Customer Service: +44 1334 657957

Overlay tables for repeat questions


The descriptions which follow refer to a simple product test consisting of 2 repeats of 2 questions and 100 respondents who all test both products:

QA

Q1P

Q1L

Q1R

Q2P

Q2L

Q2R

QP

What age are you? (Young, Old)

Which product did you try first? (list of possible products)

What did you like about the first product? (list of likes)

Please rate the first product (Liked, Indifferent, Disliked)

Which product did you try second? (list of possible products)

What did you like about the second product? (list of likes)

Please rate the second product (Liked, Indifferent, Disliked)

Which did you prefer? (preferred first tested, preferred second tested, no preference)

To produce a table of Likes by Products:

#1 is Q1L by Q1P

The following table (which also takes the name #1) is Q2L by Q2P (overlaid)

Table 1 then shows “Likes of product” down the side by “Product” across the top (breakdown), with a base of 200.

Whenever a variable is produced it is important to decide whether it is a respondent variable or a product variable.

Respondent variables contain information that relate to the respondent. They may contain information from the product questions, provided there is no ambiguity about which product is being referred to. For example it is possible to define a response in a respondent variable as “Liked product A”. The definition would be (Q1P/1 and Q1R/1) or (Q2P/1 and Q2R/1).

Product variables contain information that relates to the relevant product. For example, if we want to produce a table of “likes” with summary rows broken down by how a product was rated, we need to produce two likes variables, one for each repeat:

  • V1L from Q1L with summary rows
  • V2L from Q2L with summary rows

These variables may then by used on product based tables:

  • #2 is V1L by Q1R

The following table (which also takes the name #2) is V2L by Q2R (Overlaid)

Of course, this table does not distinguish between products. To produce the same table but only for product B (second response to Q1P and Q2P) we need to filter both tables:

#3 is V1L by Q1R filtered by Q1P/2

The following table (which also takes the name #3) is V2L by Q2R filtered by Q2P/2 (overlay)

If every respondent tried product B, the base for this table will be 100. This does not mean that it is a respondent based table, it is a filtered product based table.

Whenever product based tables are produced, and relevant product information is included in a variable, then a variable will be needed for each repeat.

Confusion often arises when different sorts of information are combined. For example, if we required a breakdown with 7 columns - Age Young, Age Old, Liked, Indifferent, Not Liked, Tried first, Tried second. Because the breakdown contains some information relevant to the “likes” question being tabulated we need a variable for each repeat.

VBR1 is QA/1, QA/2, Q1R/1, Q1R/2, Q1R/3, T, F

VBR2 is QA/1, QA/2, Q2R/1, Q2R/2, Q2R/3, F, T

#4 is V1L by VBR1

The following table (which also takes the name #4) is V2L by VBR2 (overlaid)

You will have noticed that the first columns of the variables VBR1 and VBR2 have the same definition. The last two columns are either true (everybody) or false (nobody).

It does not matter how many products are involved in the test. For product based tables, a table is produced for each repeat and these are then overlaid. For example, if each respondent tries three products from a list of seven possible then three tables are needed (one standard table and two overlays).