Methode 'Praktische Logica':

   

Methode van

Causale analyse



Voorbeeld Causale analyse:



Effect van TBS 'behandeling' - 5 jaar na vrijlating.



VRAAG: Heeft TBS een 'behandel' effect op de recidive-risiso's van ex-gedetineerden?

Vergelijking tussen recidive-scores van 'reguliere' ex-gedetineerden en ex-TBS'ers:
Vanaf 1977 resp. 1974, na een periode van 5 jaar na vrijlating.

Zie schema hieronder voor de ontnuchterende resultaten.

Enkele highlights uit deze resultaten:


We hanteren hierbij een maximale waarde voor de foutkans (alpha) van 5 procent als criterium voor de significantie van de afgeleide populatie-correlatie.
Daarnaast hanteren we een maximale waarde voor de foutmarge (epsilon) van 5 procent als criterium voor de betrouwbaarheid van de geschatte 'N=1' steekproefwaarde t.o.v. het populatiegemiddelde.
Verder hanteren we een minimale waarde voor het betrouwbaarheidsinterval (Confidence interval) van 95 procent .

Conclusies:


• De ex-tbs'ers deden het 5 jaar na vrijlating gemiddeld iets slechter dan de ex-gedetineerden (van vergelijkbare groep).
• Het gevonden verschil blijkt echter niet significant voor de totale populaties.
• Voor een willekeurig individu in de populatie maakt Tbs-'behandeling' binnen 5 jaar na detentie geen naspeurbaar verschil voor diens recidivekansen.

Zie verder ..



Overzicht

van

analyse-gegevens:



1. Table Input-Form Causal Analysis - Project data

Parameter Value
Analysis/ method Chi-square based analysis of causality
Project (title) Causal analysis; TBS Recidivism Prevalence 0-05yr
Hypothesis Crimin: (Psychop.+TBS) CE! - CE - - CE --> (less Recidive/Psychop. /+Therap.effect)
Name of 'cause' A1 (keyword) True(Psychop.+TBS)
Name of 'effect' B1 (keyword) NOT (Recidive)
Max. tolerated exception due to non-random error (alpha);
Sample Effect-difference unexplained - general value (prb.)
_________________________________0.05_____________
Testing one-sided or two-sided _________________________________2________________

2. Chi-Square table: groups and data of observations

Condition description Cases with effect present /detected (+B1) Cases with effect absent /not detected (-B1) Cases regardless of effect (±B1)
Cases with cause present /detected (+A1) ________________1062 ________________1143 ________________2205
Cases with cause absent /not detected (-A1) ________________2582 ________________2582 ________________5164
Cases regardless of cause (±A1) ________________3644 ________________3725 ________________7369

3. Test statistics - Summary.
Effect-difference explained; significance surplus.
'*' : significance of difference.

Parameter Value
 
3. 1. Chi-Square sample statistics  
3. 1. 1. Chi-square statistic  
Chi-Square (val.) _________2.085410699
Sample size N (num.) ______7369__________
3. 1. 2. Sample Correlation and related statistics  
SAMPLE CORRELATION (corr) _________0.016822538
Sample Variance explained (ppt.) _________0.000282997
3. 1. 3. Sample Significance of correlation  
Max. Popul. Effect-difference explained (prb.) _________0.925616429
Popul. Effect-difference explained; significance surplus (prb.)  

¬

_______-0.049383570
 
3. 2. Proportions to conditional probabilities  
3. 2. 1. SAMPLE PROPORTIONS  
Total Hits (ppt.) _________0.494504003
3. 2. 2. CONDITIONAL probabilities  
Conjunct Prob. (Cause,Effect) (product) (prb.) _________0.240816326
3. 2. 3. CAUSALITY QUADRANTS  
Max. Symmetrical valence (prb.) _________0.494504003
3. 2. 4. LOGICAL probabilities  
Logical subsumption, coverage of conclusion;
overall condition: symmetry (A1 B1) (prb.)
_________0.202011507
 
3. 3. Test sample proportion to population value  
3. 3. [1a.] Population is Infinite; Distribution is Continuous  
Popul. Effect-increase explained; infinite; continuous (prb.) _________0.957708005
Popul. Effect-increase expl.; infin.; contin.; significance surplus (prb.)  

¬

_______-0.017291994
3. 3. [1b.] Population is Infinite; Distribution is Discrete  
Popul. Effect-increase explained; infinite; discrete (prb.) _________0.959592825
Popul. Effect-incr. expl.; infin.; discrete; significance surplus (prb.)  

¬

_______-0.015407174
3. 3. [2a.] Population is Finite; Distribution is Continuous  
Popul. Effect-increase explained; finite; continuous (prb.) _________0.980310665
Popul. Effect-increase expl.; finite; contin.; significance surplus (prb.)  

*

________0.005310665
3. 3. [2b.] Population is Finite; Distribution is Discrete  
Popul. Effect-incr. explained; finite; discrete (prb.) _________0.981494932
Popul. Effect-incr. expl.; finite; discr.; significance surplus (prb.)  

*

________0.006494932
 
3. 4. Test difference of 2 sample proportions  
3. 4. [1a.] Population is Infinite; Distribution is Continuous  
Popul. Effect-increase explained; infinite; continuous (prb.) _________0.925699022
Popul. Effect-increase expl.; infin.; contin.; significance surplus (prb.)  

¬

_______-0.049300977
3. 4. [1b.] Population is Infinite; Distribution is Discrete  
Popul. Effect-increase explained; infinite; discrete (prb.) _________0.928172676
Popul. Effect-incr. expl.; infin.; discrete; significance surplus (prb.)  

¬

_______-0.046827323
3. 4. [2a.] Population is Finite; Distribution is Continuous  
Popul. Effect-increase explained; finite; continuous (prb.) _________0.957770987
Popul. Effect-increase expl.; finite; contin.; significance surplus (prb.)  

¬

_______-0.017229012
3. 4. [2b.] Population is Finite; Distribution is Discrete  
Popul. Effect-incr. explained; finite; discrete (prb.) _________0.959654273
Popul. Effect-incr. expl.; finite; discr.; significance surplus (prb.)  

¬

_______-0.015345726
 
3. 5. Compare sample proportion with reliability margins of population average  
3. 5. 1. TEST PROPORTIONS, DIFFERENCE SAMPLE vs. POPULATION  
Max. Popul. Effect-difference, explained (prb.) _________0.957708005
3. 5. 2. TEST 1 SAMPLE PROPORTION to POPULATION CONFIDENCE INTERVAL  
Effect in exp. group, outside of interval; significance surplus (ppt.)  

¬

_______-0.002506825
 
3. 6. Compare proportions-difference with reliability margins of population average  
3. 6. 1. TEST DIFFERENCE of 2 POPULATION PROPORTIONS  
Max. Popul. Effect-difference, explained (prb.) _________0.925699022
3. 6. 2. TEST DIFFERENCE of 2 SAMPLE PROPORTIONS to POPULATION CONFIDENCE INTERVAL  
Effect-increase in exp. group, outside of interval; significance surplus (ppt.)  

¬

_______-0.006556495
 
3. 7. Sample correlation to population prediction  
3. 7. 1. Population Correlation and related statistics  
Max. POPULATION Correlation (CORR) _________0.016822538
Min. Population Variance not explained (ppt.) _________0.999717002
3. 7. 2. Significance of population correlation  
Max. Popul. Effect-difference explained (prb.) _________0.925616429
Popul. Effect-difference explained; significance surplus (prb.)  

¬

_______-0.049383570
 
3. 8. Population correlation to individual prediction  
3. 8. 3. Population sign estimation for individuals, overall  
Max. Accuracy of popul. sign prediction, overall (ppt.) _________0.505355032
Max. Popul. sign estimation, overall, difference (ppt.) _________0.494644967
Max. Effect-difference explained; (relative) significance surplus (prb.)  

¬

_______-0.944644967
3. 8. 4. Population decile estimation for individuals, overall  
Max. Accuracy of popul. decile prediction, overall (ppt.) _________0.1________
Max. Effect-difference explained; (relative) significance surplus (prb.)  

¬

_______-0.95_______
3. 8. 5. Confidence of sign prediction for individuals  
Max. Confidence of sign prediction, indiv. (prb.) _________0._________
Max. Effect-difference explained; (relative) significance surplus (prb.)  

¬

_______-0.95_______