The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » Tanzania » Banking questions in Men's Questionnaire (DHS 2022) (Mis-routed questions in Men's Questionnaire on bank account and use of mobile phone to make mobile transactions)
Banking questions in Men's Questionnaire (DHS 2022) [message #29201] Thu, 09 May 2024 10:34 Go to next message
aviva is currently offline  aviva
Messages: 1
Registered: May 2024
Member
I am conducting analysis on the 2022 Tanzania DHS survey and have noticed an issue while using the following questions in the men's recode dataset:
Q124: "In the last 12 months, have you used a mobile phone to make financial transactions such as sending or receiving money, paying bills, purchasing goods or services, or receiving wages?" (mv169b in the men's recode)
Q125: "Do you have an account in a bank or other financial institution that you yourself use?" (mv170 in the men's recode).

There are a lot of NAs/blanks in the data for the above two questions in the men's recode (mv169b and mv170 in the men's recode) - 3072 NA out of 5763.
All the blanks are for men who have a mobile phone but do not have a smartphone.

In the women's questionnaire, it looks like all females were asked and answered Q124 and Q125 (image attached)

In the male's questionnaire, the routing for the questions on whether someone has a bank account differs:
If someone has a mobile phone, but do not have a smartphone, then they skip to question 127, skipping questions Q124 and Q125.

The percentages reported in the 2022 Tanzania DHS Report seem to have grouped the blanks in the men's record data for these questions with those who report to not have a bank account and who report to have not used a mobile phone to make financial transactions in the past 12 months. However, this assumes that those who have a mobile phone but do not have a smartphone also don't have a bank account and have not use a mobile phone to make financial transactions.

Please advise how we should proceed with this analysis, or if any corrections have been published for these questions.

[Updated on: Tue, 14 May 2024 09:19]

Report message to a moderator

Re: Banking questions in Men's Questionnaire (DHS 2022) [message #29244 is a reply to message #29201] Tue, 14 May 2024 13:31 Go to previous message
Janet-DHS is currently offline  Janet-DHS
Messages: 737
Registered: April 2022
Senior Member
Following is a response from DHS staff member, Tom Pullum:

Thank you for alerting us to this inconsistency between women and men in the questionnaire and tables (for the Tanzania 2022 survey) for using a mobile phone for banking transactions in the past 12 months. I will re-state what you said in terms of the numbers in the final report.

Both tables 15.6.1 (women) and 15.6.3 (men) have a footnote saying "Respondents were asked about use of a mobile phone for financial transactions whether or not they owned a mobile phone." However, I have checked, and this footnote only applied for women, not for men. I will paste here a table from TZMR82, in which I try to match the 21.7% in column 3 of table 15.6.3.

. tab mv169b mv169a [iweight=wt],m col

 
   last 12 |

months use |

    mobile |

telephone |

       for |

financial |     owns a mobile

transactio |       telephone

        ns |        no        yes |     Total

-----------+----------------------+----------

        no |  1,317.95  93.589638 |  1,411.54

           |     91.78       2.16 |     24.49

-----------+----------------------+----------

       yes | 118.07353  1,134.521 |1,252.5945

           |      8.22      26.22 |     21.74

-----------+----------------------+----------

         . |         0  3,098.865 | 3,098.865

           |      0.00      71.62 |     53.77

-----------+----------------------+----------

     Total | 1,436.024  4,326.976 |     5,763

           |    100.00     100.00 |    100.00
This table matches the n for men (5,673) but it confirms that there was a skip in the questionnaire for men, such that some men were NOT asked the banking question. The following table shows that men who did not have a SMART phone were not asked the banking question.

. tab mv169b mv169c [iweight=wt],m col
 

   last 12 |

months use |

    mobile |

telephone |

       for |

financial |  is respondent's mobile phone a

transactio |           smart phone

        ns |        no        yes          . |     Total

-----------+---------------------------------+----------

        no |         0  93.589638   1,317.95 |  1,411.54

           |      0.00       7.62      91.78 |     24.49

-----------+---------------------------------+----------

       yes |         0  1,134.521  118.07353 |1,252.5945

           |      0.00      92.38       8.22 |     21.74

-----------+---------------------------------+----------

         . | 3,098.865          0          0 | 3,098.865

           |    100.00       0.00       0.00 |     53.77

-----------+---------------------------------+----------

     Total | 3,098.865  1,228.111  1,436.024 |     5,763

           |    100.00     100.00     100.00 |    100.00 

Because this is a discrepancy in the questionnaire, I don't believe there is any way to modify table 15.6.3 to correspond with table 15.6.1. It WOULD be possible to modify table 15.6.1 to correspond with table 15.6.3, by re-classifying the women who did not have a smartphone as NA on the banking question (but including them in the denominator). I would not recommend that. As a general rule, NA cases should not be included in the denominator of a percentage. You could re-calculate the percentages for both women and men by removing those without a smartphone from the denominator. This would give consistency but could be misleading. I will not make a recommendation, but I hope other users will suggest something.

I have informed our data processing staff of this discrepancy and also the staff who work on questionnaire design.
Previous Topic: 2005 geocoding data
Goto Forum:
  


Current Time: Sun Jun 16 15:13:31 Coordinated Universal Time 2024