Willis (2006) re-examines the study of Cecchetti (1986) on price adjustment behavior in the magazine industry. Cecchetti assumes that a firm's pricing rules are fixed for non-overlapping three-year intervals and estimates the model using a conditional logit specification a la Chamberlain (1980). The data set consists of 38 unique magazines (mag_id) observed over the years 1953-1979. The dependent variable "pr_ch" is a dummy variable referring to whether the price changed between January of the current year and January of the following year. "y_ch" refers to the number of years since the magazine price change; "inf_ch" refers to the cumulative change in inflation since the last price change and "mags_ch" refers to the cumulative change in magazine industry sales since the previous price change. The "group_id" variable corresponds to the groups used in a Chamberlain conditional logit estimation. This can be downloaded from the Journal of Applied Econometrics web site. Replicate Table 1 of Willis (2006). Column (1) runs a logit regression; while column (2) runs the same logit regression with magazine fixed effects; column (3) runs the Chamberlain conditional logit model with "group_id" allowing three-year intervals for a price change; column (5) runs a logit with magazine and time fixed effects.