Order results by date

Order results by relevance

Factsheet: GUIDE: Understanding and reporting on opinion polls

…gest that a well-chosen, random sample of around 1,000 people is the minimum required to produce accurate results. That said, even large-scale surveys can be skewed. Therefore, the most important consideration is whether the sample is representative. If thousands of readers respond to an online survey on a popular news website, can it be said to be representative? No. At best the survey can be said to be representative of the views of the…

be the first to comment

Factsheet: GUIDE: Rape statistics in South Africa

…it to the police. 2.1% of women raped by an intimate partner had reported the crime to the police.   7.8% of women raped by a stranger or acquaintance had reported the incident to the police. The study is not nationally representative. It was conducted in South Africa’s Gauteng province. South Africa’s Demographic and Health Survey (1998) 15.2% of women aged 15 to 49 who had been forced to have sex had “sought help from the police”. The…

be the first to comment

Report: Claim that 1 in 3 South Africans are drug users based on flawed survey

…the profile of people who participated, only 5.1% were unemployed and we don’t now about the racial or gender distribution of respondents. It is highly unlikely it is reflective of the South African population. “It is not a representative sample, so you really can’t extrapolate to the general South African population from this survey.” Anti-Drug Alliance admits flaws In an interview with Africa Check, ADASA’s CEO Van Kerken admitted the…

1 comment

Report: Do 774 South Africans become jobless every day, as the DA says?

…quarter of 2014 there were 4,909,000 unemployed people in South Africa according to the narrow definition of unemployment. In the first quarter of 2015 it jumped to 5,535,000. Stats SA’s Durieux confirmed the change in their sample, saying that they previously used a sample based on 2001 census data. For 2015 they updated their sample and based it on data from the 2011 census, which was more extensive. Durieux cautioned against making…

be the first to comment

Blog post: ANALYSIS: What do we really know about international migration to & from SA?

…alf of that indicated by the 10% sample. The number reported as coming from SADC countries is 17% lower. And the number reported as “not specified” (but not born in South Africa) is nearly three times greater than in the 10% sample! Evidently, the 10% sample has been subjected to extensive cleaning and editing, in a way that was not done with the original data released to the public. Immigrant numbers from South Africa’s 2011 Census Create…

1 comment

Blog post: Tips and advice

…they reflect the true performance, over time, but to make the numbers look good, starting at the bottom of a regular cycle and ending at the top.   Was the sample large enough? Was it comprehensive? An opinion poll that samples the views of a few dozen – or even a few hundred – people is unlikely to be representative of the views of a population of millions. Most polling organisations suggest that a well-chosen sample of around 1,000 people…

be the first to comment

Report: Are 60% of sex workers in South Africa HIV positive? The data is limited

…ld, a senior technical adviser at SANAC, said there had been “other small or local studies and [testing] data” which showed an HIV prevalence of 60% or more. However, “the limited number of studies to confirm this, the small sample sizes and restricted geographic focus of the studies to date” meant that they weren’t the “most accurate and statistically sound”. Despite these limitations, Mangold said, it was “important to put a number on paper”…

be the first to comment

Report: Do a third of South Africans really suffer from mental illnesses?

…rt.) The South African Stress and Health (SASH) study referred to by SADAG was conducted between 2003 and 2004. It formed part of a world mental health survey that was initiated and funded by World Health Organisation (WHO). Representative samples of adult populations in 28 countries were canvassed, including two in Africa: South Africa and Nigeria. The South African research teams interviewed a representative sample of 4,351 adults across the…

be the first to comment

Report: Claim 2 million will need food aid in Zimbabwe exaggerates (a real) problem

…ry real problems need to be cautious about the headline statistics and understand in more detail what happens to whom and where. A further concern with recent national surveys is that they have been sampling according to old sample frames set before the land reform. For this survey ZimVac used an updated sample so it should be more reliable than past surveys but there is still much that is not captured. Conclusion – The food problem is real but…

be the first to comment

Report: Why social network surveys don’t necessarily reflect the views of SA youth

…6.5 million active users on a monthly basis. The sample sizes are significantly larger than those of traditional research suppliers, often running into tens of thousands. Data from the Census is used to weight or balance the sample, in line with standard research practice by all companies,” he said, adding that “the key to sample representivity is dispersion”. Herein lies a problem: MXit, like many social media tools, does not require people to…

12 comments

1 2 3 9