Multi-response metrics | Knowledge Base

In a multi-response (MR) question a respondent can give several answers: they know three brands, visited two stores, chose four reasons. Regular metrics count respondents. The MR metrics count mentions, the individual answers, which lets you report shares of all answers and the number of answers per person.

The MR metrics are off by default. Switch them on in Metric settings. For how MR sets are built, see Multi-response Sets.

The metrics

Metric	What it counts	Regular counterpart
Count MR (`cm_`)	Weighted number of mentions in the cell. For a single category it equals Count. For a net it adds up the mentions of every category in the net.	Count
Vert Base MR (`vbm`)	Weighted number of all mentions of the whole MR set in the column.	Vert Base
Vert% MR (`vpm`)	Count MR ÷ Vert Base MR × 100: the cell's share of all mentions. Within a column these add up to 100%.	Vert%
Volume MR (`vm_`)	Sum of weight × digit over the mentions in the cell.	Vol
Mean MR (`am_`)	Volume MR ÷ Count MR: the average digit per mention.	Mean
CpC (`cpc`), count per capita	Count MR ÷ Count: mentions per respondent who gave at least one of the cell's answers.	none
CpC.B (`cpb`), count per capita, base	Count MR ÷ Vert Base: mentions per respondent in the base, including those who gave none of the cell's answers.	none

On a single-response variable every respondent has one answer, so the MR metrics equal their regular counterparts and CpC is 1.

Worked example

Four respondents answered "Which brands do you know?":

Respondent	Brands known
R1	A, B, C
R2	A
R3	A, B
R4	C

There are 7 mentions from 4 respondents. With the brands in rows plus a net "Any brand":

Row	Count	Vert%	Count MR	Vert% MR	CpC	CpC.B
A	3	75%	3	42.9%	1.0	0.8
B	2	50%	2	28.6%	1.0	0.5
C	2	50%	2	28.6%	1.0	0.5
Any brand (net)	4	100%	7	100%	1.8	1.8

Vert% adds up to 175% because respondents know several brands. Vert% MR adds up to 100% because it divides by the 7 mentions (Vert Base MR = 7).
Brand A holds 42.9% of all brand mentions while 75% of respondents know it.
On the net, CpC = 7 ÷ 4 = 1.75: people who know any brand know 1.75 brands on average.
CpC and CpC.B differ when the base includes respondents who gave none of the cell's answers. On brand A, CpC is 1.0 (each of the three respondents who know A mentions it once) while CpC.B is 3 ÷ 4 = 0.75.

The same metrics in a crosstab of a multi-response awareness question:

Prompted awareness. Vert% adds up to more than 100% because respondents know several brands; Vert% MR shows each brand's share of all mentions; CpC is 1.0 on single brands.

Another example: six respondents gave eight answers. Regular metrics are based on the six respondents, MR metrics on the eight answers.

MR metrics basics.png — How MR-set-related metrics differ

More examples of CpC, including scale batteries, are on Count per Capita Metric.