|
本研究主要是針對國內43篇論文研究，探討以Ohio州立大學理論為
架構下的領導行為與總體工作滿足間的關係，企圖以一種較為客觀的方式
而非傳統的文獻評論法，將以往的結果做一整合，並找出領導行為在不同
領域，或不同年代會不會因研究的一些屬性變項而有差異．研究的結果發
現，整合43篇論文研究後，由相當高的平均效果量可知領導行為對工作滿
足的影響貢獻良多．不同研究年代．研究工具及失測方式對研究的影響最
為明顯：其中隨著年代愈往後，效果量會有增加的趨勢；使有LBDQ的研究
效果量要高於SBDU的研究；郵寄問卷的研究比非郵寄問卷的平均效果量要
高．其他的屬性變項也在一些特定的情境下，和論文研究的效果量有相當
的關連存在． 研究最後也對上述的這些現象，提出一個新的潛在中介
變項：配合度(accommodation；受試者的配合程度），並利用現有的資料
檢視這個新的變項，得到配合度高的研究所得的效果量也較高，反之亦然
．而且結合配合度，測量工具，是否郵寄施測三種變項間的互動，幾乎可
以相當完整的解釋這些研究的變異為何會有如此大的差異存在．希望以後
對領導行為與工作滿足感興趣的研究者在參考本研究的結果後，能以更嚴
謹的方式設計研究步驟，更詳細的記錄相關的研究資料，更一致的研究用
語來描述研究結果．
The purpose of this study was to use meta-analysis to
investigate the relationshipbetween leadership behavior based on
the theory of Ohio State University and overalljob
sattisfaction. We attempted to seek a more objective method, not
traditional literature review, to conclude former researches of
domestic disserations of master''s and doctor''sdegree and to find
out whether there was any divergence in different researchful
domains,times or some attributes of research were being
different as time went on. After integrating43 researches of
dissertation, the result of high mean effect size implicated
that leadershipbehavior''s contribution to overall job
satisfaction was very conspicuous.Particularly, theeffects of
different fields and measuring instruments affected by the
results of research were obvious: As times became nearer, so the
effect size tended higher, the effect sizes of using LBDQ were
higher then the effect sizes of using SBDQ, the effect sizes of
using mailtest were higher than the effect sizes of using non-
mail test. The results of research wereaffected by the other
characteristic variables in speacific circumstances. Finally,
we will discuss potential inferences about preceding phenomenon
and propose anew moderating variable: accommodation (the extent
of collaboration of subjects), and utilize preceding datd to
test accommodation. The result indicated higher accommodation
attained higher effect size, and vice versa. Furthermore,
combined with three variables, accommodation, measuring
instrument, mail or non-mail test, and their interaction could
fairly, completely interpret why dissertation''s effect sizes had
such variation exist? I hopesubsequent investigators, who are
interesting in leadership behavior and job satisfaction,after
reading result of this dissertation, could use more scientific
strategy to design experiment and investigation, more detailed
to record relative data, and more correspondentterm to describe
result of study.
|