8 ؽ


輳 (Experimental Design)


ܺκ ¿ ֿ ϰ ̿ Ӻ ȭ ν ڷḦ ϴ ̴.


輳 

)ѵǰ ̳ (proposition)

)

)ұԸ ȣۿ ϴ.


: ϼ ִ ?

̱翡 嵵 -> ̴.

̷ ȭ ϰ µȭ , Ȥ İ˻縦 Ѵ.


輳迡 Ҹ Ѵ.

ù°, Ӻ

°, ˻ ˻(pretesting and posttesting)

°, ܰ

ó óⰣ ܺο Ͼ ϴµ ʿϴ.


ex) ʿ伺 - 1920s, 1930s. ī electricity company - ȣս


Double-Blind Experiment - ڳ ڰ ׷ ̰ 𸣰 ó ϴ


1) Ȯǥ

2)

3) Ī: ׷ ܿ,

 

30

8

12

10

16

30-50

18

30

14

28

50̻

12

20

12

22


30 8 4 ܿ 4 ܿ ġϴ ̴.


Ī : Ī մ  ̸ Ѵ. мϴ Ҵ ϰ ִµ ̷ Ī м ŷڵ ıѴ.


輳(preexperimental designs)


1) ϻʿ(the one shot case study)

       X   O

2) ĺ(one group pretest-posttest design)

       O1     X      O2

3) ܺ(static group comparison)

       X      O1

     --------------

               O2


׷ α׷ ڵ ̻ α׷ ó Ҵٰ .


迡 ־ Ÿ絵(internal validity)


ó ⿡

Ÿ絵 ϴ

)

)

)˻: ó˻->˻. 迡 ģ

): ϴ . ְ->

)ȸ: ƹ ó ȸϴ ִ

)ڼ : ̰ ߻

)causal time order: causal ambiguity

)ó Ǵ : Ǵ . Ͼ

)ɸ - ܿ ̾ - ɸ: : ǻ

) й

) :


Ÿ絵(external validity):

迡 ܺμ迡 Ϲȭ ų


Ÿ絵 ؿ

)˻ ȿ :

˻簡 ǰ ΰ Űų ҽ ˻縦 ó ˻縦 ó ̳ ũ Ÿ

)ڼð ó ȣȿ

)óü ȿ

)ߺ ó : ܿ ó ó ȿ ִ ̴.



ĺ(pretest-posttest control group design)


R      O1     X      O2

-------------------------

R      O3             O4

ĺ(the posttest only control group design)


R      X      O1

-----------------

R              O2



ַθ 4 (the Solomon four group design)


G1     O1(60) X      O2(90)     d1  O2-O1 

-------------------------                              D1=d1-d2

G2     O3(60)         O4(70)     d2  O4-O3     

-------------------------                                             D=D1-D2

G3             X      O5(75)     d3  O5-[(O1+O3)/2] 

-------------------------                              D2=d3-d4

G4                     O6(65)     d4  O6-[(O1+O3)/2]



D=D1-D2=d1-d2-(d3-d4) = d1-d2-d3+d4

D1: ˻簡 ִ ó ũ

D2: ˻簡 ó ũ

D1-D2: ˻ ¿ ó ũ


d1=O2-O1=90-60=30

d2=O4-O3=70-60=10

d3=O5-[(O1+O3)/2]  = 75-(60+60)/2 = 15

d4=O6-[(O1+O3)/2] = 65-(60+60)/2= 5


D=d1-d2-d3+d4=30-10-15+5=10


ַθ ˻ ó ƴ϶ ȣۿ ũ ش.


:

Hard-to-get woman too-easy woman  HTGW ڵ ȣϴµ ׷?

Elaine Walster(1973)


1)л ͺ

HTGW ŷΰ? - , ڵ ڱ

ɸ̷: ؼ ǿ


2)

) show movie - teenage couple - caring much vs. not caring much

    caring very much were said to be the most desirable

) ǻ͵

    call - 3 seconds

          accept right away

    no difference.


)ٸ 򰡿 ϵ

)sexual component

) interview. not what made hard to get women desirable but advantage and disadvantage


Each man was given 5 folders of her evaluation of 5 man


  ƴ

 

 ڸ ڿ

 


--> Selectively elusive woman



:

ڱ ⸦ ٸ ΰ иϿ ִ

밳 ѵǾ ־ ð ǽ ִ


:

: Ȳ Ͼ Ȳ ȭ Ͼ ִ

ȣ effect

Logistical problem

administrative control

ethical question


ɰ ȭ Դ°?

ɰ ϸ ȭ ?

ݱ ɿ ϸ ȭߴٴ ȭ̾. , ϴ ɷ Ȯ ġϰ ־ ̴ 1 ǿ ڽĹ 忡 ߾ ȭ Դٴ Ȯ ŷ õǾ Դ.

׷ ֱ ̱ Ŀȣ(1996.10.4) ̱ ָ ġƮī ݱ ִ. Ŀ ʼ ð ȭ ̶ ̴.

ġƮī ̸ ̾.

ִ ȭ״ٸ ߻ Ҵ ʱ Ŀʹ ٸ ϰ ־ٴ ϸ ȴ. ׷ ̸ ȭ Ŵ ߰ߵ ʰ ִ.

׷ ġƮī ٹ Ͽ. Բ ߰ ݵ ڿ迡 ϴ Ź , ɷ° ɷ ̰ ִ° ̴.

ʱ⿡ ̵ ̿ ̰ ִٸ ȭ ̰ ٸ ȭ ٸ ƴٴ ǹϱ ̴.

񱳿 ̵ ̿ ũ ٸ ұϰ ʷϻ ϴû ν ̰ 巯 ̴. ̴ ξ 4 ̵ ϴ ɷ ־ ְ ִ ̴. ̿ Ŀ ʼ ð ȭ ְ ̴. (ȭϺ,1996.10.4)



ؽ (Quasi-Experimental Design) - Natural Experiment


ȸ ۿ ̷ ִ.

ȸ å. ġ


        

ɰ ظ ظ ڿؿ Ǹ ̴°?

 - (before hurricane, after hurricane)

) ùλ (Area1 - Area2)

ܺ 

         X       O1

        -----------

                 O2

)Alex Haley "Root" ٿ°?

TV ׷

TV ׷

1)time series experiment


O1 O2 O3 X O4 O5 O6

 

O1 O2   X  O3  O4


but history, seasonal variation

: ȯ , seat belt ģ


2)Non-equivalent control group design (ĺ)

        O      X      O

        ----------------

        O              O


selection and maturation bias

: ˷ ϸ Ѹ.

1977. ABC ۱

1979 Root, the next generation


1979. Scared straight

NJ. ˿ α׷

25̻ ˼ ϴ α׷ (314)

 

Two view of Reality

Falk's Fact

Finckenauer's Figures

program

Time frame of analysis

Type of sample

sample size

control group

success rate

scared straights

3

not representative

17

no

90%

scared straight

6

representative

81

yes

        

  59%           89%


delinquent-fulfilling prophecy

"By demeaning and belittling the young "tough-guys", the convicts may have compelled them to go out and prove they were not impressed"


3)Multiple time series


O      O      O      X      O      O      O

--------------------------------------------

O      O      O              O      O      



Ⱦܺм (Cross-sectional Study)


         41  - 50

         51  - 60

         61  - 70

         71  - 80

         

Cohort studies Ư ð 帧 ϴ . ƴϴ.


                1990                    2000

               41-50 (1)            41-50

               51-60 (2)            51-60 (1)

               61-70 (3)            61-70 (2)

               71-80                71-80 (3)


Panel studies cohort studies ȴٴ . (űⰣ űڰ ΰ Ǵ )

                1990            2000

               41-50*          41-50

               51-60**        51-60*

               61-70***      61-70**

               71-80           71-80***


trend studies

                1990            2000

               41-50  <------>  41-50

               51-60  <------>  51-60

               61-70  <------>  61-70

               71-80  <------>  71-80


ؽ輳

 

üȯ ð ģ ⿡ ϰ ʸ ߽ / ̻ - ȹ:v.34, n.5( 104ȣ) (1999-10)

 

ȹ->

ü ȭ ð  ġ°?

ϰ    O1             X1             O2

------------------------------

   O3             X2             O4     

ð: 

α : ȭ

̿:

ñ - ¿

ð: ٱ->ٱ