Embed Size (px)
Citation preview
The State of Female and Racial/Ethnic United Methodist Clergyin the US
Eric B. Johnson, Ph.D.April 12, 2012
1
Contents
1 Understanding Demographic Shifts in the Representation of Female andRacial/Ethnic Clergy Serving Congregations 8
1.1 Overall Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2 Change Over Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3 Annual Conference Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.4 The Size of Congregational Appointments . . . . . . . . . . . . . . . . . . . . . . . . 16
1.5 Other Attributes of Congregational Positions . . . . . . . . . . . . . . . . . . . . . . 17
1.6 Cohort Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.7 Representation in Non-Ordained Leadership Positions in 2001 . . . . . . . . . . . . . 30
2 Career Progression 32
2.1 Career Progression—Outcomes After Eight Years . . . . . . . . . . . . . . . . . . . . 32
2.2 First Appointments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.3 Rates of Leaving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.3.1 Female Clergy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.3.2 Racial/Ethnic Clergy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
A Glossary 54
B Data 54
C Methods 55
C.1 Statistical Models of Having a Female or Racial/Ethnic Pastor . . . . . . . . . . . . 55
C.2 Statistical Models of Career Progression . . . . . . . . . . . . . . . . . . . . . . . . . 56
C.3 Congregational Size at First Appointment . . . . . . . . . . . . . . . . . . . . . . . . 56
C.4 Rates of Leaving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2
List of Figures
1 Distribution of Racial/Ethnic and Female Clergy . . . . . . . . . . . . . . . . . . . . 9
2 Percentages of Racial/Ethnic Clergy Serving Ethnic Congregations . . . . . . . . . . 9
3 Increase in Female Clergy and Seniority over Time . . . . . . . . . . . . . . . . . . . 10
4 Increase in Female Clergy by Jurisdiction . . . . . . . . . . . . . . . . . . . . . . . . 10
5 Change in Percentage of Racial/Ethnic Clergy . . . . . . . . . . . . . . . . . . . . . 11
6 Percentage of Female Clergy by Jurisdiction . . . . . . . . . . . . . . . . . . . . . . . 12
7 Percentage of Female Clergy by Type of Appointment . . . . . . . . . . . . . . . . . 13
8 Percentage of Female Sole/Lead and Associate Pastors by Conference . . . . . . . . 14
9 Percentage of Racial/Ethnic Pastors by Conference . . . . . . . . . . . . . . . . . . . 15
10 Percentage of Sole/Lead Female Pastors by Congregational Membership . . . . . . . 16
11 Percentage of Sole/Lead Racial/Ethnic Pastors by Congregational Membership . . . 17
12 Change in the Odds of Having a Female Pastor Based on Characteristics of the Position 18
13 Appointment Status by Gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
14 Appointment Status by Race/Ethnicity . . . . . . . . . . . . . . . . . . . . . . . . . 20
15 Change in the Odds of Having a Racial/Ethnic Pastor Based on Characteristics ofthe Position . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
16 Cohort Structure in 2008 by Gender . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
17 Cohort Structure in 2008 by Jurisdiction . . . . . . . . . . . . . . . . . . . . . . . . . 24
18 Percent of Cohort that are Female . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
19 Percent of Cohort that are Female by Jurisdiction . . . . . . . . . . . . . . . . . . . 25
21 Percent of Cohort by Racial/Ethnic Group . . . . . . . . . . . . . . . . . . . . . . . 26
20 Cohort Structure in 2008 by Racial/Ethnic group . . . . . . . . . . . . . . . . . . . . 27
22 Average Age at First Appointment by Cohort and Gender . . . . . . . . . . . . . . . 28
23 Average Age at First Appointment by Cohort and Racial/Ethnic Status . . . . . . . 28
24 2008 Appointment Status for Clergy Serving Congregations in 2000 . . . . . . . . . . 34
3
25 Relative Change in Moving Out of a Similar-Sized Congregation for Gender andEthnicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
26 Status at First Appointment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
27 Characteristics at First Appointment . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
28 Impact of Race and Ethnicity on Size of First Appointment . . . . . . . . . . . . . . 42
29 Annual Rate of Leaving the Clergy for 2006-2008 by Gender . . . . . . . . . . . . . . 46
30 Proportion of Clergy in Ministry for at Least Y Years by Gender . . . . . . . . . . . 47
31 Annual Rate of Leaving the Clergy for 2006-2008 by Race/Ethnicity of Pastor . . . . 51
32 Proportion of Clergy in Ministry for at Least Y Years by Race/Ethnicity . . . . . . 52
List of Tables
1 Percent of Lay Leaders that are Female . . . . . . . . . . . . . . . . . . . . . . . . . 31
2 Percent of Attenders that Hold Lay Leadership Positions by Ethnicity . . . . . . . . 31
3 Direct and Indirect Exit Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4 Proportional Change in Rates of Leaving . . . . . . . . . . . . . . . . . . . . . . . . 49
5 Multinomial Regression Model Results—Transition from Position Held in 2000 to2008 (Relative to a Same-Sized Congregation) . . . . . . . . . . . . . . . . . . . . . . 58
6 Regression Model Results—Size at First Appointment . . . . . . . . . . . . . . . . . 59
4
Introduction
There is a longstanding concern with both gen-der and racial equity for clergy in the UnitedMethodist Church (UMC). Race and gender re-lations have changed dramatically since the for-mal proscriptions against women’s ordinationwere removed in 1956 and the segregated “Cen-tral Jurisdiction” was dissolved in 1968. Butgender and racial equity did not procede eas-ily from these formal steps, and despite themany changes between then and now, the degreeto which disadvantages persist for female andracial/ethnic clergy remains a pertinent ques-tion.
One reason is the lack of basic factual informa-tion on how female and racial/ethnic clergy arefaring in the UMC. It is widely known that thenumber of female clergy has increased greatly inthe past 30 years, but changes in racial/ethnicclergy have been studied infrequently. Further-more, studies that exposed the “stained glassceiling” in the clergy careers of females werelargely conducted 10 years ago or more, and atleast 70% of the female clergy currently serv-ing congregations in the UMC have entered sincethen.
This study aims to provide a set of basic quan-titative facts that will inform discussions aboutrace and gender equity by indicating where situ-ations have improved, and what areas need moreattention. To what degree are experiences of dis-advantage the legacy of earlier discrimination?Where does disadvantage persist?
Due to data constraints this study focuses on USclergy, and pays most attention to clergy servingcongregations. The definition of clergy in theUMC includes elders, deacons, and local pastors
serving congregations under appointment. In thereport, I will use the term clergy to refer to min-isters serving in both parish and non-parish ap-pointments, and the term pastor to refer specif-ically to the leader of a congregation.
To provide a basic overview of how racial/ethnicand female clergy are faring in US congregationsin the United Methodist Church, the Study isdivided into two main topics:
1. Demographics. How are racial/ethnicgroups and females represented in theclergy; where are they most and least preva-lent; what is the composition of currentclergy with respect to their cohort of en-try; and how are these groups changing overtime.
2. Career Progression. What are the careeroutcomes of female and racial/ethnic clergyover time; what are the attributes of theirfirst appointments, and at what rate do theyleave the clergy.
Data
This study will use appointment data assem-bled for a study of UMC salaries. This data setconsists of a complete record of appointmentsfrom 1997-2008 for all US pastors serving localchurches in the UMC, characteristics of the pas-toral charges they are assigned to, and a his-torical record of parish and non-parish appoint-ments for the pastors serving from 1997-2008. Arecord of which churches are assigned to multi-ple church charges is unavailable prior to 1997,so church/pastoral charge characteristics for ap-pointments prior to 1997 are not available.
5
Seniority, Cohort, and Period
The experiences of clergy may be affected bytime in three separate ways: through seniority,through the cohort with which they entered theministry, and through the period in which theyare serving.
Seniority is the amount of time a person hasbeen a pastor. For this study, it is defined asthe time since a pastor’s first recorded appoint-ment to a congregation (at any level). In general,newer pastors are more likely to be assigned tosmaller positions, be associate pastors, or havestatus other than elder. Lead pastor positionsand positions in larger/wealthier congregationsare more likely to be assigned pastors who havehigher seniority (GBHEM Salary Study 2010).
Clergy currently serving congregations are alsomembers of different cohorts. They have en-tered the ministry during different time periodsand, to some degree, their experiences have re-sulted from the common environmental condi-tions faced by these cohorts as they progressedthrough their careers. For the cohorts of fe-male pastors that entered the ministry in the1970’s, most female pastors were the first femaleassigned to their congregation, and there werevery few mid-career female pastors to serve asrole models, or who occupied administrative po-sitions within the UMC. By the end date of thisstudy, almost 30% of all pastors were female,and females represented an even greater percent-age of pastors serving in their first appointment.Characteristics of job markets are a particularlyimportant factor when considering cohorts. Forexample, there are a large number of pastors ap-proaching retirement; their departure will createmore opportunities for younger pastors just en-
tering the ministry.
Finally, the period in which a pastor isserving matters directly. Attitudes towardsracial/ethnic and female clergy have shifted no-tably of the past 40 years, so pastors who en-tered the ministry in the 1970’s have differingprospects for appointments today, regardless oftheir cohort or seniority.
The present study combines both prospectiveand retrospective elements. Since 1997, the datacontains information on all clergy assigned toUMC congregations, and tracks records of theirappointments. For pastors who entered the min-istry at or after this time period, we can de-termine exactly how many people entered theclergy and how many have left the clergy. Wedo not know how many people left the clergyprior to 1997, so for cohorts prior to 1997, wecannot know for certain how many female andracial/ethnic clergy entered in these cohorts, orwhether these pastors left the clergy at higherrates than white/male clergy. In other words, co-hort and seniority are confounded—having highseniority also implies belonging to an early co-hort within the sample.
Since racial/ethnic and female clergy have facedradically differing conditions based upon thetime in which they entered the ministry, cohortwill be a central theme of the study. The studywill document differences in the career experi-ences of racial/ethnic and female pastors fromprevious cohorts when they enter the study sam-ple in 1997, and will track how their experi-ences change between 1997 and 2008. However,it is important to recognize that these differ-ences may result from two separate processes:the appointment of these groups to differenttypes of positions over the course of their ca-
6
reers and/or differential rates at which femaleand racial/ethnic ministers leave the clergy.
Statistics about cohorts prior to 1997 are stillbe meaningful, but care needs to be taken intheir interpretation. They are not representa-tive of the entire cohort, just the fraction of thecohort that has remained in the pastorate un-til 1997. For this reason, it will be important todifferentiate between cohorts, account for senior-ity throughout the analyses, and analyze rates ofleaving the ministry.
7
1 Understanding DemographicShifts in the Representationof Female and Racial/EthnicClergy Serving Congrega-tions
How many racial/ethnic and female clergy servecongregations in the UMC? Are these pastorsmore likely to be found in some regions, posi-tions, or types of churches than others? To whatdegree are all congregations open to female andracial/ethnic clergy? The first section of partI of this report addresses these questions, usinga snapshot of data from the end of the studyperiod.1 The second and third sections thentake up the theme of time, examining the cohortstructure of the clergy and how representationlevels have changed over time.
1.1 Overall Representation
In 2008, female clergy were more than twice asprevalent as racial/ethnic clergy. Females com-prised 29% of the clergy serving congregations,while only 13% percent of pastors had a non-white racial/ethnic affiliation (fig. 1).
Of these racial/ethnic clergy, Black pastors wereby far the largest category, with 6.9% of all UMCclergy. Blacks were followed by Asian clergy(3%) and Hispanic/Latino(a) clergy (1.5%). Na-tive American, Pacific Islander, and multi-racialpastors each comprised less than 1% of theclergy. The small number of clergy in thesegroups limits the potential for statistical anal-ysis. For this reason, the report will group
1In order to achieve a reasonable sample size, a threeyear window (2006-2008) is used for this snapshot.
racial/ethnic categories with less than 1% intoa single, combined, category.
The representation of female pastors varies sub-stantially across racial/ethnic groups. Propor-tionally, there are more female pastors withinBlack and multi-racial clergy (34%) than thereare within White clergy (29%). In otherracial/ethnic categories, however, there are fewerfemale pastors than there are for White clergy.For example, only 21% of Asian pastors are fe-male.
Ethnic congregations are highly likely to have aracial/ethnic pastor—over 90% of ethnic congre-gations are led by a racial/ethnic pastor (fig. 2).2 (Conversely, less than 10% of predominantlyWhite congregations are led by non-White pas-tors).
In contrast to the experiences of ethnic con-gregations, racial/ethnic pastors are appointedto both ethnic and non-ethnic congregations.Racial/ethnic pastors were appointed to ethniccongregations in about 60% of appointments.Forty percent of their appointments were to non-ethnic congregations. Thus, from the point ofview of a racial/ethnic pastor, cross-racial/cross-cultural appointments occur frequently, but eth-nic congregations do not receive many cross-racial/cross-cultural appointments.
2This study follows the GCFA coding where congre-gations with 40% or more non-White members are con-sidered to be Ethnic Congregations.
8
Figure 1: Distribution of Racial/Ethnic and Female Clergy
0 20 40 60 80 100
0.4
0.1
0.4
0.4
1.5
3
6.9
87.4
Missing
Pacific Islander
Native American
Multi−racial
Hispanic/Latino(a)
Asian
Black
White
Percent of Pastors
0 20 40 60 80 100
41
24
21
34
24
21
34
29Percent Female
Figure 2: Percentages of Racial/Ethnic ClergyServing Ethnic Congregations
Ethnic Congregation
Non−Ethnic Congregation
Percent of Congregations with an Ethnic Pastor
Ethnic Congregation
Non−Ethnic Congregation
Percent of Ethnic Pastors at Congregations
0 20 40 60 80 100
Note: Percentages are based upon the years 2006-2008.
9
1.2 Change Over Time
Figure 3: Increase in Female Clergy and Senior-ity over Time
1998 2000 2002 2004 2006 2008
010
2030
4050
Year
Per
cent
/Yea
rs
Percent of PastorsMean Years of Seniority
Note: This figure shows the percent of the cohort thatis presently female. If female pastors from earlier cohortsleft the ministry at different rates from male pastors, thenthis percentage may not reflect the initial composition ofthe cohort.
Female Clergy Even in the 12 year periodcovered by this report, the percent of femalepastors has increased substantially. In 1997,around 20% of pastor appointments to congre-gations were held by females. By 2008, this hadincreased by 45% (to 29%, fig. 3). As the co-hort analyses below indicate, this increase re-flects the larger proportion of female pastors inrecent cohorts, and retirements among the old-est cohorts where there were few female pastors.Provided that females continue to comprise sim-ilar percentages of entering cohorts, the number
of female pastors in the UMC will continue toincrease.
This process has important implications for pas-tor seniority. Because most female pastors arerecent entrants, their seniority levels are cur-rently rather low. In 1997 female pastors had 9years of seniority, on average, and this increasedby 30% by the end of the study (fig. 3). The av-erage seniority of female pastors can be expectedto increase further as recent cohorts with largefemale contingents advance toward the middle oftheir careers.
Figure 4: Increase in Female Clergy by Jurisdic-tion
1998 2000 2002 2004 2006 2008
010
2030
4050
Year
Per
cent
Fem
ale
North CentralNortheasternSouth CentralSoutheasternWestern
Racial/Ethnic Pastors When viewed asa percent of overall pastors, the share ofracial/ethnic pastors in the UMC shifted littlefrom 1997 to 2008 (fig. 5). However, this largelyreflects the stability of Black pastors, the largest
10
racial/ethnic group in the UMC. The share ofAsian, Hispanic/Latino(a), and the combined(Native American, Pacific Islander, and multi-racial) group each increased by about 25% from1997 to 2008. This rate of change does not im-pact the overall representation of racial/ethnicpastors in the UMC by much, however, becausethese groups were small to begin with.
1.3 Annual Conference Variation
The degree of variation across annual confer-ences is one of the most striking features ofrepresentation levels for racial/ethnic and fe-male clergy. This variation occurs for bothracial/ethnic and female clergy, but follows dif-fering logic. For female clergy, the pattern isone of regional difference. Annual conferencevariation is associated with broad US regions.For racial/ethnic clergy, annual conference vari-ation exhibits a pattern of concentration—a fewannual conferences have relatively high levelsof representation for racial/ethnic clergy, whilethere are low levels in most annual conferences.3
Regional Differences in Female ClergyFigure 6 shows how the percentage of femalepastors varies by UMC jurisdiction boundaries.4
There are far fewer female pastors in the South-eastern jurisdiction than in other jurisdictions—only 20% of pastors in this jurisdiction are fe-male. By contrast, in the jurisdictions with thehighest level of female pastors, the Western andNortheastern jurisdictions, 36% and 35% of pas-
3This section uses the same three-year snapshot (2006-2008), as the earlier section.
4With a few exceptions, jurisdiction boundaries areclose to the regions defined by the US Census Bureau.
Figure 5: Change in Percentage of Racial/EthnicClergy
1998 2000 2002 2004 2006 2008
02
46
810
1214
Year
Per
cent
BlackAsianHispanic/Latino(a)Combined (<1%)
Percent of Pastors
1998 2000 2002 2004 2006 2008
−10
010
2030
Per
cent
Percent Change
11
tors serving at congregations are female. Thismeans that, in relative terms, a female is ∼45%less likely to be a congregation’s pastor in theSoutheastern than in the Western jurisdiction.
Figure 6: Percentage of Female Clergy by Juris-diction
Per
cent
010
2030
4050
6070
Nor
th C
entr
al
Nor
thea
ster
n
Sou
th C
entr
al
Sou
thea
ster
n
Wes
tern
3035
27
20
36
Note: Percentages were calculated from 2006-2008 basedupon a regression model.
However, both participation levels and re-gional variation depend heavily on the type ofposition—women are far more prevalent in as-sociate pastor positions than sole/lead positions(fig. 7). Almost half of associate pastors are fe-male in three out of five UMC jurisdictions. Fur-thermore, the regional differences in the percent-age of female pastor are more muted for associatepastors than for sole/lead pastors. Thus, even inregions with a low percentage of sole/lead femalepastors (Southeastern and South Central) almost40% of associate pastors are female. This impliesthat most of the regional variation in the per-
centage of female pastors results from differencesin the percentage of females serving in sole/leadpastor roles. Apparently, the factors that reducethe percentage of females serving as pastors inthe Southeastern jurisdiction are not applicableto associate pastors in the same degree. But,with the figures for associate pastors removed,regional differences are even more notable—forinstance, only 16% of lead/sole pastor positionsin the Southeastern jurisdiction are filled by fe-males.
Because of these differing patterns of represen-tation observed for associate and sole/lead pas-tors, the remainder of this report will examineeach separately.
Even after accounting for regional differences,there is still a remarkable amount of variationin levels of female participation between confer-ences. Figure 8 shows the percentage of femalesole/lead and associate pastors alongside juris-diction averages, ranked by jurisdiction averageand then conference within jurisdiction. Con-ferences range from a high of 46% sole/lead fe-male pastors in New England and North Cen-tral New York to a low of 9% in Kentucky.There is also a great range of variation withinjurisdictions. In some cases, the jurisdictionalboundaries are clearly not accurately capturingregional differences—for example, Pennsylvaniaconferences are closer to North Central or SouthCentral than to other conferences in the North-eastern. Region, while a major explanatory fac-tor, clearly does not tell the whole story.
Conference-level differences in female participa-tion rates are only weakly correlated with thepercentage of female sole/lead pastors, and thepercentage of associate pastors in a conferencedoes not mirror the trends for sole/lead pastors.
12
Figure 7: Percentage of Female Clergy by Type of Appointment
Per
cent
010
2030
4050
6070
Nor
th C
entr
al
Nor
thea
ster
n
Sou
th C
entr
al
Sou
thea
ster
n
Wes
tern
2734
24
16
34
Sole/Lead Pastor
Nor
th C
entr
al
Nor
thea
ster
n
Sou
th C
entr
al
Sou
thea
ster
n
Wes
tern
4854
39 39
48
Associate Pastor
If anything, there is more variation for associatepastors than there is for sole/lead pastors.
Concentration in Racial/Ethnic ClergyBecause racial/ethnic pastors are heavily ap-pointed to racial/ethnic congregations, differ-ences in regional participation levels is highlydependent on the location of ethnic congrega-tions. There is large annual conference variationin the prevalence of racial/ethnic pastors but,unlike for female pastors, this variation is notspread across jurisdictions—jurisdictional differ-ences in the percentage of racial/ethnic pastorsserving congregations are not statistically signif-icant. Rather, regional differences are more lo-calized and depend upon the characteristics ofspecific annual conferences.
Figure 9 shows the percentage of racial/ethnicpastors in each conference. For sole/lead pas-
tors, pie charts indicate the composition ofracial/ethnic pastors within that conference.Gray crosses indicate the representation ofracial/ethnic pastors among associate pastors.
The bulk of racial/ethnic pastors are concen-trated in just a few conferences, while the ma-jority of conferences have very few racial/ethnicpastors. There are twelve conferences that haveat least 20% racial/ethnic pastors. Many of theseconferences contain large, urban centers (NewYork City, Los Angeles, Chicago, San Fran-cisco, Baltimore/DC), while a few are locatedin Southern conferences (South Carolina, Mis-sissippi, and Louisiana).
The composition of the different racial/ethnicgroups is varied. As might be expected, southernconferences tend to have predominantly Blackpastors. West coast conferences, and conferencescontaining urban areas are more likely to have
13
Figure 8: Percentage of Female Sole/Lead and Associate Pastors by Conference
0 20 40 60 80Percent Female
KentuckyAlabama−West Florida
MississippiNorth Alabama
MemphisSouth Georgia
HolstonTennessee
FloridaSouth CarolinaNorth GeorgiaNorth Carolina
Western North CarolinaVirginia
Northwest TexasArkansas
Rio GrandeTexas
New MexicoNorth Texas
LouisianaOklahoma
Central TexasKansas West
Missouri Southwest Texas
NebraskaKansas East
South IndianaNorth Indiana
Illinois Great RiversWest Ohio
DakotasEast Ohio
DetroitWest Michigan
IowaNorthern Illinois
MinnesotaWisconsin
Western PennsylvaniaWest Virginia
Eastern PennsylvaniaCentral PennsylvaniaGreater New JerseyPeninsula−Delaware
Western New YorkNew York
Baltimore−WashingtonTroy
WyomingNew England
North Central New YorkDesert SouthwestCalifornia−Pacific
Missionary ConferencesRocky Mountain
YellowstonePacific Northwest
Oregon−IdahoCalifornia−Nevada
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
+++
++
++
Southeastern
South Central
North Central
Northeastern
Western
● Sole/Lead PastorAssociate Pastor
Based upon multilevel-regression results for data from 2006-2008, with jurisdiction and associate pastor status aspredictors. Lines indicate the average values for each jurisdiction.
14
Figure 9: Percentage of Racial/Ethnic Pastors by Conference
0 20 40 60 80 100Percent
++++
++++++++
+++++
++
++
++
++
+++
++
+++
+++
++
++
++
++
++
+++
++
++
++
++
++
++
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Racial/Ethnic Representation
●
Ethnic CompositionBlackAsianHispanic/Latino(a)Combined (<1%)
Ethnic RepresentationSole/Lead PastorAssociate Pastor
YellowstoneWest Virginia
Central PennsylvaniaNorthwest Texas
Western PennsylvaniaNorth Indiana
DakotasWest MichiganSouth Indiana
HolstonOregon−Idaho
WyomingKentucky
Illinois Great RiversAlabama−West Florida
MinnesotaTroy
Kansas WestNew Mexico
North Central New YorkIowa
West OhioMissouri NebraskaEast OhioOklahoma
Kansas EastWestern New York
ArkansasWestern North Carolina
MemphisSouth Georgia
VirginiaNorth AlabamaNorth Carolina
Rocky MountainSouthwest Texas
TennesseeCentral Texas
DetroitWisconsin
Desert SouthwestPacific Northwest
North TexasEastern Pennsylvania
New EnglandFlorida
North GeorgiaTexas
LouisianaMississippi
South CarolinaNorthern Illinois
Greater New JerseyCalifornia−Nevada
Peninsula−DelawareBaltimore−Washington
New YorkCalifornia−Pacific
Missionary ConferencesRio Grande
Percent0 20 40 60 80
Racial/Ethnic Composition
Bar charts give the composition of racial/ethnic groups for pastors within each conference. The Combined groupconsists of Native American, Pacific Islander, and Multi-racial affiliations. Based upon multilevel-regression resultsfor data from 2006-2008 with jurisdiction and associate pastor status as predictors.
15
a mix, often with Asian pastors as the largestgroup. Finally, the highest representation lev-els of non-ethnic pastors are found in confer-ences that were formed based on ethnic lines—Hispanic/Latino(a) pastors in the Rio Grandeconference, and Native American pastors in themissionary conferences.5
1.4 The Size of Congregational Ap-pointments
Pastors of any sex and race are unlikely to leada large congregation in their careers. For ex-ample, in the study data, only 4% of male pas-tors with 15 years of seniority had led a congre-gation with 1,000 or more members.6 Becauselarge congregations have high visibility, and thepastors of these congregations are able to directgreater levels of resources, it is important to con-sider whether females and racial/ethnic pastorsare proportionally represented in appointmentsas lead pastors to the largest congregations.
Larger congregations are, in fact, far less likelyto have female pastors than smaller congrega-tions. The percentage of female sole/lead pastorsdecreases monotonically as congregational sizeincreases—from 25-45% in small congregationsto less than 5% in congregations with 5,000+members (fig. 10). 7
Sole/lead racial/ethnic pastors are also more
5Note that the missionary conferences have beengrouped together because of the small number of pastorsin each separate missionary conference.
6Based on a simple logistic model of leading a con-gregation with 1000+ members as the senior/sole pastor,based upon seniority and gender as predictors.
7These figures are based upon a regression model.See the appendix for further information regarding themethodology.
Figure 10: Percentage of Sole/Lead Female Pas-tors by Congregational Membership
50 100 500 2000
010
2030
4050
6070
Membership (Log−Scale)
Per
cent
North CentralNortheasternSouth CentralSoutheasternWestern
likely to be found in smaller congregations, andare highly unlikely to be the leaders of the largestchurches (fig. 11).
Since pastors are more likely to be assignedto larger congregations later in their careers,to some degree this decrease is driven by thesmall numbers of female pastors in older cohorts(see below). A better comparison is to considerwhether the association between congregationalsize and gender is as large when a pastor’s co-hort is included in the analysis, or whether theassociation with size is greater in older cohortsthan for more recent ones. The magnitude ofthe association between size and the probabil-ity of having a female pastor drops by roughlyhalf when cohort is considered alongside size,indicating that the lack of more senior femalepastors contributes substantially to the paucity
16
of female pastors in the largest congregations.However, within cohorts, the negative associa-tion between size and the probability of havinga female pastor is greater for older cohorts. Inother words, there are fewer female lead pastorsin large congregations than can be attributed totheir reduced numbers in older cohorts, but thesituation is improving for more recent cohorts.
Figure 11: Percentage of Sole/LeadRacial/Ethnic Pastors by CongregationalMembership
50 100 500 2000
010
2030
4050
6070
Membership (Log−Scale)
Per
cent
1.5 Other Attributes of Congrega-tional Positions
How does the representation of female andracial/ethnic clergy vary across other character-istics of congregational positions? Multivariate,statistical models were used to assess this ques-tion. The results of these models indicate thechange in the likelihood of having female pas-
tor that is with various factors. Values over 1.0indicate factors that increase the likelihood ofhaving a female or racial/ethnic pastor, whilevalues lower than 1.0 indicate factors that re-duce this likelihood. The results for all factorscan be interpreted on an all things equal basiswith respect to the other factors discussed so far(conference, region, and size). See the appendixfor a more detailed discussion of the statisticaltechniques used and the full model results.
Female Clergy A congregation is less likelyto have a female pastor if they are located out-side of an urban area (figure 12). The odds ofhaving a female pastor are approximately 20-30% less in large-town and rural churches, andapproximately 10% less in suburban churches.These differences run counter to the impact ofsize, so while females pastors are more likely insmall churches, they are less likely in small ruralchurches than in small urban churches. Beyondsize and location, the grouping of congregationsinto multi-congregation charges matters. Theodds of having a female pastor are slightly higherin multi-congregation charges than in single-congregation charges of a similar size and loca-tion.
The picture for racial/ethnic pastors varies byracial/ethnic category. As a whole, ethnic con-gregations (those reporting 40% or more non-White membership) are about 40% less likely tohave a female pastor. The odds of being a fe-male Asian pastor are 70% less than being a fe-male White pastor. Likewise, the odds of beinga female within the group of Hispanic/Latino(a)pastors is almost 60% less than among Whitepastors. In contrast, the odds of being femaleamong Black pastors is almost 60% higher than
17
Figure 12: Change in the Odds of Having a Female Pastor Based on Characteristics of the Position
0.0 0.5 1.0 1.5 2.0 2.5 3.0Change in Odds
Ethnic Congregation
CombinedHispanic/Latino(a)
AsianBlack
Ethnicity (Reference=White): Quarter Time
Half TimeThree Quarter Time
Part−Time (Reference=Full Time): Other
DeaconLocal Pastor
Full−time Local PastorAppointment Status (Reference=Elder):
RuralLarge Town
SuburbanLocation (Reference=Urban):
Multi−Congregation Charge ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
5.8
Dots indicate the multiplicative change in the odds of having a female pastor for a number of characteristics. Avalue of 1.0 indicates no change. For variables with multiple categories, these changes are presented with respect to areference category. The Combined group consists of Native American, Pacific Islander, and Multi-racial affiliations.These results stem from a regression model.
for White pastors.
There are also notable differences in the proba-bility of being a female sole/lead pastor acrossjob characteristics. Historically, females heldleadership roles primarily as deacons and, eventoday, the odds of being female are much, muchgreater for deacons—deacons are 5.5x morelikely to be female than male. Other appoint-ment statuses are less likely to have female pas-tors compared to elders—the odds of being afemale full-time local pastor are approximately55% less and other local pastors are approxi-
mately 60% less. Part-time pastors are morelikely to be female.8
These differences in appointment status (as wellas the other comparisons in this section) aremade on an all things equal basis. In otherwords, they are the average differences one wouldexpect to see for congregational appointments
8Because the data relies on Board of Pensions datafor information on appointments and part-time pastorsare less likely to participate in the pension system, thedata is less representative of part-time pastors than it isof full-time pastors.
18
Figure 13: Appointment Status by Gender
Other
Deacon
Local Pastor
Full−time Local Pastor
Elder
FemaleMale
0 20 60 100
5%4%
0%4%
13%17%
9%10%
73%65%
that were similar for each of the other factorslisted in the table. The raw distribution of ap-pointment status positions provides a comple-mentary window into where female pastors areserving, as seen in figure 13.
There were fewer females serving as elders from2000-2008 than males, while females were morelikely to be local pastors and deacons. This dif-fers from the all things equal results becausethese appointment statuses are more likely to befound in small, part-time positions, which aremore likely to be filled by female pastors.
Racial/Ethnic Clergy Figure 15 presentsthe change in the odds that a sole/lead positionwill be filled with a racial/ethnic pastor, indi-cating how various factors are associated withthe prevalence (or lack) of racial/ethnic pastors.Conference, region, and size were also included
in the multivariate model that generated theseresults, but are not included the figure.
The enormous impact of a congregation’s ethnic-ity can be seen in the results for non-ethnic con-gregations. The change in odds approaches zero,reflecting the fact that non-ethnic congregationsare highly unlikely to have a racial/ethnic pas-tor. Being in a multi-congregation charge, how-ever, increases the odds of having a racial/ethnicpastor by more about 1.75x.
These results confirm that urban congregationsare far more likely to have a racial/ethnic pas-tor than congregations in other locations. Theodds of having a racial/ethnic pastor in subur-ban, large town, or rural congregations is lessthan half of what it is in urban congregations.This association with urban location is observedeven with congregational ethnicity included inthe model, indicating that urban location re-mains a salient predictor, even after accountingfor the propensity of ethnic congregations to befound in urban locations.
A similar pattern is observed for appointmentstatus. Racial/ethnic pastors are far more likelyto be elders than they are to be local pastorsor deacons, with a similar 50% reduction inodds for these statuses. Note, however, thatthe descriptive statistics for appointment statusshow a more complicated picture. While therewere fewer appointments as elder among Black,Hispanic/Latino(a) and pastors falling into theCombined category compared with White pas-tors, Asian pastors were more likely than Whitepastors to hold an appointment as an elder. Thisdifference suggests that cultural patterns matterfor determining the appointment patterns withinspecific race/ethnic groups.
19
The odds of being a racial/ethnic pastor areslightly less among females, but, as discussedabove, this propensity varies by racial/ethnicgroup, with higher representation of femalesamong Black pastors.
Figure 14: Appointment Status byRace/Ethnicity
Other
Deacon
Local Pastor
Full−time Local Pastor
Elder
Combined (<1%)Hispanic/Latino(a)AsianBlackWhite
0 20 40 60 80 100
4%6%6%
10%5%
1%0%0%1%0%
13%23%
4%13%
21%
9%9%
6%19%
12%
72%62%
83%57%
62%
The Combined group consists of Native American, PacificIslander, and Multi-racial affiliations.
20
Figure 15: Change in the Odds of Having a Racial/Ethnic Pastor Based on Characteristics of thePosition
0.0 0.5 1.0 1.5 2.0 2.5 3.0Change in Odds
Non−Ethnic Congregation
Female
Quarter TimeHalf Time
Three Quarter TimePart−Time (Reference=Full Time):
OtherDeacon
Local PastorFull−time Local Pastor
Appointment Status (Reference=Elder): Rural
Large TownSuburban
Location (Reference=Urban): Multi−Congregation Charge ●
●
●
●
●
●
●
●
●
●
●
●
●
21
1.6 Cohort Structure
Cohort structure is a fundamental feature ofdemographic processes that occur over time.The number of people and the representation ofgroups within each cohort reveals much aboutthe history of clergy and points toward what wecan expect in the future.
Population pyramids are a concise way of rep-resenting cohort structure. These pyramids canbe examined for insight on two separate factors.One is the balance between sides at each cohort,which indicates how well female or racial/ethnicclergy are currently represented within each co-hort. The other is the overal shape of each sideof the pyramid. A steep shape, like the one formales in figure 16, indicates the presence of asizable number of older cohorts with respect tonewer cohorts. A shape that is wide at the bot-tom and narrows towards the top, as for femalesin figure 16, signals growth (because there aremore pastors in younger cohorts than in oldercohorts).
Around 37% of UMC pastors were found in co-horts that entered the ministry prior to 1990,and thus had more than 19 years of seniority in2008. Of these older cohorts, very few (∼5%)are female, so there are few female pastors withhigh seniority. Most female pastors in the UMCbelong to newer cohorts.
There are differences in the shape of these pop-ulation pyramids between jurisdictions that re-flect the differing opportunity structure and his-torical conditions facing female pastors in eachjurisdiction. In the Southeastern and South Cen-tral jurisdictions, the graphs are more pyramidshaped—indicating a younger (in terms of se-niority) base of pastors. The graphs for the other
jurisdictions are flatter, indicating that more se-nior pastors make up a higher percentage of pas-tors in these jurisdictions. The growth that isoccurring in the South, but not in the other re-gions, is the most likely explanation for thesedifferences in shape.
In all cases, there are very small numbers of fe-male pastors until the cohorts of the mid 80’s.Females enter the pastorate in large numbersearliest in the Western jurisdiction (in the 1980-84 cohort), while the jump in female pastors inthe South Central and Southeastern jurisdictionsdoesn’t really occur until the mid 90’s. Eventhen, the jump in female pastors is less impres-sive because it occurs alongside a larger cohortof male pastors.
These graphs can help us understand whetherthe percentage of females is low because mostcongregational positions are occupied by mem-bers of older cohorts, or because of differentialparticipation within current cohorts. For theNorth Central, Northeastern, and Western ju-risdictions it is largely the former. Recent co-horts are close to gender parity, but there are alarge number of pastors from older cohorts stillin the pipeline. By contrast, gender differencesin recent cohorts are more important in the twosouthern jurisdictions because they make up alarger proportion of the pastorate in these juris-dictions.
Turning to the composition of each cohort, wesee that the percentage of females in each co-hort is successively greater for more recent co-horts (fig. 18). Female pastors comprised 40%of the three most recent cohorts. Interestingly,the increase in female representation appears toplateau in these three cohorts– the percent fe-male is not much greater in 2005-08 than it was
22
Figure 16: Cohort Structure in 2008 by Gender
Males
15 10 5 0
8
12.2
10.6
8.7
8
8.7
6.8
5.4
2.6
0.4
0.1
2005−08*
2000−04
1995−99
1990−94
1985−89
1980−84
1975−79
1970−74
1965−69
1960−64
<1960
Females
0 5 10 15
5.3
7.9
6.9
3.4
2.3
1.6
0.8
0.4
0
0
0
The percent of clergy assigned to UMC congregations in 2008, by cohort of entry. Entry cohorts are determined bythe first recorded appointment to a charge, and have been split into five year intervals.∗The 2005-2008 cohort contains one year fewer than other cohorts, so care should be taken interpreting the percentageof pastors in this cohort.
23
Figure 17: Cohort Structure in 2008 by Jurisdiction
Males
15 10 5 0
6.7
9.8
8.9
8.5
8.5
8.9
8
6.5
3
0.3
0
2005−08*
2000−04
1995−99
1990−94
1985−89
1980−84
1975−79
1970−74
1965−69
1960−64
<1960
Females
0 5 10 15
5.2
7.5
7.1
4.4
3
1.9
1.2
0.4
0.1
0
0
North Central
Males
15 10 5 0
7.2
10.4
9.9
7.8
6.9
7.9
7.1
5
3.1
0.3
0
2005−08*
2000−04
1995−99
1990−94
1985−89
1980−84
1975−79
1970−74
1965−69
1960−64
<1960
Females
0 5 10 15
6.3
9
7.6
4.7
3.1
1.9
1.2
0.5
0.1
0
0
Northeastern
Males
15 10 5 0
7.8
13.2
11
8.2
7.7
7.9
6.2
5.1
2.6
0.6
0.2
2005−08*
2000−04
1995−99
1990−94
1985−89
1980−84
1975−79
1970−74
1965−69
1960−64
<1960
Females
0 5 10 15
5.5
9.3
7.8
3.1
2
1
0.7
0.2
0
0
0
South Central
Males
15 10 5 0
9.7
14.9
12.1
9.7
8.2
9.5
6.3
4.8
2.1
0.5
0.1
2005−08*
2000−04
1995−99
1990−94
1985−89
1980−84
1975−79
1970−74
1965−69
1960−64
<1960
Females
0 5 10 15
4.9
6.9
5.5
2.1
1.3
1
0.5
0.2
0
0
0
Southeastern
Males
15 10 5 0
6
8.2
9.2
7.4
8.6
8.1
6.1
6.9
2.7
0.3
0
2005−08*
2000−04
1995−99
1990−94
1985−89
1980−84
1975−79
1970−74
1965−69
1960−64
<1960
Females
0 5 10 15
4.7
8
8.6
4.8
3.4
4.6
1.5
0.8
0
0
0
Western
∗The 2005-2008 cohort contains one year fewer than other cohorts, so care should be taken interpreting the percentageof pastors in this cohort.
24
in 95-99.9 It appears that the change in paceat which women are entering the ministry hasslowed, if not altogether stopped.
Figure 18: Percent of Cohort that are Female
020
4060
8010
0P
erce
nt o
f Coh
ort
<19
60
1960
−64
1965
−69
1970
−74
1975
−79
1980
−84
1985
−89
1990
−94
1995
−99
200
0−04
2005
−08
*
Note: This figures shows the percent of the cohort thatis presently female. If female pastors from earlier cohortsleft the ministry at different rates from male pastors, thenthis percentage may not reflect the initial composition ofthe cohort.
Examining the same figure broken out by juris-diction, the Southeastern jurisdiction is the clearoutlier. With the exception of the early bumpin female participation in the Western jurisdic-tion, the composition of each cohort is relativelysimilar for each jurisdiction, especially in recentcohorts.
9It is not clear to what degree the shifts in earliercohorts are due to female pastors leaving the ministry atgreater rates than men, rather than to the percentage ofwomen entering the ministry in each cohort.
Figure 19: Percent of Cohort that are Female byJurisdiction
020
4060
8010
0P
erce
nt o
f Coh
ort
North CentralNortheasternSouth CentralSoutheasternWestern
<19
60
1960
−64
1965
−69
1970
−74
1975
−79
1980
−84
1985
−89
1990
−94
1995
−99
200
0−04
2005
−08
*Note: This figures shows the percent of the cohort thatis presently female. If female pastors from earlier cohortleft the ministry at different rates from male pastors, thenthis percentage may not reflect the initial composition ofthe cohort.
25
Cohort Differences Between EthnicGroups Comparing population pyramidsfrom pastors in each racial/ethnic categoryto that of White pastors indicates both his-torical and current trends of racial ethnicrepresentation in the clergy (fig. 20).
Out of these population pyramids, the pyramidfor Black pastors most closely resembles that forWhite pastors. This similarity probably reflectsthe length of the historical relationship betweenBlacks and the UMC. The pyramid for Blackmales is spread out over cohorts to a similar de-gree as for Whites, indicating that there are asubstantial percentage of older, Black male pas-tors. In contrast, the bulge towards the bot-tom of the figure for Black female pastors indi-cates strong, recent growth. The bulge is greaterthan that for White females, indicating greatergrowth in the number of Black females comparedto White females.
For Asian and Hispanic/Latino(a) pastors, thereis a substantially different pattern. The distri-bution of males in each cohort is much morepyramid-like, indicating that the largest cohortsof Asian and Hispanic/Latino(a) male pastorsare recent entrants, and is indicative of growthwithin these categories. This pattern is notmirrored for females, however, suggesting thatsuch growth is occurring primarily among males.While there is some growth occurring in femalecohorts for these groups, it is not proportionalto the growth in male cohorts.
It is harder to discern a pattern within the com-bined category. There is potentially a jump inpastors in these categories in the mid 80’s formales and in the mid 90’s for females. How-ever, since there are much fewer pastors in thisgroup, especially in older cohorts, it is possible
that these jumps are due to random variation.
Overall, the representation of racial/ethnic pas-tors is greater in recent cohorts than in oldercohorts (fig. 21). As much as pastors from someracial/ethnic groups have gained ground in rel-ative terms, in terms of overall composition inthe UMC, the difference across cohorts is notdramatic.
Figure 21: Percent of Cohort by Racial/EthnicGroup
020
4060
8010
0P
erce
nt o
f Coh
ort
1965
−69
1970
−74
1975
−79
1980
−84
1985
−89
1990
−94
1995
−99
200
0−04
2005
−08
*
WhiteBlackAsianHispanic/Latino(a)Combined (<1%)
Note: The percentages given are for pastors presently ineach cohort, and may not reflect the initial compositionof the cohort. The Combined group consists of NativeAmerican, Pacific Islander, and Multi-racial affiliations.
Age at Entry An increasing number of clergyare entering the ministry as a second career, andare much older than clergy who enter the min-istry soon after college (figure 22). In the 2005-08 cohort, the average age for clergy was 45 yearsold for males and 47 years old for females. In
26
Figure 20: Cohort Structure in 2008 by Racial/Ethnic group
Males
25 20 15 10 5 0
7.9
11.8
10.2
8.4
7.9
8.8
7.2
5.9
2.8
0.4
0.1
2005−08*
2000−04
1995−99
1990−94
1985−89
1980−84
1975−79
1970−74
1965−69
1960−64
<1960
Females
0 5 10 15 20 25
5.3
7.8
6.7
3.4
2.3
1.7
0.9
0.4
0
0
0
White
Males
25 20 15 10 5 0
5.5
12.4
12.7
9.8
7.2
9.4
4.6
2.7
1.2
0.2
0
2005−08*
2000−04
1995−99
1990−94
1985−89
1980−84
1975−79
1970−74
1965−69
1960−64
<1960
Females
0 5 10 15 20 25
4.8
10.8
9.2
4.8
2.5
1.5
0.6
0.1
0
0
0
Black
Males
25 20 15 10 5 0
13.2
18.5
14.7
11.9
10.3
6.4
2.8
1
0.1
0
0
2005−08*
2000−04
1995−99
1990−94
1985−89
1980−84
1975−79
1970−74
1965−69
1960−64
<1960
Females
0 5 10 15 20 25
3.3
7.4
4.4
3.7
1.1
0.7
0.6
0
0
0
0
Asian
Males
25 20 15 10 5 0
10.5
21.8
12.1
10.2
6.5
6.8
2.3
1.7
2.3
2
0
2005−08*
2000−04
1995−99
1990−94
1985−89
1980−84
1975−79
1970−74
1965−69
1960−64
<1960
Females
0 5 10 15 20 25
5.1
6.2
6.8
1.4
3.7
0
0.6
0.3
0
0
0
Hispanic/Latino(a)
Males
25 20 15 10 5 0
9.5
16.2
14.8
10.5
11
3.3
3.3
2.4
1.4
0.5
0
2005−08*
2000−04
1995−99
1990−94
1985−89
1980−84
1975−79
1970−74
1965−69
1960−64
<1960
Females
0 5 10 15 20 25
4.8
6.7
11
1
1.4
1.9
0.5
0
0
0
0
Combined (<1%)
The Combined group consists of Native American, Pacific Islander, and Multi-racial affiliations.
27
addition, the variation (standard deviation) israther large, indicating that many pastors areeither substantially older or younger than themean. While the two year difference betweenmales and females is dwarfed by the variation inages within each gender, this difference is con-sistent across cohorts, and is statistically signif-icant. 10 These results suggest that there aremore second-career pastors among females thanmales—but only to a small degree.
Figure 22: Average Age at First Appointmentby Cohort and Gender
●●
●
●
●
●
●●
●
010
2030
4050
60A
vera
ge A
ge a
t Firs
t App
oint
men
t ● FemaleMale
<19
60
1960
−64
1965
−69
1970
−74
1975
−79
1980
−84
1985
−89
1990
−94
1995
−99
200
0−04
2005
−08
*
Error bars indicate one standard deviation.
Racial/ethnic pastors in earlier cohorts areslightly older on average (though again, not bymuch) than their White counterparts (fig. 23).
10The trend in age at first appointment in this graphis exaggerated by the fact that older pastors from earliercohorts are more likely to have died or retired, so thatonly the younger pastors from those cohorts are in thepresent sample.
However, this trend disappears for more recentcohorts and, if anything racial/ethnic pastors inmore recent cohorts are slightly younger thanWhite pastors in these cohorts.
Figure 23: Average Age at First Appointmentby Cohort and Racial/Ethnic Status
●●
●
●
●
●
●
●●
●
010
2030
4050
60A
vera
ge A
ge a
t Firs
t App
oint
men
t ● NonwhiteMale
<19
60
1960
−64
1965
−69
1970
−74
1975
−79
1980
−84
1985
−89
1990
−94
1995
−99
200
0−04
2005
−08
*
Error bars indicate one standard deviation.
Cohort Definition for this Report Thissection has examined cohorts in five year group-ings in order to provide a high resolution viewof cohort structure. However, this level of detailis not needed for the rest of the report. Fur-thermore, it is desirable to separate the cohortsfor which there is only retrospective information(prior to 1997) from those where data on all pas-tors is recorded (1997 and forward). For thisreason, I will divide cohorts into 10 year incre-ments up until 1996 (< 1960, 1960-69, 1970-79,1980-89, and 1990-96), and then split the pas-tors who join in 1997 and later into two cohorts
28
(1997-2001, and 2002-2008).
29
1.7 Representation in Non-OrdainedLeadership Positions in 2001
While the main focus of this study is on theclergy, it is important to recognize that otherleadership roles are filled by laity. Lay leadershiproles provide another site at which to considergender and racial/ethnic representation. Com-pared to the statistics and data complied onclergy, little is known about lay leadership roles.This section briefly considers the representationof women and racial/ethnic members among layleadership roles. The US Congregational LifeSurvey (2001) is currently the best source of in-formation on lay leadership. This survey wasgiven to all people attending worship on the datethat the survey was administered in 177 UMCchurches. One set of survey questions asks aboutwhether the respondent is currently serving inany leadership roles.11
Table 1 presents the percent female among re-spondents who indicated they held one of a num-ber of possible lay leadership roles.
Looking first across all racial/ethnic categories,women comprised between 55-70% of lay lead-ership roles in this sample. Serving on a con-gregation’s governing body had the lowest fe-male participation rates (55%). Since womenmade up 63% of the sample, they were under-represented in this role. By contrast, women areover-represented in small-group leadership andreligious education roles.
11While the sample size for the number of surveys re-turned is large, the participation rate for congregationswas fairly low (approximately 18%). Because of the lowparticipation rate, there is a greater chance for system-atic bias than one would think based upon the samplesize alone.
Participation in lay leadership roles is lower forAsians and Hispanic church attendees than forWhite or Black attendees, mirroring the findingsfor clergy participation rates. (Note, however,that there are a large number of Asian and His-panic women, who indicate that they participatevia some other role.)
30
Table 1: Percent of Lay Leaders that are Female
Do you currently have the followingrole:
Total White Black Asian Hispanic/Latino(a)
Combined
Member of the governing board 55 55 52 14 50 36Member of a congregational commit-tee or task force
65 65 60 57 64 42
Leading or assisting in worship 61 60 67 33 56 37Officer or leader of a small group 73 73 70 50 56 46Religious education teacher 67 67 68 60 77 60Other role 65 65 62 78 78 59Total sample 63 63 60 60 65 53
Table 2: Percent of Attenders that Hold Lay Leadership Positions by Ethnicity
Do you currently have the followingrole:
Total White Black Asian Hispanic/Latino(a)
Combined
Member of the governing board 15 16 20 9 7 14Member of a congregational commit-tee or task force
19 20 19 16 10 17
Leading or assisting in worship 10 10 18 10 8 9Officer or leader of a small group 11 11 25 10 8 9Religious education teacher 13 13 13 16 13 14Other role 10 10 11 10 8 15
Note: Participation in lay leadership positions is considered regardless of whether the respondent is in a predominantlyethnic congregation or not.
31
2 Career Progression
As a normative concept, the idea of a career canbe problematic for pastors because of its associa-tion with advancement along a career ladder andwith material achievement. But career is also adescriptive concept. The main use of the termcareer in this report is in this descriptive sense—to simply recognize the fact that pastors occupya succession of positions over their life course.
However, without claiming that status shouldbe recognized along congregational size, or thatclergy should orient themselves to advancementin these terms, it is important to recognize thatthe size of a congregational appointment mat-ters greatly in determining the material circum-stances of appointments (like clergy salaries, thenature of clergy work, the visibility of leadership,etc.). So, even in a descriptive sense, it is impor-tant to track clergy careers across congregationalsize.
For the purposes of considering career equalityacross gender and race/ethnicity there is natu-rally great interest in a set of outcomes:
• Whether female and racial/ethnic pastorsare moving into larger (or smaller) congre-gations at the same rate as white or malepastors.
• Whether female and racial/ethnic pastorsare more likely to occupy associate ratherthan lead pastor roles.
• Whether female and racial/ethnic pastorsmove from parish to non-parish ministry toa larger degree than white or male pastors.
• Whether female and racial/ethnic pastors
exit the clergy at greater rates than whiteor male pastors.
This section looks at outcomes across the fullspan of clergy careers—beginning, middle, andend—in order to ask whether the experiences offemale and racial/ethnic pastors differ substan-tially from their white/male counterparts, and toassess whether gender and/or racial/ethnic bar-riers are easing.
2.1 Career Progression—OutcomesAfter Eight Years
Over the course of the study period the careeroutcomes of female and racial/ethnic pastorsended up exhibiting a subtle pattern of disad-vantage relative to white/male pastors. Thoughnot dramatic, these differences indicate persis-tant differences in the experiences female andracial/ethnic pastors.
To measure career outcomes, this study took theset of clergy serving congregations in 2000, andrecorded the job they held eight years later, atthe end of the study period in 2008. This mea-sure of career outcome provided a relatively sim-ple means of accounting for both transitions be-tween specific positions/statuses and potentialdifferences in the length of time spent in posi-tions.12
Pastors still serving congregations in 2008 were
12Analyses of transitions between jobs (for example,comparing the first, second, and third jobs held by maleand female pastors) miss an important aspect of careers–the time spent in each job. By focusing on the job heldby pastors after a set period of time, this analysis em-phasizes the overall outcomes of specific job transitions.Note, however, that pastors who entered the ministryfrom 2001-2008 are not included in this analysis.
32
further sub-classified based on that congrega-tion’s size—whether it was smaller, larger, or ofa similar size to the congregation they served in2000. A 15% difference in either direction wasused as the threshold for determining whether acongregation was similarly sized.13
Clergy serving congregations in 2000 were mostlikely to still be serving congregations in 2008(figure 24, top three cases combined). Of thesepastors, roughly equal percentages served in con-gregations that were larger or of a similar size totheir starting point. Far fewer ended up in sub-stantially smaller congregations. This generalpattern captures the reality of career progressionover time—serving in a larger sized congregationis more likely than serving in a smaller congre-gation.
Gender and racial/ethnic splits do not departdramatically from these general trends, but theirpatterns do vary in predictable ways along gen-der and racial lines. Females were less likely toend up in larger and similar-sized congregations,and more likely to end up in smaller congrega-tions than their male counterparts. The samewas also true for many racial/ethnic pastors withrespect to whites.14 Only 6-7% of pastors wereeither on leave or working in non-parish positionsby 2008. However, in both cases, female andracial/ethnic clergy were more likely than maleand white clergy (respectively) to be in these
13Non-parish positions were classified into three broadcategories: Exits, Leave, and Non-parish work. Non-parish work included district superintendents, extensionalministry, and other services at the conference or districtlevel.
14Black and Hispanic/Latino(a) pastors were actuallyless likely than White pastors to end up in a smaller con-gregation. However, they were also less likely to end upin larger or similarly-sized congregations due to their in-creased prevalence in Leave and/or Non-parish positions.
states.15
Exits from the clergy occurred frequently—about 30% of pastors serving in 2000 had ex-ited the UMC clergy by 2008. While this mayseem high, recall that the sample includes manypastors from early cohorts (and, in fact, excludesthe most recent cohort), and tracks these pastorsover a rather lengthy period of time. Femaleswere less likely to exit than males, and Asian pas-tors and pastors in the Combined racial/ethniccategory were less likely to exit than White pas-tors. In part, this overall trend is due to greaternumbers of new pastors in these categories. Sec-tion 2.3 investigates how UMC pastors leave theclergy in more detail, accounting for such demo-graphic differences.
Statistical Models of Career OutcomesBecause the career outcomes presented in fig-ure 24 can be driven by many factors, includ-ing a pastor’s cohort, it is hard to interpret thepatterns observed in career transitions. For in-stance, women could be more likely than men totransition to smaller congregations as a result ofdifferences in the percentage of women in eachcohort, rather than gender inequality. However,this turns out not to be the case—the shifts ingender and race outcomes hold, even when ac-counting for a number of other factors.
Statistical models provide one means to assesshow career outcomes differ across gender andrace for “comparable” pastors—pastors havingthe same material conditions. This section usesstatistical models to assess whether the differ-ences shown above are due to the concentrationof female and/or racial/ethnic pastors in some
15The exception is Asian pastors, who were just aslikely as White Pastors to go on leave.
33
Figure 24: 2008 Appointment Status for Clergy Serving Congregations in 2000
0 10 20 30 40 50Percent
Non−Parish
Leave
Exit
Congregation − Smaller
Congregation − Same
Congregation − Larger
Associate Pastor●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Gender
MaleFemale
●
●
●
●
●
Race\Ethnicity
WhiteBlackAsianHispanic/Latino(a)Combined (<1%)
Gray lines indicate the overall percentage of pastors found in each state in 2008. These overall averages are overlaidby the splits between genders (diamonds vs. squares) and race (colored circles). The Combined group consists ofNative American, Pacific Islander, and Multi-racial affiliations.
34
categories of the following factors: cohort, region(jurisdictional boundaries), the size of the con-gregation the pastor was serving in 2000 (mem-bership), whether the position was an associatepastor position, the congregation’s location, thepresence of a parsonage, the age of the pastorsin 2000, whether their position was full or part-time, and their status as a UMC minister (elder,local pastor, or deacon).
In order to gauge the impact of these factors, theanalysis follows these steps:
1. Create a baseline statistical model that re-produces the descriptive findings.
2. Add the factors listed above to this model.
3. Check to see if the descriptive patterns dis-appear, or are substantially altered by in-cluding the additional factors.
The particular model used in this case assessesgender and racial differences in transition ratesrelative to a reference category. Serving in asimilar-sized congregation was used as the ref-erence category in these analyses. Thus, coeffi-cients in the analysis for gender differences forthe “clergy exit” transition, take the degree towhich pastors exit the clergy relative to servingin a similar-sized congregation and report thedifference between male and female clergy forthis statistic. Details of the modeling strategyand the full results of the model are given in theAppendix.
Coefficients for the baseline and full model arereported in figure 25. Values above zero indi-cate that females have a higher exit rate thanmales. The following patterns emerge from theseresults:
1. Coefficients for the positive outcome ofserving in a larger congregation are belowzero or statistically insignificant. There isonly a small difference for females, but themagnitude of the difference between whiteand racial/ethnic pastors is greater, espe-cially when other factors are included in themodel.
2. Coefficients for the negative and neutraltransitions are generally above zero for gen-der and race, indicating that female andracial/ethnic clergy are more likely to ex-perience these outcomes than white/maleclergy. With a few exceptions, female andracial/ethnic pastors are:
• More likely to be serving in a smallercongregation or as an associate pastor.
• More likely to be on leave or move toa non-parish position.
3. These patterns increase between the base-line and the full model rather than dimin-ish. This increase indicates that correlationsbetween career outcomes and the demo-graphic/congregational attributes includedin the full model are not responsible for thepatterns described in 1 and 2.
Female and racial/ethnic pastors are less likelyto remain in parish ministry, serving in asimilarly-sized or larger congregation. They aremore likely to move into positions that carry lessof an imprimatur of leadership—smaller congre-gations, associate pastor, and non-parish posi-tions, or to go on leave. Furthermore, this pat-tern does not appear to be the result of demo-graphic factors or systematic differences in thestarting positions of male vs. female and White
35
Figure 25: Relative Change in Moving Out of a Similar-Sized Congregation for Gender and Eth-nicity
●
●
●
●
●
●Non−Parish
Leave
Exit
Associate Pastor
Smaller Congregation
Larger CongregationFemale
●
●
●
●
●
●
Black
●
●
●
●
●
●Non−Parish
Leave
Exit
Associate Pastor
Smaller Congregation
Larger CongregationAsian
●
●
●
●
●
●
−1 0 1 2 3
Hispanic/Latino(a)
●
●
●
●
●
●Non−Parish
Leave
Exit
Associate Pastor
Smaller Congregation
Larger Congregation
−1 0 1 2 3
Combined
● BaselineWith Controls
This figure shows a measure of how much more likely female and racial/ethnic clergy are to move from a base-linecategory (serving a similar-sized congregation) to each of the job categories compared to male pastors (for gender)and White pastors (for race/ethnicity). For example, the large positive value for associate pastors in the Femaleresults indicates that females were far more likely than males to be serving as an associate pastors instead of in aSimilar-Sized congregation. The specific measure of likelihood is the multiplicative difference in the odds of servingin a similar-sized congregation. The Combined group consists of Native American, Pacific Islander, and Multi-racialaffiliations.
36
vs. racial/ethnic pastors. Since the most re-cent cohort available for inclusion in this anal-ysis was the cohort of pastors entering in 2000,this analysis does not capture the most recenttrends in appointment patterns; rather, it revealswhat the historical experience of many female orracial/ethnic pastors has been in the UMC.
37
2.2 First Appointments
The previous section considers gender andracial/ethnic differences as they develop overtime across careers, but does not consider howmuch inequality is present at the beginning ofcareers. This section takes up the theme of ini-tial inequality by examining first appointments.Are female and racial/ethnic pastors assigned todifferent types of positions in their first appoint-ments than white and/or male pastors?
Appointment Status The pathway that pas-tors take into the clergy is captured (in part) bythe status of their first appointment. Pastorswho enter as elders following seminary educa-tion are following the traditional, and in manyways, normative path to achieve full-status mem-bership in the contemporary United MethodistChurch. Yet the tradition of appointing lo-cal pastors provides an alternative pathway, onewhich is increasingly used to fill positions insmall congregations.
Local pastors are not full participants in itiner-acy however, and their first appointment is alsolikely to be their last. In combination, the in-creasing number of local pastor appointmentsand their short duration, imply that a large per-centage of the pastors appointed to a congrega-tion for the first time are likely to be local pas-tors. Full-time local pastors comprised 23% ofthe first appointments made between 2000 and2008, while other local pastors comprised 51% offirst appointments (51%; figure 26).
Elders made up a minority of first appointments(17%; figure 26). But, because they are morelikely to be reappointed multiple times and tohave longer careers as UMC clergy, the distrib-
utor of appointment statuses is reversed whenall appointments are considered. Overall, el-ders served more than 70% of congregations andmulti-congregation charges in the study, eventhough they were not the majority of first ap-pointees (see figures 13 and 14).
The high representation of local pastors amongfirst-appointees is a trend that is worth high-lighting. However, because elders are more likelyto continue on to long-term careers as UMC pas-tors, gender or race based differences in the firstappointments of elders will potentially have agreater impact for ongoing inequality than willdifferences in the appointments of local pastors.For this reason, this report will examine thecharacteristics of first appointments across bothgroups: elders and local pastors.
There are substantial gender and racial/ethnicdifferences in first appointment status. For race,the most notable trends—the high proportion ofAsian pastors who hold elder positions as firstappointments, and the high proportion of His-panic/Latino(a) pastors who hold full-time lo-cal pastor positions—are similar to overall dif-ferences in appointment status (figure 14). How-ever, the difference in male and female represen-tation among first appointments shows the re-verse of the overall pattern. Overall, there arefewer female elders than male elders, and morefemale local pastors (figure 13). However, amongfirst appointees, there are more female than maleelders, and fewer local pastors (and fewer pastorsin the positions classified as other; figure 26).
Appointment Characteristics We have al-ready seen that females are more likely thanmales to be associate pastors (figure 6), so itshould be no surprise that this trend begins with
38
Figure 26: Status at First Appointment
0 10 20 30 40 50 60 70Percent
Total
Other
Deacon
Other Local Pastor
Full−time Local Pastor
Elder●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Gender
MaleFemale
●
●
●
●
●
Race\Ethnicity
WhiteBlackAsianHispanic/Latino(a)Combined
This chart shows the appointment status at first appointment for all pastors whose first appointment took placebetween 2000 and 2008. Gray lines indicate the overall percentage of pastors found in each category. These overallnumbers are overlaid by the percentage within each gender (diamonds vs. squares) and the percentage within eachracial/ethnic group (colored circles). The Combined group consists of Native American, Pacific Islander, and Multi-racial affiliations.
39
Figure 27: Characteristics at First Appointment
32%
37%
32%
30%
45%
45%
40%
42%
Combined
Hispanic/Latino(a)
Asian
Black
White
Female
Male
TotalAssociate Pastor
71%
63%
61%
48%
64%
57%
68%
63%
Combined
Hispanic/Latino(a)
Asian
Black
White
Female
Male
TotalIn Parsonage
84%
84%
89%
88%
95%
93%
95%
94%
Combined
Hispanic/Latino(a)
Asian
Black
White
Female
Male
TotalFull−time
45%
73%
73%
81%
51%
58%
54%
56%
Combined
Hispanic/Latino(a)
Asian
Black
White
Female
Male
TotalUrban Location
0 20 40 60 80 100
25%
29%
39%
19%
20%
26%
18%
21%
39%
49%
45%
22%
44%
37%
46%
43%
36%
51%
55%
37%
48%
44%
49%
47%
45%
62%
69%
47%
27%
34%
28%
30%
0 20 40 60 80 100
A. Elders Only B. Local Pastors*
This chart shows the percentage of first appointment elders (A) and local pastors (B) having each of the followingcharacteristics: holding an associate pastor position, appointments with parsonages, holding a full-time appointment,being appointed to an urban location (see glossary for definitions). The local pastor category contains both full-timelocal pastors and other local pastors.
40
the first appointment. Associate pastor positionsare frequently treated as entry-level positions—42% of elders that entered the clergy during thestudy held associate pastor positions as theirfirst appointments. Among elders, females weremore likely to hold an associate pastor position,but the difference is relatively mild (45% for fe-males vs 40% for males; figure 27). Racial/ethnicelders, on the other hand, were less likely to holdassociate pastor positions as a first appointment.This percentage ranged between 30% and 37%for racial/ethnic pastors.16
local pastors still hold associate pastor posi-tions for first appointments, but at a much lowerrate—only 21% of non-elders held an associatepastor position for their first appointment. Gen-der differences in this rate were more notablefor local pastors than for elders. Among localpastors, 26% of females held an associate pas-tor position compared to just 18% of males. So,not only are females less likely to be a local pas-tor in the first place, they much less likely to bea local pastor who is the sole/lead pastor of acongregation.
Most elders (94%) held full-time positions at en-try, and there was little difference between males(95%) and females (93%), despite the perceptionthat female pastors may be more likely to holdpart-time roles.17 When all first-appointees are
16As with females, the greater percentage of whites thatare local pastors moderated this effect when examiningthe full sample. Whereas 27% of all entering white pas-tors held associate pastor positions, entering Black pas-tors were slightly less likely to hold associate pastor posi-tions (21%) while entering Asian and Hispanic/Latino(a)pastors were slightly more likely to be associate pastors(34% and 32% respectively).
17Note, however, that part-time pastors are likelyunder-represented in this dataset. See the Methodologysection for details.
considered, the percentage of full-time pastors ismuch lower (58%), but there is still no differencebetween male and female pastors. Racial/ethnicelders were less likely than white pastors to holda full-time position at entry (84-89% vs. 95%),but the racial/ethnic differences are not consis-tent when all pastors are examined.
Female entering pastors were more likely to bein an urban congregation (58% vs. 54%, elders;43% vs. 33%, all pastors), and less likely to livein a parsonage (57% vs. 68%, elders; 40% vs.49%, all pastors). Entering racial/ethnic pastorswere also more likely to be in an urban locationwith the exception of the combined category (73-81% vs. 51%, elders; 56-73% vs. 33%, all pas-tors), but there was no clear trend for living ina parsonage.
Congregational Size Pastors often move intolarger congregations after they accrue more ex-perience and seniority. But are there any genderor race-based differences in the size of congrega-tions to which they are first assigned? If futureassignments are based (in part) on the size ofa pastor’s current congregation, then such dif-ferences could serve to initiate long-running in-equality.
The short answer is that there is little differencebetween the size of first appointments for gen-der, and substantial differences for race. Fig-ure 28 shows the results of two statistical re-gression models. The baseline model providesbasic descriptive statistics about the differencein appointment size for race and gender, whilethe second model accounts for differences in sizeassociated with appointment status and ethniccongregation status.
41
Figure 28: Impact of Race and Ethnicity on Size of First Appointment
−80 −60 −40 −20 0 20Percentage Difference in Membership
Combined (<1%)
Hispanic/Latino(a)
Asian
Black
Female ●
●
●
●
●
Lead/Sole Pastors
● BaselineWith Controls
−80 −60 −40 −20 0 20Percentage Difference in Membership
●
●
●
●
●
Associate Pastors
The Combined group consists of Native American, Pacific Islander, and Multi-racial affiliations.
For both senior/sole pastors and associate pas-tors, the average membership of the first ap-pointment of females is, at most, just slightlyless (2-5%) than for males. By contrast, seniorand sole racial/ethnic pastors hold appointmentswhose average size ranges from 10-25% less thanthat of White pastors, while the gap is evengreater for associate pastors. Furthermore, inonly one case, Black pastors, is this gap moder-ated by accounting for ethnic congregation andappointment status. Ethnic congregations aresmaller than primarily White congregations, sothe pattern observed for Black pastors is one pos-sible outcome—after accounting for the smalleraverage size of ethnic congregations, the first ap-pointments of Black pastors are actually slightlylarger than for White pastors. However, the sizegap for Asian and Hispanic/Latino(a) pastorsare similar or even greater after accounting forthese variables, suggesting that size differencesare more deeply rooted for these groups.
42
2.3 Rates of Leaving
The level of racial/ethnic groups and femalesin the clergy is based upon a combination oftwo factors—entry rates and exit rates. Clergyburnout is notoriously high, and it is not unrea-sonable to suppose that if female or racial/ethnicclergy face a less receptive environment thanwhite/male clergy, they will leave at greaterrates. Additionally, child-raising may result inhigh rates of departure for female vs. maleclergy.
Do rates of leaving reflect these hypotheses? Inorder to address this question, the report ex-amines both direct and indirect exits from theclergy. Direct exits are transitions that directlyresult in the departure of a pastor from the clergy(table 3). These include death, retirement, andtransfers to other denominations. Indirect exitsoccur when pastors leave parish ministry for anon-parish status and then leave the clergy fromthis non-parish status (through one of the directexit events). Clergy who remain in non-parishministry, say as a seminary professor, are notcounted as exiting the clergy. Counting indirectexits recognizes the fact that ceasing to serve inparish ministry is often a precursor to leaving theclergy. The Appendix contains a more completedescription of the methodology.
Table 3 shows a reduced set of “Service Type”codes that describe the last recorded appoint-ment to a congregation for pastors in the study.Not surprisingly, Retired and Deceased events to-gether comprise a large proportion of pastoralexits from the clergy. The other large category,Discontinued, includes pastors who voluntarilydepart from the UMC clergy.
For transitions that result in an indirect exit,
both the total number of transitions are shownas well as the percent that have been classifiedas exits. This percentage indicates how manypastors left the clergy via a recorded exit transi-tion at a later date, without returning to a con-gregation in between. There are high numbersof pastors that leave the clergy without return-ing to serve a local congregation from nearly allof these indirect exit categories, reinforcing theirimportance as pathways for leaving the ministry.
The largest category of indirect exits, the Activeservice type, refers to extension appointmentsbeyond the local church. Nearly 30% of pastorswho transitioned from parish to extensional min-istry then left the clergy.18
Descriptive Statistics on Rates of LeavingIn order to summarize differences in leaving theclergy between males and females, annual ratesof leaving can be examined directly. But it isalso helpful to take a step back and consider theproportion of pastors that remain clergy aftersome number of years (through what are knownas “survival curves”). The former is better suitedto understanding how a pastor’s chances of leav-ing the clergy in any one year change over timeand/or differ between male and female clergy,while the latter provides a better means to assessthese differences in terms of the overall compo-sition of clergy.
Both cases suffer from the limitation mentionedin the introduction to the study—namely thatwith a 12 year study-window, pastors from ear-lier cohorts have higher seniority. Thus, it can-not be said with any certainty whether the differ-
18Note that these percentages actually understate thenumber of such indirect exits, since some pastors wouldhave exited the ministry after the study window ended.
43
Table 3: Direct and Indirect Exit Events
Number of Events Number of Exits Percent Exited
Direct Exit EventsRetired 7, 136 7, 136Discontinued 1, 777 1, 777Deceased 525 525Surrender Credential 338 338Withdrawn 315 315Withdrawn-Other Denomination 168 168Honorable Location 87 87Terminated 55 55
Indirect Exit EventsNon-parish Appointment 2, 753 775 28Leave∗ 1, 346 626 47No Record of Appt 1, 243 519 42Disability-Cpp 975 389 40Special Appointment 230 43 19Attend School 183 49 27Sabbatical 100 41 41
The Number of Events column counts the number of times that this code is associated with the last recorded moveout of a congregation. The Number of Exits column counts the times that an exit event is subsequently recorded forthis pastor. In other words, cases where a pastors goes on family leave and then is later reappointed to a congregationare not recorded in these counts.∗ Leave categories include: Educational, Family, Incapacity, Involuntary, Personal, and Transitional.
44
ences between male and female pastors are asso-ciated with their entering the clergy in differingcohorts or due to changes in the occupationalchances that arise once a certain level of senior-ity is reached.
2.3.1 Female Clergy
Annual Rates of Leaving As a first look athow pastors leave the clergy, I examined an-nual rates of leaving for a three year windowin the data (2006-2008), grouping them by sexand years of seniority (fig. 29). For male pas-tors with less than 30 years of tenure leavingrates are fairly constant— about 4.5% leave theclergy per year regardless of seniority. Above 30years, leaving rates shoot up dramatically, dueto retirement and death.19 In many other linesof work, new-entrants are at considerably morerisk of leaving a field, leading to higher rates ofleaving at low tenures. At present, this appearsnot to be the case for UMC clergy—leaving ratesfor current entrants are not any higher than forestablished pastors.
But another way of viewing this pattern is that4.5% is a rather high annual rate of leaving, andthat the consistency of this rate over all levelsof seniority reinforces the subjective view thatthe clergy is a difficult calling, where burnout iscommon.
Leaving rates for female pastors also shoot updramatically at 30 years, while remaining rela-tively level below this mark. There are sometrends, however, that are not seen for male pas-tors. Among older cohorts, rates of leaving arehigher for females than they are for males. Fe-
19In fact, leaving rates continue to increase beyond thebounds of the graph, up to 55 years.
male pastors who had between 10 and 30 years ofseniority in 2006-08, left the clergy at a greaterrate than male pastors (∼5.5% per year). Theselevels of seniority correspond to the cohorts offemale pastors who entered the clergy from 1976-1998. However, female pastors with 0 to 10 yearsof seniority were less likely than male pastors toleave the clergy (∼3.0% per year).
The Proportion of Pastors Remaining inthe Clergy Over Time How do these ratesof leaving matter over time? Do the differencesbetween sexes greatly affect the overall compo-sition of the ministry, or not much? Survivalcurves, which show the proportion (i.e. percent-age) of pastors who have remained clergy at leastY years following the beginning of the time pe-riod are better suited to answering these ques-tions. Survival curves also add an additional el-ement to the analysis. While the rates of leav-ing represent a snapshot taken across a cross-section of seniority/cohort levels for one momentin time, survival curves trace the pattern of leav-ing over time within cohorts.
The survival curves (figure 30) illustrate the cu-mulative toll that annual rates of leaving takeon the ranks of the clergy. Since annual rates ofleaving are similar across seniority levels, the sur-vival curves for the 1970-79 cohort and above arerelatively similar—at the 10 year mark, ∼ 40%of clergy present at the beginning of the periodhave exited the clergy, with ∼ 60% remaining.Since it is likely that each cohort experiencedbroadly similar rates of loss for each preceding10 year period, these departure rates from theclergy seem large indeed.
Against this broader backdrop, the differencesbetween male and female pastors are more sub-
45
Figure 29: Annual Rate of Leaving the Clergy for 2006-2008 by Gender
05
1015
20A
nnua
l Exi
t Rat
e (P
erce
nt)
0 10 20 30 40(2006−2008) (1996−1998) (1986−1988) (1976−1978) (1966−1968)
Years of Seniority (Year of First Appointment)
FemaleMale
This figure shows the annual rate at which pastors left the UMC clergy for the years 2006-2008. Rates of leaving arecalculated separately by year of seniority (points). A pastor’s years of seniority correspond closely to their cohort.Corresponding three-year cohort windows are listed beneath the years of seniority.
46
Figure 30: Proportion of Clergy in Ministry for at Least Y Years by Genderyl
im0.
00.
20.
40.
60.
81.
0 <1960
ylim
0.0
0.2
0.4
0.6
0.8
1.01960−69
ylim
0.0
0.2
0.4
0.6
0.8
1.0 1970−79
ylim
0.0
0.2
0.4
0.6
0.8
1.01980−89
ylim
0.0
0.2
0.4
0.6
0.8
1.0 1990−96
ylim
0 2 4 6 8 10 12
0.0
0.2
0.4
0.6
0.8
1.01997−01
ylim
0 2 4 6 8 10 12
0.0
0.2
0.4
0.6
0.8
1.0 2002−08
Pro
port
ion
of P
asto
rs
Year
FemaleMale
This figure shows “survival curves” for pastors in each cohort. Survival curves plot the proportion of pastors who arestill active clergy over the passage of time. For the two cohorts after 1997, seniority is used directly as the measure oftime. For cohorts prior to the study’s start date (1997), these curves are conditioned upon survival until 1997—thatis, they show the departure rates only for pastors who have remained in the ministry until 1997. The passage of timeis captured as the number of years since 1997. The Combined group consists of Native American, Pacific Islander,and Multi-racial affiliations.
47
tle. Just as for annual rates of leaving, femalepastors do better than male pastors in recentcohorts, and worse in older cohorts. The cu-mulative impact of differential leaving rates islargest in the 1970-79 and 1980-89 cohorts. Forthe 1980-89 cohort, there is an 11 percentagepoint difference in the percent of pastors remain-ing in the clergy at 10 years (57% of females and68% of males). By contrast in the 2002-08 co-hort, the difference (at 5 years) is 5 percentagepoints in the other direction (87% of females and82% of males).
Despite the differences between the results fromleaving rates and survival curves, the main sub-stantive findings are relatively similar. Femalepastors in recent cohorts are less likely to leavethe clergy than their male counterparts, whilefemales in older cohorts are more likely to leave.The difference between males and females is notdramatic, but it does add up over time to im-pact the composition of pastors. Apart from thecross-over between leaving rates of males and fe-males, the differences between cohorts are rel-atively uninteresting—leaving rates are remark-ably similar across cohorts/seniority levels untilretirement becomes a major factor.
Factors Associated with High Rates ofLeaving The results of the descriptive analysisindicated that women who entered the clergy in1997 and later are less likely to leave the ministryin a given year than males. This section asks howdeparture rates vary across regions (UMC Juris-dictions), whether the types of positions that fe-male pastors are assigned to can account for theirlower departure rates, and whether the rate atwhich females leave the clergy is greater in somepositions than in others.
Multivariate models were used to model leavingrates in each cohort. The cohorts were modeledseparately, but the results are presented togetherin table 4 for ease of comparison. Cohorts priorto 1970 are not modeled because leaving ratesare dominated by retirement, and our interestlies primarily in understanding why pastors leavethe clergy prior to retirement.
The coefficients in these models indicate the pro-portional change in the rate of leaving associatedwith various predictors. Coefficients that are notstatistically significant are omitted from the ta-ble. The table presents two sets of models—abaseline set, with only gender and race presentin addition to seniority, and a full model wherea number of predictors are included simultane-ously.
The baseline model essentially reproduces thedescriptive analysis presented above, and the re-sults reflect this design. For older cohorts (1970-79 and 1980-89), the annual rate at which fe-males left the clergy was higher than for malesacross the study period. In the 1990-96 cohort,the gender difference is statistically insignificant.For cohorts since 1996 females leave the clergyat a lower rate than males.
The shift in the coefficients between the baselineand full model is highly revealing. In the fullmodel, the Female coefficients for the 1970-79and 1980-89 cohorts are greatly reduced, to thepoint where the 1970-79 coefficient (0.16) is nolonger statistically significant, while the 1980-89coefficient (0.11) is less than a third of what itwas in the baseline model. This reduction indi-cates that female pastors from these cohorts weremore likely to occupy positions in which pastorsare more likely to leave the ministry. Looking atthe other coefficients in the full model, we can
48
Table 4: Proportional Change in Rates of Leaving
1970-79 1980-89 1990-96 1996+
Baseline ModelFemale 0.25 0.46 −0.23Race/Ethnicity (White Omitted)
Black 0.81 0.37Asian 1.99 0.39 −0.36 −0.46Hispanic/Latino(a) 1.01Other
Full ModelFemale −0.19
Black −0.26AsianHispanic/Latino(a)Other
Jurisdiction (Midwest Omitted)Northeastern −0.17 −0.16 −0.15South CentralSoutheastern −0.16Western −0.43
Age (10’s of Years) 20.30 6.25 2.69 0.55Associate Pastor −0.56 −0.41 −0.29 −0.26Log of Membership −0.15 −0.10Part-time Status (Full Time Omitted)
Three Quarter Time 1.00 0.89 0.70Half Time −0.29 −0.17Quarter Time
Appointment Type (Elder Omitted)Full-time Local Pastor 0.65 0.58 0.68 1.40Part-time/Other Local Pastor −0.46 0.38 1.23Deacon 6.38 1.55Other Appointment Status 0.34 2.65
Non-ethnic Congregation 0.49
49
see that these positions are: age, membership,part-time status, and full-time local pastor sta-tus. Thus, female pastors in these cohorts werein less advantageous positions and left the clergyat a higher rate because of this.
By contrast, there is little difference between thebaseline and full models for the coefficients for fe-males in the 1996+ cohort, suggesting that hold-ing more advantageous positions does not ex-plain the reduced rate of female departures inthis cohort.
The predictors of leaving rates are generally sim-ilar across cohorts. Associate pastors and pas-tors at larger congregations are less likely toleave the clergy, while part-time and non-elderpastors are more likely to leave the clergy. Olderpastors are also more likely to leave the pas-torate, and age has the biggest impact in oldercohorts, where pastors are closest to retirementage. While the coefficients for jurisdiction arestatistically significant in a number of cases, itis difficult to discern a coherent story for thesedifferences.
There is little evidence that associate pastorstatus, congregational membership, jurisdiction,appointment category, or ethnicity are associ-ated with leaving rates in different ways formales and females. The potential for such dif-ferential associations was tested by including in-teraction terms between gender and each factor.Most of these interaction terms were statisticallyinsignificant and the few that were not did notfollow a pattern.
2.3.2 Racial/Ethnic Clergy
Annual Rates of Leaving The annual ratesof leaving for racial/ethnic pastors are summa-rized in figure 31. Because of the smaller sam-ple size for racial/ethnic pastors, rates of leav-ing were calculated in five-year increments, andit was only possible to include the two largestracial/ethnic groups: Black and Asian pastors.
As with female pastors, there is little differencein rates of leaving among recent cohorts, butracial/ethnic pastors from earlier, high senioritycohorts are leaving at a higher rate than Whitepastors from these cohorts (fig. 31).
The Proportion of Pastors Remaining inthe Clergy Over Time The consequencesof higher rates of leaving in earlier cohortscan be seen by comparing the proportion ofracial/ethnic pastors remaining at the end of thestudy period to the same proportion for Whitepastors (fig. 32). In recent cohorts, there is lit-tle consistent difference in this proportion. How-ever, for the 1980-89 cohort and, in particular,the 1970-79 cohort, more racial/ethnic pastorsleft the clergy than White pastors. Only half ofAsian and Hispanic pastors from these cohortsremained in the clergy in 2008, compared to over70% for White pastors in this cohort.
Factors Associated with High Rates ofLeaving The statistical model confirms thepattern seen in the descriptive statistics. In thebaseline model (table 4), Black, Asian, and His-panic/Latino(a) pastors from the 1970-79 and1980-89 cohorts are more likely to leave theclergy than are White pastors from these co-horts. However, when examining the full model,
50
Figure 31: Annual Rate of Leaving the Clergy for 2006-2008 by Race/Ethnicity of Pastor
05
1015
20A
nnua
l Exi
t Rat
e (P
erce
nt)
0 10 20 30 40(2006−2008) (1996−1998) (1986−1988) (1976−1978) (1966−1968)
Years of Seniority (Year of First Appointment)
WhiteBlackAsian
This figure shows the annual rate at which pastors left the UMC clergy for the years 2006-2008. Only Black andAsian clergy had enough sample size for inclusion in the graph. Rates of leaving are calculated separately by yearof seniority (points). Because of the narrow time frame for the data, a pastor’s years of seniority correspond closelyto their cohort. The corresponding three-year windows of start dates are listed beneath the axis for seniority. Trendlines were calculated using the lowess smoother function.
51
Figure 32: Proportion of Clergy in Ministry for at Least Y Years by Race/Ethnicityyl
im0.
00.
20.
40.
60.
81.
0 <1960
ylim
0.0
0.2
0.4
0.6
0.8
1.01960−69
ylim
0.0
0.2
0.4
0.6
0.8
1.0 1970−79
ylim
0.0
0.2
0.4
0.6
0.8
1.01980−89
ylim
0.0
0.2
0.4
0.6
0.8
1.0 1990−96
ylim
0 2 4 6 8 10 12
0.0
0.2
0.4
0.6
0.8
1.01997−01
ylim
0 2 4 6 8 10 12
0.0
0.2
0.4
0.6
0.8
1.0 2002−08
Pro
port
ion
of P
asto
rs
Year
Combined (<1%)Hispanic/Latino(a)AsianBlackWhite
This figure shows “survival curves” for pastors in each cohort. Survival curves plot the proportion of pastors who arestill active clergy over the passage of time. For the two cohorts after 1997, seniority is used directly as the measure oftime. For cohorts prior to the study’s start date (1997), these curves are conditioned upon survival until 1997—thatis, they show the departure rates only for pastors who have remained in the ministry until 1997. The passage of timeis captured as the number of years since 1997.
52
these trends become statistically insignificant forthe 1970-79 and 1980-89 cohorts. This shiftsuggests that the higher leaving rates amongracial/ethnic pastors in early cohorts is due todifferences in the types of positions occupied bythese pastors, and/or their age relative to Whitepastors. Thus, this study provides little evidenceto suggest that racial/ethnic pastors are experi-encing greater rates of burnout, even in earliercohorts where there is an observed difference inrates of leaving.
53
A Glossary
Clergy elders, deacons, and local pastors affil-iated with the UMC, regardless of whetherthey are actively serving a congregation, onleave, or in a non-parish appointment.
Cohort Clergy who received their first congre-gational appointment during the same timewindow.
Congregational Size The membership of acongregation at which a pastor is appointed.If the pastor is appointed to a multi-congregational charge, then it is the to-tal membership of all congregations in thecharge.
Location Urban,Suburban,Large-Town, andRural. These designations were as-signed based on census tract, using RuralUrban Commuting Area (RUCA) des-ignations, http://www.ers.usda.gov/
Data/RuralUrbanCommutingAreaCodes/.The four-tier classification is based on asuggested simplification of RUCA codesgiven at http://depts.washington.edu/
uwruca/ruca-uses.php.
Multi-Congregation Charge A pastoralcharge that consists of multiple congrega-tions.
Pastor The appointed leader of a congregationor multi-congregation charge.
Seniority The number of years from a clergy’sfirst appointment to a congregation.
B Data
The data used in this report combines dataon attributes and annual statistics of congrega-tions maintained by the General Council on Fi-nance and Administration (GCFA), with data onclergy appointments maintained by the GeneralBoard of Pensions (GBOP). The data was ini-tially assembled to produce a report on clergysalaries, Salaries for United Methodist Clergy inthe US ContextQuantitative Analysis. For thepresent report, this data set has been extendedto include GBOP data on clergy appointmentsstatus beyond the local church. See the Ap-pendix of the salary report for a list of availablevariables and a discussion of missing values.
The GCFA data tracks information reported an-nually by congregations. Smaller congregationsare often grouped together under one pastoralcharge, forming multi-congregation charges. Be-cause pastor appointments can include multi-ple congregations, GCFA data need to be ag-gregated in order to reflect the characteristicsof pastoral charges rather than congregations.GCFA data contain a number of variables withinformation on specific congregations, includingcharge-church alignments. Charge-church align-ments extend only to 1997, thus setting thestarting point of this study. The GCFA alsotracks information about the pastors assignedto local churches each year including: gender,race, birthdate, appointment status, and contri-butions made by the congregation for the pas-tor’s salary and housing expenses.
The GBOP tracks people who are enrolledin the UMC pension system based upon re-ports provided by conference administrative of-ficials. Most, but not all, people who lead
54
UMC churches fall into this category. Excep-tions include: ministers from other denomina-tions, unpaid local pastors, some part-time pas-tors, ministers who have retired, and supply min-isters. Because of these exclusions, the GCFAdata provides a more complete sampling frame-work. However, for the pastors it does follow,the GBOP holds data on appointment historiesand the level of part-time work. Data on centralconference pastors is not available across the en-tire period, so this report focuses on the US.
I used GCFA records to establish the samplingframe of US pastoral charges that existed from1997-2008 (i.e. the universe of which potentialappointments), and then matched GBOP ap-pointment and salary information to the chargesin this sampling frame. Finally, I used theGBOP’s historical record of appointments togenerate statistics on the career path of pas-tors (e.g. seniority). This procedure exploitedthe strengths of both datasets. The GBOP doesnot collect information on the race/ethnicity ofpastors, so GCFA records were used for this at-tribute. When there were multiple congregationsin a pastoral charge, GCFA information on thelocal churches was aggregated to the charge levelfor analysis, and the location was based on thatof the largest church in the charge.
Each row in the resulting dataset contains infor-mation for one pastor in one year. Since GCFAdata is reported on a calendar year basis, chargeswere matched to pastor information at the end ofthat calendar year. For example, a pastor whosesalary changed in June of 2006 from $35,000 to$38,000 would have $35,000 recorded for 2005and $38,000 for 2006.
I combined conferences that merged during thestudy period, so that conference boundaries were
the same across the entire period. The mergedconferences are: Greater New Jersey, Arkansas,and Missouri. The following missionary confer-ences had very few pastors and were pooled intoa single category, “Missionary Conferences”, foranalysis: Oklahoma Indian Missionary, Red BirdMissionary, and Alaska Missionary.
C Methods
C.1 Statistical Models of Having a Fe-male or Racial/Ethnic Pastor
Logistic regression was used to provide de-scriptive statistics for how often females orracial/ethnic clergy are appointed to congrega-tions. Multi-level models were used to accountfor, and describe, differences in representationrates between conferences. A number of modelswere fit in order to accomodate different analyticgoals. The descriptive results in the paper arebased on models with few explanatory variables,so that logistic regression is being used to mea-sure differences (i.e. between regions, appoint-ment statuses, and annual conferences). Vari-ables were added to these models in order toinvestigate potential causes of gender and racegaps in representation. When the added vari-ables are also correlated with the frequency ofhaving a female or racial/ethnic clergy, then theinitial difference can change providing insightinto the source of gender and race gaps.
Results from these models can be summarizedin two ways. First, they can be used to calcu-late the percent female (or racial/ethnic) undervarious conditions. Second, the impact of vari-ous factors can be presented in terms of coeffi-cients (often referred to as effects) based on how
55
much they shift these percentages with respectto a baseline category. The exponent of logis-tic regression coefficients is interpretable as thechange in odds of an event occurring. While oddsare frequently associated with games of chance,in this case they merely represent an alterna-tive way of measuring the frequency with whichracial/ethnic or female clergy occur with respectto white/male clergy. Unfortunately, however,the change in odds is not the same as the per-cent differences in the percentage of pastors thatare female, which is a more intuitive statistic formost people.
C.2 Statistical Models of Career Pro-gression
The models used to represent career progressionare multinomial logit models where the data con-sists of pastors serving at congregations in 2000,and the outcome is the job status in 2008. Oneimportant aspect of this strategy is that individ-ual job transitions between 2000 and 2008 arenot considered, only outcomes after 8 years.
I present two models, a baseline-model, with noother variables besides gender and race, and afull-model with all of the factors listed aboveincluded. The baseline model shows how thesimple descriptive transition percentages, shownabove in figure 24, translate into the modelframework, and indicate whether these differ-ences are statistically significant.20 Since thereference category is serving in a similar-sizedcongregation, to interpret the model coefficientsfirst note that women are less likely than mento serve in similarly-sized congregations. The
20Comparing the model results with this figure is a goodway to understand what the model indicates.
multinomial model results tell us whether thegender gap is similar for other outcomes. So,in the model results (table 5), for females wesee that two transitions with similar-sized gen-der gaps (serving in a larger congregation andexiting the clergy) are statistically insignificant.The cases where there are positive, significant co-efficients indicate the change in the gender gap.
Model results are presented in table 5. Figure 25graphs ecoefficient − 1. This indicates the factordifference in the odds of males versus females ex-periencing that outcome relative to the referenceoutcome.
C.3 Congregational Size at First Ap-pointment
The models of congregational size (membership)at first appointment looks only at the size of thecongregation or multi-congregation charge thatclergy were assigned to in their first appoint-ment. The model uses standard linear regres-sion with the log of congregational size as theoutcome variable.
C.4 Rates of Leaving
Capturing departures from the ministry is trick-ier than it might first appear. There are a num-ber of relatively unambiguous “exit” transitionsthat directly follow an appointment to a congre-gation. Pastors do re-enter the clergy followingsuch events—for instance, pastors who are dis-missed can be reinstated; and those who leavefor other denominations may return. However,with one exception (retirements) re-entry is un-common and the decision to leave the clergy is,in any case, significant, even when it turns out
56
to be temporary. For these reasons, this studyexamines only the time until a pastor’s first exitevent (ignoring any subsequent re-entries).
Retired pastors do frequently return to the pul-pit. The decision to treat retirement like otherexit events, counting the date of retirement asan exit from the clergy is, in part, a prag-matic decision. The data on pastoral appoint-ments following retirement is less reliable thanfor appointments prior to retirement. In addi-tion, however, retirement marks a substantialchange in the relationship between pastors andthe denomination—most notably, the end of itin-eracy. Retired UMC pastors can choose the ap-pointments they will participate in (if any), andare more likely to engage with congregations ona part-time basis, and to serve in smaller and/orless-demanding positions than prior to retire-ment.
In addition to explicit exit events, there are alsotransitions that, while not exits themselves, aremore likely to lead to an exit event. For in-stance, pastors who are struggling may shift tonon-parish work, or take extended leaves, andthen exit the clergy from these positions. Ignor-ing such indirect exits would miss an importantpathway through which pastors leave the clergy;but both the uncertainty of whether pastors willreturn to parish ministry and the delay betweenthe last congregational appointment and leav-ing the clergy serve to create problems for anal-ysis. Transitions to non-parish work are ex-plored in more detail in section 2.1. For the pur-poses of providing insight into clergy exit rates,this section focuses on transitions to non-parishwork as “indirect” exit events. Transitions fromparish ministry to non-parish work or leave sta-tus are counted as exit events when (A) the pas-
tor does not return to parish work during thestudy period, and (B) there is an exit transitionwithin the study period. Effectively, this defi-nition counts the exit day of pastors as the dayon which they leave a parish appointment ratherthan the day they leave the clergy. This pro-cedure introduces a minor source of error; sinceindirect exits are more likely to be unobservedclose to 2008 (the end of the study window), ex-its are undercounted.
57
Table 5: Multinomial Regression Model Results—Transition from Position Held in 2000 to 2008(Relative to a Same-Sized Congregation)
LargerCongre-gation
SmallerCongre-gation
AssociatePastor
Exit Leave Non-Parish
Baseline ModelConstant -0.13*** -0.72*** -2.19*** 0.32*** -1.46*** -1.52***Female 0.02 0.65*** 1.39*** 0.07 0.96*** 0.64***Race/Ethnicity (White Omitted)
Black -0.07 0.13 -0.60*** 0.07 0.14 0.47***Asian 0.03 0.21 0.14 -0.22 -0.07 0.49***Hispanic/Latino(a) 0.00 0.14 0.59 0.20 0.60* 0.98***Other -0.15 0.41 0.36 0.02 0.49 0.23
Full ModelConstant -0.21* -0.63*** -0.96*** -0.67*** -0.79*** -0.87***Female -0.07 0.52*** 1.92*** 0.16* 1.23*** 1.23***
Black -0.26*** 0.38*** -0.38* -0.07 0.00 0.95***Asian -0.34*** 0.19 0.42 -0.33* -0.38* 0.88***Hispanic/Latino(a) -0.42* 0.39 0.80 -0.07 0.12 2.06***Other -0.20 0.30 0.07 0.05 0.27 0.46
CohortPre-1960 Cohort 0.08 -0.88 0.77 1.86* 0.40 3.44*1960-69 Cohort -0.27* -0.51*** -0.02 1.69*** -0.09 1.77***1970-79 Cohort -0.13 -0.47*** -0.23 0.15 -0.24* 0.72***1980-89 Cohort 0.00 -0.27*** -0.23* 0.13 -0.10 0.04Jurisdiction (Midwest Omitted)
Northeastern 0.00 -0.11 -0.56*** -0.12 0.08 -0.13South Central 0.70*** -0.03 0.43*** 0.11 0.39*** 0.34***Southeastern 0.58*** -0.21*** 0.01 -0.06 0.25* 0.02Western -0.12 0.52*** 0.21 -0.10 0.22 0.11
Log of Membership -0.56*** 0.99*** -0.10 -0.27*** -0.37*** 0.11Log of Membership2 0.23*** 0.01 0.36*** 0.19*** 0.22*** 0.17***Associate Pastor 0.09 2.06*** 5.89*** 0.99*** 1.23*** 0.95***Location (Urban Omitted)
Suburban 0.08 -0.11 0.13 0.14 -0.18* -0.27***Large Town 0.02 0.12 -0.01 0.08 -0.14 -0.10Rural -0.10 -0.03 -0.16 0.12 -0.14 -0.20*
Presence of Parsonage -0.06 0.00 -0.24* -0.05 -0.06 -0.19*Age (10’s of Years) -0.32*** 0.06 -0.25*** 4.64*** -0.17*** -0.37***Age2 -0.03 0.19*** 0.28*** 1.94*** 0.15*** 0.13*Part-time Status (Full Time Omitted)
Three Quarter Time -0.24* 0.40* 0.20 1.34*** 1.12*** 0.17Half Time -0.60*** -0.23 0.82* -0.37* -0.32 -0.11Quarter Time -0.79*** -0.05 0.80 -0.48* 0.16 -0.33
Appointment Type (Elder Omitted)Full-time Local Pastor -0.52*** 0.00 0.39* 0.40*** -0.20 -0.42***Part-time/Other Local Pastor -0.52*** -0.05 0.06 -0.39*** -0.64*** -0.30*Deacon -1.00 2.78 56.40*** 24.03*** 12.46* 15.28*Other Appointment Status -0.39*** -0.17 -0.37 0.31* -0.05 -0.50***
Note: The following interaction terms were included in m3.1, but are not reported for the sake of brevity: EthnicCharge*Ethnicity, Ethnicity*Appointment Status, Female*Associate Pastor, Associate Pastor*Ethnicity. Statisticalsignificance: *, p < 0.05; **, p < 0.01; ***, p < 0.001.
Table 6: Regression Model Results—Size at First Appointment
m1 m2 m3 m3.1
Constant 4.77** 4.79** 5.28** 5.30**Female -0.02 -0.03 -0.05* -0.05*Associate Pastor 1.82** 1.82** 1.62** 1.61**Race/Ethnicity (White Omitted)
Black -0.13** 0.17** 0.14** 0.09Asian -0.15* 0.03 -0.22** -0.40**Hispanic/Latino(a) -0.24** -0.01 -0.17* -0.36*Other -0.10 0.11 0.05 0.00
Ethnic Charge -0.45** -0.38** -0.31**Appointment Type (Elder Omitted)
Full-time Local Pastor -0.12** -0.13**Part-time/Other Local Pastor -0.76** -0.78**Deacon 0.19** 0.19**Other Appointment Status -0.39** -0.41**
Female*Associate Pastor 0.07* 0.07* 0.03 0.03Black*Associate Pastor -0.21* -0.28** -0.16* -0.16*Asian*Associate Pastor -0.32** -0.26* -0.03 -0.03Hispanic/Latino(a)*Associate Pastor -0.55** -0.71** -0.47** -0.45**Combined*Associate Pastor -0.65** -0.71** -0.56** -0.55**R2 0.49 0.5 0.58 0.58N 13769 13769 13769 13769
Note: The following interaction terms were included in m3.1, but are not reported for the sake of brevity: EthnicCharge*Ethnicity, Ethnicity*Appointment Status, Female*Associate Pastor, Associate Pastor*Ethnicity. Statisticalsignificance: *, p < 0.05; **, p < 0.01; ***, p < 0.001.
59
