_AGREES",
  y_var_type = "categorical",
  predictor_variable = "CONTACT_ATTEMPTS"
)