Latentinės nusikalstamumo dalies prognozavimo klausimu

The official data on registered crimes or criminal offenders is one of the sources on criminological information. For various reasons, not all of the offences become known to law enforcement authorities. The authorities may not qualify the offences properly. Therefore, there is the unknown, “invisible” side of the criminality, i.e. unregistered part of this phenomenon. The ratio of latent criminality to registered criminality is also uncertain at a particular point in time. Likewise, the fluctuation of this ratio time-wise or structure-wise is uncertain, therefore, it may be concluded that the instability of this ratio could have a negative impact to the error margin associated with the results of the analysis of criminological indicators. The above-mentioned ratio is compound and one of its elements is the dynamically shifting time-lag between the commitment of criminal offences and their official registration. The first part of the article is dedicated to the analysis of the impact of this ratio to the bias associated with the results of the analysis of criminological indicators as well as to describing the model, which helps to minimize these error margins. Unfortunately, the described method requires that rigorous initial conditions are satisfied, otherwise the obtained results can not be considered reliable, i.e. this method is not suitable when the structure of data varies due to internal or external reasons or when the amount of criminal activity to be analyzed is nil, or when the criminal offence is retried under a different article of the Criminal Code during the pre-trial investigation. Therefore, the mathematical model used to minimize the margin of error encountered due to distortion of statistical data caused by variable time-shift between the commitment of criminal act and their registration is to be improved. While a certain part of the above-mentioned issues may be solved by using statistical data in respect of criminal acts in respect of which the court has already convicted the suspect or by simply prolonging the time-period of the analysis, the remaining issues can be solved only by revising the described model. The article proposes several modifications, which allow to eliminate the disadvantages of the original method, because they require application of less stringent conditions, thus enabling application thereof in volatile conditions of social reality more flexibly. Using these variations of the model should assist in fine-tuning of the assessment of distribution or structure of criminality as well as other criminological indicators, allowing obtaining more reliable criminological results.Šio straipsnio pirmojoje dalyje buvo aprašytas metodas, kurį taikant, esant patenkintoms reikalaujamoms prielaidoms, galima būtų įvertinti vieno iš veiksnių, suponuojančių kriminologinių rodiklių analizės rezultatų paklaidas, įtaką ir šias paklaidas sumažinti, tai yra gauti patikimesnius nusikalstamumo rodiklių duomenis. Deja, aprašytajam modeliui būtina, kad būtų patenkintos gana griežtos pradinės (taikymo) sąlygos, priešingu atveju gauti rezultatai negalėtų būti laikomi patikimais. Šioje dalyje bus aptariamos kelios minėto metodo modifikacijos, kurios padėtų išvengti vienos ar kitos taikymo prielaidos, t. y. būtų galima išplėsti aprašyto metodo taikymo atvejų aibę