Navigation auf uzh.ch

Suche

Department of Informatics s.e.a.l

Work Unit Study Data

Lab Study

Project

https://github.com/google/gson

Change Tasks 

https://github.com/google/gson/issues/153

https://github.com/google/gson/issues/42

Instruction Question

Please report in the plugin whenever you start working on something new. Plase think-aloud while working on the change task.

 

Regression Model Details: Work Unit Type Detection

NOMREG

CRITERIA CIN(95) DELTA(0) MXITER(100) MXSTEP(5) CHKSEP(20) LCONVERGE(0) PCONVERGE(0.000001) SINGULAR(0.00000001)

MODEL=| FORWARD

STEPWISE=PIN(.05) POUT(0.1) MINEFFECT(0) RULE(SINGLE) ENTRYMETHOD(LR) REMOVALMETHOD(LR) 
INTERCEPT=INCLUDE

 

Model:

Model -2 Log-Likelihood Chi-Quadrat Degree of Freedom
constant 126.6    
final 58.6 68.0 12

Regression Model Details: Work Unit Boundaries Detection

LOGISTIC REGRESSION

METHOD=FSTEP(LR)

CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.15)

Variables in the equation:

Effect Coefficient Significance
str_step -.89 .05
sameClass_step -.97 .02
constant -2.40 .000

 

Regression Model Details: Detecting relevant code elements:

LOGISTIC REGRESSION

METHOD=FSTEP(LR)

Variables in the equation (ALL):

Effect Coefficient Significance
freq 6.86 .000
constant -2.49 .000

 

Variables in the equation (LC):

 

Effect Coefficient Significance
#ofUniquePrecMethods 8.70 .001
str_step 21.125 .012
sameField_step 1.9 .09
constant -2.735 .000
 

 

Variables in the equation (SC):

 

Effect Coefficient Significance
freq 9.8 .000
rec 2.5 .05
constant -4.9 .000

 

 

Variables in the equation (CF):

 

 

Effect Coefficient Significance
freq 6.2 .05
str_step 28.5 .018
constant -3.1 .003

 

Variables in the equation (TC):

 

Effect Coefficient Significance
freq 8.5 .004
constant -2.9 .000
 

 

 

 

Field Study

Project

All participants worked on their usual projects.

Change Tasks

All participants worked on their usual change tasks.

Instruction Questions

Please write down on what you were just working on.

Please identify within this timeline of source code interactions when you started to work on what you just described.

Which of these code elements which you identified for the work you just worked on are actually relevant?

 

Regression Model Details: Work Unit Type Detection

NOMREG

CRITERIA CIN(95) DELTA(0) MXITER(100) MXSTEP(5) CHKSEP(20) LCONVERGE(0) PCONVERGE(0.000001) SINGULAR(0.00000001)

MODEL=| FORWARD

STEPWISE=PIN(.05) POUT(0.1) MINEFFECT(0) RULE(SINGLE) ENTRYMETHOD(LR) REMOVALMETHOD(LR) 
 INTERCEPT=INCLUDE

Model: 

Model -2 Log-Likelihood Chi-Quadrat Degree of Freedom Significance
constant 87.8      
final 31.6 56.2 12 .000

Regression Model Details: Work Unit Boundaries Detection

LOGISTIC REGRESSION

METHOD=FSTEP(LR)

CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.15)

Variables in the equation:

Effect Coefficient Significance
field_step -1.7 .033
t 0.008 .000
constant -3.319 .000

 

Regression Model Details: Detecting relevant code elements:

LOGISTIC REGRESSION

METHOD=FSTEP(LR)

Variables in the equation (ALL):

Effect Coefficient Significance
str_step 7.98 .002
lex_step 2.62 .003
constant -1.47 .000
 

 

Variables in the equation (LC):

Effect Coefficient Significance
rec -5.9 .005
constant 2.4 .020

 

Variables in the equation (SC):

 

Effect Coefficient Significance
sameClass_step 3.99 .014
constant -1.59 .001

 

Variables in the equation (CF):

 

Effect Coefficient Significance
rec -5.21 .006
str_step 17.53 .007
lex_step 8.14 .07
constant -5.92 .000

 

Variables in the equation (TC):

Effect Coefficient Significance
#ofUniquePrecMethods .60 .036
constant -3.95 .003