ñòàòüå ïðåäïðèíÿòà ïîïûòêà íà ýïèñòåìîëîãè÷åñêîì óðîâíå ïðîÿñíèòü ìåòîäîëîãè÷åñêèå ïðîáëåìû, ñâÿçàííûå ñ èñïîëüçîâàíèåì êîíâåíöèîíàëüíîãî êîëè÷åñòâåííîãî àíàëèçà â ïîëèòè÷åñêîé íàóêå. Äëÿ ýòîãî êîíâåíöèîíàëüíûå êîëè÷åñòâåííûå ìåòîäû ðàññìàòðèâàþòñÿ ñêâîçü ïðèçìó òåçèñà Äþýìà-Êóàéíà. Îñîáîå âíèìàíèå óäåëÿåòñÿ òðåì èññëåäîâàòåëüñêèì ïðàêòèêàì êîëè÷åñòâåííîãî àíàëèçà: 1) òåñòèðîâàíèþ íóëåâûõ ãèïîòåç êàê ñàìîäîâëåþùåãî àíàëèòè÷åñêîãî èíñòðóìåíòà; 2) íåâíèìàíèþ ê êàóçàëüíîé íåîäíîðîäíîñòè ÿâëåíèé â áîëüøèõ âûáîðêàõ; 3) ïðèìåíåíèþ îïåðàöèîíàëèçàöèé ïåðåìåííûõ, ïîäâåðæåííûõ áîëüøîìó ÷èñëó èíòåðïðåòàöèé. Ðàñïðîñòðàíåííîñòü ýòèõ ïðàêòèê âåäåò ê óñóãóáëåíèþ: à) ñòðóêòóðíîé íåäîäåòåðìèíèðîâàííîñòè òåîðèé; á) âîçìîæíîñòè ôîðìóëèðîâàòü áîëüøîå ÷èñëî òåîðèé ïðè óñëîæíåíèè èõ ýìïèðè÷åñêîé ïðîâåðêè; â) òðóäíîñòåé ïî äîñòèæåíèþ êîíâåíöèè î ãðàíèöå ìåæäó ôàêòàìè è òåîðèÿìè. Ôîðìóëèðóþòñÿ âîïðîñû îá óñëîâèÿõ ïðîãðåññà â ïîëèòè÷åñêîé íàóêå ñ òî÷êè çðåíèÿ òåçèñà Äþýìà-Êóàéíà.

The paper applies the Duhem-Quine thesis to conventional quantitative methods in political science. As a result, the discussion of methodological problems associated with these methods is implanted into the epistemological issues highlighted by Duhem-Quine thesis. Special attention is devoted to the widespread research practices, such as 1) null hypothesis significance testing, 2) large-N analysis and 3) dealing with phenomena with a very broad and general character. The paper argues that, partly due to these practices, some epistomological problems are aggravated: a) structural underdetermination of theories is exacerbated; b) the value of new theories tested via old data becomes unclear and doubtful; c) the convention about the boundary between theories and facts is harder to achieve. Some questions about the conditions of the progress in political science are posed.

ìåòîäîëîãèÿ ïîëèòè÷åñêîé íàóêè; êîëè÷åñòâåííûå ìåòîäû; òåçèñ Äþýìà-Êóàéíà.

political science methodology; quantitative methods; Duhem-Quine thesis.