/ 9 January 2008

Caution on survey is standard practice, says stats council

The Statistics Council has not questioned the scientific basis of last year’s Community Survey, council head Howard Gabriels said on Tuesday.

He said the fact that it had recommended that Stats South Africa (Stats SA) issue a cautionary notice on the reliability of some of the data was standard scientific practice.

Gabriels was reacting to a newspaper report claiming that the statutory council had said irregularities in the survey made it questionable whether the document could provide a scientific basis for government policy.

The Democratic Alliance on Tuesday called for the dismissal of Statistician General Pali Lehohla, while the South African Communist Party said unreliable statistics were ”tantamount to lying to our people”.

The results of the survey, a mini-census, were hailed by President Thabo Mbeki as proof that his government was delivering on its promises.

Gabriels said in a statement that the 24-member council, charged with the oversight of Stats SA, had been closely involved with the survey.

It was happy that Stats SA had made every effort to follow scientific methodology, and had recommended the survey’s public release.

He said the council had recommended that Stats SA issue a ”caution” on data on the population of institutions such as prisons and hospitals, unemployment, social grants, income and provincial distribution of households.

”This is standard practice in the scientific world and shows the extent to which both council and Stats SA value transparency,” he said.

”The analogy is that warnings issued with medicines do not mean that the medicine is rendered useless or ineffectual.”

On the issue of employment, for example, survey data showed a higher percentage of people unemployed than in an earlier labour force survey.

However, the data on employment compared well with the labour survey.

”There is a difference in the formulation of questions on unemployment, which may explain why the data differs,” he said.

Likewise, it was suspected that respondents might have reported private pension as social pensions, which could explain an anomaly in figures there.

”The fact that the council recommended that Stats SA issue caution with the data does not in anyway suggest that the scientific basis of the Community Survey can be questioned,” he said.

”I deny that the council has questioned the scientific basis for the CS.”

Stats SA itself said in a separate statement that the cautionary note was transparent practice.

”Stats SA maintain that the survey results are reliable and trustworthy,” it said.

DA deputy finance spokesperson Kobus Marais said earlier the ”errors” had resulted in a serious crisis of confidence in Stats SA and were a major embarrassment for the government.

Marais said Finance Minister Trevor Manuel should replace Lehohla, who did not inspire the confidence needed to ensure a rapid return to credibility for Stats SA.

The SACP said if the media report was true, it was ”indeed a disgrace”, and posed a serious threat to South Africa’s development agenda.

Many of the concerns the SACP had raised on what it said was the failure of current macro-economic policies had been dismissed using this ”blatantly unreliable” data.

”Unreliable statistics are tantamount to lying to our people and failing them in their justified aspirations,” it said. – Sapa