Project

General

Profile

Support #180

Sports Club membership among Scottish adults

Added by Chris Martin over 10 years ago. Updated over 10 years ago.

Status:
Closed
Priority:
Normal
Assignee:
-
Category:
Data inconsistency
Start date:
08/07/2013
% Done:

100%

Description

Hi,

I have a query relating to the variable b_club, (sports club membership).

I am wanting to calculate two figures for Scotland only:
a) The %age of adults who have played a sports/exercise in last 12 months who are a member of a sports club,(ie the same definition as when question is asked and
b) the %age of all adults who are a member of a sports club (assuming rightly or wrong that those who have not done a sports/exercise in last year are NOT a member of a sports club).

I think I might be doing something wrong. (Output attached)

First, I exclude anyone outside Scotland and those who don't have a weight. This leaves me with 2,908 cases. (All figures unweighted for ease here).

Then, I exclude cases with only proxy info (these cases have no info on sports participation or membership questions) and the one case where all the relevant data is missing. This leave me with 2,682. I thought this should be my denominator, representative of all adults in Scotland.

However, see the table below (also row 247 in the attached output) of whether undertaken any sports in last year by sports club membership.

b_club sports club member      Total
        -9    -8    1    2    
sports_any2 Yes    0    654    0    3    657
        No    345    0    471    1209    2025
Total        345    654    471    1212    2682

For sports club membership among adult population, the numerator would be the 471 who are sports club members (giving 18%: 471/2682*100). (Or 17% if I applied the appropriate weight.)

However, there are 345 cases where the sports club member question is missing even when they have done a sport in the last year and 3 cases where they have been asked the sports membership question even though they’ve not participated in any sport in the last year. Moreover, I don’t think the 345 are a random sub-group, which makes me worry about the estimates.

I wonder if you could tell me how to deal with the 345 + 3 cases in terms of my estimates, or whether I'm making a more fundamental mistake.

Many thanks,
Chris

Also available in: Atom PDF