Generate a data quality report for survey responses
Source:R/quality_report.R
sframe_quality_results.RdEvaluates collected response data against the instrument specification and produces a structured quality report. The report covers attention check performance, completion time, straight-lining within scale blocks, item-level missingness, respondent-level missingness, and duplicate respondent IDs where supplied.
Usage
sframe_quality_results(
data,
instrument,
respondent_id = NULL,
submitted_at = NULL,
started_at = NULL,
time_min = NULL,
straightline_scales = TRUE,
straightline_min_items = 4L,
missing_threshold = 0.2
)Arguments
- data
A
tibbleordata.frameof responses, typically produced byread_responses().- instrument
An
sframeobject created bysframe_instrument().- respondent_id
Character or NULL. The column name holding unique respondent identifiers. Used for duplicate detection.
- submitted_at
Character or NULL. The column name holding submission timestamps. Used for completion time analysis.
- started_at
Character or NULL. The column name holding survey start timestamps. When
NULL,sframe_quality_results()looks for a recognised start-time column automatically.- time_min
Numeric or NULL. Minimum acceptable completion time in seconds. Respondents with a submission time below this threshold are flagged as speeders when timing data are available.
- straightline_scales
Logical. Whether to check for straight-lining within each defined scale block. Defaults to
TRUE.- straightline_min_items
Integer. The minimum number of items a scale must have before it is checked for straight-lining. Defaults to
4. A respondent who gives the identical response to every item in a 2-item scale has done exactly what a genuinely consistent respondent does, this is not evidence of inattention on its own, and checking scales that short flags a large share of honest respondents (see the worked example invignette("surveyframe"), where 3 scales of 2 items alone drove a 91 percent flag rate before this threshold existed). Set to2to restore the previous, more permissive behaviour. Many published scales have 3 items, and the default leaves them unchecked. Set3to check them, and read a flag on a 3-item scale as a prompt to look at the respondent, since identical answers to 3 items are common among attentive respondents. The printed report names each scale it skipped and the setting that would check it. In an analysis plan, pass it on the quality block asoptions = list(straightline_min_items = 3).- missing_threshold
Numeric. The proportion of missing item responses above which a respondent is flagged. Defaults to
0.2.
Value
An object of class sframe_quality_report, a named list with
elements: summary, attention, timing, straightline, missing,
and duplicates. Use print() for a formatted summary.
Details
Timing analysis is available when the data contain a submission timestamp
column and either an explicit started_at column or one of the recognised
defaults: started_at, start_time, started, or .started_at.
Examples
instr <- read_sframe(
system.file("extdata", "tourism_services_demo.sframe",
package = "surveyframe")
)
responses <- read_responses(
system.file("extdata", "tourism_services_responses.csv",
package = "surveyframe"),
instr,
respondent_id = "respondent_id",
submitted_at = "submitted_at",
meta_cols = "started_at"
)
qr <- sframe_quality_results(
responses,
instr,
respondent_id = "respondent_id",
submitted_at = "submitted_at",
started_at = "started_at",
straightline_scales = FALSE
)
print(qr)