Department of Labor Logo United States Department of Labor
Dot gov

The .gov means it's official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you're on a federal government site.


The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Consumer Price Index

Quality Adjustment in the CPI: Girls’ Tops

Girl's tops image

Item definition

The Entry Level Item (ELI) category of Girls’ Tops is part of the index for Girls’ Apparel, and includes all types and styles of shirts, sweaters, cardigans, vests, activewear tops, and one-piece leotards and bodysuits in preschool and grade school girls' sizes. In this factsheet, we examine the regression model used in the quality adjustment of price data for girls’ tops.

CPI sample

The sample used for the new hedonic model included approximately 770 observations. The model was constructed using web-based information due to restrictions on data collection during the COVID-19 pandemic, unlike models from the past that had a mix of store and website data. There were more than 150 unique brands in the sample, and the top 10 brands accounted for 49 percent of observations. There were more than 50 distinct outlets in the sample, and the top five outlets accounted for 53 percent of observations.  

Current Model

Table 1 presents the details of the estimation of the current model used to quality adjust girls’ tops.

Table 1. Parameter estimates for the hedonic regression for girls’ tops
Variable Name DF Parameter Estimate Standard Error t Value Pr > |t| Tolerance Statistic


1 3.23801 0.0546 59.3 <.0001

Style: Partial Closure Pullover

1 0.1714 0.05313 3.23 0.0013 0.86813

Style: Cardigan

1 0.43989 0.06076 7.24 <.0001 0.75139

Style: Sweater

1 0.40204 0.06885 5.84 <.0001 0.76221

Style: Open Front Shirt

1 0.30707 0.06327 4.85 <.0001 0.85853

Style: Activewear Top

1 0.15675 0.05491 2.85 0.0044 0.85538

Style: Sweatshirt

1 0.37332 0.0515 7.25 <.0001 0.72616

Style: No Closure Pullover


Type of Brand: Luxury

1 1.62855 0.22191 7.34 <.0001 0.86466

Type of Brand: National

1 0.25361 0.03381 7.5 <.0001 0.68902

Type of Brand: Private


Sleeve Length: Long Sleeves

1 0.19434 0.03177 6.12 <.0001 0.65823

Sleeve Length: Short Sleeve & Sleeveless


Features: Trim

1 0.11688 0.04034 2.9 0.0039 0.86943

Features: Screen Print

1 -0.10627 0.03071 -3.46 0.0006 0.73823

Features: No Features


Fiber (continuous): Cotton

1 -0.00125 0.00042 -3.01 0.0027 0.77554

Fiber (continuous): Cashmere/Wool

1 0.01036 0.0035 2.96 0.0032 0.92586

Fiber (continuous): Polyester


NOTE: This model contains control variables for type of business, type of store, and geographic region, but they are not used in adjustments.

Summary notes

  • Quality adjustments derived from hedonic regressions for girls’ tops were first incorporated into the CPI in December 1995. Since then, the model for girls’ tops has been updated six times.
  • Characteristics that received the most focus during the model evaluation include item style, type of brand, sleeve length, item features, and fiber content.
  • Two variables that surfaced as price differentiators were activewear tops and sweatshirts. No previous hedonic model for girls’ tops included an adjustment for these variables. In fact, all styles of tops but vests and bodysuits are accounted for in the model.
  • Fiber variables in the new model were not as significant as in previous models. For example, the last model iteration had three additional adjustments for synthetic fibers. This model found that only natural fibers were statistically significant.
  • Applique, embroidery, or applique-embroidery combination variables were included in many past models. In the new model, however, these feature characteristics were replaced by two others, trim and screen print, which were not as present in previous models but turned out to be significant in the new iteration.
  • Applique-embroidery and partial closure pullover were two prevalent interaction variables in previous models, but these only involved either style or feature characteristics. Until now, fiber content was not used in this way. The cashmere-wool interaction was calculated as a summation of fibers, and a parameter estimate was developed accordingly for use in adjusting both cashmere and wool fiber changes.


Access data for Girls' Apparel in our online database.

Additional information

Additional information may be obtained from the Consumer Price Index Information Office by email or calling 202-691-7000. Information on the CPI's overall methodology can be found in the BLS Handbook of Methods.

Last Modified Date: June 27, 2024