Limits...
The effect of selection bias in studies of fads and fashions.

Denrell J, Kovács B - PLoS ONE (2015)

Bottom Line: Most studies of fashion and fads focus on objects and practices that once were popular.Through simulations and the analysis of a data set that has previously not been used to analyze the rise and fall of cultural practices, the New York Times text archive, we show that studying a whole range of cultural objects, both popular and less popular, is essential for understanding the drivers of popularity.In particular, we show that estimates of statistical models of the drivers of popularity will be biased if researchers use only trajectories of those practices that once were popular.

View Article: PubMed Central - PubMed

Affiliation: Warwick Business School, Coventry, United Kingdom.

ABSTRACT
Most studies of fashion and fads focus on objects and practices that once were popular. We argue that limiting the sample to such trajectories generates a selection bias that obscures the underlying process and generates biased estimates. Through simulations and the analysis of a data set that has previously not been used to analyze the rise and fall of cultural practices, the New York Times text archive, we show that studying a whole range of cultural objects, both popular and less popular, is essential for understanding the drivers of popularity. In particular, we show that estimates of statistical models of the drivers of popularity will be biased if researchers use only trajectories of those practices that once were popular.

Show MeSH

Related in: MedlinePlus

David and Strang (2006)’s data: Comparison of the observed data and the predicted values from Poisson models using parameters from Table 2.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4401772&req=5

pone.0123471.g004: David and Strang (2006)’s data: Comparison of the observed data and the predicted values from Poisson models using parameters from Table 2.

Mentions: The proposed model describes the data reasonably well: the Heckman pseudo-R2 values are 0.48 and 0.78 for the two data sets. To better show model fit, we plot the observed counts as well as five randomly generated trajectories based on the coefficient estimates. Fig 3 shows the result for Abrahamson’s data [2] on quality circles and Fig 4 shows the result for David and Strang’s data [19] on the spread of Total Quality Management. These randomly generated trajectories describe the observed counts well, although there is a tendency for the model to overshoot. Moreover, the model fails to predict the peak, which is not surprising: the exact location and timing of the peak is very hard to predict using diffusion models. As soon as the peak is reached, the model traces the data again.


The effect of selection bias in studies of fads and fashions.

Denrell J, Kovács B - PLoS ONE (2015)

David and Strang (2006)’s data: Comparison of the observed data and the predicted values from Poisson models using parameters from Table 2.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4401772&req=5

pone.0123471.g004: David and Strang (2006)’s data: Comparison of the observed data and the predicted values from Poisson models using parameters from Table 2.
Mentions: The proposed model describes the data reasonably well: the Heckman pseudo-R2 values are 0.48 and 0.78 for the two data sets. To better show model fit, we plot the observed counts as well as five randomly generated trajectories based on the coefficient estimates. Fig 3 shows the result for Abrahamson’s data [2] on quality circles and Fig 4 shows the result for David and Strang’s data [19] on the spread of Total Quality Management. These randomly generated trajectories describe the observed counts well, although there is a tendency for the model to overshoot. Moreover, the model fails to predict the peak, which is not surprising: the exact location and timing of the peak is very hard to predict using diffusion models. As soon as the peak is reached, the model traces the data again.

Bottom Line: Most studies of fashion and fads focus on objects and practices that once were popular.Through simulations and the analysis of a data set that has previously not been used to analyze the rise and fall of cultural practices, the New York Times text archive, we show that studying a whole range of cultural objects, both popular and less popular, is essential for understanding the drivers of popularity.In particular, we show that estimates of statistical models of the drivers of popularity will be biased if researchers use only trajectories of those practices that once were popular.

View Article: PubMed Central - PubMed

Affiliation: Warwick Business School, Coventry, United Kingdom.

ABSTRACT
Most studies of fashion and fads focus on objects and practices that once were popular. We argue that limiting the sample to such trajectories generates a selection bias that obscures the underlying process and generates biased estimates. Through simulations and the analysis of a data set that has previously not been used to analyze the rise and fall of cultural practices, the New York Times text archive, we show that studying a whole range of cultural objects, both popular and less popular, is essential for understanding the drivers of popularity. In particular, we show that estimates of statistical models of the drivers of popularity will be biased if researchers use only trajectories of those practices that once were popular.

Show MeSH
Related in: MedlinePlus