Weighting the Arab Opinion Index Syrian data moves the share saying it makes no difference to them from 65.4 to 66.5 per cent, and the weight variation implies a design effect of about 1.05, giving an effective sample of about 1,177 on that question.
Claim ID
CH4-AOI-002
Exact or summary claim
Weighting the Arab Opinion Index Syrian data moves the share saying it makes no difference to them from 65.4 to 66.5 per cent, and the weight variation implies a design effect of about 1.05, giving an effective sample of about 1,177 on that question.
Destination
Chapter 4, the survey section
Status
AUTHOR_ESTIMATE
Source
The published microdata, Weight variable
Source type
Author computation on primary microdata
Published
2025
Accessed
2026-09-06
Measure and scope
FORMULA: Kish design effect, n * sum(w^2) / (sum w)^2, computed on the Syrian rows only and on the Q426 answering base only. Syrian weights take SEVEN values between 0.3000 and 1.0000. Q426 base n = 1,235, deff = 1.0489, effective n = 1,177. ASSUMPTION AND SENSITIVITY: this accounts for weight variation alone and NOT for clustering; the file carries no cluster or stratum variable, so the true design effect for a multi-stage cluster sample is higher and the effective sample smaller. The same statistic computed on the whole 40,130-row seventeen-country file is 1.2210, which is a property of that file and NOT of the Syrian sample; CHAPTER 4 v27 PRINTED THAT FIGURE IN ERROR AND v28 CORRECTS IT.
Known conflict
None. The codebook documents no weighting at all.
Specialist
Survey methods
Downstream
Chapter 4 line 211
Action
Print the design effect only with its formula and with the clustering caveat, or not at all.