⍆ᰰ ⊹ ᮑ⊯ᯱᢤ⍆ᛵዷ፲⍹Ꮒ ᶋ᤺ἡᯱᢤ⍆ᗦ≨ 1ጃⓘ◘᪇ᤲ᥉₈᪸◘⍆ᱡ 2ᐉℙ⟠ጙ⍆Ồᡡ⍆᪇៝ᛓ⍆ᖗ Logistic regression model ܗCox proportional hazard model๊Ȃ፞Ӽӫມளཽᡲ೩Ӽ፤НᎨ ԥ፹ȂࠔέగϟࠔؐޠᏱऌȄΚૢདྷژಜॏ စᡜܗःفስစᡜၷЎޠᚃٲসདژΚ ᏱȂѠ൸ཽޣញᖓདྷ࢞ژᕎнޠکᏱϵԓ ᓟ᜵ЬІณܛᎍȄԫѵȂ౫ϭภ࿂ᅗҭІन ۢܗ౪Ȃܗፓᚖϛܿ၍ޠёಜॏൣߓȄ ࡤᗵ֥ା౪ޤᜌ൩ஆޠಜॏРึݳȂԄ կٲᄃαȂಜॏᏱึޠȂଷΠᏱ౪፤ޠ൩ Ӽᡑ໕ϸݚMANOVAȃfactor analysisȃstruc- ஆϟѵȂڐᄃԥӼᐤѭڸНޠ८Ӫڸଔ tural equation model (SEM)ȂܗGEEȃ ᝧȄҐНതၑშпΤ఼яޠ፤ख़Ȃᙐ्ᇴ mixed model๊ᎍңܼᚃᏱα੬րޠძ-֊२ ݃НڸᐤѭึԄեኈಜॏᏱޠᅌໍȄ ᙮ข໕І๗ݏᡑ݉ߩٯளᄙϸո೪ޠଠ ᘫಜॏРݳȂࣲளُܼᏱёІःفስȄ ӕђαಜॏᡞޠτ໕я౫ȂԄSPSSȃRȃ ಜॏᏱȂ౫ϭӶөঐᏱስȂϑፁ STATAȃPythonȌ๊๊Ȃ੬րSASІR๊ ࠐึȄसп౪፤ڸᔗңୣϸȂѠ྆ϸ࣐౪፤ ᡞᄈܼߒᏱٯϛΫϸЅȂ܂܂ᡲณݳ ಜॏ(౪ಜॏ)ІᔗңಜॏȇՅᔗңಜॏέѠ ץഁαКȄαख़ණޠژᆎᆎӱளᏳयτঢ় τϸ࣐୧ңಜॏȃϏཿಜॏȃఁيᇅЗ౪ಜ ளளӶᎨё፤НਣȂѬᜱݨpঅϊܼ ॏІҢޑಜॏ๊ስȄڐϜҢޑಜॏ֊ڸᚃ 0.05(ᡘ)џ౪၍ȇܗӶٲःفਣȂ ᏱਁਁࣻᜱȄՅҢޑಜॏᏱөᆎ౪፤ץޠഁึ ಜॏഌϸளېଋಜॏঢ়Ӭձпᗘϛཽᐈձ ȃђпႬဟ౪Ꮳၽᆘഁ࡚ޠऐॵౢໍІಜ ಜॏᡞܗϛᔘಜॏ౪፤֩ޠᘚȄ ॏᔗңࣻᜱᡞޠСུУȂࣲழҢޑಜ ॏᏱץޠഁໍؐᇅึϟӱȄ ౫ϭϛ፤ᎨᚃᏱःفёНതȂ ܗኦቹᚃᏱᏱ፤НȂҢޑಜॏഎϛѠ ޠુܗϏڏȄՅᄈܼःفስߒޠᏱȂளᄈ ܼΚஃϥ߇ΥߟޠಜॏӫມІРݳདژగՅҢ ᮑ⊯ᯱᢤ⍆ᛵዷ፲⍹Ꮒᶋ᤺ἡᯱᢤ⍆ᗦ≨ ᅛ ᚃᏱಜॏᏱȂΚߟᡲ೩ӼѠញூ ᯱᢤ⍆ᑋᛦ◘⒱ᠾᛵ⊖ᝥᶋ᤺ ⍆ᰰ⊹ ᖦ ᎊᤲ⚠1 ṗ⯥᷒2 ҐНതȂ์्ߩٯၑშᇅτঢ়ϸٵಜ ॏޠ२्ᢏܗ܉ЗூȂՅདྷНِ࡚ޠ ϹΤȂוᏖᡲτঢ়ᄈಜॏԥϛӤޠᢏདȂС ࡤӶᏱಭಜॏαԥᐡཽࡼϛӤࣽݳȄ ᯱᢤ⍆ᛵዷ፲⍹Ꮒᶋ᤺ ࣨȄᙐޠᡑҕРݳϸݚԄt-testȃ ಜॏᏱȂ൸Ґ፵αٿᇴȂ֊Κߟ౪ one-way ANOVAȃPearson ͬ2 testȂژளُ ȶϛጃۢܓȷޠऌᏱȄ࣐Ϩቅቅᇴȉӱ࣐ ޠᡑณҕϸݚԄWilcoxon signed-rank ಜॏᏱҐ፵α౪ႇӼၦଊޠऌᏱȄՅ17з testȃWilcoxon rank sum testȃKruskal-Wallis खϜဩۗȂҦܼᄈȶ፱ീᏱȷޠःفȂ test๊Ȃӕژτঢ়ளңηൊңޠᡑ(ܗ ЩၷҔԓݨӶϛጃۢܓঐ྆ޠ܉ Ӽᡑ)ଠᘫРݳԄ linear regression modelȃ ۗȄ࿌ਣޠಜॏᏱݨӶᏱα௷ޠӗಣӬॏ 2018ԑ62ڢ8 Ѯѕѿᚂৱϴོོѐ 41 ⍆ᰰ⊹ ᆘ(֊ᚕයޠᐡ౦ॏᆘȂԄሾφя౫࢛ᘉޠ ٯϛញூԥϨቅ੬րޠȂկᄈ࿌ਣ19зखԒх ᐡ౦࣐ե)ȄࡤٿȂ17Վ18зखȂಜॏᏱࠍ۾ ਣޫίޠᏱȂڐᄃȶ҂ְȷٯ܉྆ޠϛ իژΠЉНᏱІข໕Ᏹ๊ስȂԄข໕ڐуЉ ܛԥഎѠп௦ڨȄ ᮑ⊯ᯱᢤ⍆ᛵዷ፲⍹Ꮒᶋ᤺ἡᯱᢤ⍆ᗦ≨ ᅛ ᡞڸӵ౩ޠᚕȂܗข໕ӵםύޠା๊࡚ Յ19зखίљဩࡤȂҢڽऌᏱޠಜॏτ໕ ๊ȄܛпڐᄃಜॏᏱึޠȂڸҢࣁਁ ึȂڐስєࢃၼᏱȃෛޑᏱІᚃᏱ๊Ȅ ਁࣻᜱޠȂٯྜึڐϛՍҢޠᏱϵԓІ ഷԥӫپޠφϟΚȂ֊Francis Galtonңಜॏ ၽᆘϜҢȄ РݳȂᇴ݃ΠࡿવᒲᜌऌᏱ፤ख़ޠಜॏᏱ ՅಜॏᏱޠକྜȂഷԟя౫Ӷ՚ϰ1662Ԓ ऽᏱJohn GrauntޠਫϜȂ࿌ਣхߓޠཏࡧ ȶݾϟȷ(statecraft) 1Ȃࡿਗ਼Κঐঢ়ޠ ಜݾȂሰ्ᄈπޠяҢ౦ȃԬκ౦ȂࣦՎ ೳᄍࢻ๊ಜॏԆ्ԥΚۢޠΠ၍Р ࣐ٸޠ๋؛ᐄȄܛпȂڐᄃӶᚕ౫Ӷߗ400Ԓ ࠊޠзࣩȂȶಜॏᏱȷঐӫມࣻ࿌ܼȶ࢈ ๋ऌᏱȷȂΚߟᆔ౪ᏱޠᏱୱȄՅӶ18з खϜဩ(՚ϰ1749Ԓ)ȂѫΚ՞ٿՍޠᏱ Gottfried AchenwallࠍࡿᆏಜॏᏱȂΚߟє ֥Ᏽ࢈ѭȃ࢈ݾစᔽᏱȃݾϟࣦՎݳল ᏱޠᏱୱ 2 Ȅαख़ᐤѭनෂึȂϛᜳΠ၍ ಜॏึޠྜȂڸНަཽᏱୱਁਁࣻᜱȂ ஞϛѠϸȄ ἡᯱᢤ⍆ᗦ≨ᛵឰᐉ-ỻᛵ᫆◶ᅘⅈ᥅ፍ ᩿ᅘ☽ᛵ᫆◶ᩱ⍶ᮻ5 ࿌хಜॏᏱڑጓ࡛ޠҴᇅȶྦޠୱᚡȷ ஞϛѠϸȄӶଇಜॏᏱྦޠୱᚡਣȂ्Ӓԥ Κঐ྆܉Ȃ൸ӶτӼݸίȂҕᡞޠၦଊ എϛѠூޠȂࢉȂಜॏޠᆡઢ൸्џң စᡜαϑึҢޠኻҐၦଊȂџ௱զҕᡞޠᄃ ၦଊȄਣȂᗵᙡΚঐࠊဋୱᚡȂفഥ ౪ȂޠۢڿȉЩԄשউདྷःفᑦ ఋထޠԞᕼᔇ130mmHgȉ٦שউ҇ ӒܜᇰȂᑦҕထޠԞᕼᔇঅȂΚঐ ޠۢڿঅȂ֊ȶ౪୳ΚܓȷޠԇӶȂη ൸ᇴȂनࡤޠၦਠҢᐡ(ښdata generating 19зखۗȂಜॏᏱӶަཽऌᏱึޠ mechanism, DGM)୳ΚޠȂՅߩᓎᐡя౫ ຺ึᡘȄ՚ϰ1820ԒȂAdolphe Queteletණ ޠȄඳѰၘᇴȂ֊౪Ѭԥઢ(ܗѠདྷԚՍ яΠܛᒞޠȶަཽޑ౪Ᏹȷ(social physics)Ȃ ดAlmighty)ޤၿȂ൸࣐եѡڑᐡ౦፤ޠ ٻңΠτ໕ඣख़ܓಜॏІϸඣख़ޠРݳȂٿ ඣख़ަཽޠ౫ຬ 3 ȄഷӫޠΚঐ྆܉ ȶ҂ְȷ܉྆ޠȂЩԄџඣख़Κঐঢ়ܗΚ ঐყᡞȂȶ҂ְٿᇴȷԥࡪቅኻޠӈȄαख़ ޠᇴݳȂᄈܼשউ౫хՅّញூಭп࣐ளȂ 42 ІᏱஆᙄ4Ȅ 2018ԑ62ڢ8 Ѯѕѿᚂৱϴོོѐ Ᏹᇰ࣐Ȃ౪ޠۢڿȂߩٯઢ(ܗՍ ด)ᘙሾφџޠۢ؛ȄӶಜॏᐤѭᅌᡑޠዙࢻ ϜȂשউѠпึ౫ᄈઢޠིȂ൸ࣻ߭ ౪ӓޠઢϘޤၿޠȂՅȶᇰޤԥ ४ܓȷޠȂܛпಜॏึޠڐᄃڸᄈۡఁޠ ѵܼᇰᜌлᡞࡋޠᢏԇӶȂ൸ȶࡋᢏᐡ౦ অनࡤޠȶᓎᐡᇳৰȷفഥߞϨቅኻφȉ๊ ፤ȷޠҐ೪ȄՎܼϭСᑀስॴឤޠлᢏᐡ ᜟήِםϸପȉഀ៊ޠљ༬םϸପȉᚗࡿϸ ౦፤ȂԄٕЫ(Bayesian)ಜॏȂڐ൸ϛሰ्α ପܗளᄙϸପȉծоηءԥዀྦ๏ 6 Ȃӱ ख़ၦਠҢᐡޠۢڿښ೪Ȅ ࣐ᓎᐡᇳৰ൸ۢဏαȂ൸౪ܓᇰޤαϛ ӲژȶྦޠୱᚡȷȂڐᄃ൸ޠȶᇰ ޤԥ४ܓȷܛᏳयȄη൸ᇴȂ౪Ѭԥઢ ѠޤၿᇳৰԄեึҢޠȂηณޤݳၿᇳৰ ޠঅҔ॓ܗτϊȄ (Սด)ޤၿȂשউҘሊӶ྄ᅿשউޠऌᏱःف ಒήঐୱᚡȶᛨޠୱᚡȷȂη൸ᇴȂ Іข໕РݳȂџΠ၍τՍดޠ౪Ȅկޠᇰ དྷϜשউוగᢏขঅޠϸයโ࡚Ȃ֊ݣโ ޤІРݳᖃԥᇳৰȂܛпӶ18зखϟࠊȂऌ ຺࡚ϊשউᇰۢȶЩၷԂȷޠၦਠܓ፵Ȅᛨ Ᏹঢ়ΚޣӶདྷفഥϨቅঅϘхߓȶྦޠୱ ۢޠୱᚡȂ྆܉α֊Ԅեۢဏȶᡑȷڸ ᚡȷޠ๏Ȅܛпᆘ҂ְȃϜ՞ܗಁ ȶѷړȷޠୱᚡȄ൸пᡑՅّȂѻޠ ๊τঢ়ዤޠޤಜॏ໕Ȃഎစೞᇰ࣐Ѡ ۢဏԥѠ๙ᄈᚕৰڸȃഷτ๙ᄈᚕৰܗ ྦޠୱᚡޠ๏Ȅԥ፹ޠȂԄݏџࢦᎨಜॏ ᚕৰ҂РڸȌ๊ȂહӶଇᛨޠୱᚡਣȂծ ึޠᐤѭНӈȂשউѠпึ౫ȶ҂ְȷ оηءԥᒳݳցᘟ٦ΚঐঅЩၷԂޠᚕය (mean)ঐऽНԆȂӶ18зखϜဩпࠊȂѬ զॏ໕ࡿޠዀȄ ঐϜ໕࡚ޠӫມȂѻѠхߓᆘ҂ְȂ ՅಜॏѭึޠӶαख़ήঐୱᚡࣲءԥ ηѠхߓಁȃϜ՞ȃϜ๊๊ޠཏဏȂ ዀྦ๏ޠݸίȂя౫Π२τᙾޠȂ֊ ՅӶ18зखϜဩϟࡤȂτӼޠಜॏᏱঢ় ᜟܛᆏޠȶಜॏᏱڑጓȷȄᔗңଜᐡ౦(in- ȶлᢏȷαӤཏᆘ҂ְȶྦޠୱᚡȷޠ verse probability)Іஆܼഷτ྆ծݳࠍȂ֊ ഷٺ๏Ȃܛп౫хᏱಭႇಜॏᏱשޠউȂϘ f (͢,ࢽʜx1, x2, Ȍ, xn) ڸf (x1, x2, Ȍ, xnʜ͢,ࢽ) ཽmeanޠཏဏ๊Ӥܼᆘ҂ְȄկفഥ Ԛ๊Щޠپᜱ߾ȂӕପӬளᄙϸପІᚗࡿ ᆘ҂ְഷޠٺȶྦޠୱᚡȷޠ๏ ϸପޠᏱړȂܝܝල(Laplace)ڸାල ȂڐᄃณݳጃۢȄՅኻޠᐤѭึᝓᙲ (Gauss)ϸրւңଜᐡ౦ڸཌྷϸޠᏱ ޠ౪௱Ᏻႇโ༞ȉڐᄃϛดȂΠ၍ᐤѭ൸ཽ ѾȂϸրԚѓ௱ᏳяΠڎঐପޠಜॏᏱڑ ึ౫щᅗΠୌดܓȂӓϛ҇ดޠ๗ݏȄ ጓȄпܝܝල࣐хߓޠᏱᇴȂӶႲ೪ᓎᐡ ಒΡঐୱᚡȶᇳৰϸପȷȂಜॏᏱঢ়ึ ᡑ໕ٸඊᚗࡿϸପޠ೪ίȂ௵ңଜᐡ౦ޠ ౫ྦޠୱᚡϟࡤȂಒΡঐ८ᖞޠୱᚡȂࢌด ௱౪ȂࠍѠпӤਣ௱яϜ՞ڸ҂ְ๙ᄈᚕৰ ޠᇰޤԥ४ܓȂٻூשউҘሊณݳ๏ᄈ ϸ࣐࣐ȶྦȷڸȶᛨȷޠୱᚡϟഷٺ๏ȄՅ 2018ԑ62ڢ8 Ѯѕѿᚂৱϴོོѐ ᮑ⊯ᯱᢤ⍆ᛵዷ፲⍹Ꮒᶋ᤺ἡᯱᢤ⍆ᗦ≨ ᅛ অޠҔጃ๏ȂܛпಜॏᏱঢ়ᙾՅ൷ؒȂզॏ ⍆ᰰ⊹ ၗԥτޠᜱ߾ȂՅлၦਠҢᐡڏښԥ 43 ⍆ᰰ⊹ пାල࣐хߓޠᏱᇴࠍᇰ࣐ȂӶႲ೪ᓎᐡᡑ໕ ᗃᗃ༂ఁ௳రณߴӵఁ௳ࣻשᜱᏱୱȂІڐ ٸඊளᄙϸପޠ೪ίȂ௵ңଜᐡ౦ޠ௱౪Ȃ ᄈ์ޠఃᇍ⽱⽳ڸេദȄ ࠍѠпӤਣ௱яᆘ҂ְְڸРৰϸր࣐ ᮑ⊯ᯱᢤ⍆ᛵዷ፲⍹Ꮒᶋ᤺ἡᯱᢤ⍆ᗦ≨ ᅛ ȶྦȷڸȶᛨȷޠୱᚡϟഷٺ๏Ȅԥ፹ޠ ȂӶᐤѭࢺࢻึޠίȂܝܝලӶ࿌ਣᗷ ดΠӬ౪ޠᏱ௱ᏳȂկуЗϜڐᄃԟཱུՍ ᇰۢᆘ҂ְϘഷԂޠȶྦޠୱᚡȷޠզ ॏ໕ȂܼуܺతΠՍϐ௱ᏳяޠٿᏱᇴȄᐤ ѭึޠ൸ቅԥ፹ȂᏳयΠשউ౫Ӷпள ᄙϸପȃᆘ҂ְІְРৰ࣐ಜॏлࢻ౪፤ ึޠӱȄདྷདྷᐤѭᄈಜॏޠኈȂसܝ ܝල࿌ਣࡼᚗࡿϸପ࣐ഷޠٺᇳৰϸପ лȂη೩౫ϭಜॏޠөᆎ౪፤ڸዂཽשڸ উ౫ӶژࣽܛІདྷޠτ࣐ϛӤȄ 1. Graunt, John. (1662). Natural and Political Observations Made upon the Bills of Mortality. London: Martyn. 2. Pearson, Karl. (1978).ɆIntroduction: The early history of statistics,ɇ in The History of Statistics in the 17th and 18th Centuries, E. S. Pearson (Ed.), pp. 8-9. London: Charles Griffin & Company Limited. 3. Stiger, Stephen M. (1999).ɆThe average man is 168 years old,ɇ in Statistics on the Table: The History of Statistics Concepts ᶬ⊹ and Methods. (chapter 2) Cambridge, Mass: ಜॏึޠȂᗷሰ्ཌྷᑗϸȃጤܓх Harvard University Press, pp. 51-65. ๊٘ޠ౪ಜॏनෂ࿌౪፤ଽ࢝Ȃկٲᄃ 4. Stiger, Stephen M. (1999).ɆGalton and αȂӶಜॏᐤѭึޠϜȂНࡧདྷޠϮΤȂ identification by fingerprints,ɇ in Statistics ԄಜॏᏱഷԟೞ࿌ձ࢈ݾᏱޠᏱୱȂژѡڑಜ on the Table: The History of Statistics ॏ౪፤ᗵ֥ᄈઢ(Սด)ޠིȂڐᄃഎΚӕ Concepts and Methods. (chapter 6) ᇴ݃ಜॏᏱӶᏱስϜȂΚߟڸНऌᏱ Cambridge, Mass: Harvard University Press, ഷ࣐ᒒߗޠᏱߟȄ์ӶಜॏᏱስϜȂϬԥ pp. 131-40. ೩Ӽ्ӕᆡໍϟȂկኦቹҐНȂܨᑓЖ ҞȂЎτঢ়ᄈಜॏᏱិ৽ޠȂСࡤӶᏱಭಜ ॏᏱਣȂηϛӤޠНࡧདྷِ࡚џࡧՄಜ ॏᏱȄΠ၍ಜॏᏱޠᐤѭึȂӶᏱಭಜॏᏱ ޠၰαηΫϸԥ፹ޠΚӈٲȄ ഷࡤདᗃᇄᢋτᏱަཽऌᏱଲ࢈ݾᏱق༂ ᱞఁ௳ӶಜॏѭІࣻڐᜱޤᜌޠΚϹࡿᏳȂ 44 ᪺ᒙ፲⚼ 2018ԑ62ڢ8 Ѯѕѿᚂৱϴོོѐ 5. ༂ᱞ(2017)ȄಜॏᏱࡧޠၰ-፤౪ᇅᔗң (ߒ)ޏȄᇄіȈϥࠓȄ 6. Eisenhart, Churchill. (1983).ɆLaw of error I: Development of the concept,ɇ in Encyclopedia of Statistics Science, S. Kotz, N. L. Johnson, and C. B. Read (Eds.), Vol. 4, pp. 532-3. (Figure 1) New York: Wiley.