NYT's statistical sleight of hand

The New York Times has a pattern of reporting facts and statistics in manner leading to erroneous conclusions favorable to management's interests. March 6's report on the consequences of parental notification onb abortion rates is a classic case in point, and is already debunked on the internet.

Michael New of the Heritage Foundation shows in "A Lesson in data and analysis for the New York Times," how the Times studied only half of the states which implemented parent notification laws, and then chose to track changes in the percentage of teen pregnancies ending in abortion, instead of the more significant overall rate of abortion.

Because relatively few teens give birth each year, the Times' measure can fluctuate dramatically, making the data difficult to analyze properly.

Furthermore, the Times ignores the likelihood that the presence of parental—involvement laws may reduce abortions not only by influencing the decisions of girls who are already pregnant, but also by reducing the number of teenage girls who become pregnant at all. An analysis of the percentage of total teen pregnancies that end in abortion will not lend any great insight into the effectiveness of parental—involvement laws in reducing abortions by changing minors' sexual behavior.

Last week AT exposed the masking of the decline of the metropolitan print edition in the company's SEC filings. In both cases, the numbers reported were true, but led to a mistaken understanding of the underlying reality.

Hat tip: Paul Shlichta 

Thomas Lifson  3 09 06

The New York Times has a pattern of reporting facts and statistics in manner leading to erroneous conclusions favorable to management's interests. March 6's report on the consequences of parental notification onb abortion rates is a classic case in point, and is already debunked on the internet.

Michael New of the Heritage Foundation shows in "A Lesson in data and analysis for the New York Times," how the Times studied only half of the states which implemented parent notification laws, and then chose to track changes in the percentage of teen pregnancies ending in abortion, instead of the more significant overall rate of abortion.

Because relatively few teens give birth each year, the Times' measure can fluctuate dramatically, making the data difficult to analyze properly.

Furthermore, the Times ignores the likelihood that the presence of parental—involvement laws may reduce abortions not only by influencing the decisions of girls who are already pregnant, but also by reducing the number of teenage girls who become pregnant at all. An analysis of the percentage of total teen pregnancies that end in abortion will not lend any great insight into the effectiveness of parental—involvement laws in reducing abortions by changing minors' sexual behavior.

Last week AT exposed the masking of the decline of the metropolitan print edition in the company's SEC filings. In both cases, the numbers reported were true, but led to a mistaken understanding of the underlying reality.

Hat tip: Paul Shlichta 

Thomas Lifson  3 09 06