2015 OSR RI Margins of Error

---

Overview

10,000 Sample
5,000 Sample
10,000 Split

---

Sheet 1: 10,000 Sample

Margins of Error for Estimating a Proportion for the Savannah Reinterview
Cells in pink can be changed
Assumptions
Sample Size	10,000
Confidence Level	95%
z value	1.96
Design Effect	1
Eligibility Rate	1
Response Rate	75%
Sample Group	Sample Allocation
Non-ID Panel	45%
Non-ID Nonpanel	45%
ID	10%
Sample Group	Sample Size
Non-ID Panel	4,500
Non-ID Nonpanel	4,500
ID	1,000
Non-ID Panel Margins of Error
Estimate Proportion	Overall	75% characteristic	50% characteristic	20% characteristic	10% characteristic
0.9	0.0101	0.0117	0.0143	0.0226	0.0320			Margin of error of .03 or less
0.8	0.0135	0.0156	0.0191	0.0302	0.0427			Margin of error greater than .03
0.7	0.0155	0.0179	0.0219	0.0346	0.0489
0.6	0.0165	0.0191	0.0234	0.0370	0.0523
0.5	0.0169	0.0195	0.0239	0.0377	0.0533
For example, if the estimate is 90% of the reinterview households in the Non-ID panel can verify their reponse, the confidence interval is expected to be 0.9 +/- 0.0101.
If we want to consider a characteristic that represents 20% of the population (such as Hispanics), an estimate of 90% would have a confidence interval of .9 +/- .0226 for the Non-ID panel.
Non-ID Nonpanel Margins of Error
Estimate Proportion	Overall	75% characteristic	50% characteristic	20% characteristic	10% characteristic
0.9	0.0101	0.0117	0.0143	0.0226	0.0320
0.8	0.0135	0.0156	0.0191	0.0302	0.0427
0.7	0.0155	0.0179	0.0219	0.0346	0.0489
0.6	0.0165	0.0191	0.0234	0.0370	0.0523
0.5	0.0169	0.0195	0.0239	0.0377	0.0533
ID Margins of Error
Estimate Proportion	Overall	75% characteristic	50% characteristic	20% characteristic	10% characteristic
0.9	0.0215	0.0248	0.0304	0.0480	0.0679
0.8	0.0286	0.0331	0.0405	0.0640	0.0905
0.7	0.0328	0.0379	0.0464	0.0733	0.1037
0.6	0.0351	0.0405	0.0496	0.0784	0.1109
0.5	0.0358	0.0413	0.0506	0.0800	0.1132

---

Sheet 2: 5,000 Sample

Margins of Error for Estimating a Proportion for the Savannah Reinterview
Cells in pink can be changed
Assumptions
Sample Size	5,000
alpha	0.05
z value	1.96
Design Effect	1
Eligibility Rate	1
Response Rate	75%
Sample Group	Sample Allocation
Non-ID Panel	40%
Non-ID Nonpanel	40%
ID	20%
Sample Group	Sample Size
Non-ID Panel	2,000
Non-ID Nonpanel	2,000
ID	1,000
Non-ID Panel Margins of Error
Estimate Proportion	Overall	75% characteristic	50% characteristic	20% characteristic	10% characteristic
0.9	0.0152	0.0175	0.0215	0.0339	0.0480			Margin of error of .03 or less
0.8	0.0202	0.0234	0.0286	0.0453	0.0640			Margin of error greater than .03
0.7	0.0232	0.0268	0.0328	0.0519	0.0733
0.6	0.0248	0.0286	0.0351	0.0554	0.0784
0.5	0.0253	0.0292	0.0358	0.0566	0.0800
For example, if the estimate is 90% of the reinterview households in the Non-ID panel can verify their reponse, the confidence interval is expected to be 0.9 +/- 0.0152.
If we want to consider a characteristic that represents 20% of the population (such as Hispanics), an estimate of 90% would have a confidence interval of .9 +/- .0339 for the Non-ID panel.
Non-ID Nonpanel Margins of Error
Estimate Proportion	Overall	75% characteristic	50% characteristic	20% characteristic	10% characteristic
0.9	0.0152	0.0175	0.0215	0.0339	0.0480
0.8	0.0202	0.0234	0.0286	0.0453	0.0640
0.7	0.0232	0.0268	0.0328	0.0519	0.0733
0.6	0.0248	0.0286	0.0351	0.0554	0.0784
0.5	0.0253	0.0292	0.0358	0.0566	0.0800
ID Margins of Error
Estimate Proportion	Overall	75% characteristic	50% characteristic	20% characteristic	10% characteristic
0.9	0.0215	0.0248	0.0304	0.0480	0.0679
0.8	0.0286	0.0331	0.0405	0.0640	0.0905
0.7	0.0328	0.0379	0.0464	0.0733	0.1037
0.6	0.0351	0.0405	0.0496	0.0784	0.1109
0.5	0.0358	0.0413	0.0506	0.0800	0.1132

---

Sheet 3: 10,000 Split

Margins of Error for Estimating a Proportion for the Savannah Reinterview
Cells in pink can be changed
Assumptions
Sample Size	10,000
Confidence Level	95%
z value	1.96
Design Effect	1
Eligibility Rate	1
Response Rate	75%
Data Collection Mode	Sample Group	Sample Allocation
CATI	Non-ID Panel	36%
	Non-ID Nonpanel	36%
	ID	8%
In-person	Non-ID Panel	9%
	Non-ID Nonpanel	9%
	ID	2%
Data Collection Mode	Sample Group	Sample Size
CATI	Non-ID Panel	3600
	Non-ID Nonpanel	3600
	ID	800
In-person	Non-ID Panel	900
	Non-ID Nonpanel	900
	ID	200
CATI Non-ID Panel Margins of Error
Estimate Proportion	Overall	75% characteristic	50% characteristic	20% characteristic	10% characteristic
0.9	0.0113	0.0131	0.0160	0.0253	0.0358			Margin of error of .03 or less
0.8	0.0151	0.0174	0.0213	0.0337	0.0477			Margin of error greater than .03
0.7	0.0173	0.0200	0.0244	0.0387	0.0547
0.6	0.0185	0.0213	0.0261	0.0413	0.0584
0.5	0.0189	0.0218	0.0267	0.0422	0.0596
For example, if the estimate is 90% of the reinterview households in the CATI Non-ID panel can verify their reponse, the confidence interval is expected to be 0.9 +/- 0.0113.
If we want to consider a characteristic that represents 20% of the population (such as Hispanics), an estimate of 90% would have a confidence interval of .9 +/- .0253 for the Non-ID panel.
CATI Non-ID Nonpanel Margins of Error
Estimate Proportion	Overall	75% characteristic	50% characteristic	20% characteristic	10% characteristic
0.9	0.0113	0.0131	0.0160	0.0253	0.0358
0.8	0.0151	0.0174	0.0213	0.0337	0.0477
0.7	0.0173	0.0200	0.0244	0.0387	0.0547
0.6	0.0185	0.0213	0.0261	0.0413	0.0584
0.5	0.0189	0.0218	0.0267	0.0422	0.0596
CATI ID Margins of Error
Estimate Proportion	Overall	75% characteristic	50% characteristic	20% characteristic	10% characteristic
0.9	0.0240	0.0277	0.0339	0.0537	0.0759
0.8	0.0320	0.0370	0.0453	0.0716	0.1012
0.7	0.0367	0.0423	0.0519	0.0820	0.1160
0.6	0.0392	0.0453	0.0554	0.0877	0.1240
0.5	0.0400	0.0462	0.0566	0.0895	0.1265
In-person Non-ID Panel Margins of Error
Estimate Proportion	Overall	75% characteristic	50% characteristic	20% characteristic	10% characteristic
0.9	0.0226	0.0261	0.0320	0.0506	0.0716
0.8	0.0302	0.0348	0.0427	0.0675	0.0954
0.7	0.0346	0.0399	0.0489	0.0773	0.1093
0.6	0.0370	0.0427	0.0523	0.0826	0.1169
0.5	0.0377	0.0436	0.0533	0.0843	0.1193
In-person Non-ID Nonpanel Margins of Error
Estimate Proportion	Overall	75% characteristic	50% characteristic	20% characteristic	10% characteristic
0.9	0.0226	0.0261	0.0320	0.0506	0.0716
0.8	0.0302	0.0348	0.0427	0.0675	0.0954
0.7	0.0346	0.0399	0.0489	0.0773	0.1093
0.6	0.0370	0.0427	0.0523	0.0826	0.1169
0.5	0.0377	0.0436	0.0533	0.0843	0.1193
In-person ID Margins of Error
Estimate Proportion	Overall	75% characteristic	50% characteristic	20% characteristic	10% characteristic
0.9	0.0480	0.0554	0.0679	0.1074	0.1518
0.8	0.0640	0.0739	0.0905	0.1431	0.2024
0.7	0.0733	0.0847	0.1037	0.1640	0.2319
0.6	0.0784	0.0905	0.1109	0.1753	0.2479
0.5	0.0800	0.0924	0.1132	0.1789	0.2530