Full replication code for data access, variable construction, cell sizes and distribution, chi-squared tests, and data visualization creation is available in a Github repository: https://github.com/jrpepin/NCHAT_Money/blob/master/data_viz_script.R.

Data ————————————————————————

This visualization uses population-based U.S. data from the 2020–2021 National Couples’ Health and Time Study, downloaded from ICPSR (study #: 38417). The Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) supported the data collection of 3,639 partnered individuals through the Gallup Panel. Sexual and gender diverse families were oversampled in order to provide high-quality, population-representative data of same- and different-gender couples. When weighted, the survey provides nationally representative data on cohabiting and married individuals (aged 20–60).

Measures ——————————————————————

Money arrangement

Respondents were told: “Couples handle their money differently. Which of the following do you do?” The response choices were: (1) We keep all of our money separately, (2) Put some money together, or (3) Put all of our money together. Only 6 (.2%) respondents refused to answer the question item.

Marital status

This visualization uses legal marriage status to differentiate married and cohabiting couples. Respondents who answered affirmatively when asked “Have you and your current spouse/partner ever experienced any of the following with each other? Legal marriage” were considered legally married. Four respondents refused to answer the question.

Parental status

The direct parent identifiers are only available through a restricted Virtual Data Enclave, requiring institutional IRB approval and an application. With reproducibility a priority, parents were instead identified using the skip logic of a publicly available variable on respondents’ stress over dividing childcare with their partner. Respondents who were not shown the question (C18D_B) were determined to be not-parenting and those shown the question were coded as parenting.

Couple type

Two variables were used to identify the couple type: D2 (“Which of the following best describes your gender?”) and HHR5 (“How would [Field-HHP1] describe their gender?”). Women who reported their partner was a man (and vice versa) were coded as different-gender couples. Men who reported their partner was a man and women who reported their partner was a woman were coded as same-gender couples. Respondents who chose another gender for themselves or their partners were not included in this data visualization.

Age group

The survey contained a variable (AGE_SD) indicating the respondents’ age was between one of three ranges: (1) 18–34, (2) 35–50, or (3) 51–69. The codebook and another variable (AGE_SD_2CAT) indicate the possible age range for main respondents is 20–60. Therefore, the categories were recoded to reflect the codebook top/bottom ages (partner’s age may be younger/older). The refusal rate for this item was 1.6% (n=57).

Sample ——————————————————————–

This data visualization is based on same- and different-gender U.S. couples (aged 20–60) who had no missing information regarding their money arrangements, marital and parental status, and age group (243 cased dropped). The final sample included 3,396 partnered individuals. More details about the sample recruitment can be found online: https://osf.io/preprints/socarxiv/p8a34/.

Table 1.

Summary statistics of the analytic sample (unweighted)
Variable N = 33961
Money Treatment
Put all money together 1,614 (48%)
Put some money together 950 (28%)
Keep all money separate 832 (24%)
Marital Status
Married 2,534 (75%)
Cohabiting 862 (25%)
Parental Status
Parenting 1,149 (34%)
Not parenting 2,247 (66%)
Couple Type
Different-gender 2,444 (72%)
Woman-woman 371 (11%)
Man-man 581 (17%)
Age
Age: 20-34 688 (20%)
Age: 35-50 1,473 (43%)
Age: 51-60 1,235 (36%)
1 n (unweighted) (% (unweighted))

CHI-SQUARED TESTS ——————————————–

A series of chi-squared tests were conducted to test for statistical differences within and between married and cohabiting respondents. The tables show unweighted and weighted frequencies to provide information about the cell sizes. The chi-squared tests use the survey weight, accounting for the oversampling in recruitment of respondents for the survey.

Table 2.

Chi-squared tests within marital groups
unweighted n [weighted n]
Variable Married Cohabiting
Put all money together1 Put some money together1 Keep all money separate1 p-value2 Put all money together1 Put some money together1 Keep all money separate1 p-value2
parent 0.020 0.3
Parenting 661 [976] 230 [274] 108 [150] 36 [54] 46 [51] 68 [87]
Not parenting 819 [926] 470 [322] 246 [219] 98 [98] 204 [121] 410 [258]
couple 0.040 0.7
Different-gender 1,224 [1,885] 458 [585] 254 [362] 88 [148] 140 [165] 280 [333]
Woman-woman 126 [9] 111 [7] 31 [4] 15 [2] 31 [3] 57 [5]
Man-man 130 [8] 131 [4] 69 [3] 31 [3] 79 [5] 141 [7]
age 0.8 >0.9
Age: 20-34 209 [348] 104 [120] 48 [60] 45 [72] 89 [75] 193 [158]
Age: 35-50 678 [942] 323 [295] 166 [196] 47 [53] 93 [70] 166 [125]
Age: 51-60 593 [612] 273 [181] 140 [114] 42 [28] 68 [27] 119 [62]
1 n (unweighted) [n]
2 chi-squared test adjusted by a design effect estimate

Table 3.

Chi-squared tests within demographic group by marital status
unweighted n [weighted n]
Variable Put all money together1 Put some money together1 Keep all money separate1 p-value2
Parenting
<0.001
Married 661 [976] 230 [274] 108 [150]
Cohabiting 36 [54] 46 [51] 68 [87]
Not parenting
<0.001
Married 819 [926] 470 [322] 246 [219]
Cohabiting 98 [98] 204 [121] 410 [258]
Different-gender
<0.001
Married 1,224 [1,885] 458 [585] 254 [362]
Cohabiting 88 [148] 140 [165] 280 [333]
Woman-woman
0.022
Married 126 [9] 111 [7] 31 [4]
Cohabiting 15 [2] 31 [3] 57 [5]
Man-man
<0.001
Married 130 [8] 131 [4] 69 [3]
Cohabiting 31 [3] 79 [5] 141 [7]
Age: 20-34
<0.001
Married 209 [348] 104 [120] 48 [60]
Cohabiting 45 [72] 89 [75] 193 [158]
Age: 35-50
<0.001
Married 678 [942] 323 [295] 166 [196]
Cohabiting 47 [53] 93 [70] 166 [125]
Age: 51-60
<0.001
Married 593 [612] 273 [181] 140 [114]
Cohabiting 42 [28] 68 [27] 119 [62]
1 n (unweighted) [n]
2 chi-squared test adjusted by a design effect estimate

Figure Data —————————————————————

The bivariate statistics used in the figure were constructed using the survey-provided variable (WEIGHT_MAINRESPONDENT) to correct for over- or undersampling of specific target populations.