睡眠見守りセンサーデータの構造方程式モデリングによる因果分析

⇃ᮏ㧗➼ᑓ㛛Ꮫᰯ ◊✲⣖せ ➨ ྕ㸦㸧

╧╀ぢᏲࡾࢭࣥࢧ࣮ࢹ࣮ࢱࡢᵓ㐀᪉⛬ᘧࣔࢹࣜࣥࢢ࡟ࡼࡿᅉᯝศᯒ

኱▼ ಙᘯ

1,*

_{ ᒣᮏ ┤ᶞ}

1

_{ ▼⏣ ᫂⏨}

2

_{ ᮧୖ ⣧}

1

A Causal Analysis by Structural Equation Modeling of Sleep Monitoring Sensor Data

Nobuhiro Oishi1,*_{, Naoki Yamamoto}2_{, Akio Ishida}3_{, Jun Murakami}2

In this paper, structural equation modeling (SEM) is used to analyze the causal relationship between sleep stages and environmental data. The data used for the analysis is obtained by a care support device used in an elderly care facility. By applying the stepwise method to this data, we were able to find four observation variables that affect sleep stages. The latent variables are then determined by a scree plot. We proposed a causal model in which four observed variables and two latent variables affect sleep levels. Statistical analysis environment R and the lavaan package were used for SEM analysis in this paper. The results show that SEM can be used to build a rational model for the effects of vital signs and environmental conditions on good sleep.

࣮࣮࢟࣡ࢻ㸸╧╀ぢᏲࡾࢭࣥࢧ࣮ࠊᅉᯝศᯒࠊᵓ㐀᪉⛬ᘧࣔࢹࣜࣥࢢࠊR ゝㄒ

.H\ZRUGV㸸sleep monitoring sensor, causal analysis, structural equation modeling, R language

㸯㸬ࡣࡌࡵ࡟

㏆ᖺࠊ኱㔞ࡢࢹ࣮ࢱࡀᐜ᫆࡟ᡭ࡟ධࡿࣅࢵࢢࢹ࣮ࢱࡢ᫬ ௦࡟࡞ࡗࡓ(1)_{ࠋ฼⏝ྍ⬟࡞ࢹ࣮ࢱ࡜ࡋ࡚ࡣࠊID ௜ POS ࢹ࣮} ࢱࡸࢿࢵࢺ㏻㈍ࡢ㉎㈙ࢹ࣮ࢱࠊ࢙࢘ࣈࢧ࢖ࢺࡢ㜀ぴ⤫ィࡸ SNS ࡛ࡢⓎಙ⤫ィࠊ஌⏝㌴ࡢ⮬ື㐠㌿ࡸ࢙࢘࢔ࣛࣈࣝ➃ᮎ ➼࡟ᦚ㍕ࡉࢀࡓྛ✀ࢭࣥࢧ࣮࡟ࡼࡿ≀⌮㔞ࢹ࣮ࢱࡸ఩⨨᝟ ሗࠊ⏕య᝟ሗࠊN ࢩࢫࢸ࣒࡟௦⾲ࡉࢀࡿᅛᐃ࣓࢝ࣛࡢ⏬ീ ࠿ࡽᚓࡽࢀࡿྛ✀ࡢ᝟ሗ࡞࡝ࡀ࠶ࡿࠋ ほ ࡉࢀࡿ᝟ሗ㛫ࡢ㛵ಀᛶࢆ⤫ィⓗ࡟ศᯒࡍࡿࡇ࡜࡛ࠊ ௒ࡲ࡛᭕᫕ࡔࡗࡓほ ኚᩘ㛫ࡢ㛵ಀᛶࢆࡣࡗࡁࡾ࡜♧ࡍࡇ ࡜ࡀ࡛ࡁࡿࡼ࠺࡟࡞ࡗࡓࠋ౛࠼ࡤࠊ㔜ᅇᖐศᯒࠊࡉࡽ࡟ࡣ ₯ᅾኚᩘࢆ⤌ࡳ㎸ࢇ࡛ࣔࢹࣝ໬ࡋࡓ᥈⣴ⓗᅉᏊศᯒ㸦EFA ; Exploratory Factor Analysis 㸧 ࡸ ☜ ㄆ ⓗ ᅉ Ꮚ ศ ᯒ 㸦 CFA ; Confirmatory Factor Analysis㸧࡞࡝ࡀᣲࡆࡽࢀࡿࠋ

ほ ࡉࢀࡓ᝟ሗ㛫࡟ࡣࠊ㛵ಀᛶࡀ࠶ࡿࡤ࠿ࡾ࡛࡞ࡃࠊほ  ኚᩘ࡜ほ ኚᩘࡢ㛫࡟ᅉᯝ㛵ಀࡀᏑᅾࡍࡿࡶࡢࡶ࠶ࡿࠋ ࡇࡢᅉᯝ㛵ಀࢆ᫂ࡽ࠿࡟ࡍࡿศᯒἲ࡜ࡋ࡚ࠊࢢࣛࣇ࢕࢝ࣝ ࣔࢹࣜࣥࢢ㸦GM; Graphical Modeling) (3)_{ࡸᵓ㐀᪉⛬ᘧࣔࢹࣜ}

ࣥࢢ㸦SEM; Structural Equation Modeling) (3, 6-8)_ࡀ࠶ࡿࠋ

⩻ࡗ࡚ࡇࢀ࠿ࡽࡢ᪥ᮏࡀ┤㠃ࡍࡿ♫఍ⓗ࡞ၥ㢟ࢆ⪃࠼࡚ ࡳࡿ࡜ࠊᮏ᱁ⓗ࡞㧗㱋໬♫఍࡜ྠ᫬࡟ฟ⏕⋡పୗࡀᘬࡁ㉳ ࡇࡍປാຊேཱྀࡢῶᑡၥ㢟ࡀ࠶ࡿࠋࡇࡢၥ㢟ࡀࡍ࡛࡟⏕ࡌ ࡚࠸ࡿ⌧ሙࡶ࠶ࡾࠊ≉࡟⪁ே௓ㆤ᪋タ࡟࠾ࡅࡿ௓ㆤ⚟♴ኈ ࡢேᡭ୙㊊ࡣ῝้࡛ࠊᑵປ⪅࡬ࡢ㐣ᗘ࡞ປാ㈇ᢸ࡟ࡼࡿධ ᡤ⪅࡬ࡢࢧ࣮ࣅࢫపୗ࡞࡝ࡢၥ㢟ࡀ㉳ࡁ࡚࠸ࡿ(2)_ࠋࡑࡢ౛ ࡜ࡋ࡚ࠊධᡤ⪅ࡀኪ㛫╀ࢀࡎ࡟࣋ࢵࢻࢆ㞳ᗋࡍࡿ࡜ࠊᙜ┤ ໅ົࡢ௓ㆤኈࡀࡑࡢᑐᛂ࡟㏣ࢃࢀࠊᮏ᮶௚ࡢධᡤ⪅ࢆୡヰ ࡍࡿࡓࡵࡢ᫬㛫ࡀ๐ࡽࢀࡿ࡜࠸࠺ၥ㢟ࡀᣲࡆࡽࢀࡿࠋ ࡑࡇ࡛ࠊධᡤ⪅࡟ࡼࡾࡼ࠸╧╀ࢆ࡜ࡗ࡚ࡶࡽ࠼ࢀࡤࠊ௓ ㆤ⚟♴ኈࡢປാ㈇ᢸࢆᑡࡋ࡛ࡶ㍍ࡃ࡛ࡁࡿ࡜⪃࠼ࡓࠋᮏㄽ ᩥ࡛ࡣࠊ⪁ே௓ㆤ᪋タ࡛౑⏝ࡉࢀ࡚࠸ࡿ௓ㆤࢧ࣏࣮ࢺᶵჾ ࠿ࡽᚓࡽࢀࡿ⎔ቃࢹ࣮ࢱ࡜╧╀ࡢ῝ࡉ࡜ࡢᅉᯝ㛵ಀࢆศᯒ ࡍࡿࡇ࡜࡛ࠊࡼࡾࡼ࠸╧╀ࢆᚓࡿࡓࡵࡢ᮲௳ࢆぢࡘࡅࡿࡇ ࡜ࢆ┠ⓗ࡜ࡋࡓࠋࡑࢀ࡟ࡼࡾࠊධᡤ⪅ྛேࡀࡼࡾࡼ࠸╧╀ ࢆᚓࡽࢀࡿࡼ࠺࡟࡞ࡾࠊ೺ᗣ࡞⏕άࢆ㏦ࡿࡇ࡜ࡀ࡛ࡁࡿ࡜ ࡜ࡶ࡟ࠊධᗋᚋࡢᕠᅇࢱ࢖࣑ࣥࢢࢆ㐺ษ࡟࡛ࡁࠊ୍ேᙜࡓ ࡾࡢ௓ㆤ⚟♴ኈ࡟࠿࠿ࡿປാ㈇ᢸࢆ㍍ῶࡍࡿࡇ࡜ࡀྍ⬟࡟ ࡞ࡿ࡜ᮇᚅ࡛ࡁࡿࠋ ᮏㄽᩥ࡟࠾ࡅࡿࢹ࣮ࢱࡢศᯒ࡟ࡣࠊ⤫ィゎᯒ⎔ቃ R(4)_ࢆ ⏝࠸ࡓࠋR ࡣࣇ࣮࢙ࣜ࢘࢔࡛࠶ࡾࠊᩍ⫱ⓗ࡟ࡶࡇࡢゝㄒࡢ ౑⏝ࡀ᭷┈࡛࠶ࡿ࡜⪃࠼ࡽࢀࡿ(5)_{ࠋᅉᯝ㛵ಀࡢศᯒᡭἲ࡜} ࡋ࡚ࡣᵓ㐀᪉⛬ᘧࣔࢹࣜࣥࢢ㸦SEM㸧ࢆ᥇⏝ࡋࠊR ࡢ lavaan ࣃࢵࢣ࣮ࢪ(9)_{ࢆ⏝࠸࡚ᐇ⿦ࡋࡓࠋࡇࡢࣃࢵࢣ࣮ࢪࡣ⌧ᅾࡶ} 㛤Ⓨࡀ⥆࠸࡚࠸ࡿࣃࢵࢣ࣮ࢪ࡛ࡣ࠶ࡿࡀࠊศᯒ⤖ᯝࡣཝᐦ ࡛ṇ☜࡛࠶ࡿ࡜ศᯒᛶ⬟࡟ࡘ࠸࡚ࡣᐃホࡀ࠶ࡿ(7)_ࠋࡲࡓࠊ ࡇࡢࣃࢵࢣ࣮ࢪ࡛᥇⏝ࡉࢀ࡚࠸ࡿࣔࢹࣝグ㏙ᩥἲࡶ⡆༢࡛ ࠶ࡿ࡜࡜ࡶ࡟ࠊᵝࠎ࡞ࣔࢹࣝࢆ⾲⌧࡛ࡁࡿỗ⏝ᛶࡀ࠶ࡿࠋ

㸰㸬ᅉᯝศᯒ࡟⏝࠸ࡿࢹ࣮ࢱࡢ㑅ᢥ

ᮏㄽᩥ࡛ࡣࠊASD ♫〇ࡢ௓ㆤ᪋タྥࡅ╧╀ぢᏲࡾࢭࣥࢧ ࣮ࠕࡲࡶࡿ㹼ࡢࠖ(10)_{ࢆᐇ㝿࡟᪋タ࡛฼⏝ࡋࡓࢹ࣮ࢱࡢᥦ౪}  1 _{㟁Ꮚ᝟ሗࢩࢫࢸ࣒ᕤᏛ⣔} ࠛ861-1102 ⇃ᮏ┴ྜᚿᕷ㡲ᒇ 2659-2

Faculty of Electronics and Information Systems Engineering, 2659-2 Suya, Koshi-shi, Kumamoto, Japan 861-1102   2 _{ࣜ࣋ࣛࣝ࢔࣮ࢶ⣔}

ࠛ861-1102 ⇃ᮏ┴ྜᚿᕷ㡲ᒇ 2659-2 Faculty of Liberal Arts,

2659-2 Suya, Koshi-shi, Kumamoto, Japan 861-1102  * Corresponding author:

E-mail address: oishi@kumamoto-nct.ac.jp (N. Oishi).

㏿ ሗ

― 83 ―

╧╀ぢᏲࡾࢭࣥࢧ࣮ࢹ࣮ࢱࡢᵓ㐀᪉⛬ᘧࣔࢹࣜࣥࢢ࡟ࡼࡿᅉᯝศᯒ㸦኱▼࣭ᒣᮏ࣭▼⏣࣭ᮧୖ㸧

  5HVHDUFK5HSRUWVRI 1,7.XPDPRWR&ROOHJH 9RO  ࢆཷࡅ࡚ศᯒࢆ⾜ࡗࡓࠋࡇࡢ⿦⨨ࡣ௓ㆤ᪋タࡢ࣋ࢵࢻ࡟ྲྀ ࡾ௜ࡅࠊẼ ࠊ‵ᗘࠊẼᅽࠊ↷ᗘࠊ⏕య᝟ሗ㸦ᚰᢿᩘࠊ࿧ ྾ᩘࠊ╧╀ࣞ࣋ࣝ㸧ࠊධᗋࢹ࣮ࢱ㸦㉳ᗋࠊධᗋࠊ㞳ᗋ㸧ࢆほ  ࡋࠊࢡࣛ࢘ࢻ࡬࢔ࢵࣉ࣮ࣟࢻࡍࡿࠋ ᐃ㛫㝸ࡣ1 ศ㛫㝸 ࣮ࣔࢻ࡜5 ศ㛫㝸࣮ࣔࢻࡢ 2 ࡘࡢ ᐃ࣮ࣔࢻࡀ࠶ࡾࠊ1 ࣧ ᭶ศࡢࢹ࣮ࢱࡀࢡࣛ࢘ࢻୖ࡟⮬ືⓗ࡟ಖᏑ࣭ಖ⟶ࡉࢀࡿࠋ ௒ᅇࡢศᯒࡢࡓࡵ࡟⏝ពࡋࡓࢹ࣮ࢱࡣ1 ேศ࡟㛵ࡍࡿࠊ 5 ศ㛫㝸ࡢ ᐃ࣮ࣔࢻ࡛ᚓࡽࢀࡓᐊ 㸦WP㸧ࠊ‵ᗘ㸦KP㸧ࠊ Ẽᅽ㸦DW㸧ࠊ↷ᗘ㸦LO㸧ࠊ1 ศ๓ࡢ࿧྾ᩘ㸦UU㸧ࠊ1 ศ๓ࡢᚰᢿ ᩘ㸦KU㸧࠾ࡼࡧ╧╀ࣞ࣋ࣝ㸦VV㸧ࡢ 7 ಶࡢほ ኚᩘ࡛࠶ࡿ 㸦ᣓᘼෆ࡟ほ ኚᩘྡࢆ♧ࡍ㸧ࠋVV ࡣࣞ࣋ࣝ 1 ࠿ࡽࣞ࣋ࣝ 4 ࡢ4 ್ࢆྲྀࡿࠋࡑࡢ௚ࡢほ ኚᩘࡣࠊṇࡢᐇᩘ࡛࠶ࡿࠋศ ᯒ࡟⏝࠸ࡓ᫬⣔ิࢹ࣮ࢱࡣḞᦆ್ࡢ࠶ࡿ᫬Ⅼࢆ㝖ࡃ 2926 ᫬Ⅼ㸦⣙10 ᪥ศ㸧ࡢࡶࡢ࡛࠶ࡿࠋ ᅗ㸯࡟᫬⣔ิࢹ࣮ࢱࡢ୍㒊ࡢほ ኚᩘࡢ࠺ࡕࠊ࠶ࡿ᫬้ ࠿ࡽ480 ศ㛫㸦8 ᫬㛫㸧ࡢࢹ࣮ࢱࢆᢤ⢋ࡋ࡚♧ࡍࠋVV ࡸ UU ࠾ࡼࡧKU ࡣ್ࡢኚ໬ࡀ⃭ࡋࡃࠊ୍ぢࡍࡿ࡜ࣛࣥࢲ࣒࡛࠶ ࡿ࠿ࡢࡼ࠺࡟ࡉ࠼ぢ࠼ࡿࠋ୍᪉ࠊKP ࡣኚ໬ࡀᑡ࡞ࡃࠊ࠶ࡿ ᫬้ࢆቃ࡟ࡋ࡚㧗್࠿ࡽప್࡬࡜ኚࢃࡗ࡚࠸ࡿࡇ࡜ࡀศ࠿ ࡿࠋ 㸵ಶࡢほ ࢹ࣮ࢱ㛫ࡢ㛵ಀࢆぢࡿࡓࡵ࡟ࠊ┦㛵ᅗ⾜ิࢆ ᅗ㸰࡟♧ࡍࠋUU ࡟ࡣ 0 ࡢ್ࡀ࠶ࡿࡀࠊࡇࢀࡣ╧╀᫬↓࿧྾ ⑕࡞࡝ࡢ඙ೃࢆ♧ࡍ࡜⪃࠼ࡽࢀࡿࡓࡵࠊእࢀ್࡜ࡋ࡚ࡢฎ ⌮ࡣࡋ࡚࠸࡞࠸ࠋྛኚᩘ㛫࡟ࡣᙉ࠸┦㛵ࢆぢࡿࡇ࡜ࡀ࡛ࡁ ࡎࠊVV ࡜ྛኚᩘ㛫ࡢ┦㛵㛵ಀࢆࡇࡢ⾜ิ࠿ࡽㄞࡳྲྀࡿࡇ࡜ ࡣ㞴ࡋ࠸ࠋ ࡇࡢࡼ࠺࡟ VV ࡜ࡢ㛵ಀᛶࡀุ↛࡜ࡋ࡞࠸ 7 ಶࡢほ ኚ ᩘࡢࡍ࡭࡚ࢆ౑ࡗ࡚ᅉᯝศᯒࢆࡋࡓ࡜ࡇࢁࠊࣔࢹࣝࡢ㐺ྜ ᗘᣦᶆࡀⰋ㐺ྜࢆ♧ࡉ࡞࠿ࡗࡓࠋ㸦౛࠼ࡤᚋ㏙ࡢᣦᶆ࡛ࡣࠊ AGFI = 0.936ࠊRMSEA = 0.103 ࡞࡝㸧ࠋࡑࡢࡓࡵࠊศᯒ࡟౑ ⏝ࡍࡿኚᩘࢆῶࡽࡍࡇ࡜࡟ࡋࡓࠋኚᩘ㑅ᢥ(11)_{࡟ࡣࢫࢸࢵࣉ} ࣡࢖ࢬἲ(12) _{ࢆ⏝࠸ࠊḟᘧࡢ㉥ụ᝟ሗ㔞ᇶ‽㸦AIC ; Akaike’s} Information Criterion㸧ࢆࣔࢽࢱ࣮ࡋ࡞ࡀࡽࠊ᥇⏝ࡍࡿほ ኚ ᩘࢆỴᐃࡋࡓ(12)_ࠋ AIC ൌ m ൈ lnσ ቄ௬ሺ೔ሻି௬ಶೄ೅ሺ೔ሻቅ మ ೙ ೔సభ ୫ ൅ 2 ... (㸯) ݊:ࢧࣥࣉࣝᩘ m:ㄝ᫂ኚᩘࡢᩘ ݕሺ௜ሻ_{:i ␒┠ࡢࢧࣥࣉࣝ࡟࠾ࡅࡿ┠ⓗኚᩘࡢ್} ݕ_ாௌ்ሺ௜ሻ: i ␒┠ࡢࢧࣥࣉࣝ࡟࠾ࡅࡿ┠ⓗኚᩘࡢ᥎ᐃ್ ᐇ㝿࡟ࡣࠊ┠ⓗኚᩘࢆVV ࡜ࡋ࡚ࠊR ࡢ step 㛵ᩘࢆ⏝࠸࡚ ᐇ᪋ࡋࡓࠋࡑࡢ⤖ᯝࠊ⾲㸯࡟♧ࡍࡼ࠺࡟ࠊLO ࠊKP ࠊKU ࠾ ࡼࡧ UU ࡢ 4 ಶࡢኚᩘࢆ᥇⏝ࡋ࡚ VV ࢆ⾲ࡋࡓሙྜ࡟ AIC ࡀ᭱ᑠ࡜࡞ࡿࡇ࡜ࡀศ࠿ࡗࡓࠋࡇࡢ⤖ᯝ࠿ࡽࠊ┠ⓗኚᩘࢆ ྵࡵ࡚5 ಶࡢほ ኚᩘࢆ౑⏝ࡋ࡚ศᯒࢆ⾜࠺ࡇ࡜࡟ࡋࡓࠋ

㸱㸬₯ᅾኚᩘࡢᑟධ

ୖ㏙ࡢ5 ಶࡢኚᩘࡢ⫼ᚋ࡟࠶ࡿᵓᡂᴫᛕࢆ⾲ࡍࡓࡵ࡟ࠊ ₯ᅾኚᩘࢆᑟධࡍࡿࠋࡑࡢಶᩘࢆ᳨ウࡍࡿࡓࡵ࡟ࠊほ ኚ ᩘ㛫ࡢ┦㛵ಀᩘ⾜ิࡢᅛ᭷್ࢆồࡵࠊᅗ㸱ࡢࢫࢡ࣮ࣜࣉࣟ ࢵࢺࢆᚓࡓࠋࡇࡢᅗ࠿ࡽࠊ➨3 ᅛ᭷್௨㝆ࡣ࡯ࡰ࡞ࡔࡽ࠿ ࡞᭤⥺࡜࡞ࡿࡢ࡛ࠊ₯ᅾኚᩘࡢಶᩘࡣ2 ಶ࡜ࡋࡓࠋ࠾࠾ࡼ ࡑ1 ௨ୖࡢ್ࢆᣢࡘᅛ᭷್ࡢಶᩘࡶ 2 ಶ࡛࠶ࡿࡢ࡛ࠊ࢝࢖ ࢨ࣮ᇶ‽(7)_{࡟ࡶྜ⮴ࡋ࡚࠸ࡿࠋ} ࡑࡇ࡛ࠊ2 ಶࡢ₯ᅾኚᩘࡀ⾲ࡍᵓᡂᴫᛕ࡜ࠊࡑࢀࡒࢀࡢ ₯ᅾኚᩘ࡟⤖ࡧ௜ࡅࡿほ ኚᩘࢆ௨ୗࡢࡼ࠺࡟ࡍࡿࠋ I㸦ᐊෆ⎔ቃ㸧㸸↷ᗘ㸦LO㸧࠾ࡼࡧ‵ᗘ㸦KP㸧 I㸦ࣂ࢖ࢱࣝࢧ࢖ࣥ㸧㸸ᚰᢿᩘ㸦KU㸧࠾ࡼࡧ࿧྾ᩘ㸦UU㸧  ⾲㸯 ࢫࢸࢵࣉ࣡࢖ࢬἲ࡟ࡼࡿኚᩘ㑅ᢥࡢ㐣⛬ ⟬ἲ 㑅ᢥࡋࡓኚᩘ AIC ⟬ἲ ኚᩘῶ ᑡἲ Ў KUUUWPKPLODW 775.55 KUUUWPKPLO 774.22 KUUUKPLO 773.84 ኚᩘቑ ῶἲ Ĺ LOKUUU 780.23 LOKU 817.61 LO 829.55 ࡞ࡋ 990.38 ᅗ㸯 ほ ࢹ࣮ࢱࡢ᫬⣔ิ⾲♧㸦ᢤ⢋㸧              WLPH>K@ VO HHS VW DJ H         WLPH>K@ KHDU W UD WH > EHDW V  P @         WLPH>K@ UH VS LUD WLR Q UD WH > WLP HV  P @         WLPH>K@ KX P LG LW\ >  @ ᅗ㸰 ほ ࢹ࣮ࢱࡢᩓᕸᅗ⾜ิ VV                   KU UU     _WP KP       LO                  DW ― 84 ―

⇃ᮏ㧗➼ᑓ㛛Ꮫᰯ ◊✲⣖せ ➨ ྕ㸦㸧

㸲㸬ᥦ᱌ࣔࢹࣝ࡜ࣃࢫಀᩘ

ᅉᯝ㛵ಀࢆ⾲ࡍ࡟ࡣࠊከ㔜ᣦᶆࣔࢹࣝࡸࠊከ㔜ᣦᶆከ㔜 ཎᅉ㸦MIMIC ; Multiple Indicator Multiple Cause㸧ࣔࢹࣝࠊ㒊 ศⓗ᭱ᑠ஧஌㸦PLS ; Partial Least Squares㸧ࣔࢹࣝ࡞࡝ࡀ࠶ ࡿ(7)_{ࠋ௒ᅇ᫂ࡽ࠿࡟ࡋࡓ࠸ࡢࡣࠊほ ኚᩘ࡛࠶ࡿ╧╀ࣞ࣋} ࣝ㸦VV㸧࡟ࠊ௚ࡢ₯ᅾኚᩘࡸほ ኚᩘࡀ࡝࠺ᙳ㡪ࢆཬࡰࡋ࡚ ࠸ࡿ࠿࡜࠸࠺ࡇ࡜࡛࠶ࡿࡢ࡛ࠊMIMIC ࣔࢹࣝࡸ PLS ࣔࢹ ࣝࡀ㐺ࡋ࡚࠸ࡿ࠿ࡽࠊᅗ㸲࡟♧ࡍࣔࢹࣝࢆᥦ᱌ࡍࡿࠋᐊෆ ⎔ቃ㸦I㸧ࡀࣂ࢖ࢱࣝࢧ࢖ࣥ㸦I㸧࡟ᙳ㡪ࢆཬࡰࡍࡇ࡜ࡣ࠶ ࡲࡾ⪃࠼ࡽࢀࡎࠊࡑࡢ㏫᪉ྥࡢᙳ㡪ࡶ⪃࠼࡟ࡃ࠸ࠋࡑࡢࡓ ࡵࠊࡇࡢࣔࢹࣝ࡟࠾࠸࡚ࠊI ࡜ I ࡣ஫࠸࡟┤஺ࡋ࡚࠸ࡿ࡜ ௬ᐃࡋࡓࠋࡇࡢࣔࢹࣝࡢࣃࢫಀᩘ࠾ࡼࡧVV ࡟⤖ࡧ௜࠸࡚࠸ ࡿㄗᕪศᩓࢆồࡵࡿࡓࡵࡢ ᐃ᪉⛬ᘧ࡜ᵓ㐀᪉⛬ᘧࡣḟࡢ ࡼ࠺࡟࡞ࡿࠋ ቐ ݂1 ൌ ߙଵή ݈݅ ൅ ߙଶή ݄݉ ݂2 ൌ ߙଷή ݄ݎ1 ൅ ߙସή ݎݎ1 ݏݏ ൌ ߛଵή ݂1 ൅ ߛଶή ݂2 ൅ ݁ଵ ... (㸰) ࡇࡇ࡛ࠊI ࡜ I ࡣෆ⏕ኚᩘ࡛ࡣ࠶ࡿࡀࠊㄗᕪኚᩘࢆకࡗ ࡚࠸࡞࠸ࠋࡇࢀࡣPLS ࣔࢹ࡛ࣝࡣᕥ㎶࡟⨨࠿ࢀࡓෆ⏕ኚᩘ ࡛࠶ࡿ₯ᅾኚᩘࡀྑ㎶ࡢእ⏕ኚᩘ࡟ࡼࡗ࡚ᐃ⩏ࡉࢀࡿ࡜⪃ ࠼ࡿࡓࡵ࡛࠶ࡿࠋࡘࡲࡾࠊᐊෆ⎔ቃ㸦I㸧ࡣ↷ᗘ㸦LO㸧࡜‵ ᗘ㸦KP㸧࡟ࡼࡾᐃ⩏ࡉࢀ࡚࠸ࡿᵓᡂᴫᛕ࡜ᤊ࠼ࠊྠᵝ࡟ࠊ ࣂ࢖ࢱࣝࢧ࢖ࣥ㸦I㸧ࡣᚰᢿᩘ㸦KU㸧࡜࿧྾ᩘ㸦UU㸧࡟ࡼ ࡗ࡚ᐃ⩏ࡉࢀࡿᵓᡂᴫᛕ࡜ᤊ࠼࡚࠸ࡿࠋࡑࡢࡓࡵࠊI ࡶ I ࡶㄗᕪኚᩘࢆకࢃ࡞࠸ࠋ (㸰)ᘧࢆࠊlavaan ࡢࣔࢹࣝグ㏙ᩥἲ࡛グ㏙ࡋࠊlavaan ࣃࢵ ࢣ࣮ࢪࡢsem 㛵ᩘࢆ⏝࠸࡚ẕᩘࢆ᥎ᐃࡋࡓࠋ᥎ᐃ࡟ࡣ᭱ᑬ 㸦ML ; Maximum Likelihood㸧᥎ᐃἲࢆ⏝࠸ࡓࠋ࡞࠾ࠊゎᯒ ࡟⏝࠸ࡓほ ࢹ࣮ࢱࡣᶆ‽໬ࡋࡓࡶࡢࢆ⏝࠸ࡓࠋࡇࡢࢫࢡ ࣜࣉࢺࢆᅗ㸳࡟♧ࡍࠋᶆ‽໬ࡋࡓẕᩘࡢ᥎ᐃ್ࡣ⾲㸰ࡢࡼ ࠺࡟࡞ࡗࡓࠋ⾲୰࡛ࠊࣃࢫಀᩘȘ1࡜Ș3ࡀ㈇ࡢ್࡜࡞ࡗ࡚ ࠸ࡿࠋࡇࢀࡣࠊ↷ᗘ㸦LO㸧ࡀᙉࡃ㸦᫂ࡿࡃ㸧࡞ࢀࡤᐊෆ⎔ቃ ࡢ್㸦I㸧ࡀᑠࡉࡃ࡞ࡾࠊ㏫࡟ࠊᬯࡃ࡞ࢀࡤᐊෆ⎔ቃࡢ್ࡀ ኱ࡁࡃ࡞ࡿࡇ࡜ࢆព࿡ࡍࡿࠋྠᵝ࡟ࠊᚰᢿᩘ㸦KU㸧ࡀ኱ࡁ ࡃ㸦᪩ࡃ㸧࡞ࢀࡤࣂ࢖ࢱࣝࢧ࢖ࣥࡢ್㸦I㸧ࡀᑠࡉࡃ࡞ࡾࠊ ᚰᢿᩘࡀᑠࡉࡃ㸦㐜ࡃ㸧࡞ࢀࡤࣂ࢖ࢱࣝࢧ࢖ࣥࡢ್ࡀ኱ࡁ ࡃ࡞ࡿࡇ࡜ࢆព࿡ࡋ࡚࠸ࡿࠋࣃࢫಀᩘࡢ኱ࡁࡉࡣࠊࡑࡢ್ ࡀ኱ࡁ࠸࡯࡝ࠊᙳ㡪ࡀ኱ࡁ࠸ࡇ࡜ࢆ⾲ࡋ࡚࠸ࡿࠋࡇࢀࡽࡢ ࣃࢫಀᩘࡣᅗ㸲࡟ࡶグධࡋࡓࠋ ᥦ᱌ࣔࢹࣝࡢ㐺ྜᗘᣦᶆ(7)_{ࢆ⾲㸱࡟♧ࡍࠋGFI㸦Goodness}

of Fit Index㸧ࠊAGFI㸦Adjusted GFI㸧࠾ࡼࡧ CFI㸦Comparative Fit Index㸧ࡣほ ኚᩘࡢศᩓ࡟ᑐࡍࡿࣔࢹࣝࡢㄝ᫂⋡࡜࠸ ࠺ほⅬ࠿ࡽศᯒࡢ⢭ᗘࢆホ౯ࡍࡿᣦᩘ࡛ࠊ0 ࠿ࡽ 1 ࡲ࡛ࡢ ್ࢆྲྀࡾࠊ0.95 ࡼࡾ኱࡛࠶ࡿ࡜Ⰻ㐺ྜ࡜ุ᩿ࡉࢀࡿࠋࡇࡢ ್ࡀ1.000 ࡜࡞ࡗ࡚࠸ࡿࡢࡣࠊᥦ᱌ࣔࢹࣝࡢ⮬⏤ᗘࡀ 0 ࡛ ࠶ ࡿ ࡓ ࡵ ࡛ ࠶ ࡿ ࠋRMSEA 㸦 Root Mean Square Error of Approximation㸧࡜ SRMR㸦Standardized Root Mean Square

 ⾲㸰 ᶆ‽໬ࡋࡓẕᩘࡢ᥎ᐃ್ Ș1 Ș2 Ș3 Ș4 Ț1 Ț2 H ࡢศᩓ -0.923 0.238 -0.959 0.274 0.223 0.117 0.928 ᅗ㸱  ኚᩘ㑅ᢥ᫬ࡢࢫࢡ࣮ࣜࣉࣟࢵࢺ            QXPEHURIODWHQWYDULDEOHV HLJH Q YDOXH ᅗ㸲 ᥦ᱌ࣔࢹࣝࡢࣃࢫᅗ ODYDDQ ࣃࢵࢣ࣮ࢪࢆ⏝࠸࡚ࠊᥦ᱌ࣔࢹࣝ࡟ 6(0 ࢆ㐺⏝ࡍࡿ ኚᩘ㑅ᢥᚋࠊᶆ‽໬ࡋࡓࢹ࣮ࢱࡀGDWDGDW࡟᱁⣡ࡉࢀ࡚࠸ࡿ 㐺ྜᗘᣦᶆ㸸*),$*),506($&),6505$,&%,&  OLEUDU\ODYDDQ  PRGHOᅗ  ࡢᥦ᱌ࣔࢹࣝࢆ ODYDDQ ࡢᩥἲ࡛グ㏙ PRGHO IaLOKP IaKUUU II aVV Iaa I Iaa I Iaa I  ILWODYDDQVHPPRGHOGDWD GDWDGDWRUWKRJRQDO 7 IL[HG[ 7VWGOY ) VXPPDU\ILWVWDQGDUGL]HG 7 ILW0HDVXUHVILWILWPHDVXUHV  FJILDJILUPVHDFILVUPUDLFELF  ᅗ㸳 ODYDDQ ࡢࢫࢡࣜࣉࢺ ― 85 ― ⇃ᮏ㧗➼ᑓ㛛Ꮫᰯࠉ◊✲⣖せࠉ➨11ྕ㸦2019㸧

╧╀ぢᏲࡾࢭࣥࢧ࣮ࢹ࣮ࢱࡢᵓ㐀᪉⛬ᘧࣔࢹࣜࣥࢢ࡟ࡼࡿᅉᯝศᯒ㸦኱▼࣭ᒣᮏ࣭▼⏣࣭ᮧୖ㸧

  5HVHDUFK5HSRUWVRI 1,7.XPDPRWR&ROOHJH 9RO  Residual㸧ࡣᐇ㝿ࡢࢹ࣮ࢱ࡜ࣔࢹࣝ࡟ࡼࡿண ್࡜ࡢࡎࢀ ࡢᑠࡉࡉ࡟╔┠ࡋࡓホ౯ᣦᩘ࡛ࠊ0.05 ࡼࡾᑠ࡛࠶ࡿ࡜Ⰻ㐺 ྜ࡜ุ᩿ࡉࢀࡿࠋࡇࡢ್ࡀ0.000 ࡜࡞ࡗ࡚࠸ࡿࡢࡶࠊ⮬⏤ ᗘࡀ0 ࡛࠶ࡿࡓࡵ࡛࠶ࡿࠋ

㸳㸬ศᯒ⤖ᯝࡢゎ㔘

ศᯒࡢ⤖ᯝᚓࡽࢀࡓࣃࢫಀᩘࢆࡶ࡜࡟ࠊᅗ㸲ࡢᥦ᱌ࣔࢹ ࣝࡢࣃࢫᅗࢆゎ㔘ࡍࡿࠋᐊෆ⎔ቃ㸦I㸧ࡣ↷ᗘ㸦LO㸧࠾ࡼࡧ ‵ᗘ㸦KP㸧࡛ᐃ⩏ࡉࢀ࡚࠸ࡿࡀࠊࡑࡢ኱ࡁࡉࡣ-0.923 ࡜ 0.238 ࡛࠶ࡾࠊᐊෆ⎔ቃࡣ↷ᗘ࡟኱ࡁࡃᕥྑࡉࢀࡿࡇ࡜ࡀศ࠿ࡿࠋ ࣂ࢖ࢱࣝࢧ࢖ࣥ㸦I㸧ࡣᚰᢿᩘ㸦KU㸧࠾ࡼࡧ࿧྾ᩘ㸦UU㸧 ࡛ᐃ⩏ࡉࢀ࡚࠸ࡿࡀࠊࡑࡢ኱ࡁࡉࡣ-0.595 ࡜ 0.274 ࡛࠶ࡾࠊ ࣂ࢖ࢱࣝࢧ࢖ࣥࡢ್ࡣᚰᢿᩘ࡟ᕥྑࡉࢀࡿࠋ╧╀ࣞ࣋ࣝ 㸦VV㸧ࡣᐊෆ⎔ቃ࠿ࡽ 0.223 ࡢᙳ㡪ࢆཷࡅ࡞ࡀࡽࠊࣂ࢖ࢱࣝ ࢧ࢖ࣥࡢ್࠿ࡽࡶ0.117 ࡢᙳ㡪ࢆཷࡅ࡚࠸ࡿࡇ࡜࠿ࡽࠊ╧ ╀ࣞ࣋ࣝࡣࣂ࢖ࢱࣝࢧ࢖ࣥࡼࡾࡶᐊෆ⎔ቃ࡟ࡼࡾ኱ࡁࡃ౫ Ꮡࡋ࡚࠸ࡿ࡜ゝ࠼ࡿࠋ ྛほ ኚᩘLOࠊKPࠊKU ࠾ࡼࡧ UU ࡀ VV ࡟ཬࡰࡍᙳ㡪ࡢ ⥲ྜຠᯝࢆ௨ୗ࡟ぢ✚ࡶࡿࠋ LO Ѝ VV ࡢ⥲ྜຠᯝ㸸-0.923™0.223 = -0.206 KP Ѝ VV ࡢ⥲ྜຠᯝ㸸0.238™0.223 = 0.053 KU Ѝ VV ࡢ⥲ྜຠᯝ㸸-0.959™0.117 = -0.112 UU Ѝ VV ࡢ⥲ྜຠᯝ㸸0.274™0.117 = 0.032 ╧╀ࣞ࣋ࣝ࡟᭱ࡶᙳ㡪ࢆ୚࠼ࡿࡢࡣ↷ᗘ࡛ࠊࡑࡢḟ࡟኱ ࡁ࡞ᙳ㡪ࢆ୚࠼ࡿࡢࡣᚰᢿᩘ࡛࠶ࡾࠊ࡝ࡕࡽࡶ㈇ࡢ್࡛࠶ ࡿࠋࡘࡲࡾࠊ㒊ᒇࢆᬯࡃࡍࡿࡇ࡜ࡀ╧╀ࣞ࣋ࣝࢆୖࡆࡿࡓ ࡵ࡟ࡣ᭱ࡶຠᯝࡀ࠶ࡾࠊḟ࡟ᚰᢿᩘࢆᑠࡉࡃࡍࡿࡇ࡜ࡀຠ ᯝⓗ࡛࠶ࡿ࡜ゝ࠼ࡿࠋ

㸴㸬ࡲ࡜ࡵ

╧╀ࣞ࣋ࣝ࡟ᙳ㡪ࢆཬࡰࡍࡓࡃࡉࢇࡢほ ኚᩘࡢ୰࠿ ࡽࠊࢫࢸࢵࣉ࣡࢖ࢬἲ࡟ࡼࡾほ ኚᩘࢆ㑅ᢥࡍࡿࡇ࡜࡛ࠊ ᵓ㐀᪉⛬ᘧࣔࢹࣜࣥࢢ࡟ࡼࡿᅉᯝ㛵ಀࡢศᯒࢆ⾜࠺ࡇ࡜ࡀ ࡛ࡁࡓࠋศᯒ⤖ᯝ࠿ࡽࡣࠊࣂ࢖ࢱࣝࢧ࢖ࣥࡼࡾࡶᐊෆ⎔ቃ ࡀࡼࡾ኱ࡁࡃ╧╀࡟ᙳ㡪ࡍࡿ࡜࡜ࡶ࡟ࠊ῝࠸╧╀ࢆᚓࡿࡓ ࡵ࡟ࡣࠊ㒊ᒇࢆᬯࡃࡋࡓୖ࡛ࠊⴠࡕ╔࠸࡚ࡺࡗࡓࡾ࡜ࡋࡓ 㞺ᅖẼ࡟ࡍࡿ࡞࡝࡟ࡼࡾᚰᢿᩘࢆᑠࡉࡃࡍࡿ࡜ࡼ࠸ࡇ࡜ࡀ ศ࠿ࡗࡓࠋ ௒ᅇࡣࣃࢵࢣ࣮ࢪlavaan ࢆ⏝࠸࡚ศᯒࢆ⾜ࡗࡓࡀࠊᵓ㐀 ᪉⛬ᘧࡢグ㏙ࡀᐜ࡛᫆ࠊR ࡟ࡼࡾᅉᯝศᯒࡀ๭ྜ⡆༢࡟ᐇ ⾜ྍ⬟࡛࠶ࡿࡇ࡜ࡶ☜ㄆ࡛ࡁࡓࠋR ࡟ࡼࡾࢹ࣮ࢱศᯒࢆ⾜ ࠺㝿࡟ࡣࠊlavaan ࡟ࡼࡿᅉᯝศᯒࡶᗈࡃᬑཬࡍࡿ࡜ࡼ࠸࡜ ⪃࠼࡚࠸ࡿࠋ ௒ᚋࡢㄢ㢟࡜ࡋ࡚ࠊ௒ᅇᥦ᱌ࡋࡓࣔࢹ࡛ࣝࡣ╧╀ࣞ࣋ࣝ ࡀ῝ࡃ࡞ࡗ࡚࠸ࡃ㐣⛬࡟ࡘ࠸࡚ࡣ᫂ࡽ࠿࡟ࡋ࡚࠸࡞࠸ࡓ ࡵࠊᡂ㛗᭤⥺ࣔࢹࣝ࡟ࡼࡾ╧╀ࣞ࣋ࣝࡀ῝ࡃ࡞ࡗ࡚࠸ࡃ㐣 ⛬ࢆ᫂ࡽ࠿࡟ࡋࡓ࠸ࠋ 㸦௧࿴ඖᖺ9 ᭶ 19 ᪥ཷ௜㸧 㸦௧࿴ඖᖺ12 ᭶ 5 ᪥ཷ⌮㸧 ཧ⪃ᩥ⊩ (1) ⥲ົ┬㸸ࠕᖹᡂ29 ᖺ∧᝟ሗ㏻ಙⓑ᭩ࠖ, pp.52-62㸦2018㸧 (2) ཌ⏕ປാ┬㸸ࠕᖹᡂ 30 ᖺ∧ཌ⏕ປാⓑ᭩㸫㞀ᐖࡸ⑓ Ẽ࡞࡝࡜ྥࡁྜ࠸ࠊ඲࡚ࡢேࡀά㌍࡛ࡁࡿ♫఍࡟㸫ࠖ, pp.369-398㸦2019㸧 (3) ᑠᓥ㝯▮ࠊᒣᮏᑗྐ㸸ࠕExcel ࡛Ꮫࡪ ඹศᩓᵓ㐀ศᯒ ࡜ࢢࣛࣇ࢕࢝ࣝࣔࢹࣜࣥࢢࠖ, pp89-179, ࣮࣒࢜♫ (2013).

(4) “The R Project for Statistical Computing”, https://www.r-project.org/ , Retrieved Sep. 10, 2019㸬

(5) ▼⏣᫂⏨,ᒣᮏ┤ᶞ,኱▼ಙᘯ,ᮧୖ⣧㸸ࠕከḟඖࢹ࣮ࢱ ศゎࡢᡭἲࢆ⏝࠸ࡓ❧యࣃࢬࣝࡢゎἲ(ࡑࡢ 4)ࠖ㸪ึ➼ ᩘᏛ㸪➨86 ྕ㸪pp.20-24(2019)㸬 㸦⥅⥆୰㸧 (6) ᒣ⏣๛ྐࠊᮡ⃝Ṋಇࠊᮧ஭⣧୍㑻㸸ࠕR ࡟ࡼࡿࡸࡉࡋ ࠸⤫ィᏛࠖ㸪pp.309-319, ࣮࣒࢜♫ (2008)㸬 (7) ㇏⏣⚽ᶞ㸸ࠕඹศᩓᵓ㐀ศᯒ [R ೫] 㸫ᵓ㐀᪉⛬ᘧࣔ ࢹࣜࣥࢢ㸫ࠖ㸪pp.18-195, ᮾிᅗ᭩ (2014)㸬 (8) ㇏⏣⚽ᶞࠊ๓⏣ᛅᙪࠊᰗ஭ᬕኵ㸸ࠕཎᅉࢆࡉࡄࡿ⤫ィ Ꮫ  㸫 ඹ ศ ᩓ ᵓ 㐀 ศ ᯒ ධ 㛛 ࠖ㸪pp.99-132, ㅮ ㄯ ♫ (1992)㸬

(9) Yves Rosseel : “The lavaan tutorial”, pp.8-15, Ghent University(2019). (10)ࠕࡲࡶࡿ㹼ࡢࠖ, http://mamoruno.miel.care/ , Retrieved Sep. 10, 2019. (11)Ỉ㔝ḯྖ㸸ࠕከኚ㔞ࢹ࣮ࢱゎᯒㅮ⩏ࠖ, pp.61-69, ᮅ಴ ᭩ᗑ㸦1996㸧 (12)ࠕAIC ࢆ౑ࡗࡓኚᩘ㑅ᢥࠖ, http://www.hnami.or.tv/d/ index.php?radvance , Retrieved Sep. 10, 2019㸬 

⾲㸱 ᥦ᱌ࣔࢹࣝࡢ㐺ྜᗘᣦᶆ

GFI AGFI CFI RMSEA SRMR

1.000 1.000 1.000 0.000 0.000

― 86 ―