X hits on this document

62 views

0 shares

0 downloads

0 comments

17 / 25

386

SYMONS AND JOHNSON

Table 6 Effect Sizes as a Function of Taskfor SR-OR Manipulation

Variable and class

Stimulus type Trait Noun

Familiarity of rated oth High LOW Unable to rate Intimacy of rated other High Low Unable to rate Type of SR taskb Self-descriptiveness Imagery Other Type of other taskb Descriptiveness Imagery Nonspecific other Other Dependent variable Free recall Recognition Distractor task Present Absent Expectation of test Expect Do not expect Mode of stimulus p r e s e n t a t i o nc Monitor Projector

er

Read (by experimenter)

Participant population Undergraduates Other Experimental design Within subjects Between subjects

Between- class effect (QB)

57.81 ***

0.69

7.57*

54.46***

57.17"**

10.54"*

15.69"*

19.72"**

26.96***

10.07"*

1.25

35

0.58a

0.49

0.67

13

-0.01b

-0.16

0.13

14

0.19b

0.07

0.32

5

0.25b

0.02

0.47

61

0.36a

0.30

0.43

4

0.24~

-0.04

0.53

4

0.34a

0.16

0.51

44

0.42~

0.34

0.50

15

-0.02b

-0.18

0.14

8

0.64a

0.44

0.84

Mean weighted

95% CI for d÷

effect size

(d÷)

Lower Upper

0.07

0.33

0.34

0.49

0.16

0.51

0.21 0.39

0.36 0.61

0.39 0.14

0.56 0.32

-0.14 0.35

0.18 0.48

0.29

0.41

0.16

1.06

0.45

0.60

-0.08

0.13

17

0.20a

48

0.41~

4

0.34ac

28

0.48

41

0.23

65

0.35

4

0.61

43

0.53

26

0.03

39

0.53a

0.45

0.61

13

--0.01b

--0.16

0.13

15

0.12b

-0.02

0.26

51

0.28

18

0.50

0.22

0.36

0.40

0.62

k

15

0.02

54

0.41

52

0.29

17

0.51

Homogeneity within each class

(Qwi) a

    • 75.95

      **

    • 96.64

      ***

    • 225.89

      ***

1.38 2.44

18.26 202.12"** 2.44

    • 72.78

      **

    • 27.96

      ***

75.00***

    • 64.23

      **

    • 27.96

      **

22.87

    • 57.95

      ***

    • 149.35

      "**

    • 70.51

      "**

    • 75.66

      ***

    • 139.05

      "**

    • 46.90

      ***

    • 163.77

      "**

    • 129.58

      "**

    • 37.82

      ***

    • 18.08

      "

    • 156.89

      "**

    • 63.44

      ***

226.45*** 2.70

Note. Mean effect sizes sharing the same subscript do not differ (p > .05, post hoc). CI = confidence

interval; OR = other reference; SR = self-reference. Significance indicates rejection of the hypothesis of homogeneity,

bOne study, Bower and Gilligan (1979,

Experiment 2), included two SR versus OR comparisons that were omitted from this model test because

they

involved

the

only

autobiographical-biographical

tasks

in

the

SR-OR

manipulation

class,

c Two

studies (Brown et al., 1986, Experiment 3;

Wells

et

al.,

1984)

were

omitted

from

this model

test because

they did not *p < .05.

fit any coding class.

**p

<

.01.

***p

<

.001.

gardless of manipulation class. Finally, we draw conclusions based on significant model tests that are unique to each particu- lar manipulation class.

SRE in Memory Is Robust

One important finding of our study is that the SRE does occur with highly significant regularity. Although study findings are inconsistent, SR was superior to semantic and OR encoding in facilitating memory in the studies reviewed. It may seem to some readers that this is a foregone conclusion of any review

of the SRE, but there is some disagreement among researchers in this literature as to whether SR tends to promote memory better overall than other kinds of processes. For example, as we noted earlier, Higgins and Bargh's (1987) important narrative review concludes that "self-reference is neither necessary nor sufficient for memory of input to be facilitated" (p. 392). Our meta-analytic review suggests that, in fact, S R - - a l t h o u g h not necessary to promote good r e c a l l - - i s sufficient, by virtue of its ability to promote both item-specific and relational processing as well as its ability to promote compatible encoding and retrieval conditions.

Document info
Document views62
Page views62
Page last viewedThu Dec 08 08:25:52 UTC 2016
Pages25
Paragraphs1482
Words20141

Comments