SRHS Simple Models part II

coded_df <- readRDS("C:/Users/court/Google Drive/Research/SRHS/SRHS_coded.RDS")
coded_df <- coded_df %>%
  mutate_if(
        is.labelled,
        ~(as_factor(.)))

df <- coded_df %>% select(
   ll_finrole2,
   ll_size,
   ll_size2,
   ll_comm_unit,
   bp_advert_b,
   bp_screen_b,
   bp_dec_flex_b,
   bp_dec_standard_b,
   bp_dec_offer_b,
   bp_vouch_b,
   bp_crim_b,
   bp_dis_b,
   bp_keep_pct_num,
   bp_keep_rsn,
   bp_install_b,
   bp_comm_unit_fee_app_b,
   bp_comm_unit_fee_sec_b,
   bp_comm_unit_fee_clean_b,
   bp_comm_unit_fee_park_b,
   bp_fair_b,
   bp_comm_unit_fee_incl_b,
   bp_comm_unit_deposit_b,
   bp_rent_raise_b,
   bp_rent_raise_num,
   bp_rent_raise_rising,
   bp_rent_raise_regulation,
   bp_rent_raise_market,
   bp_rent_raise_s_tax,
   bp_rent_raise_s_repair,
   bp_rent_raise_s_new,
   bp_rent_raise_s_market,
   bp_rent_raise_s_reg,
   bp_terminate_b,
   bp_terminate_num,
   bp_terminate_paylate,
   bp_terminate_pay,
   bp_court_b,
   bp_income_b,
   bp_income_num,
   ll_remain_1or5yr_b,
   ll_remain_reg
   )

df <- df %>% mutate(
  ll_size = relevel(as.factor(ll_size), ref="small"),
  ll_comm_unit = relevel(as.factor(ll_comm_unit), ref="2 bedroom"),
  ll_finrole2 = as.factor(ll_finrole2)
)

df <- df %>% rename(
                    "Landlord_size_" = ll_size2,
                    "Financial_role_" = ll_finrole2,
                    "Common_unit_" = ll_comm_unit,
                    "Advertising" = bp_advert_b,
                    "Screening" = bp_screen_b,
                    "Flexible_decision" = bp_dec_flex_b,
                    "Standard_criteria" = bp_dec_standard_b,
                    "Offers_to_nonstandard" = bp_dec_offer_b,
                    "Rent_to_vouchers" = bp_vouch_b,
                    "Rent_to_criminal" = bp_crim_b,
                    "Disability_accom." = bp_dis_b,
                    "Deposit_pct_kept" = bp_keep_pct_num,
                    "Deposit_kept_reason" = bp_keep_rsn,
                    "Allow_installments" = bp_install_b,
                    "Fair_market_price" = bp_fair_b,
                    "App_fee"   =   bp_comm_unit_fee_app_b,
                    "Security_deposit" =    bp_comm_unit_fee_sec_b,
                    "Cleaning_fee"  =   bp_comm_unit_fee_clean_b,
                    "Parking_fee"   =   bp_comm_unit_fee_park_b,
                    "Utilities_included" = bp_comm_unit_fee_incl_b,
                    "Last_month_required" = bp_comm_unit_deposit_b,
                    "Raised_rent" = bp_rent_raise_b,
                    "Raised_rent_pct" = bp_rent_raise_num,
                    "Raised_rent_rising_costs" = bp_rent_raise_rising,
                    "Raised_rent_regulations" = bp_rent_raise_regulation,
                    "Raised_rent_market" =  bp_rent_raise_market,
                    "Raised_rent_only_tax" = bp_rent_raise_s_tax,
                      "Raised_rent_only_repairs" =bp_rent_raise_s_repair,
                     "Raised_rent_only_newproperty" = bp_rent_raise_s_new,
                      "Raised_rent_only_market" =bp_rent_raise_s_market,
                      "Raised_rent_only_regulation" =bp_rent_raise_s_reg,
                     "Ever_terminated" = bp_terminate_b,
                     "Num_of_terminations" = bp_terminate_num,
                     "Terminated_paid_late" = bp_terminate_paylate,
                     "Terminated_non_payment" = bp_terminate_pay,
                     "Terminated_use_courts" = bp_court_b,
                     "Renterincome_above_avg_Seattle" = bp_income_b,
                     "Renter_income_estimate" = bp_income_num,
                     "Still_mngr_1_5yrs" = ll_remain_1or5yr_b,
                     "Stop_mngr_burdensome_regs" = ll_remain_reg)
                   


# --------------------------

Business practices, landlord size, and financial role

This document shows simple crosstabs and models for:

• each business practice and landlord size
• each business practice, landlord size, financial role, and most common unit

Most variables were forced into a binary (likert scale included) and odds ratios are presented. Where possible, numbers were left numeric.

What’s new from our last discussion:

• landlord size now consists of 4 categories:
  • small single family homes
  • small non-single family homes
  • big single family homes
  • big non-single family homes
• most common unit has been added to models
• new business practices have been added that reflect those who raised rent exclusively due to one reason
  • note: “other” wasn’t considered a reason for raising rent (e.g. if landlord cited “taxes” and “other” as their reasons for raising rent, that landlord was considered to only have “taxes” as their sole reason. This is because a large proportion of landlords cited "other)

Cross tab of covariate variables:

df %>% 
  tbl_summary(
    by = Financial_role_,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Landlord size & financial role**")

Characteristic	Landlord size & financial role		p-value1
	Other	Primary Income
Landlord_size_, n (%)			<0.001
small (single-fam)	1,639 (95%)	82 (4.8%)
small (non single-fam)	1,128 (93%)	87 (7.2%)
big (single-fam)	40 (74%)	14 (26%)
big (non single-fam)	554 (67%)	274 (33%)
1 Pearson's Chi-squared test

Leasing

Advertising

Outcome:

• 0: word-of-mouth
• 1: non word-of-mouth

Crosstabs:

df %>% 
  tbl_summary(
    by = Advertising,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Advertising**")

Characteristic	Advertising		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	1,506 (87%)	221 (13%)
small (non single-fam)	1,097 (90%)	122 (10%)
big (single-fam)	50 (89%)	6 (11%)
big (non single-fam)	806 (97%)	28 (3.4%)
Financial_role_, n (%)			<0.001
Other	2,997 (89%)	364 (11%)
Primary Income	445 (97%)	12 (2.6%)
Common_unit_, n (%)			0.018
2 bedroom	1,325 (92%)	117 (8.1%)
0 bedroom (studio/efficiency)	138 (93%)	10 (6.8%)
1 bedroom	805 (91%)	83 (9.3%)
3 bedroom	751 (89%)	95 (11%)
4 bedroom	192 (86%)	31 (14%)
5+ bedroom	65 (87%)	10 (13%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Advertising~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Advertising~1 + Landlord_size_ + Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.76 (0.60 to 0.96)	0.020
big (single-fam)	0.82 (0.31 to 1.79)	0.65
big (non single-fam)	0.24 (0.16 to 0.35)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate = TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.66 (0.49 to 0.87)	0.004
big (single-fam)	1.00 (0.38 to 2.21)	>0.99
big (non single-fam)	0.26 (0.16 to 0.40)	<0.001
Financial_role_
Other	—
Primary Income	0.32 (0.16 to 0.57)	<0.001
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.38 (0.65 to 2.64)	0.37
1 bedroom	1.54 (1.13 to 2.09)	0.006
3 bedroom	1.08 (0.80 to 1.45)	0.62
4 bedroom	1.32 (0.84 to 2.02)	0.21
5+ bedroom	1.41 (0.66 to 2.73)	0.34
1 OR = Odds Ratio, CI = Confidence Interval

Screening

Outcome: use of third party screening service

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Screening,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Screening**")

Characteristic	Screening		p-value1
	no	yes
Landlord_size_, n (%)			<0.001
small (single-fam)	1,096 (63%)	631 (37%)
small (non single-fam)	714 (59%)	505 (41%)
big (single-fam)	30 (54%)	26 (46%)
big (non single-fam)	403 (48%)	431 (52%)
Financial_role_, n (%)			<0.001
Other	2,001 (60%)	1,360 (40%)
Primary Income	230 (50%)	227 (50%)
Common_unit_, n (%)			0.76
2 bedroom	839 (58%)	603 (42%)
0 bedroom (studio/efficiency)	83 (56%)	65 (44%)
1 bedroom	514 (58%)	374 (42%)
3 bedroom	497 (59%)	349 (41%)
4 bedroom	129 (58%)	94 (42%)
5+ bedroom	50 (67%)	25 (33%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Screening~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Screening~1 + Landlord_size_ + Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.23 (1.06 to 1.43)	0.007
big (single-fam)	1.51 (0.88 to 2.57)	0.13
big (non single-fam)	1.86 (1.57 to 2.20)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.40 (1.17 to 1.67)	<0.001
big (single-fam)	1.43 (0.82 to 2.46)	0.20
big (non single-fam)	2.11 (1.73 to 2.59)	<0.001
Financial_role_
Other	—
Primary Income	1.15 (0.93 to 1.43)	0.19
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	0.83 (0.58 to 1.18)	0.31
1 bedroom	0.87 (0.73 to 1.04)	0.13
3 bedroom	1.22 (1.01 to 1.47)	0.036
4 bedroom	1.29 (0.96 to 1.74)	0.091
5+ bedroom	0.85 (0.51 to 1.38)	0.51
1 OR = Odds Ratio, CI = Confidence Interval

Flexible decision

Outcome: criteria used to make leasing decisions is flexible

• 1: agree
• 0: disagree or neutral

Crosstabs:

df %>% 
  tbl_summary(
    by = Flexible_decision,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Flexible decision**")

Characteristic	Flexible decision		p-value1
	disagree/neutral	agree
Landlord_size_, n (%)			0.82
small (single-fam)	676 (43%)	907 (57%)
small (non single-fam)	474 (42%)	644 (58%)
big (single-fam)	19 (38%)	31 (62%)
big (non single-fam)	297 (41%)	426 (59%)
Financial_role_, n (%)			0.16
Other	1,311 (43%)	1,762 (57%)
Primary Income	149 (39%)	234 (61%)
Common_unit_, n (%)			0.63
2 bedroom	550 (42%)	763 (58%)
0 bedroom (studio/efficiency)	44 (34%)	84 (66%)
1 bedroom	343 (42%)	467 (58%)
3 bedroom	325 (43%)	432 (57%)
4 bedroom	88 (43%)	116 (57%)
5+ bedroom	29 (41%)	41 (59%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Flexible_decision~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Flexible_decision~1 + Landlord_size_ + Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.01 (0.87 to 1.18)	0.87
big (single-fam)	1.22 (0.69 to 2.21)	0.51
big (non single-fam)	1.07 (0.89 to 1.28)	0.46
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.00 (0.83 to 1.19)	0.96
big (single-fam)	1.22 (0.68 to 2.26)	0.51
big (non single-fam)	0.99 (0.80 to 1.23)	0.95
Financial_role_
Other	—
Primary Income	1.14 (0.90 to 1.45)	0.27
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.36 (0.93 to 2.02)	0.12
1 bedroom	0.98 (0.82 to 1.17)	0.82
3 bedroom	0.95 (0.78 to 1.15)	0.59
4 bedroom	0.95 (0.70 to 1.29)	0.73
5+ bedroom	1.02 (0.62 to 1.68)	0.95
1 OR = Odds Ratio, CI = Confidence Interval

Standard criteria

Outcome: decisions based on a standard set of criteria

• 1: agree
• 0: disagree or neutral

Crosstabs:

df %>% 
  tbl_summary(
    by = Standard_criteria,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Standard criteria**")

Characteristic	Standard criteria		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	502 (29%)	1,208 (71%)
small (non single-fam)	337 (28%)	867 (72%)
big (single-fam)	17 (31%)	38 (69%)
big (non single-fam)	162 (20%)	668 (80%)
Financial_role_, n (%)			<0.001
Other	924 (28%)	2,403 (72%)
Primary Income	89 (20%)	365 (80%)
Common_unit_, n (%)			0.83
2 bedroom	369 (26%)	1,057 (74%)
0 bedroom (studio/efficiency)	39 (27%)	108 (73%)
1 bedroom	235 (27%)	650 (73%)
3 bedroom	233 (28%)	605 (72%)
4 bedroom	54 (24%)	167 (76%)
5+ bedroom	23 (31%)	52 (69%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Standard_criteria~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Standard_criteria~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.07 (0.91 to 1.26)	0.42
big (single-fam)	0.93 (0.53 to 1.70)	0.80
big (non single-fam)	1.71 (1.40 to 2.10)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.21 (1.00 to 1.46)	0.054
big (single-fam)	0.81 (0.46 to 1.50)	0.48
big (non single-fam)	1.77 (1.40 to 2.25)	<0.001
Financial_role_
Other	—
Primary Income	1.39 (1.07 to 1.82)	0.015
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	0.76 (0.52 to 1.15)	0.18
1 bedroom	0.84 (0.69 to 1.03)	0.089
3 bedroom	1.05 (0.86 to 1.29)	0.63
4 bedroom	1.29 (0.92 to 1.82)	0.14
5+ bedroom	0.88 (0.54 to 1.50)	0.64
1 OR = Odds Ratio, CI = Confidence Interval

Offers to nonstandard

Outcome: how often offer tenancy to low-resource applicants who do not meet standard requirements

• 1: always or often
• 0: sometimes or rarely

Crosstabs:

df %>% 
  tbl_summary(
    by = Offers_to_nonstandard,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Offers to nonstandard**")

Characteristic	Offers to nonstandard		p-value1
	0	1
Landlord_size_, n (%)			0.009
small (single-fam)	1,535 (92%)	136 (8.1%)
small (non single-fam)	1,084 (92%)	100 (8.4%)
big (single-fam)	50 (91%)	5 (9.1%)
big (non single-fam)	723 (88%)	100 (12%)
Financial_role_, n (%)			<0.001
Other	2,998 (92%)	269 (8.2%)
Primary Income	379 (84%)	71 (16%)
Common_unit_, n (%)			<0.001
2 bedroom	1,295 (92%)	113 (8.0%)
0 bedroom (studio/efficiency)	117 (80%)	30 (20%)
1 bedroom	786 (90%)	84 (9.7%)
3 bedroom	749 (92%)	66 (8.1%)
4 bedroom	200 (91%)	19 (8.7%)
5+ bedroom	61 (82%)	13 (18%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Offers_to_nonstandard~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Offers_to_nonstandard~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.04 (0.79 to 1.36)	0.77
big (single-fam)	1.13 (0.39 to 2.62)	0.80
big (non single-fam)	1.56 (1.19 to 2.05)	0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.99 (0.72 to 1.36)	0.95
big (single-fam)	1.02 (0.35 to 2.40)	0.97
big (non single-fam)	1.21 (0.85 to 1.70)	0.28
Financial_role_
Other	—
Primary Income	1.83 (1.32 to 2.49)	<0.001
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	2.54 (1.58 to 3.99)	<0.001
1 bedroom	1.10 (0.81 to 1.50)	0.53
3 bedroom	1.09 (0.77 to 1.52)	0.62
4 bedroom	1.19 (0.69 to 1.98)	0.51
5+ bedroom	2.46 (1.25 to 4.55)	0.006
1 OR = Odds Ratio, CI = Confidence Interval

Rent to vouchers

Outomce: ever rented to tenant using vouchers

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Rent_to_vouchers,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Rent to vouchers**")

Characteristic	Rent to vouchers		p-value1
	no, dk, missing	yes
Landlord_size_, n (%)			<0.001
small (single-fam)	1,589 (92%)	138 (8.0%)
small (non single-fam)	1,061 (87%)	158 (13%)
big (single-fam)	39 (70%)	17 (30%)
big (non single-fam)	404 (48%)	430 (52%)
Financial_role_, n (%)			<0.001
Other	2,823 (84%)	538 (16%)
Primary Income	259 (57%)	198 (43%)
Common_unit_, n (%)			<0.001
2 bedroom	1,165 (81%)	277 (19%)
0 bedroom (studio/efficiency)	98 (66%)	50 (34%)
1 bedroom	629 (71%)	259 (29%)
3 bedroom	746 (88%)	100 (12%)
4 bedroom	197 (88%)	26 (12%)
5+ bedroom	64 (85%)	11 (15%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Rent_to_vouchers~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Rent_to_vouchers~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.71 (1.35 to 2.18)	<0.001
big (single-fam)	5.02 (2.70 to 8.96)	<0.001
big (non single-fam)	12.3 (9.85 to 15.3)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.93 (1.47 to 2.54)	<0.001
big (single-fam)	4.76 (2.53 to 8.60)	<0.001
big (non single-fam)	11.3 (8.64 to 14.8)	<0.001
Financial_role_
Other	—
Primary Income	1.64 (1.28 to 2.09)	<0.001
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.09 (0.72 to 1.63)	0.69
1 bedroom	1.12 (0.90 to 1.40)	0.32
3 bedroom	1.15 (0.86 to 1.54)	0.34
4 bedroom	1.25 (0.77 to 1.97)	0.36
5+ bedroom	1.35 (0.63 to 2.66)	0.41
1 OR = Odds Ratio, CI = Confidence Interval

Rent to criminal

Outcome: ever rented to tenant you knew had criminal record

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Rent_to_criminal,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Rents to criminal hx**")

Characteristic	Rents to criminal hx		p-value1
	no, dk, missing	yes
Landlord_size_, n (%)			<0.001
small (single-fam)	1,561 (90%)	166 (9.6%)
small (non single-fam)	1,089 (89%)	130 (11%)
big (single-fam)	43 (77%)	13 (23%)
big (non single-fam)	527 (63%)	307 (37%)
Financial_role_, n (%)			<0.001
Other	2,923 (87%)	438 (13%)
Primary Income	286 (63%)	171 (37%)
Common_unit_, n (%)			<0.001
2 bedroom	1,226 (85%)	216 (15%)
0 bedroom (studio/efficiency)	103 (70%)	45 (30%)
1 bedroom	699 (79%)	189 (21%)
3 bedroom	735 (87%)	111 (13%)
4 bedroom	198 (89%)	25 (11%)
5+ bedroom	63 (84%)	12 (16%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Rent_to_criminal~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Rent_to_criminal~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.12 (0.88 to 1.43)	0.35
big (single-fam)	2.84 (1.44 to 5.25)	0.001
big (non single-fam)	5.48 (4.43 to 6.79)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.30 (0.98 to 1.71)	0.067
big (single-fam)	2.51 (1.25 to 4.72)	0.006
big (non single-fam)	4.93 (3.77 to 6.47)	<0.001
Financial_role_
Other	—
Primary Income	2.15 (1.69 to 2.73)	<0.001
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.47 (0.97 to 2.22)	0.066
1 bedroom	1.04 (0.82 to 1.31)	0.75
3 bedroom	1.41 (1.07 to 1.87)	0.015
4 bedroom	1.24 (0.76 to 1.95)	0.37
5+ bedroom	1.57 (0.77 to 2.97)	0.19
1 OR = Odds Ratio, CI = Confidence Interval

Disability accomodations

Outcome: ever had tenant request disability accomodations

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Disability_accom.,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Disability accommodations**")

Characteristic	Disability accommodations		p-value1
	no, dk, missing	yes
Landlord_size_, n (%)			<0.001
small (single-fam)	1,668 (97%)	59 (3.4%)
small (non single-fam)	1,134 (93%)	85 (7.0%)
big (single-fam)	50 (89%)	6 (11%)
big (non single-fam)	604 (72%)	230 (28%)
Financial_role_, n (%)			<0.001
Other	3,098 (92%)	263 (7.8%)
Primary Income	343 (75%)	114 (25%)
Common_unit_, n (%)			<0.001
2 bedroom	1,307 (91%)	135 (9.4%)
0 bedroom (studio/efficiency)	121 (82%)	27 (18%)
1 bedroom	756 (85%)	132 (15%)
3 bedroom	791 (93%)	55 (6.5%)
4 bedroom	211 (95%)	12 (5.4%)
5+ bedroom	74 (99%)	1 (1.3%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Disability_accom.~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Disability_accom.~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	2.12 (1.51 to 2.99)	<0.001
big (single-fam)	3.39 (1.26 to 7.66)	0.007
big (non single-fam)	10.8 (8.02 to 14.7)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	2.32 (1.57 to 3.43)	<0.001
big (single-fam)	2.96 (1.09 to 6.77)	0.018
big (non single-fam)	10.5 (7.29 to 15.3)	<0.001
Financial_role_
Other	—
Primary Income	1.83 (1.38 to 2.41)	<0.001
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.09 (0.66 to 1.76)	0.72
1 bedroom	1.08 (0.82 to 1.42)	0.59
3 bedroom	1.59 (1.09 to 2.31)	0.016
4 bedroom	1.44 (0.72 to 2.68)	0.28
5+ bedroom	0.23 (0.01 to 1.09)	0.15
1 OR = Odds Ratio, CI = Confidence Interval

Fees

Deposit pct. kept

Outcome: what percentage of security/pet deposits did you retain

• 0: 0%
• 25: less than 25%
• 50: about 50%
• 75: more than 75%
• 100: 100%

Crosstabs:

df %>% 
  tbl_summary(
    by = Deposit_pct_kept,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Deposit pct kept**")

Characteristic	Deposit pct kept					p-value1
	0	25	50	75	100
Landlord_size_, n (%)
small (single-fam)	775 (57%)	341 (25%)	60 (4.4%)	38 (2.8%)	153 (11%)
small (non single-fam)	640 (62%)	228 (22%)	48 (4.6%)	26 (2.5%)	92 (8.9%)
big (single-fam)	25 (47%)	14 (26%)	3 (5.7%)	3 (5.7%)	8 (15%)
big (non single-fam)	359 (45%)	278 (35%)	66 (8.3%)	18 (2.3%)	78 (9.8%)
Financial_role_, n (%)						<0.001
Other	1,611 (57%)	699 (25%)	138 (4.9%)	71 (2.5%)	293 (10%)
Primary Income	185 (43%)	157 (37%)	36 (8.4%)	14 (3.3%)	35 (8.2%)
Common_unit_, n (%)
2 bedroom	697 (55%)	337 (27%)	76 (6.0%)	23 (1.8%)	127 (10%)
0 bedroom (studio/efficiency)	79 (61%)	27 (21%)	9 (6.9%)	3 (2.3%)	12 (9.2%)
1 bedroom	440 (56%)	202 (26%)	43 (5.5%)	22 (2.8%)	75 (9.6%)
3 bedroom	349 (51%)	198 (29%)	30 (4.4%)	27 (4.0%)	74 (11%)
4 bedroom	103 (55%)	48 (26%)	10 (5.4%)	4 (2.2%)	21 (11%)
5+ bedroom	35 (54%)	21 (32%)	1 (1.5%)	3 (4.6%)	5 (7.7%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Deposit_pct_kept~1 + Landlord_size_,
          family=gaussian(link="identity"), data=df)

m2 <- glm(Deposit_pct_kept~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=gaussian(link="identity"), data=df)



tbl_regression(m1)

Characteristic	Beta (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	-3.1 (-5.6 to -0.54)	0.018
big (single-fam)	7.1 (-1.6 to 16)	0.11
big (non single-fam)	2.6 (-0.19 to 5.3)	0.068
1 CI = Confidence Interval

tbl_regression(m2)

Characteristic	Beta (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	-2.4 (-5.3 to 0.57)	0.11
big (single-fam)	6.9 (-2.0 to 16)	0.13
big (non single-fam)	3.6 (0.34 to 7.0)	0.031
Financial_role_
Other	—
Primary Income	1.2 (-2.2 to 4.7)	0.48
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	-2.3 (-8.1 to 3.5)	0.43
1 bedroom	-0.86 (-3.8 to 2.0)	0.56
3 bedroom	2.4 (-0.72 to 5.6)	0.13
4 bedroom	1.0 (-4.0 to 6.1)	0.69
5+ bedroom	-1.1 (-9.0 to 6.9)	0.79
1 CI = Confidence Interval

Deposit kept reason

Outcome: why did you retain some or all of deposit

• 0: cover damages
• 1: cover unpaid rent or fees

Crosstabs:

df %>% 
  tbl_summary(
    by = Deposit_kept_reason,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Deposit kept reason**")

Characteristic	Deposit kept reason		p-value1
	0	1
Landlord_size_, n (%)			0.055
small (single-fam)	429 (79%)	111 (21%)
small (non single-fam)	266 (81%)	63 (19%)
big (single-fam)	18 (78%)	5 (22%)
big (non single-fam)	261 (73%)	97 (27%)
Financial_role_, n (%)			0.049
Other	819 (79%)	220 (21%)
Primary Income	148 (73%)	56 (27%)
Common_unit_, n (%)			0.007
2 bedroom	364 (78%)	101 (22%)
0 bedroom (studio/efficiency)	25 (58%)	18 (42%)
1 bedroom	228 (78%)	64 (22%)
3 bedroom	234 (79%)	63 (21%)
4 bedroom	68 (87%)	10 (13%)
5+ bedroom	18 (67%)	9 (33%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Deposit_kept_reason~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Deposit_kept_reason~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.92 (0.65 to 1.29)	0.62
big (single-fam)	1.07 (0.35 to 2.76)	0.89
big (non single-fam)	1.44 (1.05 to 1.96)	0.023
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.84 (0.56 to 1.25)	0.40
big (single-fam)	1.21 (0.39 to 3.19)	0.71
big (non single-fam)	1.27 (0.85 to 1.89)	0.25
Financial_role_
Other	—
Primary Income	1.13 (0.77 to 1.65)	0.53
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	2.18 (1.10 to 4.25)	0.022
1 bedroom	0.93 (0.64 to 1.35)	0.72
3 bedroom	1.01 (0.68 to 1.48)	0.98
4 bedroom	0.54 (0.25 to 1.08)	0.10
5+ bedroom	1.84 (0.76 to 4.18)	0.16
1 OR = Odds Ratio, CI = Confidence Interval

Allow installments

Outcome: prior to 2017, allow fees or deposits in installments

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Allow_installments,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Allow installments**")

Characteristic	Allow installments		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	972 (60%)	661 (40%)
small (non single-fam)	640 (56%)	508 (44%)
big (single-fam)	21 (40%)	32 (60%)
big (non single-fam)	302 (37%)	510 (63%)
Financial_role_, n (%)			<0.001
Other	1,742 (55%)	1,446 (45%)
Primary Income	186 (42%)	257 (58%)
Common_unit_, n (%)			0.008
2 bedroom	683 (49%)	700 (51%)
0 bedroom (studio/efficiency)	77 (55%)	63 (45%)
1 bedroom	455 (53%)	396 (47%)
3 bedroom	462 (57%)	342 (43%)
4 bedroom	120 (56%)	94 (44%)
5+ bedroom	33 (47%)	37 (53%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Allow_installments~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Allow_installments~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.17 (1.00 to 1.36)	0.047
big (single-fam)	2.24 (1.29 to 3.97)	0.005
big (non single-fam)	2.48 (2.09 to 2.96)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.31 (1.10 to 1.56)	0.003
big (single-fam)	2.46 (1.39 to 4.46)	0.002
big (non single-fam)	2.81 (2.28 to 3.47)	<0.001
Financial_role_
Other	—
Primary Income	1.15 (0.92 to 1.44)	0.22
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	0.56 (0.39 to 0.81)	0.002
1 bedroom	0.68 (0.56 to 0.81)	<0.001
3 bedroom	0.92 (0.76 to 1.11)	0.37
4 bedroom	0.98 (0.72 to 1.33)	0.90
5+ bedroom	1.40 (0.85 to 2.29)	0.18
1 OR = Odds Ratio, CI = Confidence Interval

Utilities included

Outcome: do you include utilities in rent for most common unit size

• 0: no
• 1: yes (any)

Crosstabs:

df %>% 
  tbl_summary(
    by = Utilities_included,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
 modify_spanning_header(starts_with("stat_") ~ "**Utilities included**")

Characteristic	Utilities included		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	1,375 (81%)	329 (19%)
small (non single-fam)	386 (39%)	610 (61%)
big (single-fam)	46 (85%)	8 (15%)
big (non single-fam)	341 (42%)	475 (58%)
Financial_role_, n (%)			0.024
Other	1,900 (61%)	1,219 (39%)
Primary Income	241 (55%)	195 (45%)
Common_unit_, n (%)			<0.001
2 bedroom	839 (59%)	582 (41%)
0 bedroom (studio/efficiency)	52 (36%)	91 (64%)
1 bedroom	299 (34%)	579 (66%)
3 bedroom	696 (84%)	135 (16%)
4 bedroom	202 (91%)	21 (9.4%)
5+ bedroom	60 (81%)	14 (19%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Utilities_included~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Utilities_included~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	6.60 (5.55 to 7.88)	<0.001
big (single-fam)	0.73 (0.32 to 1.47)	0.41
big (non single-fam)	5.82 (4.85 to 7.00)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	4.37 (3.62 to 5.29)	<0.001
big (single-fam)	0.72 (0.29 to 1.57)	0.44
big (non single-fam)	3.48 (2.80 to 4.33)	<0.001
Financial_role_
Other	—
Primary Income	0.66 (0.52 to 0.84)	<0.001
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.62 (1.13 to 2.36)	0.010
1 bedroom	2.46 (2.05 to 2.97)	<0.001
3 bedroom	0.45 (0.36 to 0.57)	<0.001
4 bedroom	0.27 (0.16 to 0.42)	<0.001
5+ bedroom	0.59 (0.31 to 1.06)	0.094
1 OR = Odds Ratio, CI = Confidence Interval

Security deposit

Outcome: average security deposit fee for most common unit size

• 0: below or equal to the median fee by most common unit size
• 1: above the median fee by most common unit size

Crosstabs:

df %>% 
  tbl_summary(
    by = Security_deposit,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
  modify_spanning_header(starts_with("stat_") ~ "**Security deposit**")

Characteristic	Security deposit		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	994 (63%)	591 (37%)
small (non single-fam)	581 (61%)	378 (39%)
big (single-fam)	39 (75%)	13 (25%)
big (non single-fam)	564 (71%)	234 (29%)
Financial_role_, n (%)			0.016
Other	1,866 (63%)	1,080 (37%)
Primary Income	300 (69%)	133 (31%)
Common_unit_, n (%)			<0.001
2 bedroom	824 (61%)	536 (39%)
0 bedroom (studio/efficiency)	105 (77%)	31 (23%)
1 bedroom	505 (60%)	333 (40%)
3 bedroom	585 (75%)	193 (25%)
4 bedroom	118 (56%)	93 (44%)
5+ bedroom	41 (58%)	30 (42%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Security_deposit~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Security_deposit~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.09 (0.93 to 1.29)	0.28
big (single-fam)	0.56 (0.29 to 1.03)	0.074
big (non single-fam)	0.70 (0.58 to 0.84)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.90 (0.75 to 1.09)	0.28
big (single-fam)	0.54 (0.27 to 1.02)	0.070
big (non single-fam)	0.55 (0.44 to 0.69)	<0.001
Financial_role_
Other	—
Primary Income	0.91 (0.72 to 1.15)	0.45
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	0.54 (0.35 to 0.82)	0.005
1 bedroom	1.14 (0.95 to 1.37)	0.16
3 bedroom	0.45 (0.36 to 0.55)	<0.001
4 bedroom	1.06 (0.78 to 1.44)	0.70
5+ bedroom	1.01 (0.61 to 1.64)	0.97
1 OR = Odds Ratio, CI = Confidence Interval

App fee

Outcome: average application fee for most common unit size

• 0: below or equal to the median fee by most common unit size
• 1: above the median fee by most common unit size

Crosstabs:

df %>% 
  tbl_summary(
    by = App_fee,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**App fee**")

Characteristic	App fee		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	733 (67%)	362 (33%)
small (non single-fam)	429 (63%)	254 (37%)
big (single-fam)	17 (44%)	22 (56%)
big (non single-fam)	311 (48%)	334 (52%)
Financial_role_, n (%)			<0.001
Other	1,322 (63%)	790 (37%)
Primary Income	162 (48%)	176 (52%)
Common_unit_, n (%)			0.023
2 bedroom	620 (62%)	380 (38%)
0 bedroom (studio/efficiency)	54 (52%)	49 (48%)
1 bedroom	350 (57%)	268 (43%)
3 bedroom	347 (63%)	206 (37%)
4 bedroom	96 (67%)	47 (33%)
5+ bedroom	23 (51%)	22 (49%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(App_fee~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(App_fee~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.20 (0.98 to 1.46)	0.075
big (single-fam)	2.62 (1.38 to 5.06)	0.003
big (non single-fam)	2.17 (1.78 to 2.66)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.27 (1.01 to 1.58)	0.038
big (single-fam)	2.29 (1.18 to 4.50)	0.014
big (non single-fam)	2.10 (1.65 to 2.68)	<0.001
Financial_role_
Other	—
Primary Income	1.34 (1.04 to 1.72)	0.022
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.13 (0.74 to 1.72)	0.58
1 bedroom	1.05 (0.85 to 1.29)	0.67
3 bedroom	1.21 (0.96 to 1.53)	0.11
4 bedroom	1.02 (0.69 to 1.50)	0.93
5+ bedroom	1.90 (1.02 to 3.51)	0.040
1 OR = Odds Ratio, CI = Confidence Interval

Cleaning fee

Outcome: average cleaning fee for most common unit size

• 0: below or equal to the median fee by most common unit size
• 1: above the median fee by most common unit size

Crosstabs:

df %>% 
  tbl_summary(
    by = Cleaning_fee,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Cleaning fee**")

Characteristic	Cleaning fee		p-value1
	0	1
Landlord_size_, n (%)			0.010
small (single-fam)	592 (72%)	234 (28%)
small (non single-fam)	378 (79%)	99 (21%)
big (single-fam)	19 (76%)	6 (24%)
big (non single-fam)	308 (78%)	87 (22%)
Financial_role_, n (%)			0.61
Other	1,139 (75%)	378 (25%)
Primary Income	152 (77%)	46 (23%)
Common_unit_, n (%)			<0.001
2 bedroom	539 (78%)	155 (22%)
0 bedroom (studio/efficiency)	26 (50%)	26 (50%)
1 bedroom	353 (85%)	60 (15%)
3 bedroom	281 (70%)	123 (30%)
4 bedroom	66 (58%)	47 (42%)
5+ bedroom	32 (68%)	15 (32%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Cleaning_fee~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Cleaning_fee~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.66 (0.51 to 0.86)	0.003
big (single-fam)	0.80 (0.29 to 1.92)	0.64
big (non single-fam)	0.71 (0.54 to 0.94)	0.020
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.80 (0.59 to 1.08)	0.14
big (single-fam)	0.67 (0.24 to 1.66)	0.41
big (non single-fam)	0.93 (0.66 to 1.30)	0.67
Financial_role_
Other	—
Primary Income	1.02 (0.69 to 1.48)	0.94
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	3.67 (2.03 to 6.61)	<0.001
1 bedroom	0.61 (0.43 to 0.84)	0.003
3 bedroom	1.46 (1.09 to 1.97)	0.012
4 bedroom	2.39 (1.54 to 3.69)	<0.001
5+ bedroom	1.53 (0.78 to 2.88)	0.20
1 OR = Odds Ratio, CI = Confidence Interval

Parking fee

Outcome: average parking fee for most common unit size

• 0: below or equal to the median fee by most common unit size
• 1: above the median fee by most common unit size

Crosstabs:

df %>% 
  tbl_summary(
    by = Parking_fee,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Parking fee**")

Characteristic	Parking fee		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	193 (96%)	9 (4.5%)
small (non single-fam)	118 (85%)	21 (15%)
big (single-fam)	8 (89%)	1 (11%)
big (non single-fam)	150 (74%)	52 (26%)
Financial_role_, n (%)			<0.001
Other	396 (89%)	50 (11%)
Primary Income	69 (68%)	33 (32%)
Common_unit_, n (%)
2 bedroom	190 (91%)	18 (8.7%)
0 bedroom (studio/efficiency)	20 (74%)	7 (26%)
1 bedroom	136 (71%)	56 (29%)
3 bedroom	87 (99%)	1 (1.1%)
4 bedroom	27 (96%)	1 (3.6%)
5+ bedroom	9 (100%)	0 (0%)
1 Fisher's exact test; Pearson's Chi-squared test

Models:

m1 <- glm(Parking_fee~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Parking_fee~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	3.82 (1.74 to 9.03)	0.001
big (single-fam)	2.68 (0.14 to 17.1)	0.38
big (non single-fam)	7.43 (3.72 to 16.6)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.54 (0.65 to 3.91)	0.34
big (single-fam)	1.91 (0.09 to 15.5)	0.59
big (non single-fam)	1.91 (0.82 to 4.79)	0.15
Financial_role_
Other	—
Primary Income	2.12 (1.19 to 3.75)	0.010
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	2.56 (0.88 to 6.97)	0.072
1 bedroom	3.45 (1.92 to 6.45)	<0.001
3 bedroom	0.17 (0.01 to 0.94)	0.10
4 bedroom	0.54 (0.03 to 3.20)	0.58
5+ bedroom	0.00 (NA to 1,518,902,281,235,844,825,128)	0.99
1 OR = Odds Ratio, CI = Confidence Interval

Rent

Fair market price

Outcome: average monthly rent for most common unit size above or below fair market price

• 0: below or equal to fair market price
• 1: above fair market price
• note: here are 2018 king county fair market prices
  • 1br=$1,108
  • 2br=$1,243
  • 3br=$1,527
    
  • 4br=$2,211
  • 5br=$2,617

Crosstabs:

df %>% 
  tbl_summary(
    by = Fair_market_price,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Fair market price**")

Characteristic	Fair market price		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	1,132 (66%)	595 (34%)
small (non single-fam)	697 (57%)	522 (43%)
big (single-fam)	36 (64%)	20 (36%)
big (non single-fam)	671 (80%)	163 (20%)
Financial_role_, n (%)			0.073
Other	2,204 (66%)	1,157 (34%)
Primary Income	319 (70%)	138 (30%)
Common_unit_, n (%)			<0.001
2 bedroom	1,252 (87%)	190 (13%)
0 bedroom (studio/efficiency)	60 (41%)	88 (59%)
1 bedroom	560 (63%)	328 (37%)
3 bedroom	488 (58%)	358 (42%)
4 bedroom	176 (79%)	47 (21%)
5+ bedroom	0 (0%)	75 (100%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Fair_market_price~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Fair_market_price~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.42 (1.23 to 1.66)	<0.001
big (single-fam)	1.06 (0.60 to 1.82)	0.85
big (non single-fam)	0.46 (0.38 to 0.56)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.83 (0.67 to 1.01)	0.065
big (single-fam)	1.02 (0.55 to 1.86)	0.94
big (non single-fam)	0.29 (0.22 to 0.38)	<0.001
Financial_role_
Other	—
Primary Income	1.19 (0.91 to 1.55)	0.19
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	14.8 (10.1 to 22.0)	<0.001
1 bedroom	4.96 (4.00 to 6.18)	<0.001
3 bedroom	3.96 (3.20 to 4.92)	<0.001
4 bedroom	1.41 (0.97 to 2.01)	0.068
5+ bedroom	91,212,350 (4,852 to 2,791,366,407,498,063,121,664,642,028,482,684,482,006,442,840,846,224)	0.95
1 OR = Odds Ratio, CI = Confidence Interval

Last month required

Outcome: require last month’s rent before move in for most common unit size

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Last_month_required,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Last month required**")

Characteristic	Last month required		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	932 (55%)	773 (45%)
small (non single-fam)	587 (59%)	402 (41%)
big (single-fam)	38 (69%)	17 (31%)
big (non single-fam)	655 (80%)	164 (20%)
Financial_role_, n (%)			<0.001
Other	1,866 (60%)	1,240 (40%)
Primary Income	334 (75%)	112 (25%)
Common_unit_, n (%)			<0.001
2 bedroom	872 (61%)	549 (39%)
0 bedroom (studio/efficiency)	95 (68%)	45 (32%)
1 bedroom	622 (71%)	254 (29%)
3 bedroom	467 (56%)	368 (44%)
4 bedroom	119 (54%)	103 (46%)
5+ bedroom	37 (50%)	37 (50%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Last_month_required~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Last_month_required~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.83 (0.70 to 0.97)	0.018
big (single-fam)	0.54 (0.29 to 0.95)	0.037
big (non single-fam)	0.30 (0.25 to 0.37)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.89 (0.75 to 1.06)	0.19
big (single-fam)	0.54 (0.29 to 0.97)	0.047
big (non single-fam)	0.36 (0.29 to 0.45)	<0.001
Financial_role_
Other	—
Primary Income	0.79 (0.62 to 1.00)	0.056
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.04 (0.70 to 1.53)	0.84
1 bedroom	0.77 (0.64 to 0.93)	0.007
3 bedroom	1.04 (0.86 to 1.25)	0.71
4 bedroom	1.12 (0.83 to 1.50)	0.46
5+ bedroom	1.36 (0.84 to 2.20)	0.20
1 OR = Odds Ratio, CI = Confidence Interval

Raised rent

Outcome: raised rent in past year

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Raised_rent,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Raised rent**")

Characteristic	Raised rent		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	779 (45%)	934 (55%)
small (non single-fam)	457 (38%)	732 (62%)
big (single-fam)	11 (20%)	43 (80%)
big (non single-fam)	111 (13%)	712 (87%)
Financial_role_, n (%)			<0.001
Other	1,268 (38%)	2,042 (62%)
Primary Income	82 (18%)	369 (82%)
Common_unit_, n (%)			<0.001
2 bedroom	496 (35%)	932 (65%)
0 bedroom (studio/efficiency)	35 (24%)	111 (76%)
1 bedroom	271 (31%)	608 (69%)
3 bedroom	343 (41%)	495 (59%)
4 bedroom	97 (44%)	125 (56%)
5+ bedroom	22 (29%)	53 (71%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Raised_rent~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Raised_rent~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.34 (1.15 to 1.55)	<0.001
big (single-fam)	3.26 (1.73 to 6.69)	<0.001
big (non single-fam)	5.35 (4.30 to 6.70)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.54 (1.29 to 1.84)	<0.001
big (single-fam)	2.86 (1.50 to 5.90)	0.002
big (non single-fam)	5.28 (4.12 to 6.82)	<0.001
Financial_role_
Other	—
Primary Income	1.61 (1.23 to 2.13)	<0.001
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	0.98 (0.65 to 1.52)	0.94
1 bedroom	0.87 (0.72 to 1.06)	0.16
3 bedroom	1.09 (0.90 to 1.31)	0.39
4 bedroom	0.99 (0.73 to 1.34)	0.94
5+ bedroom	1.75 (1.05 to 3.01)	0.035
1 OR = Odds Ratio, CI = Confidence Interval

Raised rent pct.

Outcome: raised rent by how much percent

• 0: have not raised rent
• 3: 1-5%
• 7: 6-10%
• 13: 11-15%
• 20: 16-25%
• 30: more than 25%

Crosstabs:

df %>% 
  tbl_summary(
    by = Raised_rent_pct,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Raised rent pct**")

Characteristic	Raised rent pct						p-value1
	0	3	7	13	20	30
Landlord_size_, n (%)
small (single-fam)	779 (45%)	503 (29%)	281 (16%)	61 (3.6%)	56 (3.3%)	33 (1.9%)
small (non single-fam)	457 (38%)	368 (31%)	259 (22%)	47 (4.0%)	37 (3.1%)	21 (1.8%)
big (single-fam)	11 (20%)	22 (41%)	16 (30%)	3 (5.6%)	1 (1.9%)	1 (1.9%)
big (non single-fam)	111 (13%)	316 (38%)	301 (37%)	64 (7.8%)	21 (2.6%)	10 (1.2%)
Financial_role_, n (%)							<0.001
Other	1,268 (38%)	1,035 (31%)	704 (21%)	149 (4.5%)	97 (2.9%)	57 (1.7%)
Primary Income	82 (18%)	167 (37%)	151 (33%)	25 (5.5%)	18 (4.0%)	8 (1.8%)
Common_unit_, n (%)
2 bedroom	496 (35%)	469 (33%)	325 (23%)	61 (4.3%)	53 (3.7%)	24 (1.7%)
0 bedroom (studio/efficiency)	35 (24%)	46 (32%)	55 (38%)	9 (6.2%)	1 (0.7%)	0 (0%)
1 bedroom	271 (31%)	288 (33%)	235 (27%)	46 (5.2%)	24 (2.7%)	15 (1.7%)
3 bedroom	343 (41%)	252 (30%)	168 (20%)	36 (4.3%)	23 (2.7%)	16 (1.9%)
4 bedroom	97 (44%)	72 (32%)	34 (15%)	7 (3.2%)	7 (3.2%)	5 (2.3%)
5+ bedroom	22 (29%)	32 (43%)	12 (16%)	5 (6.7%)	3 (4.0%)	1 (1.3%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Raised_rent_pct~1 + Landlord_size_,
          family=gaussian(link="identity"), data=df)

m2 <- glm(Raised_rent_pct~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=gaussian(link="identity"), data=df)



tbl_regression(m1)

Characteristic	Beta (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.40 (-0.01 to 0.81)	0.059
big (single-fam)	1.2 (-0.28 to 2.7)	0.11
big (non single-fam)	1.9 (1.4 to 2.3)	<0.001
1 CI = Confidence Interval

tbl_regression(m2)

Characteristic	Beta (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.58 (0.10 to 1.1)	0.017
big (single-fam)	1.2 (-0.38 to 2.7)	0.14
big (non single-fam)	1.8 (1.3 to 2.4)	<0.001
Financial_role_
Other	—
Primary Income	0.64 (0.05 to 1.2)	0.033
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	-0.55 (-1.5 to 0.42)	0.27
1 bedroom	-0.17 (-0.65 to 0.30)	0.47
3 bedroom	0.10 (-0.40 to 0.61)	0.69
4 bedroom	-0.10 (-0.91 to 0.71)	0.81
5+ bedroom	0.49 (-0.81 to 1.8)	0.46
1 CI = Confidence Interval

Raised rent - rising costs

Outcome: raised rent due to rising costs of taxes, repairs (not exclusive)

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Raised_rent_rising_costs,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Raised rent rising costs**")

Characteristic	Raised rent rising costs		p-value1
	0	1
Landlord_size_, n (%)			0.47
small (single-fam)	160 (17%)	774 (83%)
small (non single-fam)	142 (19%)	590 (81%)
big (single-fam)	6 (14%)	37 (86%)
big (non single-fam)	119 (17%)	593 (83%)
Financial_role_, n (%)			0.45
Other	365 (18%)	1,677 (82%)
Primary Income	60 (16%)	309 (84%)
Common_unit_, n (%)			0.19
2 bedroom	145 (16%)	787 (84%)
0 bedroom (studio/efficiency)	25 (23%)	86 (77%)
1 bedroom	115 (19%)	493 (81%)
3 bedroom	93 (19%)	402 (81%)
4 bedroom	22 (18%)	103 (82%)
5+ bedroom	13 (25%)	40 (75%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Raised_rent_rising_costs~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Raised_rent_rising_costs~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.86 (0.67 to 1.10)	0.23
big (single-fam)	1.27 (0.57 to 3.40)	0.59
big (non single-fam)	1.03 (0.79 to 1.34)	0.82
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.78 (0.59 to 1.04)	0.094
big (single-fam)	1.46 (0.61 to 4.32)	0.44
big (non single-fam)	1.00 (0.73 to 1.38)	>0.99
Financial_role_
Other	—
Primary Income	1.10 (0.80 to 1.53)	0.57
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	0.64 (0.40 to 1.06)	0.073
1 bedroom	0.79 (0.60 to 1.04)	0.088
3 bedroom	0.77 (0.56 to 1.04)	0.090
4 bedroom	0.81 (0.49 to 1.38)	0.42
5+ bedroom	0.53 (0.28 to 1.07)	0.064
1 OR = Odds Ratio, CI = Confidence Interval

Raised rent - regulations

Outcome: raised rent due to regulations (not exclusive)

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Raised_rent_regulations,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Raised rent regulations**")

Characteristic	Raised rent regulations		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	762 (82%)	172 (18%)
small (non single-fam)	601 (82%)	131 (18%)
big (single-fam)	26 (60%)	17 (40%)
big (non single-fam)	503 (71%)	209 (29%)
Financial_role_, n (%)			<0.001
Other	1,619 (79%)	423 (21%)
Primary Income	264 (72%)	105 (28%)
Common_unit_, n (%)			0.13
2 bedroom	726 (78%)	206 (22%)
0 bedroom (studio/efficiency)	84 (76%)	27 (24%)
1 bedroom	458 (75%)	150 (25%)
3 bedroom	407 (82%)	88 (18%)
4 bedroom	95 (76%)	30 (24%)
5+ bedroom	40 (75%)	13 (25%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Raised_rent_regulations~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Raised_rent_regulations~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.97 (0.75 to 1.24)	0.79
big (single-fam)	2.90 (1.51 to 5.42)	<0.001
big (non single-fam)	1.84 (1.46 to 2.32)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.97 (0.73 to 1.28)	0.82
big (single-fam)	2.76 (1.41 to 5.26)	0.002
big (non single-fam)	1.75 (1.32 to 2.33)	<0.001
Financial_role_
Other	—
Primary Income	1.13 (0.86 to 1.49)	0.37
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	0.94 (0.58 to 1.50)	0.81
1 bedroom	1.01 (0.79 to 1.30)	0.92
3 bedroom	0.85 (0.63 to 1.14)	0.29
4 bedroom	1.22 (0.76 to 1.92)	0.39
5+ bedroom	1.28 (0.64 to 2.42)	0.46
1 OR = Odds Ratio, CI = Confidence Interval

Raised rent - market

Outcome: raised rent due to housing market (not exclusive)

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Raised_rent_market,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Raised rent market**")

Characteristic	Raised rent market		p-value1
	0	1
Landlord_size_, n (%)			0.47
small (single-fam)	641 (69%)	293 (31%)
small (non single-fam)	524 (72%)	208 (28%)
big (single-fam)	28 (65%)	15 (35%)
big (non single-fam)	488 (69%)	224 (31%)
Financial_role_, n (%)			0.28
Other	1,425 (70%)	617 (30%)
Primary Income	247 (67%)	122 (33%)
Common_unit_, n (%)			0.88
2 bedroom	645 (69%)	287 (31%)
0 bedroom (studio/efficiency)	71 (64%)	40 (36%)
1 bedroom	421 (69%)	187 (31%)
3 bedroom	348 (70%)	147 (30%)
4 bedroom	85 (68%)	40 (32%)
5+ bedroom	37 (70%)	16 (30%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Raised_rent_market~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Raised_rent_market~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.87 (0.70 to 1.07)	0.19
big (single-fam)	1.17 (0.60 to 2.20)	0.63
big (non single-fam)	1.00 (0.81 to 1.24)	0.97
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.85 (0.67 to 1.09)	0.20
big (single-fam)	1.23 (0.63 to 2.34)	0.53
big (non single-fam)	0.91 (0.70 to 1.17)	0.46
Financial_role_
Other	—
Primary Income	1.12 (0.86 to 1.44)	0.40
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.29 (0.84 to 1.96)	0.24
1 bedroom	1.00 (0.80 to 1.26)	0.98
3 bedroom	0.91 (0.70 to 1.17)	0.45
4 bedroom	0.99 (0.65 to 1.49)	0.96
5+ bedroom	0.91 (0.48 to 1.65)	0.77
1 OR = Odds Ratio, CI = Confidence Interval

Raised rent - only taxes

Outcome: raised rent ONLY due to increased property taxes

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Raised_rent_only_tax,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Raised rent only due to taxes**")

Characteristic	Raised rent only due to taxes		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	628 (67%)	306 (33%)
small (non single-fam)	525 (72%)	207 (28%)
big (single-fam)	39 (91%)	4 (9.3%)
big (non single-fam)	575 (81%)	137 (19%)
Financial_role_, n (%)			0.18
Other	1,481 (73%)	561 (27%)
Primary Income	280 (76%)	89 (24%)
Common_unit_, n (%)			0.044
2 bedroom	678 (73%)	254 (27%)
0 bedroom (studio/efficiency)	84 (76%)	27 (24%)
1 bedroom	463 (76%)	145 (24%)
3 bedroom	337 (68%)	158 (32%)
4 bedroom	90 (72%)	35 (28%)
5+ bedroom	43 (81%)	10 (19%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Raised_rent_only_tax~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Raised_rent_only_tax~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.81 (0.65 to 1.00)	0.049
big (single-fam)	0.21 (0.06 to 0.53)	0.003
big (non single-fam)	0.49 (0.39 to 0.61)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.80 (0.63 to 1.02)	0.077
big (single-fam)	0.21 (0.06 to 0.54)	0.004
big (non single-fam)	0.46 (0.35 to 0.61)	<0.001
Financial_role_
Other	—
Primary Income	1.18 (0.88 to 1.57)	0.26
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.06 (0.65 to 1.67)	0.81
1 bedroom	0.96 (0.75 to 1.23)	0.77
3 bedroom	1.04 (0.80 to 1.34)	0.77
4 bedroom	0.88 (0.57 to 1.35)	0.58
5+ bedroom	0.50 (0.23 to 0.98)	0.057
1 OR = Odds Ratio, CI = Confidence Interval

Raised rent - only repairs

Outcome: raised rent ONLY due to increased cost of repairs

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Raised_rent_only_repairs,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Raised rent only due to repairs**")

Characteristic	Raised rent only due to repairs		p-value1
	0	1
Landlord_size_, n (%)			0.22
small (single-fam)	914 (98%)	20 (2.1%)
small (non single-fam)	718 (98%)	14 (1.9%)
big (single-fam)	42 (98%)	1 (2.3%)
big (non single-fam)	705 (99%)	7 (1.0%)
Financial_role_, n (%)			0.14
Other	2,003 (98%)	39 (1.9%)
Primary Income	366 (99%)	3 (0.8%)
Common_unit_, n (%)			0.57
2 bedroom	917 (98%)	15 (1.6%)
0 bedroom (studio/efficiency)	111 (100%)	0 (0%)
1 bedroom	594 (98%)	14 (2.3%)
3 bedroom	486 (98%)	9 (1.8%)
4 bedroom	122 (98%)	3 (2.4%)
5+ bedroom	53 (100%)	0 (0%)
1 Fisher's exact test; Pearson's Chi-squared test

Models:

m1 <- glm(Raised_rent_only_repairs~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Raised_rent_only_repairs~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.89 (0.44 to 1.76)	0.74
big (single-fam)	1.09 (0.06 to 5.42)	0.94
big (non single-fam)	0.45 (0.18 to 1.03)	0.074
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.85 (0.39 to 1.82)	0.68
big (single-fam)	1.16 (0.06 to 5.92)	0.88
big (non single-fam)	0.44 (0.15 to 1.13)	0.10
Financial_role_
Other	—
Primary Income	0.55 (0.13 to 1.65)	0.35
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	0.00 (0.00 to 77,847,770,599,214)	0.99
1 bedroom	1.79 (0.83 to 3.84)	0.13
3 bedroom	0.92 (0.37 to 2.17)	0.85
4 bedroom	1.18 (0.27 to 3.81)	0.80
5+ bedroom	0.00 (0.00 to 116,787,813,358,625,468,976,866)	>0.99
1 OR = Odds Ratio, CI = Confidence Interval

Raised rent - only new

Outcome: raised rent ONLY due to new propoerty or new ownership

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Raised_rent_only_newproperty,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Raised rent only due to new property**")

Characteristic	Raised rent only due to new property		p-value1
	0	1
Landlord_size_, n (%)			0.29
small (single-fam)	932 (100%)	2 (0.2%)
small (non single-fam)	727 (99%)	5 (0.7%)
big (single-fam)	43 (100%)	0 (0%)
big (non single-fam)	706 (99%)	6 (0.8%)
Financial_role_, n (%)			>0.99
Other	2,031 (99%)	11 (0.5%)
Primary Income	367 (99%)	2 (0.5%)
Common_unit_, n (%)			0.25
2 bedroom	928 (100%)	4 (0.4%)
0 bedroom (studio/efficiency)	109 (98%)	2 (1.8%)
1 bedroom	603 (99%)	5 (0.8%)
3 bedroom	494 (100%)	1 (0.2%)
4 bedroom	124 (99%)	1 (0.8%)
5+ bedroom	53 (100%)	0 (0%)
1 Fisher's exact test

Models:

This model doesn’t have enough cases with the outcome=1, hence the messy results.

m1 <- glm(Raised_rent_only_newproperty~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Raised_rent_only_newproperty~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	3.20 (0.69 to 22.4)	0.16
big (single-fam)	0.00 (NA to 705,970,280,972,727,745,122,408,442)	>0.99
big (non single-fam)	3.96 (0.91 to 27.1)	0.092
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	3.20 (0.57 to 25.8)	0.21
big (single-fam)	0.00 (NA to 128,447,357,876,125,552,880,428,264,644,846,228,488,004,048,064)	>0.99
big (non single-fam)	3.56 (0.60 to 30.6)	0.19
Financial_role_
Other	—
Primary Income	0.64 (0.09 to 2.70)	0.58
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	3.29 (0.44 to 17.7)	0.18
1 bedroom	1.65 (0.43 to 6.84)	0.46
3 bedroom	0.78 (0.04 to 5.96)	0.84
4 bedroom	3.41 (0.16 to 27.4)	0.30
5+ bedroom	0.00 (NA to 115,759,794,265,682,191,952,884,604,684,088,644,426,466)	>0.99
1 OR = Odds Ratio, CI = Confidence Interval

Raised rent - only market

Outcome: raised rent ONLY due to changes in local housing market

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Raised_rent_only_market,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Raised rent only due to market**")

Characteristic	Raised rent only due to market		p-value1
	0	1
Landlord_size_, n (%)			0.30
small (single-fam)	847 (91%)	87 (9.3%)
small (non single-fam)	672 (92%)	60 (8.2%)
big (single-fam)	42 (98%)	1 (2.3%)
big (non single-fam)	659 (93%)	53 (7.4%)
Financial_role_, n (%)			0.46
Other	1,869 (92%)	173 (8.5%)
Primary Income	342 (93%)	27 (7.3%)
Common_unit_, n (%)			0.92
2 bedroom	859 (92%)	73 (7.8%)
0 bedroom (studio/efficiency)	101 (91%)	10 (9.0%)
1 bedroom	555 (91%)	53 (8.7%)
3 bedroom	449 (91%)	46 (9.3%)
4 bedroom	116 (93%)	9 (7.2%)
5+ bedroom	48 (91%)	5 (9.4%)
1 Fisher's exact test; Pearson's Chi-squared test

Models:

m1 <- glm(Raised_rent_only_market~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Raised_rent_only_market~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.87 (0.61 to 1.22)	0.42
big (single-fam)	0.23 (0.01 to 1.08)	0.15
big (non single-fam)	0.78 (0.55 to 1.11)	0.18
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.90 (0.61 to 1.32)	0.59
big (single-fam)	0.25 (0.01 to 1.17)	0.17
big (non single-fam)	0.75 (0.48 to 1.16)	0.21
Financial_role_
Other	—
Primary Income	0.93 (0.58 to 1.45)	0.75
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.28 (0.60 to 2.50)	0.50
1 bedroom	1.18 (0.80 to 1.73)	0.40
3 bedroom	1.13 (0.74 to 1.70)	0.57
4 bedroom	0.86 (0.39 to 1.72)	0.69
5+ bedroom	1.12 (0.38 to 2.70)	0.81
1 OR = Odds Ratio, CI = Confidence Interval

Raised rent - only regulations

Outcome: raised rent ONLY due to new regulations

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Raised_rent_only_regulation,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Raised rent only due to new regulations**")

Characteristic	Raised rent only due to new regulations		p-value1
	0	1
Landlord_size_, n (%)			0.080
small (single-fam)	920 (99%)	14 (1.5%)
small (non single-fam)	723 (99%)	9 (1.2%)
big (single-fam)	40 (93%)	3 (7.0%)
big (non single-fam)	700 (98%)	12 (1.7%)
Financial_role_, n (%)			0.22
Other	2,008 (98%)	34 (1.7%)
Primary Income	366 (99%)	3 (0.8%)
Common_unit_, n (%)			0.18
2 bedroom	919 (99%)	13 (1.4%)
0 bedroom (studio/efficiency)	109 (98%)	2 (1.8%)
1 bedroom	603 (99%)	5 (0.8%)
3 bedroom	484 (98%)	11 (2.2%)
4 bedroom	121 (97%)	4 (3.2%)
5+ bedroom	52 (98%)	1 (1.9%)
1 Fisher's exact test; Pearson's Chi-squared test

Models:

m1 <- glm(Raised_rent_only_tax~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Raised_rent_only_tax~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.81 (0.65 to 1.00)	0.049
big (single-fam)	0.21 (0.06 to 0.53)	0.003
big (non single-fam)	0.49 (0.39 to 0.61)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.80 (0.63 to 1.02)	0.077
big (single-fam)	0.21 (0.06 to 0.54)	0.004
big (non single-fam)	0.46 (0.35 to 0.61)	<0.001
Financial_role_
Other	—
Primary Income	1.18 (0.88 to 1.57)	0.26
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.06 (0.65 to 1.67)	0.81
1 bedroom	0.96 (0.75 to 1.23)	0.77
3 bedroom	1.04 (0.80 to 1.34)	0.77
4 bedroom	0.88 (0.57 to 1.35)	0.58
5+ bedroom	0.50 (0.23 to 0.98)	0.057
1 OR = Odds Ratio, CI = Confidence Interval

Terminations

Ever terminated

Outcome: in 2017, did you terminate tenancies

• 0: none
• 1: 1 or more

Crosstabs:

df %>% 
  tbl_summary(
    by = Ever_terminated,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Ever terminated**")

Characteristic	Ever terminated		p-value1
	none	1 or more/dk
Landlord_size_, n (%)			<0.001
small (single-fam)	1,601 (93%)	123 (7.1%)
small (non single-fam)	1,104 (92%)	97 (8.1%)
big (single-fam)	40 (73%)	15 (27%)
big (non single-fam)	608 (73%)	221 (27%)
Financial_role_, n (%)			<0.001
Other	2,999 (90%)	337 (10%)
Primary Income	342 (75%)	114 (25%)
Common_unit_, n (%)			<0.001
2 bedroom	1,267 (88%)	173 (12%)
0 bedroom (studio/efficiency)	107 (73%)	40 (27%)
1 bedroom	753 (85%)	130 (15%)
3 bedroom	785 (93%)	58 (6.9%)
4 bedroom	198 (89%)	25 (11%)
5+ bedroom	64 (85%)	11 (15%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Ever_terminated~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Ever_terminated~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.14 (0.87 to 1.51)	0.34
big (single-fam)	4.88 (2.55 to 8.90)	<0.001
big (non single-fam)	4.73 (3.73 to 6.02)	<0.001
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.02 (0.74 to 1.41)	0.88
big (single-fam)	4.67 (2.41 to 8.64)	<0.001
big (non single-fam)	3.79 (2.82 to 5.11)	<0.001
Financial_role_
Other	—
Primary Income	1.58 (1.20 to 2.06)	<0.001
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	1.86 (1.21 to 2.84)	0.004
1 bedroom	0.92 (0.71 to 1.20)	0.54
3 bedroom	0.75 (0.53 to 1.05)	0.094
4 bedroom	1.31 (0.80 to 2.08)	0.27
5+ bedroom	1.66 (0.79 to 3.19)	0.15
1 OR = Odds Ratio, CI = Confidence Interval

Num of terminations

Ouctome: in 2017, how many tenancies did you terminate

• 3: 1-5
• 7: 6-10
• 13: 11-20
• 25: more than 20

Crosstabs:

df %>% 
  tbl_summary(
    by = Num_of_terminations,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Number of terminations**")

Characteristic	Number of terminations				p-value1
	3	7	15	25
Landlord_size_, n (%)					0.59
small (single-fam)	104 (98%)	2 (1.9%)	0 (0%)	0 (0%)
small (non single-fam)	81 (95%)	3 (3.5%)	0 (0%)	1 (1.2%)
big (single-fam)	15 (100%)	0 (0%)	0 (0%)	0 (0%)
big (non single-fam)	189 (92%)	9 (4.4%)	5 (2.4%)	3 (1.5%)
Financial_role_, n (%)					0.002
Other	289 (96%)	7 (2.3%)	5 (1.7%)	0 (0%)
Primary Income	96 (91%)	6 (5.7%)	0 (0%)	4 (3.8%)
Common_unit_, n (%)
2 bedroom	154 (97%)	4 (2.5%)	0 (0%)	0 (0%)
0 bedroom (studio/efficiency)	33 (87%)	2 (5.3%)	3 (7.9%)	0 (0%)
1 bedroom	104 (90%)	6 (5.2%)	2 (1.7%)	3 (2.6%)
3 bedroom	51 (94%)	2 (3.7%)	0 (0%)	1 (1.9%)
4 bedroom	21 (100%)	0 (0%)	0 (0%)	0 (0%)
5+ bedroom	10 (100%)	0 (0%)	0 (0%)	0 (0%)
1 Fisher's exact test

Models:

m1 <- glm(Num_of_terminations~1 + Landlord_size_,
          family=gaussian(link="identity"), data=df)

m2 <- glm(Num_of_terminations~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=gaussian(link="identity"), data=df)



tbl_regression(m1)

Characteristic	Beta (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.32 (-0.41 to 1.1)	0.39
big (single-fam)	-0.08 (-1.5 to 1.3)	0.92
big (non single-fam)	0.71 (0.10 to 1.3)	0.022
1 CI = Confidence Interval

tbl_regression(m2)

Characteristic	Beta (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.51 (-0.36 to 1.4)	0.25
big (single-fam)	-0.08 (-1.5 to 1.4)	0.92
big (non single-fam)	0.51 (-0.28 to 1.3)	0.21
Financial_role_
Other	—
Primary Income	0.49 (-0.16 to 1.1)	0.14
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	0.79 (-0.18 to 1.8)	0.11
1 bedroom	0.74 (0.08 to 1.4)	0.028
3 bedroom	0.69 (-0.21 to 1.6)	0.13
4 bedroom	0.29 (-1.0 to 1.6)	0.66
5+ bedroom	0.12 (-1.6 to 1.9)	0.89
1 CI = Confidence Interval

Terminated - paid late

Outcome: any terminated due to late payment

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Terminated_paid_late,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Terminated - paid late**")

Characteristic	Terminated - paid late		p-value1
	0	1
Landlord_size_, n (%)			0.029
small (single-fam)	5 (19%)	22 (81%)
small (non single-fam)	8 (27%)	22 (73%)
big (single-fam)	1 (14%)	6 (86%)
big (non single-fam)	3 (5.3%)	54 (95%)
Financial_role_, n (%)			0.35
Other	15 (16%)	77 (84%)
Primary Income	2 (7.1%)	26 (93%)
Common_unit_, n (%)			0.70
2 bedroom	6 (13%)	40 (87%)
0 bedroom (studio/efficiency)	0 (0%)	14 (100%)
1 bedroom	5 (16%)	26 (84%)
3 bedroom	2 (14%)	12 (86%)
4 bedroom	1 (17%)	5 (83%)
5+ bedroom	0 (0%)	2 (100%)
1 Fisher's exact test

Models:

m1 <- glm(Terminated_paid_late~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Terminated_paid_late~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.62 (0.17 to 2.17)	0.47
big (single-fam)	1.36 (0.17 to 28.7)	0.79
big (non single-fam)	4.09 (0.93 to 21.4)	0.068
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.86 (0.18 to 4.00)	0.85
big (single-fam)	1.42 (0.17 to 30.8)	0.77
big (non single-fam)	3.67 (0.67 to 23.7)	0.14
Financial_role_
Other	—
Primary Income	1.11 (0.21 to 8.64)	0.91
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	8,833,860 (0.00 to NA)	>0.99
1 bedroom	0.69 (0.18 to 2.76)	0.59
3 bedroom	1.16 (0.19 to 9.66)	0.88
4 bedroom	1.17 (0.13 to 26.1)	0.90
5+ bedroom	26,815,835 (0.00 to NA)	>0.99
1 OR = Odds Ratio, CI = Confidence Interval

Terminated - non payment

Outcome: any terminated due to late payment

• 0: no
• 1: yes

Crosstabs:

df %>% 
  tbl_summary(
    by = Terminated_non_payment,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Terminated - non payment**")

Characteristic	Terminated - non payment		p-value1
	0	1
Landlord_size_, n (%)			<0.001
small (single-fam)	5 (14%)	32 (86%)
small (non single-fam)	8 (22%)	28 (78%)
big (single-fam)	1 (14%)	6 (86%)
big (non single-fam)	3 (2.1%)	138 (98%)
Financial_role_, n (%)			0.037
Other	15 (11%)	127 (89%)
Primary Income	2 (2.6%)	74 (97%)
Common_unit_, n (%)			0.43
2 bedroom	6 (8.5%)	65 (92%)
0 bedroom (studio/efficiency)	0 (0%)	32 (100%)
1 bedroom	5 (6.7%)	70 (93%)
3 bedroom	2 (8.0%)	23 (92%)
4 bedroom	1 (14%)	6 (86%)
5+ bedroom	0 (0%)	4 (100%)
1 Fisher's exact test; Pearson's Chi-squared test

Models:

m1 <- glm(Terminated_non_payment~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Terminated_non_payment~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.55 (0.15 to 1.83)	0.34
big (single-fam)	0.94 (0.12 to 19.6)	0.96
big (non single-fam)	7.19 (1.68 to 36.5)	0.009
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	0.94 (0.21 to 4.12)	0.94
big (single-fam)	0.98 (0.12 to 20.9)	0.99
big (non single-fam)	6.98 (1.32 to 45.1)	0.027
Financial_role_
Other	—
Primary Income	1.29 (0.24 to 9.86)	0.78
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	11,134,000 (0.00 to 153,429,362,178,518,012,976,266,608,028,224,042,482,820,240,642,668,086,282,620,684,062,044,822,224,680,680,220,046,226,202,400,222,480,680,080,600,622,042,206,826,608,440,424,086,600,266,826,480,228,686,606,228,888,082,466,226,402,480,886,686,806,084,468,068,224,604,642,042,606,080,046,866,002,222,288,646,226,222,604,442,648,224)	>0.99
1 bedroom	0.83 (0.22 to 3.25)	0.79
3 bedroom	1.67 (0.31 to 13.1)	0.57
4 bedroom	1.05 (0.13 to 22.7)	0.97
5+ bedroom	47,354,877 (0.00 to NA)	>0.99
1 OR = Odds Ratio, CI = Confidence Interval

Terminated - use courts

Ouctome: of the terminations, were these handled with the courts

• 0: always or usually handle with tenants
• 1: always or usually handle with the courts

Crosstabs:

df %>% 
  tbl_summary(
    by = Terminated_use_courts,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Terminated - use of courts**")

Characteristic	Terminated - use of courts		p-value1
	0	1
Landlord_size_, n (%)			0.97
small (single-fam)	10 (62%)	6 (38%)
small (non single-fam)	12 (60%)	8 (40%)
big (single-fam)	3 (50%)	3 (50%)
big (non single-fam)	63 (60%)	42 (40%)
Financial_role_, n (%)			0.99
Other	52 (60%)	34 (40%)
Primary Income	35 (60%)	23 (40%)
Common_unit_, n (%)			0.92
2 bedroom	31 (60%)	21 (40%)
0 bedroom (studio/efficiency)	14 (61%)	9 (39%)
1 bedroom	29 (59%)	20 (41%)
3 bedroom	6 (46%)	7 (54%)
4 bedroom	3 (75%)	1 (25%)
5+ bedroom	1 (100%)	0 (0%)
1 Fisher's exact test; Pearson's Chi-squared test

Models:

m1 <- glm(Terminated_use_courts~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Terminated_use_courts~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)
tbl_regression(m2, exponentiate=TRUE)

Renter Profile, Future

Renter income estimate

Outcome: average income of renter by most common unit size (used midpoints)

• 13: less than 25,000
• 37: 25,000-50,000
• 62: 50,000-75,000
• 100: 75,000-150,000
• 200: more than 150,000

Crosstabs:

df %>% 
  tbl_summary(
    by = Renter_income_estimate,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Renter income estimate (unit=1000)**")

Characteristic	Renter income estimate (unit=1000)					p-value1
	13	37	62	100	200
Landlord_size_, n (%)
small (single-fam)	64 (4.0%)	254 (16%)	396 (25%)	677 (42%)	204 (13%)
small (non single-fam)	60 (5.4%)	264 (24%)	359 (32%)	364 (33%)	72 (6.4%)
big (single-fam)	2 (4.1%)	6 (12%)	22 (45%)	16 (33%)	3 (6.1%)
big (non single-fam)	42 (5.5%)	290 (38%)	297 (39%)	133 (17%)	8 (1.0%)
Financial_role_, n (%)						<0.001
Other	130 (4.2%)	685 (22%)	919 (30%)	1,085 (35%)	274 (8.9%)
Primary Income	33 (7.8%)	127 (30%)	150 (35%)	100 (24%)	13 (3.1%)
Common_unit_, n (%)
2 bedroom	37 (2.8%)	281 (21%)	443 (33%)	490 (37%)	83 (6.2%)
0 bedroom (studio/efficiency)	19 (13%)	65 (45%)	40 (28%)	19 (13%)	1 (0.7%)
1 bedroom	32 (3.9%)	265 (33%)	284 (35%)	199 (25%)	31 (3.8%)
3 bedroom	24 (3.1%)	117 (15%)	206 (26%)	332 (42%)	103 (13%)
4 bedroom	11 (5.3%)	34 (16%)	33 (16%)	77 (37%)	52 (25%)
5+ bedroom	25 (37%)	12 (18%)	11 (16%)	11 (16%)	9 (13%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Renter_income_estimate~1 + Landlord_size_,
          family=gaussian(link="identity"), data=df)

m2 <- glm(Renter_income_estimate~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=gaussian(link="identity"), data=df)



tbl_regression(m1)

Characteristic	Beta (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	-15 (-18 to -12)	<0.001
big (single-fam)	-12 (-24 to 0.31)	0.056
big (non single-fam)	-32 (-36 to -28)	<0.001
1 CI = Confidence Interval

tbl_regression(m2)

Characteristic	Beta (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	-7.8 (-12 to -4.0)	<0.001
big (single-fam)	-14 (-27 to -1.7)	0.026
big (non single-fam)	-23 (-27 to -18)	<0.001
Financial_role_
Other	—
Primary Income	-3.1 (-7.8 to 1.6)	0.20
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	-20 (-27 to -12)	<0.001
1 bedroom	-7.0 (-11 to -3.2)	<0.001
3 bedroom	7.4 (3.4 to 11)	<0.001
4 bedroom	20 (13 to 26)	<0.001
5+ bedroom	-19 (-30 to -8.5)	<0.001
1 CI = Confidence Interval

Still be prop mngr in 1, 5 yr

Outcome: do you expect to remain a property manager

• 0: unsure, definitely or propably not in 1 or 5 yrs
• 1: definitely or probably in 1 or 5 yrs

Crosstabs:

df %>% 
  tbl_summary(
    by = Still_mngr_1_5yrs,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Still be manager in 1 or 5 yrs**")

Characteristic	Still be manager in 1 or 5 yrs		p-value1
	0	1
Landlord_size_, n (%)			0.70
small (single-fam)	1,398 (81%)	326 (19%)
small (non single-fam)	971 (81%)	228 (19%)
big (single-fam)	47 (84%)	9 (16%)
big (non single-fam)	684 (83%)	143 (17%)
Financial_role_, n (%)			0.87
Other	2,714 (81%)	618 (19%)
Primary Income	370 (81%)	86 (19%)
Common_unit_, n (%)			0.14
2 bedroom	1,160 (81%)	276 (19%)
0 bedroom (studio/efficiency)	125 (85%)	22 (15%)
1 bedroom	738 (83%)	148 (17%)
3 bedroom	674 (80%)	170 (20%)
4 bedroom	184 (83%)	39 (17%)
5+ bedroom	67 (89%)	8 (11%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Still_mngr_1_5yrs~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Still_mngr_1_5yrs~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.01 (0.83 to 1.21)	0.94
big (single-fam)	0.82 (0.37 to 1.61)	0.59
big (non single-fam)	0.90 (0.72 to 1.11)	0.32
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.01 (0.81 to 1.25)	0.95
big (single-fam)	0.84 (0.38 to 1.66)	0.63
big (non single-fam)	0.93 (0.71 to 1.20)	0.56
Financial_role_
Other	—
Primary Income	1.12 (0.85 to 1.47)	0.41
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	0.74 (0.45 to 1.17)	0.21
1 bedroom	0.85 (0.68 to 1.06)	0.15
3 bedroom	1.05 (0.83 to 1.31)	0.70
4 bedroom	0.89 (0.60 to 1.29)	0.54
5+ bedroom	0.49 (0.22 to 0.98)	0.064
1 OR = Odds Ratio, CI = Confidence Interval

Stop being prop mngr burdensome regs

Outcome: why might you stop being property manager

• 0: other reasons
• 1: burdensome regulatory environment

Crosstabs:

df %>% 
  tbl_summary(
    by = Stop_mngr_burdensome_regs,
    percent = "row",
    missing = "no",
    include = c(Landlord_size_, Financial_role_, Common_unit_),
    statistic = list(all_categorical() ~ "{n} ({p}%)")) %>%
  add_p() %>%
  bold_labels() %>%
  modify_header(update = all_stat_cols() ~ "**{level}**") %>% # Remove the Ns from the header row
modify_spanning_header(starts_with("stat_") ~ "**Stop managing due to regulations**")

Characteristic	Stop managing due to regulations		p-value1
	0	1
Landlord_size_, n (%)			0.97
small (single-fam)	1,481 (86%)	246 (14%)
small (non single-fam)	1,044 (86%)	175 (14%)
big (single-fam)	47 (84%)	9 (16%)
big (non single-fam)	711 (85%)	123 (15%)
Financial_role_, n (%)			0.16
Other	2,885 (86%)	476 (14%)
Primary Income	381 (83%)	76 (17%)
Common_unit_, n (%)			0.11
2 bedroom	1,224 (85%)	218 (15%)
0 bedroom (studio/efficiency)	135 (91%)	13 (8.8%)
1 bedroom	767 (86%)	121 (14%)
3 bedroom	713 (84%)	133 (16%)
4 bedroom	194 (87%)	29 (13%)
5+ bedroom	69 (92%)	6 (8.0%)
1 Pearson's Chi-squared test

Models:

m1 <- glm(Stop_mngr_burdensome_regs~1 + Landlord_size_,
          family=binomial(link="logit"), data=df)

m2 <- glm(Stop_mngr_burdensome_regs~1 + Landlord_size_ +
            Financial_role_ + Common_unit_,
          family=binomial(link="logit"), data=df)



tbl_regression(m1, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.01 (0.82 to 1.24)	0.93
big (single-fam)	1.15 (0.52 to 2.27)	0.70
big (non single-fam)	1.04 (0.82 to 1.31)	0.73
1 OR = Odds Ratio, CI = Confidence Interval

tbl_regression(m2, exponentiate=TRUE)

Characteristic	OR (95% CI)1	p-value
Landlord_size_
small (single-fam)	—
small (non single-fam)	1.05 (0.82 to 1.34)	0.68
big (single-fam)	1.15 (0.52 to 2.29)	0.70
big (non single-fam)	1.09 (0.83 to 1.45)	0.53
Financial_role_
Other	—
Primary Income	1.24 (0.92 to 1.65)	0.15
Common_unit_
2 bedroom	—
0 bedroom (studio/efficiency)	0.50 (0.27 to 0.88)	0.024
1 bedroom	0.86 (0.67 to 1.09)	0.22
3 bedroom	1.08 (0.84 to 1.39)	0.54
4 bedroom	0.87 (0.56 to 1.32)	0.53
5+ bedroom	0.49 (0.19 to 1.07)	0.11
1 OR = Odds Ratio, CI = Confidence Interval