Families on TANF in Illinois:  Employment Assets and Liabilities

Appendix E.
Methodology for Multivariate Analysis

[ Main page of Report | Contents of Report ]

To examine the affect of liabilities on employment in a multivariate context, we use a logit models to estimate whether the number or specific liabilities affects whether a TANF case head is employed 30 hours or more per week. The analysis sample includes the 372 survey respondents who had no missing data on employment status, selected background characteristics, or any of the 16 liability measures. A list of variables included in the models and their means and standard deviations are presented in Table E-1. The logit estimation results are presented in Tables E-2 and E-3.

First, we estimate equations (1) and (2) to determine whether the number of liabilities affects employment status. These equations express employment status as a function of the number of employment liabilities and a set of background characteristics, including a measure of prior welfare receipt. We specify the number of barriers as a series of seven dummy variables in equation (1) and as a series of three dummy variables in equation (2).


where:

EMP30i =1 if working 30 or more hours per week; 0 otherwise
N1ji =1 if the number of liabilities is j; 0 otherwise; j 1,…, 6
N17i =1 if the number of liabilities is 7 or more; 0 otherwise
Xki =set of background control variables,(1) k = 1,…, K
Ui =random disturbance term
=parameters to be estimated
i = index for survey respondents, i = 1,…, 372


where:

EMP30i =1 if working 30 or more hours per week; 0 otherwise
N21i =1 if the number of liabilities is 1, 2, or 3; 0 otherwise
N22i =1 if the number of liabilities is 4, 5, or 6; 0 otherwise
N23i = 1 if the number of liabilities is 7 or more; 0 otherwise
Xki =set of background control variables, k = 1,…, K
Ui =random disturbance term
=parameters to be estimated
i = index for survey respondents, i = 1,…, 372

Next, we estimate equation (3) to determine how each individual liability affects employment status. This equation expresses employment as a function of 14 employment liabilities, prior welfare receipt, and the same set of background characteristics as in the previous two equations.


where:

EMP30i =1 if working 30 or more hours per week; 0 otherwise
Lji =1 if specific liability j is present; 0 otherwise; j 1,…, 16
Xki =set of background control variables, k = 1,…, K
Ui =random disturbance term
=parameters to be estimated
i = index for survey respondents, i = 1,…, 372

Finally, for comparison purposes, we estimate equation (4) to determine how the background characteristics alone affect employment.


where:

EMP30i =1 if working 30 or more hours per week; 0 otherwise
Xki =set of background control variables, k = 1,…, K
Ui =random disturbance term
=parameters to be estimated
i = index for survey respondents, i = 1,…, 404

Table E.1
Variable Descriptions
Variable Mean Standard Deviation
Dependent Variable
Case head works 30 or more hours per week .30 .46
Independent Variables
Human Capital Liabilities
No high school diploma or GED .44 .50
Fewer than four quarters of recent work experience .59 .49
Performed fewer than four common job tasks .28 .45
Personal Challenges
Physical health problem .21 .41
Mental health problem .25 .43
Chemical dependence .03 .17
Severe physical domestic violence in past year .13 .33
Signs of learning disability .12 .33
Multiple arrests .16 .37
Difficulty with English language .02 .38
Logistical and Situational Challenges
Child/other family member/friend w/health problem or special need .35 .48
Pregnant .08 .80
Child under age one in household .28 .45
Transportation barrier .21 .54
Child care .31 .46
Unstable housing .23 .42
Counts of Liabilities
One .12 .33
Two .16 .37
Three .21 .41
Four .17 .37
Five .13 .33
Six .07 .26
Seven or more .10 .30
Two to three .37 .48
Four to six .36 .48
Seven or more .10 .30
Background Characteristics
Age 25-34 .34 .48
Age 34 and over .30 .46
African American .82 .38
Non-African American, Non-White .07 .26
Never married .82 .39
Separated, divorced, or widowed .14 .35
County unemployment rate 5.80 .50
80% or more of zip code is non-Hispanic African American .48 .50
Percentage of past 25 months received TANF 69.64 32.47
Percentage of past 25 months received TANF squared 5901.21 3959.92
Have child between age one and five .61 .49
Have two children .28 .45
Have three children .47 .50
Source: 2001-02 survey of Illinois TANF cases, N=416.

Table E.2
Effects of Multiple Liabilities on the Probability that a TANF Case Head Works 30 or More Hours Per Week
  Model 1 Model 2
Coefficient Std Error t P>|t| Coefficient Std Error t P>|t|
Number of Liabilities
One -1.09 0.66 -1.65 0.10 -1.09 0.66 -1.66 0.10
Two -1.00 0.65 -1.54 0.12        
Three -1.38 0.63 -2.17 0.03        
Four -1.75 0.67 -2.63 0.01        
Five -1.72 0.71 -2.41 0.02        
Six -1.76 0.75 -2.35 0.02        
Seven or more -3.26 0.90 -3.61 0.00        
Grouped Liabilities
Two to three         -1.21 0.61 -1.98 0.05
Four to six         -1.74 0.62 -2.82 0.01
Seven or more         -3.25 0.90 -3.62 0.00
Background Characteristics
Age 25-34 0.79 0.35 2.25 0.03 0.78 0.35 2.24 0.03
Age 34 and over 1.12 0.39 2.84 0.01 1.08 0.39 2.80 0.01
African American -0.27 0.49 -0.54 0.59 -0.28 0.50 -0.57 0.57
Non-African American, Non- White 1.09 0.62 1.76 0.08 1.04 0.61 1.70 0.09
Never married -0.20 0.63 -0.32 0.75 -0.17 0.64 -0.27 0.79
Separated, divorced, widowed 0.12 0.70 0.18 0.86 0.18 0.70 0.25 0.80
County unemployment rate -0.44 0.24 -1.86 0.06 -0.43 0.24 -1.82 0.07
80 percent or more of zip code is non-Hispanic African American 0.28 0.29 0.96 0.34 0.27 0.29 0.95 0.35
Percentage of past 25 months received TANF 0.01 0.02 0.24 0.81 0.01 0.02 0.24 0.81
Percentage of past 25 months received TANF squared 0.00 0.00 0.02 0.98 0.00 0.00 0.03 0.98
Have child between age one and five -0.21 0.34 -0.61 0.54 -0.21 0.34 -0.63 0.53
Have two children 0.31 0.39 0.80 0.42 0.33 0.39 0.84 0.40
Have three children 0.69 0.36 1.93 0.06 0.71 0.36 1.97 0.05
Constant 1.90 1.65 1.15 0.25 1.83 1.64 1.12 0.27
F-Statistic 2.43       2.75      
Prob > F 0.0006       0.0003      
Number of Observations 375       375      
Source: Results of multinomial logit models predicting the probability of working 30+ hours per week using data from 2001-02 survey of Illinois TANF cases.

Table E.3
Effects of Specific Liabilities on the Probability that a TANF Case
Head Works 30 or More Hours Per Week
  Model 3 Model 4
  Coefficient Std Error t P>|t| Coefficient Std Error t P>|t|
Human Capital Liabilities
No high school diploma or GED -0.17 0.29 -0.59 0.56        
Limited work experience -0.70 0.29 -2.46 0.01        
Performed fewer than four common job tasks -0.11 0.32 -0.36 0.72        
Personal Challenges
Physical health problem -0.90 0.43 -2.12 0.03        
Mental health problem -0.13 0.39 -0.34 0.73        
Chemical dependence -0.13 0.95 -0.14 0.89        
Severe physical domestic violence in past year 0.58 0.42 1.37 0.17        
Signs of a learning disability 0.23 0.47 0.49 0.63        
Multiple arrests -0.80 0.45 -1.79 0.07        
Difficulty with English -0.69 0.89 -0.78 0.44        
Logistical and Situational Challenges
Child/family member/friend w/ health problem or special need 0.14 0.31 0.43 0.67        
Pregnant -0.59 0.46 -1.28 0.20        
Child under age one in household -0.18 0.35 -0.51 0.61        
Transportation -0.46 0.39 -1.18 0.24        
Child care -1.00 0.38 -2.66 0.01        
Unstable housing -0.05 0.39 -0.13 0.90        
Background Characteristics
Age 25-34 0.61 0.39 1.58 0.11 0.83 0.32 2.56 0.01
Age 34 and over 0.97 0.42 2.30 0.02 0.97 0.36 2.70 0.01
African American 0.16 0.51 0.32 0.75 -0.13 0.46 -0.27 0.79
Non-African American, Non-White 1.58 0.66 2.42 0.02 0.62 0.57 1.09 0.28
Never married -0.49 0.78 -0.62 0.54 -0.25 0.58 -0.43 0.67
Separated, divorced, widowed -0.18 0.82 -0.22 0.82 0.09 0.63 0.14 0.89
County unemployment rate -0.48 0.25 -1.97 0.05 -0.46 0.27 -1.75 0.08
80 percent or more of zip code is non-Hispanic African American 0.06 0.31 0.21 0.84 0.14 0.27 0.51 0.61
Percentage of past 25 months received TANF 0.02 0.02 0.75 0.45 0.01 0.02 0.53 0.60
Percentage of past 25 months received TANF squared 0.00 0.00 -0.37 0.71 0.00 0.00 -0.04 0.97
Have child between age one and five -0.23 0.31 -0.73 0.47        
Have child under age five         -0.33 0.30 -1.09 0.28
Have two children 0.41 0.41 1.00 0.32 0.32 0.37 0.87 0.39
Have three children 0.56 0.39 1.45 0.15 0.66 0.35 1.87 0.06
Constant 1.46 1.64 0.89 0.37 0.40 1.71 0.24 0.81
F-Statistic 2.40       2.98      
Prob > F 0.0001       0.0004      
Number of Observations 374       404      
Source: Results of multinomial logit models predicting the probability of working 30+ hours per week using data from 2001-02 survey of Illinois TANF cases

Endnote

1. The background control variables are: age, race, marital status, number of children, presence of young children, percent of time on welfare in past 25 months, county unemployment rate, neighborhood racial concentration (i.e., 80 percent or more African-American).

Where to?

Top of Page

Main Page of Report | Contents of Report

Home Pages:
Human Services Policy (HSP)
Assistant Secretary for Planning and Evaluation (ASPE)
U.S. Department of Health and Human Services (HHS)