Chapter 9 Data Quality Issues and Datasets

Respondent Error

Committed by the respondent. Can be unintentional or intentional and can happen at anytime.

Intentional Error

Respondent willingly misrepresents themselves. Nonresponsive, falsehoods, and speeding.

Unintentional Error

Respondent gives invalid responses but believes they are true. Misunderstanding, guessing, distractions, and attention loss.

Nonresponsive Safeguards

Ensure anonymity, incentives, 3rd person technique and forced response option

Falsehood Safeguards

Ensuring anonymity, incentives, 3rd person technique, and validation checks

Speeding Safeguards

Reversal of scale endpoints and attention checks

Misunderstanding Safeguards

Direct questions

Quessing Safeguards

Response options

Attention Loss Safeguards

Reversal of scale endpoints and attention check

Distraction Safeguards

Short questionaire and clean design questionnaire

Dataset

Created by data coding, defined as the identification of code values.

Problematic respondent types

Incomplete responses, nonresponses, yea-saying pattern, nay-saying pattern, middle of the road patterns

Solutions

1. Break off: keep survey short

2. Item omission: force response

3. Yea pattern: attention check

4. Nay pattern: attention check

5. Middle of the road: attention check