detect_missing_items

Flags respondents whose proportion of missing item responses exceeds
<code>max_prop</code>. Like all detectors, it only finds; exclusion happens
in the Execute stage.

Rule-driven, auditable cleaning of survey and assessment
response data, implementing the WeianData Detect-Execute-Report
workflow. Provides a multi-format, multi-encoding input layer
(CSV, 'Excel', 'SPSS', 'Stata', 'SAS', Parquet, JSON), the
dcc_data container with a provenance chain, level-0 structural
diagnostics, five built-in response-quality detectors (missing
items, straight-lining, response time, trap items, score
anomalies), a declarative YAML rule engine, an execution engine
with a cell-level audit log, answer-key scoring, multi-form to
master item bank mapping, a normalized report model rendered as bilingual
staff workbooks and HTML, complete statistical bundles, and versioned
machine JSON/JSONL with findings-to-changes reconciliation,
cell-level lineage tracing, and
manifest-based one-command reproduction. Includes a protected bilingual
strict project workbook and matching JSON contract with cell-addressed
validation, non-mutating preflight, preview-first execution, and localized
staff guidance. All formally supported input backends install with the
package; PDF is optional rather than a fixed report output.

Kunxiang Ma

Data Cleaning Center for Survey and Assessment Data

WEIAN DATA TECH (Beijing) Co., Ltd. 

detect_missing_items function

<dl><dt>x</dt>
<dd>A <code>dcc_data</code> object or data.frame.</dd>
<dt>items</dt>
<dd>Character vector of item column names.</dd>
<dt>max_prop</dt>
<dd>Maximum tolerated missing proportion (default 0.5).</dd>
<dt>id_var</dt>
<dd>Name of the record-id column, or <code>NULL</code> for row
numbers.</dd>
<dt>severity</dt>
<dd>Severity assigned to findings (default <code>"warn"</code>).</dd>
<dt>structural</dt>
<dd>Optional logical matrix (rows aligned to the data,
columns to <code>items</code>) marking cells that were not administered
-- e.g. from a <code>skip_logic</code> rule. Not-administered cells are
excluded from both the numerator and denominator of the missing
proportion. <code>NULL</code> (default) treats every item as administered and
is byte-identical to the pre-1.1.0 behaviour.</dd></dl>

Arguments

Detect excessive item nonresponse per respondent — detect_missing_items

<dl>

<dt>x</dt>
<dd>A <code>dcc_data</code> object or data.frame.</dd>


<dt>items</dt>
<dd>Character vector of item column names.</dd>


<dt>max_prop</dt>
<dd>Maximum tolerated missing proportion (default 0.5).</dd>


<dt>id_var</dt>
<dd>Name of the record-id column, or <code>NULL</code> for row
numbers.</dd>


<dt>severity</dt>
<dd>Severity assigned to findings (default <code>"warn"</code>).</dd>


<dt>structural</dt>
<dd>Optional logical matrix (rows aligned to the data,
columns to <code>items</code>) marking cells that were not administered
-- e.g. from a <code>skip_logic</code> rule. Not-administered cells are
excluded from both the numerator and denominator of the missing
proportion. <code>NULL</code> (default) treats every item as administered and
is byte-identical to the pre-1.1.0 behaviour.</dd>

</dl>

detect_missing_items: Detect excessive item nonresponse per respondent

Description

Usage

Value

Arguments

Examples