Statistical Difference

In an energy balance, the statistical difference is the numerical difference between the total supply of an energy product and the total use of it. It is presented in line 2 of the balance, as displayed in tables 8.1 and 8.2, and is calculated by subtracting the total use of energy (the sum of lines 3 to 7) from the total supply of energy products (line 1) [IRES, Ch. VIII, para. 8.45, PDF p. 122, 2018]. Structurally, the row is placed between the top (supply) and middle blocks of the balance [IRES, Ch. VIII, para. 8.15, PDF p. 118, 2018].

Formula

Statistical difference = Total supply of energy products (line 1) − Total use of energy (sum of lines 3 to 7)

[IRES, Ch. VIII, para. 8.45, PDF p. 122, 2018]

Causes

Statistical difference arises from various practical limitations and problems related to the collection of the data that make up supply and demand. For example, the data may be subject to sampling or other collection errors, and/or be taken from different data sources that use different time periods, different spatial coverage, different fuel specifications, or different conversions from volume to mass or from mass to energy content on the supply and demand sides of the balance [IRES, Ch. VIII, para. 8.45, PDF p. 122, 2018].

Diagnostic role for data quality

The reasons for a large statistical difference should be examined, because a large difference indicates that the input data are inaccurate and/or incomplete [IRES, Ch. VIII, para. 8.45, PDF p. 122, 2018]. Statistical differences in commodity balances can help explain large statistical differences in the (overall) energy balance:

• If commodity balances show negligible statistical differences, this may indicate that the conversion factors to energy units should be investigated, since they may be the reason for a large statistical difference in the energy balance.
• If the statistical difference for a specific product’s commodity balance is large, this may indicate that data collection for that specific product needs investigation.

It is acknowledged that countries’ experiences vary with respect to the presentation and treatment of statistical differences; the forthcoming Energy Statistics Compilers Manual (ESCM) will provide an overview of the issues involved and identify good practices that countries might wish to follow [IRES, Ch. VIII, para. 8.46, PDF p. 122, 2018] — recommendations tracker row VIII/8.45.

As a quality indicator

IRES Chapter IX names the statistical difference explicitly as a quality indicator: since it does not measure quality directly but is a by-product of the statistical process, comparing energy consumption data with energy supply figures in the statistical difference flow can be used to validate the data and flag potential problem areas — a practical illustration of the general quality-measures-vs.-quality-indicators distinction in Quality Measurement and Reporting [IRES, Ch. IX, para. 9.19, PDF p. 134, 2018].

Related

• IRES Chapter VIII — Energy Balances — the chapter digest this page supports
• Energy Balance — the parent accounting framework; see its structure section for where this row sits between the top and middle blocks
• Commodity Balance — the single-product balance whose own statistical difference row can help diagnose a large statistical difference in the overall energy balance
• Quality Measurement and Reporting — Chapter IX names the statistical difference as a quality indicator (para 9.19)
• Energy Statistics Compilers Manual — the forthcoming companion guidance on good practices for presenting and treating statistical differences