FOI reference: FOI-2024-2251
You asked
I am writing to request the ONS provides additional information on sampling and index construction for the following item RPI > Housing and household expenditure > Household Services > Telephone > MOBILE PHONE APPLICATIONS (Item Code 440408). In particular I wanted to see if you could please help with the below:
For the top 30 apps for a given shop, to collect the sub index for that shop are you weighing the apps using a given set of weights or taking a simple average.
For the top 30 apps - are those 30 specific apps selected at the beginning of the year or are they allowed to change to reflect what the actual most popular apps are for a given month
How and when is the data sampled
We said
Thank you for your queries about how mobile phone applications (apps) are treated in the consumer price indices.
Mobile phone apps’ prices are collected and indices compiled based on the best-seller chart positions of the most popular apps. As the composition of the charts changes, so does the sample of products selected each month. This type of methodology is used to avoid biases that could occur if we followed specific apps as they decreased in price with age each month.
The mobile phone apps subindex for each shop is calculated using equal weights, that is, no distinct weights are applied. A geometric mean is used to combine prices for the Consumer Prices Index (and Consumer Prices Index including owner occupiers’ housing costs) and an arithmetic mean is used for the Retail Prices Index.
The apps change each month depending on their inclusion in app best-seller charts.
The prices are collected on “index day” (the main price collection day each month) according to the chart positions at that point.