DalWeck'6: MLR with lupliteli 2/omte
=) Inteduction to dummy variable
Qualitative into long gendu ,union statius or maritalstatus
is capptead using dommy variabe
-Dummy variablir elso kaaan as binary vaiobe C zero- um)
becaisse they Daly toke two ralrs : l it a cundition is mut andо
othirnne
•eg. to indirote gender, temok 10l it person lemoh,=0 11mal)
- using O and I is mathe matically conveaient for intrpreting ditllerews
=? SUR wth dummy varcobl os ngressor
- lonsider: woge= Bo1. Sotemale tu
where we assume SURY holds: Elal femol) =0
Thentora
Bo feroh =0
Eluoge.Ifemole) o t So temole =Cs
Botбo femcl)
. avy woge for min = pr
•avy wage for wonin il S0
☑
Base group: thegroy ussigud t ra o (eis. mals
-Intirapd (B1) : represents the arg wage for base grou
- lostficends (80): represents difterene in dry outams bitnen gop
bo is not a slope in tradiliinal seases M 1s a shittin the alsp
So= Elwage Itamal=1) - Elwoge Iterel =0)
The poynletion relationship is mimicked in ous estimctes
fo wage, So = woge mage
whire wige m (wagers is the ary wage for oa
mon (womun) in the sarpl
-Putting femce and mae in eg
집 is iedurdant -anly eed Imcops bntso
Lithisis called dummy varwbl tray
=7 MLR th dumry veriabu патрl of poteed collmoris
Consid woye fo & So temcle + pierper
So = ElwogeI temaleal, enper ]- Eluoge | terde 0, epи
So nua meusures the gende diff wkon me hold fixed esper
, Date.
Dummy varcablos for mltiple atego1
-whin didding meltipl gpalitatre umatles (eg. marrioge ,gended.
cosote dirtiact caligories to allow tor specitee grrop othacts.
- in a woge modl lnvolving geade Ltemol) ond marital states Cmoни)
pup con be dividdd into 4 whanstive ond mutually aelusite grte
•married males (marrma4): marrind (1- feman)
marrind famol Cmairtem): married Clamоп(
slagh moli Csingmah) (emorind). (1 femal)
siagh Csing fem)
temae (lumarrieds Ctemal)
Ettect ut gende and marital status on woge
• ckoose a base group ond nllice dunng tos other 3 gops to
awid dumny vocabe trap
it choose singmal os buse, mlud maimol, murten ond sinyten
in Hgession. The coetficients on these vorcobes are alertirs to
sngh miл.
wth Iwage =log luoges as dipeadnt aiabl, me con give
them a perantoge changt inlerpratetien
Eromplu: Wages ond morital stotus
Juaye = 321t 213 marcmdle 198 martem -110 sing ter
i 0.79 educ 10.027 1pu . 00054p1 079 tmure
-.00053 вnиre 2
n= 526 R²= 401
cebris paribuiCon exper and tesuraI
maried womon euras 19.8to uss thon singh mоn
single woman earns oboat 1h0ti Gss than sinou me
Usmytan 1.98, statutirally sig 0t 5t Cpe D.047)
Mak marrioge pamum: manmole earns 213% merd then snge
•Hou abon't female: slope for mair fem =. 321-.198
Siagfen .321-.110
dit=.148-C-110) = -.088
marlied winea era 8.74% uss than snglluonen
, Date. /
7 Usiny dummy vriebus to race poratl ordisal informativ
- ordinal variablis ( whurd orde mathers s1aa 1-5) should he
Converhd into dummy varcablis rather then trealid os singu
wntinuoes regrasso.
- Teating each rond es its eun coligory allows for nom Inrtteee
- Елampa (beant)
belowarg Clooks 1 or 2) orgC looks 3 : aboveory Cloilis 2h
.fer peopu ad extrem values lond 5 L c/t0 each,grruping
is bettи
use ang as bast
reg lwage belary obrag : lwaye Ko + "So beloug tS, abvog
rey fo conted for morde varcubos: lway f. tSobelarg tSiobu & Biedoct
Ly coxfficiands tor belevg lovers when wsE
63 wpe
e
moreAthepe 1 2 otemd
aboviary still statistically rasig and guts smalle in megail
- 11 1s aluays good practee to look ot oll cuttiuunes tosee
it signs ond mogaituch maka senve
=7 Model flee.bility with dummias
- The idia to replaa looks bo dummies for addd tleribilt alro
opnlias to quantilative discale variobes
- Consider nork experrana exper E E0, 122,3-303
Dlinior model: Ivage po 1B,erper tu
restrictin oppraco i ene mon yeor ot exper soow afleit
bo most as Zmo
① Quodratic med I: Iwager fit fp 1 Ppс21и
in ollow more flexblrss coptuca diminishingl Mreasiny tturas to cipe
mlaning slope changes os esper
Dummy Vriekh modl. laoge = fot fidit Brdetin 4 Piedis tn
dk=1 if exper =ky kzl2. 30
oad oblem mox Plritlitg
withint ossumingang specitic muttimaticel strpe (stiacgted o poreb
-Why use dummias for deecate nимвes
•non uniturm it: cgnse ditf 9rddie in woges froom O.tel gear to 241ь
replacl ordinal lcuugorienl lreic : stops model assuming dirlonn betuecs rosks
, Date.
7 Inbractions omong минп duwwi
- whon entigencal grorps ditind by Dor more bindry charactnstes
Cog gender and marital stertus), we tar use ntraction teims
Iage 32/a. 110 kameli 1. 213 mared -301 (ewel.m
-Why wafficrents timalı. marred dittera?
bacause to estimate martam temol =1 marird
- 321 -. 110 C1) 1.213011-301C1)CI) =-.168
The intraction meckl is otten praferrad becaure it allas yon to see
ditt in betmeon geaders drailly
merriage preminm
For mus: marniege pramior1s.213
For womun, premioon is sum ot morcind a thicnt ond mbrection tem
C0.213-0.301= -0.088)
=7 Alleming for ditt slopes Loloye vntroctions)
- By intracting a dumny varcable with quantitative varcebh ly exper),
не сол снак a mochl where grooys have beoth ditt inteiupts and dift
slopes
•problem with simple modils: twage = Bo + Se femalet Bieper ta
ossurnus the return to esperena CB.) is commen across genders.
-Estndd mial
Inoye = (f.t Sitema)&(p.t.Sitemab exper t
= S.1 So tamale Ipierper t Sitemali.exper tu
when use S, emphasini the Si and Si are ditferenais
Jnlrcep Sliga
Moe Cfemol-0) B
Femch Cfemol-1]
o1 50 pc15i
Gund ditt (femah So δι
- Exomnu : Inoge = 2.048- 518 temal 1. 0233 exper t. 0274 terd. eper
napt for men: 2.098 slope fon men: 0283
помел: 1·58 wimn02831-0234=.0517
75
L 00283
idatico1 slope:.0517
me
=) Inteduction to dummy variable
Qualitative into long gendu ,union statius or maritalstatus
is capptead using dommy variabe
-Dummy variablir elso kaaan as binary vaiobe C zero- um)
becaisse they Daly toke two ralrs : l it a cundition is mut andо
othirnne
•eg. to indirote gender, temok 10l it person lemoh,=0 11mal)
- using O and I is mathe matically conveaient for intrpreting ditllerews
=? SUR wth dummy varcobl os ngressor
- lonsider: woge= Bo1. Sotemale tu
where we assume SURY holds: Elal femol) =0
Thentora
Bo feroh =0
Eluoge.Ifemole) o t So temole =Cs
Botбo femcl)
. avy woge for min = pr
•avy wage for wonin il S0
☑
Base group: thegroy ussigud t ra o (eis. mals
-Intirapd (B1) : represents the arg wage for base grou
- lostficends (80): represents difterene in dry outams bitnen gop
bo is not a slope in tradiliinal seases M 1s a shittin the alsp
So= Elwage Itamal=1) - Elwoge Iterel =0)
The poynletion relationship is mimicked in ous estimctes
fo wage, So = woge mage
whire wige m (wagers is the ary wage for oa
mon (womun) in the sarpl
-Putting femce and mae in eg
집 is iedurdant -anly eed Imcops bntso
Lithisis called dummy varwbl tray
=7 MLR th dumry veriabu патрl of poteed collmoris
Consid woye fo & So temcle + pierper
So = ElwogeI temaleal, enper ]- Eluoge | terde 0, epи
So nua meusures the gende diff wkon me hold fixed esper
, Date.
Dummy varcablos for mltiple atego1
-whin didding meltipl gpalitatre umatles (eg. marrioge ,gended.
cosote dirtiact caligories to allow tor specitee grrop othacts.
- in a woge modl lnvolving geade Ltemol) ond marital states Cmoни)
pup con be dividdd into 4 whanstive ond mutually aelusite grte
•married males (marrma4): marrind (1- feman)
marrind famol Cmairtem): married Clamоп(
slagh moli Csingmah) (emorind). (1 femal)
siagh Csing fem)
temae (lumarrieds Ctemal)
Ettect ut gende and marital status on woge
• ckoose a base group ond nllice dunng tos other 3 gops to
awid dumny vocabe trap
it choose singmal os buse, mlud maimol, murten ond sinyten
in Hgession. The coetficients on these vorcobes are alertirs to
sngh miл.
wth Iwage =log luoges as dipeadnt aiabl, me con give
them a perantoge changt inlerpratetien
Eromplu: Wages ond morital stotus
Juaye = 321t 213 marcmdle 198 martem -110 sing ter
i 0.79 educ 10.027 1pu . 00054p1 079 tmure
-.00053 вnиre 2
n= 526 R²= 401
cebris paribuiCon exper and tesuraI
maried womon euras 19.8to uss thon singh mоn
single woman earns oboat 1h0ti Gss than sinou me
Usmytan 1.98, statutirally sig 0t 5t Cpe D.047)
Mak marrioge pamum: manmole earns 213% merd then snge
•Hou abon't female: slope for mair fem =. 321-.198
Siagfen .321-.110
dit=.148-C-110) = -.088
marlied winea era 8.74% uss than snglluonen
, Date. /
7 Usiny dummy vriebus to race poratl ordisal informativ
- ordinal variablis ( whurd orde mathers s1aa 1-5) should he
Converhd into dummy varcablis rather then trealid os singu
wntinuoes regrasso.
- Teating each rond es its eun coligory allows for nom Inrtteee
- Елampa (beant)
belowarg Clooks 1 or 2) orgC looks 3 : aboveory Cloilis 2h
.fer peopu ad extrem values lond 5 L c/t0 each,grruping
is bettи
use ang as bast
reg lwage belary obrag : lwaye Ko + "So beloug tS, abvog
rey fo conted for morde varcubos: lway f. tSobelarg tSiobu & Biedoct
Ly coxfficiands tor belevg lovers when wsE
63 wpe
e
moreAthepe 1 2 otemd
aboviary still statistically rasig and guts smalle in megail
- 11 1s aluays good practee to look ot oll cuttiuunes tosee
it signs ond mogaituch maka senve
=7 Model flee.bility with dummias
- The idia to replaa looks bo dummies for addd tleribilt alro
opnlias to quantilative discale variobes
- Consider nork experrana exper E E0, 122,3-303
Dlinior model: Ivage po 1B,erper tu
restrictin oppraco i ene mon yeor ot exper soow afleit
bo most as Zmo
① Quodratic med I: Iwager fit fp 1 Ppс21и
in ollow more flexblrss coptuca diminishingl Mreasiny tturas to cipe
mlaning slope changes os esper
Dummy Vriekh modl. laoge = fot fidit Brdetin 4 Piedis tn
dk=1 if exper =ky kzl2. 30
oad oblem mox Plritlitg
withint ossumingang specitic muttimaticel strpe (stiacgted o poreb
-Why use dummias for deecate nимвes
•non uniturm it: cgnse ditf 9rddie in woges froom O.tel gear to 241ь
replacl ordinal lcuugorienl lreic : stops model assuming dirlonn betuecs rosks
, Date.
7 Inbractions omong минп duwwi
- whon entigencal grorps ditind by Dor more bindry charactnstes
Cog gender and marital stertus), we tar use ntraction teims
Iage 32/a. 110 kameli 1. 213 mared -301 (ewel.m
-Why wafficrents timalı. marred dittera?
bacause to estimate martam temol =1 marird
- 321 -. 110 C1) 1.213011-301C1)CI) =-.168
The intraction meckl is otten praferrad becaure it allas yon to see
ditt in betmeon geaders drailly
merriage preminm
For mus: marniege pramior1s.213
For womun, premioon is sum ot morcind a thicnt ond mbrection tem
C0.213-0.301= -0.088)
=7 Alleming for ditt slopes Loloye vntroctions)
- By intracting a dumny varcable with quantitative varcebh ly exper),
не сол снак a mochl where grooys have beoth ditt inteiupts and dift
slopes
•problem with simple modils: twage = Bo + Se femalet Bieper ta
ossurnus the return to esperena CB.) is commen across genders.
-Estndd mial
Inoye = (f.t Sitema)&(p.t.Sitemab exper t
= S.1 So tamale Ipierper t Sitemali.exper tu
when use S, emphasini the Si and Si are ditferenais
Jnlrcep Sliga
Moe Cfemol-0) B
Femch Cfemol-1]
o1 50 pc15i
Gund ditt (femah So δι
- Exomnu : Inoge = 2.048- 518 temal 1. 0233 exper t. 0274 terd. eper
napt for men: 2.098 slope fon men: 0283
помел: 1·58 wimn02831-0234=.0517
75
L 00283
idatico1 slope:.0517
me
