Processing math: 100%
Rodhouse, Joseph B, and Kathy Ott. 2022. “Respondent Perceptions of Previously Reported Data.” Survey Practice, June. https://doi.org/10.29115/SP-2022-0006.
• Table 1. PRD experiment groups after stratified random assignment.
• Figure 1. Example of PRD presented to respondents within the web survey.
• Table 2. Agreement statements (AS) gauging respondent attitudes and perceptions of PRD.
• Table 3. Distribution of responses to PRD agreement statements.
• Table 4. Testing the independence of PRD perceptions and PRD characteristics (amount and recency).
• Figure A1. Example mailing for Group A2B2 emphasizing PRD use in the web survey
• Figure A2. Introductory screen for those with PRD in their web survey link (A2B1 and A2B2)
• Figure A3. Example of what respondents in A1B1 saw (and those in A2B1 and A2B2 when they did not have PRD for these items)
• Table A1. Distribution of responses to PRD agreement statements.
• Table A2. “It was clear to me that NASS pre-filled information in some answer cells in advance.”
• Table A3. “I recall providing the pre-filled information used in the answer cells on a previous NASS survey.”
• Table A4. “The pre-filled information in the answer cells was accurate.”
• Table A5. “The pre-filled information made it easier for me to complete the survey.”
• Table A6. “The pre-filled information helped me finish the survey faster.”
• Table A7. “I would have liked to see pre-filled information in more answer cells.”
• Table A8. “Overall, I have a positive reaction to pre-filled information being used in the survey.”
• Table A9. Design-based analysis of the association between PRD characteristics and awareness that PRD was present before beginning survey.
• Table A10. Design-based analysis of the association between PRD characteristics and perceptions that PRD used in the survey was accurate.
• Table A11. Design-Based Analysis of the association between PRD characteristics and awareness that PRD was present before beginning survey.
• Table A12. Design-based analysis of the association between PRD characteristics and perceptions that the survey was easier to complete.
• Table A13. Design-based analysis of the association between PRD characteristics and perceptions that PRD aided faster survey completion.
• Table A14. Design-based analysis of the association between PRD characteristics and desire that more PRD be used.
• Table A15. Design-based analysis of the association between PRD characteristics and overall positive reaction to PRD use in the survey.

## Abstract

Respondents’ answers on surveys they have previously completed are often referred to as previously reported data or PRD. This type of historical data is sometimes used when conducting a current survey, often in an effort to increase data quality and reduce respondent burden. Using PRD in surveys has been thoroughly researched and has been shown to have both positive and negative impacts to measures of data quality and burden. However, most often these studies focus on objective measures, and subjective measures of respondent perceptions of their experience are largely missing. To address this, we present data from an experiment using PRD in the Census of Agriculture’s Content Test, which utilized a set of attitudinal questions to gauge respondent’s views on the use of PRD in the survey, including measures capturing burden perceptions and overall reactions to PRD use. We find that the majority of respondents view PRD as aiding in faster and easier survey completion, as well as having an overall positive reaction to its use. We also find evidence that attitudes toward PRD are impacted by certain PRD metadata, such as the amount and recency of the data that are used in respondents’ surveys.

Accepted: May 31, 2022 EDT

# Appendix

Example of the 2022 Census of Agriculture Form can be found here:
2022_CoA_Questionnaire_Final.pdf (usda.gov)

Figure A1. Example mailing for Group A2B2 emphasizing PRD use in the web survey
Figure A2. Introductory screen for those with PRD in their web survey link (A2B1 and A2B2)
Figure A3. Example of what respondents in A1B1 saw (and those in A2B1 and A2B2 when they did not have PRD for these items)
Table A1. Distribution of responses to PRD agreement statements.
 PRD Agreement Statements (AS) Estimated Percent (Standard Error of Estimated Percent) Str. Dis. 2 3 4 Str. Agr. DK AS1. It was clear to me that NASS pre-filled information in some answer cells in advance 3.4 (0.6) 1.9 (0.5) 7.8 (0.9) 16.0 (1.3) 53.5 (1.7) 17.5 (1.3) AS2. I recall providing the pre-filled information used in the answer cells on a previous NASS survey 5.0 (0.8) 2.5 (0.5) 10.7 (1.1) 15.8 (1.3) 41.4 (1.7) 24.6 (1.5) AS3. The pre-filled information in the answer cells was accurate 2.8 (0.6) 2.7 (0.6) 13.2 (1.2) 18.2 (1.3) 49.3 (1.7) 13.8 (1.2) AS4. The pre-filled information made it easier for me to complete the survey 3.9 (0.7) 2.9 (0.6) 11.9 (1.1) 18.3 (1.3) 50.2 (1.7) 12.8 (1.2) AS5. The pre-filled information helped me finish the survey faster 4.5 (0.7) 4.1 (0.7) 13.4 (1.2) 16.9 (1.3) 48.5 (1.7) 12.6 (1.2) AS6. I would have liked to see pre-filled information in more answer cells 4.8 (0.8) 2.5 (0.5) 15.3 (1.2) 15.5 (1.2) 46.2 (1.7) 15.7 (1.3) AS7. Overall, I have a positive reaction to pre-filled information being used in the survey 1.6 (0.4) 2.5 (0.5) 13.1 (1.2) 21.7 (1.4) 49.8 (1.7) 11.3 (1.1)

Notes: Row percentages may not sum to 100 due to rounding.

Table A2. “It was clear to me that NASS pre-filled information in some answer cells in advance.”
 1 - Strongly Disagree 2 3 4 5 – Strongly Agree Don’t Know A2B1 2.97 1.97 7.95 16.93 51.47 18.71 A2B2 3.71 1.78 7.69 15.14 55.32 16.35

Notes: Numbers displayed are percentages.
Total n = 2,551; A2B1 n = 1,253; A2B2 n = 1,298.
Chi-square T1*T2 = 2.07, p = 0.83.

Table A3. “I recall providing the pre-filled information used in the answer cells on a previous NASS survey.”
 1 - Strongly Disagree 2 3 4 5 – Strongly Agree Don’t Know A2B1 4.79 2.64 10.74 16.74 40.62 24.47 A2B2 5.11 2.41 10.71 14.94 42.14 24.68

Notes: Numbers displayed are percentages.
Total n = 2,535; A2B1 n = 1,246; A2B2 n = 1,289.
Chi-square T1*T2 = 0.63, p = 0.98.

Table A4. “The pre-filled information in the answer cells was accurate.”
 1 - Strongly Disagree 2 3 4 5 – Strongly Agree Don’t Know A2B1 3.04 2.27 13.46 17.73 49.60 13.90 A2B2 2.63 3.16 12.92 18.68 48.91 13.69

Notes: Numbers displayed are percentages.
Total n = 2,540; A2B1 n = 1,245; A2B2 n = 1,295.
Chi-square T1*T2 = 0.90, p = 0.96

Table A5. “The pre-filled information made it easier for me to complete the survey.”
 1 - Strongly Disagree 2 3 4 5 – Strongly Agree Don’t Know A2B1 3.79 2.92 12.81 17.04 51.71 11.74 A2B2 3.93 2.91 10.97 19.6 48.81 13.77

Notes: Numbers displayed are percentages.
Total n = 2,535; A2B1 n = 1,245; A2B2 n = 1,290.
Chi-square T1*T2 = 2.33, p = 0.80

Table A6. “The pre-filled information helped me finish the survey faster.”
 1 - Strongly Disagree 2 3 4 5 – Strongly Agree Don’t Know A2B1 4.75 4.34 13.25 15.83 49.39 12.40 A2B2 4.30 3.76 13.55 17.98 47.56 12.86

Notes: Numbers displayed are percentages.
Total n = 2,531; A2B1 n = 1,239; A2B2 n = 1,292.
Chi-square T1*T2 = 1.04, p = 0.95

Table A7. “I would have liked to see pre-filled information in more answer cells.”
 1 - Strongly Disagree 2 3 4 5 – Strongly Agree Don’t Know A2B1 4.58 2.99 15.44 13.97 47.33 15.69 A2B2 5.05 1.99 15.21 16.96 45.07 15.71

Notes: Numbers displayed are percentages.
Total n = 2,530; A2B1 n = 1,238; A2B2 n = 1,292.
Chi-square T1*T2 = 2.40, p = 0.79

Table A8. “Overall, I have a positive reaction to pre-filled information being used in the survey.”
 1 - Strongly Disagree 2 3 4 5 – Strongly Agree Don’t Know A2B1 2.06 2.51 13.44 20.54 50.11 11.34 A2B2 1.12 2.39 12.87 22.78 49.51 11.34

Notes: Numbers displayed are percentages.
Total n = 2,536; A2B1 n = 1,240; A2B2 n = 1,296.
Chi-square T1*T2 = 1.75, p = 0.88

Table A9. Design-based analysis of the association between PRD characteristics and awareness that PRD was present before beginning survey.
 PRD Characteristics (Amount & Recency) Agreement Statement 1. – It was clear to me that NASS pre-filled information in some answer cells in advance. Row Percentages (Linearized SE) Strongly Disagree, 2, 3, DK 4, Strongly Agree Total More & Recent 29.63 (6.83) 70.37 (6.83) 100.00 More & Older 25.62 (2.30) 74.38 (2.30) 100.00 Less & Older 33.72 (2.28) 66.28 (2.28) 100.00 Less & Recent 27.45 (6.81) 72.55 (6.81) 100.00 Tests of Independence Rao-Scott Chi-Square $P$ value 32.46 0.0056 Parameters of the Rao-Scott Design-Adjusted Test $n_{AS1}$ = 2,551 Design df = 15 Design Correction = 1.63

Notes: Percentages binned for ease of interpretation with PRD Characteristics. Rao-Scott (1984) Test of Independence calculated on un-binned distribution of responses.

Table A10. Design-based analysis of the association between PRD characteristics and perceptions that PRD used in the survey was accurate.
 PRD Characteristics (Amount & Recency) Agreement Statement 2. – I recall providing the pre-filled information used in the answer cells on a previous NASS survey. Row Percentages (Linearized SE) Strongly Disagree, 2, 3, DK 4, Strongly Agree Total More & Recent 29.71 (6.44) 70.29 (6.44) 100.00 More & Older 38.21 (2.58) 61.79 (2.58) 100.00 Less & Older 47.02 (2.39) 52.98 (2.39) 100.00 Less & Recent 25.74 (4.79) 74.26 (4.79) 100.00 Tests of Independence Rao-Scott Chi-Square $P$ value 43.00 0.0002 Parameters of the Rao-Scott Design-Adjusted Test $n_{AS2}$ = 2,535 Design df = 15 Design Correction = 1.67

Notes: Percentages binned for ease of interpretation with PRD Characteristics. Rao-Scott (1984) Test of Independence calculated on un-binned distribution of responses.

Table A11. Design-Based Analysis of the association between PRD characteristics and awareness that PRD was present before beginning survey.
 PRD Characteristics (Amount & Recency) Agreement Statement 3. – The pre-filled information in the answer cells was accurate. Row Percentages (Linearized SE) Strongly Disagree, 2, 3, DK 4, Strongly Agree Total More & Recent 41.83 (6.99) 58.17 (6.99) 100.00 More & Older 30.33 (2.44) 69.67 (2.44) 100.00 Less & Older 33.83 (2.31) 66.17 (2.31) 100.00 Less & Recent 23.17 (5.56) 76.83 (5.56) 100.00 Tests of Independence Rao-Scott Chi-Square $P$ value 29.14 0.0154 Parameters of the Rao-Scott Design-Adjusted Test $n_{AS3}$ = 2,540 Design df = 15 Design Correction = 1.70

Notes: Percentages binned for ease of interpretation with PRD Characteristics. Rao-Scott (1984) Test of Independence calculated on un-binned distribution of responses.

Table A12. Design-based analysis of the association between PRD characteristics and perceptions that the survey was easier to complete.
 PRD Characteristics (Amount & Recency) Agreement Statement 4. – The pre-filled information made it easier for me to complete the survey. Row Percentages (Linearized SE) Strongly Disagree, 2, 3, DK 4, Strongly Agree Total More & Recent 34.80 (6.83) 65.20 (6.83) 100.00 More & Older 29.14 (2.41) 70.86 (2.41) 100.00 Less & Older 33.26 (2.29) 66.74 (2.29) 100.00 Less & Recent 17.71 (2.82) 82.29 (2.82) 100.00 Tests of Independence Rao-Scott Chi-Square $P$ value 25.88 0.0393 Parameters of the Rao-Scott Design-Adjusted Test $n_{AS4}$ = 2,535 Design df = 15 Design Correction = 1.51

Notes: Percentages binned for ease of interpretation with PRD Characteristics. Rao-Scott (1984) Test of Independence calculated on un-binned distribution of responses.

Table A13. Design-based analysis of the association between PRD characteristics and perceptions that PRD aided faster survey completion.
 PRD Characteristics (Amount & Recency) Agreement Statement 5. – The pre-filled information helped me finish the survey faster. Row Percentages (Linearized SE) Strongly Disagree, 2, 3, DK 4, Strongly Agree Total More & Recent 37.27 (6.78) 62.73 (6.78) 100.00 More & Older 32.80 (2.51) 67.20 (2.51) 100.00 Less & Older 36.31 (2.33) 63.69 (2.33) 100.00 Less & Recent 18.48 (2.94) 81.52 (2.94) 100.00 Tests of Independence Rao-Scott Chi-Square $P$ value 19.65 0.1859 Parameters of the Rao-Scott Design-Adjusted Test $n_{AS5}$ = 2,531 Design df = 15 Design Correction = 1.61

Notes: Percentages binned for ease of interpretation with PRD Characteristics. Rao-Scott (1984) Test of Independence calculated on un-binned distribution of responses.

Table A14. Design-based analysis of the association between PRD characteristics and desire that more PRD be used.
 PRD Characteristics (Amount & Recency) Agreement Statement 6. – I would have liked to see pre-filled information in more answer cells. Row Percentages (Linearized SE) Strongly Disagree, 2, 3, DK 4, Strongly Agree Total More & Recent 42.29 (6.84) 57.71 (6.84) 100.00 More & Older 36.25 (2.56) 63.75 (2.56) 100.00 Less & Older 40.03 (2.36) 59.97 (2.36) 100.00 Less & Recent 24.01 (5.51) 75.99 (5.51) 100.00 Tests of Independence Rao-Scott Chi-Square $P$ value 32.43 0.0056 Parameters of the Rao-Scott Design-Adjusted Test $n_{AS6}$ = 2,530 Design df = 15 Design Correction = 1.66

Notes: Percentages binned for ease of interpretation with PRD Characteristics. Rao-Scott (1984) Test of Independence calculated on un-binned distribution of responses.

Table A15. Design-based analysis of the association between PRD characteristics and overall positive reaction to PRD use in the survey.
 PRD Characteristics (Amount & Recency) Agreement Statement 5. – Overall, I have a positive reaction to prefilled information being used in the survey. Row Percentages (Linearized SE) Strongly Disagree, 2, 3, DK 4, Strongly Agree Total More & Recent 33.03 (6.75) 66.97 (6.75) 100.00 More & Older 28.73 (2.41) 71.27 (2.41) 100.00 Less & Older 28.48 (2.19) 71.52 (2.19) 100.00 Less & Recent 21.21 (5.39) 78.79 (5.39) 100.00 Tests of Independence Rao-Scott Chi-Square $P$ value 16.17 0.3705 Parameters of the Rao-Scott Design-Adjusted Test $n_{AS4}$ = 2,536 Design df = 15 Design Correction = 1.52

Notes: Percentages binned for ease of interpretation with PRD Characteristics. Rao-Scott (1984) Test of Independence calculated on un-binned distribution of responses.