The International Republican Institute (IRI) said that information gathered from nationwide poll in Pakistan between September 27 and October 26, 2016, was based upon international standards for market and social research methodologies.
“IRI has worked in Pakistan since 2002, and has conducted numerous polls for more than a decade. Our current programme in Pakistan focuses on work with political parties from across the political spectrum and relies on polling to inform decision-makers about the priorities of Pakistani citizens,” it said in a statement.
The IRI’s statement said the survey was undertaken to provide a tool for political parties to develop sound policy solutions that reflect public priorities.
Giving information about its methodology, the IRI said the survey sample consisted of 4,978 voting-age residents of Pakistan from 75 districts across the provinces of Punjab, Khyber Pakhtunkhwa, Sindh and Balochistan (excluding the Federally Administered Tribal Areas and Chitral).
It further said in recent months IRI briefed representatives from political parties including Pakistan Muslim League Nawaz (PML-N), Pakistan Tehreek-e-Insaf (PTI), Pakistan People’s Party (PPP), Qaumi Watan Party (QWP), Awami National Party (ANP), Jamiat Ulema-e-Islam-F (JUI-F), National Party (NP), and Pakhtunkhwa Milli Awami Party (PkMAP) on the results of a nationwide survey of nearly 5,000 residents.”
“As such, IRI does not intend to release the information publicly but rather seeks to have it inform policy development undertaken by Pakistan’s political parties,” it added.
The probability-proportional-to-size (PPS) method was used to select localities.
Households were selected using the “random walk method” to select a starting point and then using the Left Hand Rule to start a skip pattern to select households for an interview.
The Kish Grid was used to select respondents of 18 years and older within these randomly-selected households.
It said interviews were conducted through in-person, in-home interviews with trained Pakistani enumerators.
Female respondents were interviewed by female interviewers and male respondents by male interviewers.
The response rate was approximately 66 percent (based on the response rate formula developed by the American Research Association) with a confidence interval of plus or minus 1.39 percent at the mid-range and a confidence level of 95 percent, the IRI said.
