Skip to main content

Table 2 Quality control indicators. The UNIVERSAL (University and Mental Health) project

From: Predictive models for suicidal thoughts and behaviors among Spanish University students: rationale and methods of the UNIVERSAL (University & mental health) project

Indicator

Operational indicator

Action required

Participation and completion Report:

  

- Registered respondents: Individuals registered to the study

n (% over eligible)

 

- Not started surveys: Respondents that have not accessed the link to the survey, among registered

  

- Incomplete surveys: Respondents that have entered the survey but have not completed it, among registered

n (%)

 

- Completed surveys: Respondents that have finalized the whole interview, among registered

n (%)

 

Validation status:

  

- Valid respondents: Registered respondents that fulfill eligibility criteria

n (%)

 

- Excluded respondents: Registered respondents that do not satisfy eligibility

n (%)

 

- Validation Pending: Registered respondents pending to be validated

n (%)

 

Data quality indicators:

  

- Survey duration table: Descriptive information on duration time by university and overall.

Mean (SDa) Median P25b

P75c min. max.

 

- Assessment of problematic respondents based on time: students with survey duration time below and over specific thresholdsd

n (%)

➔ The respondents with duration below and above the threshold are listed

• Assessment of problematic respondents based on item completeness: Surveys with number of items left unanswered or with response “I do not want to answer” over percentile 95

P95e

➔ The respondents with number of items left unanswered and with higher number of responses “I do not want to answer” over Percentile 95 are listed. The proportion of items left unanswered for each respondent is evaluated to determine acceptability of the survey.

• Assessment of problematic items: In order to identify problematic items or possible skip errors of the on-line questionnaire, those items with a higher proportion of empty responses and those not willing to answer are listed.

P95e

➔ The items with high proportion of incompleteness (if any) are evaluated in terms of understandability, sensitive question level or possible skip errors.

  1. aStandard deviation, bPercentile 25, cPercentile 75, dthat are considered short and large enough based on the number of items administered to all respondents, ePercentile 95