Understanding Society User Support

Dear Sir,

I am in the process of running a query on ‘satisfaction levels with leisure time’
in both the British Household Panel and Understanding Society surveys.
I have pasted in an example of the responses below, as you will see there
is a significant variation in the distribution of the responses from the British
Household Panel survey— which tends to follow a more standard distribution—
compared with the responses yielded from the Understanding Society survey, where
the proportion of responses show an increase with the level of satisfaction.

In terms of weighting and filters, for British Household Panel wave 17, I have used:

Weight: QXRWTUK1
Filter: QJBSEMP = 1 (employee)
Variable: QLFSAT7

For Understanding Society wave 1, I have used:

Weight: a_indpxus_xw.
Filter: a_jlsemp = 1 (employee)
Variable: a_sclfsat7

As you can see in the results below, there is a significant change in the
results given for the same question between the two surveys.
I would appreciate any help deciphering why the responses
vary quite so much.

British Household Panel wave 17

Frequency Percent
Missing or wild 86 1.2
Not satisfied at all 133 1.8
2 489 6.8
3 1019 14.1
Not satis/dissat 1608 22.3
5 2105 29.2
6 1264 17.5
Completely satisfied 509 7.1
Total 7212 100.0

Understanding Society wave 1

Frequency Percent
missing 2456 16.6
completely dissatisfied 548 3.7
mostly dissatisfied 558 3.8
somewhat dissatisfied 793 5.4
neither satisfied
or dissatisfied 1546 10.5
somewhat satisfied 1818 12.3
mostly satisfied 3576 24.2
completely satisfied 3464 23.5
Total 14759 100.0

Many thanks

Ray Storry