The PANAS X: Manual For Positive And Negative Affect Schedule Expanded Form X

User Manual: Pdf

Open the PDF directly: View PDF .
Page Count: 28

8-1-1999
The PANAS-X: Manual for the Positive and Negative Affect Schedule - Expanded Form
- David Watson
- Lee Anna Clark
  - Comments

Department of Psychology Publications

8-1-1999

e PANAS-X: Manual for the Positive and

Negative Aect Schedule - Expanded Form

David Watson

University of Iowa

Lee Anna Clark

University of Iowa

Comments

Please contact the authors regarding questions of reusing the PANAS-X.

Hosted by Iowa Research Online. For more information please contact: lib-ir@uiowa.edu.

updated 8/99

THE PANAS-X

Manual for the

Positive and Negative Affect Schedule - Expanded Form

The University of Iowa

PANAS-X Manual

Table of Contents

List of Tables................................................................................................................................ ii

I. The Hierarchical Structure of Self-Rated Affect................................................................................. 1

II. The Higher Order Scales................................................................... .......................................... 2

A. Construction of the Original Positive and Negative Affect Scales ................................................ 2

B. Normative and Internal Consistency Data................................................................................ 3

Between-subjects data....................................................................................................... 3

Aggregated within-subject data.................................................................................................. 5

C. Construct Validity.............................................................................................................. 6

Factorial validity............................................................................................................. 6

Convergence with peer ratings ........................................................................................... 7

III. The Specific Affect Scales.......................................................................................................... 8

A. Construction of the Fear, Sadness, Guilt, Hostility, Shyness

Fatigue, and Surprise Scales........................................................................................... 8

B. Construction of the Joviality, Self-Assurance, Attentiveness, and Serenity Scales ....................... 11

C. Normative and Internal Consistency Data................................................................................ 12

Basic between-subjects data ............................................................................................ 12

Analyses of gender........................................................................................................ 13

Within-subject data....................................................................................................... 14

D. Construct Validity.............................................................................................................. 14

Comparison analyses with the POMS scales....................................................................... 14

Convergence with peer ratings ........................................................................................ 15

IV. General Issues Concerning the PANAS-X Scales ............................................................................. 17

A. Testing the Hierarchical Arrangement of the PANAS-X Scales...................................................... 17

B. Using the PANAS-X Scales as State Measures .......................................................................... 18

Relations with other measures of mood and symptomatology ............................................... 18

Studies of intraindividual mood fluctuation.................................................................................. 18

C. Using the PANAS-X scales as Measures of Trait Affect............................................................... 19

Test-retest stability data............................................................................................................ 19

Convergence with aggregated state measures...................................................................... 20

Relations with measures of personality and emotionality ..................................................... 20

Summary of trait validity evidence.............................................................................................. 22

V. References ............................................................................................................................... 22

PANAS-X Manual

List of Tables

Table 1 Sample PANAS-X Protocol Illustrating "Past Few Weeks" Time Instructions................................................1

Table 2 Item Composition of the PANAS-X Scales...............................................................................................2

Table 3 Means and Standard Deviations for the General Positive Affect (PA) and Negative Affect (NA) Scales as a

Function of Rated Time Frame and Subject Population ...........................................................................3

Table 4 Internal Consistency Reliabilities (Coefficient a) and Intercorrelations of the General PA and NA Scales ......... 4

Table 5 Descriptive Statistics for the General PA and NA Scales Derived from Mean Daily Mood Scores ...................5

Table 6 Correlations between the General PA and NA Scales and Regression-Based Scores on the First Two

Varimax Factors in Six Samples Assessed with Tellegen's Set of 60 Mood Descriptors ................................. 6

Table 7 Correlations between the General PA and NA Scales and Regression-Based Scores on the First Two

Varimax Factors in Ten Samples Assessed with the 60 PANAS-X Mood Descriptors ......................................7

Table 8 Correlations between the General Positive Affect and Negative Affect Scales and Regression-Based Scores

on the First Two Varimax Factors Emerging in Within-Subject Data .........................................................8

Table 9 Convergence between Self- and Peer-ratings on the Two Higher Order PANAS-X Scales .................................. 8

Table 10 Varimax-Rotated Factor Loadings of the 60 Tellegen Mood Terms in the Past Few Weeks Solution ..............9

Table 11 Varimax-Rotated Factor Loadings of the Positive Mood Terms Defining the Joviality, Self-Assurance,

Attentiveness, and Serenity Factors (Past Week Instructions, N = 607) ................................................... 11

Table 12 Descriptive Statistics and Internal Consistency Reliabilities (Coefficient a) for the 11 Lower Order Scales

................................................................................................................................................... 12

Table 13 Gender-Specific Normative Statistics for the Self-Assurance, Hostility, and Serenity Scales ........................ 14

Table 14 Descriptive Statistics for the 11 Lower Order Scales Derived from Mean Daily Mood Scores ...................... 15

Table 15 Correlations among the PANAS-X Scales and Corresponding Scales from the Profile of Mood States ............ 15

Table 16 Self- versus Peer-Ratings on Seven Lower Order PANAS-X Scales (Dormitory Study) ............................... 16

Table 17 Self- versus Peer-Ratings on the 11 Lower Order PANAS-X Scales (Dating Study) .................................... 16

Table 18 Varimax-Rotated Loadings of the Lower Order PANAS-X Scales with Different Time Instructions ................. 17

Table 19 Correlations between the General Positive Affect and Negative Affect Scales and Regression-Based Scores

on the First Two Varimax Factors Defined by the 11 Lower Order PANAS-X Scales ................................... 18

Table 20 Test-Retest Reliabilities of the PANAS-X Scales (2-Month Retest Interval) ............................................... 19

Table 21 Long-term Retest Stabilities of the Higher Order PANAS-X scales (M Retest Interval = 72.4 Mos.) .............. 20

Table 22 Correlations Between General Trait Ratings on the Higher Order Scales and Aggregated State Ratings .......... 20

Table 23 Varimax-Rotated Factor Loadings of Personality, Emotionality, and Higher Order PANAS-X Scales

(General Instructions) in a Sample of SMU Undergraduates ................................................................... 21

Table 24 Varimax-Rotated Factor Loadings of Personality, Emotionality and Lower-Order PANAS-X Scales (General

Instructions) in Sample of SMU Undergraduates ................................................................................. 22

PANAS-X Manual

I. The Hierarchical Structure of Self-Rated Affect

In recent research, two broad, general factors--typically

labeled Positive Affect (PA) and Negative Affect (NA)--have

emerged reliably as the dominant dimensions of emotional

experience. These factors have been identified in both intra-

and interindividual analyses, and they emerge consistently

across diverse descriptor sets, time frames, response

formats, languages, and cultures (Almagor & Ben-Porath,

1989; Mayer & Gaschke, 1988; Meyer & Shack, 1989;

Tellegen, 1985; Watson, 1988b; Watson, Clark, &

Tellegen, 1984; Watson & Tellegen, 1985; Zevon &

Tellegen, 1982). To measure these factors, Watson,

Clark, and Tellegen (1988) developed the Positive and

Negative Affect Schedule (PANAS), which consists of two

10-item scales for PA and NA, respectively.

These two general dimensions account for most of the

variance in self-rated affect--together they account for

roughly one-half to three-quarters of the common variance in

mood terms (see Watson, 1988b; Watson & Tellegen,

1985). Nevertheless, specific emotional states can also be

identified in the same data. In fact, on the basis of earlier

work by Tellegen , Watson and Tellegen (1985) proposed a

hierarchical taxonomic scheme in which the two broad,

higher order dimensions are each composed of several

correlated, yet ultimately distinguishable affective states

(see also Watson & Clark, 1989, 1992a). In this model,

the higher level reflects the valence of the mood descriptors

(i.e., whether they represent negative or positive states),

whereas the lower level reflects their specific content (i.e.,

the distinctive qualities of the individual affects).

To assess these specific emotional states, we have created a

60-item, expanded version of the PANAS (the PANAS-X).

In addition to the two original higher order scales, the

PANAS-X measures 11 specific affects: Fear, Sadness,

Guilt, Hostility, Shyness, Fatigue, Surprise, Joviality,

Self-Assurance, Attentiveness, and Serenity. The

PANAS-X thus provides for mood measurement at two

different levels.

The PANAS-X is simple and easy to administer. Most

subjects complete the entire 60-item schedule in 10 minutes

or less. However, investigators facing more severe time

constraints can select and assess only those scales that are

most relevant to their research. A sample PANAS-X

protocol is shown in Table 1. In addition, the terms

comprising each of the PANAS-X scales are shown in

Table 2.

The PANAS-X was created in three relatively distinct stages.

As mentioned, the two higher order scales were developed

first. Seven specific affect scales primarily involving

___________________________________________________________________________________

Table 1 Sample PANAS-X Protocol Illustrating "Past Few Weeks" Time Instructions

___________________________________________________________________________________

This scale consists of a number of words and phrases that describe different feelings and emotions. Read each item and then

mark the appropriate answer in the space next to that word. Indicate

to what extent you have felt this way during the past

few weeks

. Use the following scale to record your answers:

1 2 3 4 5

very slightly a little moderately quite a bit extremely

or not at all

______ cheerful ______ sad ______ active ______ angry at self

______ disgusted ______ calm ______ guilty ______ enthusiastic

______ attentive ______ afraid ______ joyful ______ downhearted

______ bashful ______ tired ______ nervous ______ sheepish

______ sluggish ______ amazed ______ lonely ______ distressed

______ daring ______ shaky ______ sleepy ______ blameworthy

______ surprised ______ happy ______ excited ______ determined

______ strong ______ timid ______ hostile ______ frightened

______ scornful ______ alone ______ proud ______ astonished

______ relaxed ______ alert ______ jittery ______ interested

______ irritable ______ upset ______ lively ______ loathing

______ delighted ______ angry ______ ashamed ______ confident

______ inspired ______ bold ______ at ease ______ energetic

______ fearless ______ blue ______ scared ______ concentrating

______ disgusted ______ shy ______ drowsy ______ dissatisfied

with self with self

___________________________________________________________________________________

PANAS-X Manual

____________________________________________________________________________________________________

Table 2 Item Composition of the PANAS-X Scales

____________________________________________________________________________________________________

General Dimension Scales

Negative Affect (10) afraid, scared, nervous, jittery, irritable, hostile, guilty, ashamed, upset, distressed

Positive Affect (10) active, alert, attentive, determined, enthusiastic, excited, inspired, interested,

proud, strong

Basic Negative Emotion Scales

Fear (6) afraid, scared, frightened, nervous, jittery, shaky

Hostility (6) angry, hostile, irritable, scornful, disgusted, loathing

Guilt (6) guilty, ashamed, blameworthy, angry at self, disgusted with self, dissatisfied with self

Sadness (5) sad, blue, downhearted, alone, lonely

Basic Positive Emotion Scales

Joviality (8) happy, joyful, delighted, cheerful, excited, enthusiastic, lively, energetic

Self-Assurance (6) proud, strong, confident, bold, daring, fearless

Attentiveness (4) alert, attentive, concentrating, determined

Other Affective States

Shyness (4) shy, bashful, sheepish, timid

Fatigue (4) sleepy, tired, sluggish, drowsy

Serenity (3) calm, relaxed, at ease

Surprise (3) amazed, surprised, astonished

Note. The number of terms comprising each scale is shown in parentheses.

____________________________________________________________________________________________________

different negative affects were constructed next. Finally,

four specific positive affect scales were created. In this

manual, we recap briefly the development of the original

PANAS scales, and present their basic reliability and validity

data (the interested reader is referred to Watson et al., 1988,

for more detail). We then describe the development of the

specific affect (PANAS-X) scales, and present data supporting

their reliability and validity. Finally, we discuss several

important issues regarding the instrument as a whole.

II. The Higher Order Scales

A. Construction of the Original Positive and Negative

Affect Scales

The goal in developing these scales was to create reliable and

valid measures that were also brief and simple to administer.

The primary concern was to select descriptors that were

relatively pure markers of either Negative Affect or Positive

Affect; that is, terms that had a substantial loading on one

factor but a near-zero loading on the other. As a starting

point, we used the 60 terms included in the factor analyses

reported by Zevon and Tellegen (1982). Tellegen

constructed this set from an initial pool of 117 affective

words and phrases derived from the earlier studies of Izard

(1972), Nowlis (1965), Zuckerman and Lubin (1965), and

Ekman (1971). A principal components analysis of content

sortings of this larger pool of items identified 20 synonym

groups, and the final list of 60 terms was constructed by

choosing three marker terms from each content group (see

Zevon & Tellegen, 1982, Table 1). Thus, these terms

provide a comprehensive assessment of the affective lexicon.

From this list of 60 terms we selected those descriptors that

had an average loading of .40 or greater on the relevant factor

across both the within- and between-subjects analyses

reported in Zevon and Tellegen (1982). Twenty Positive

Affect and 30 Negative Affect markers met this initial

criterion. However, as noted previously, we were also

concerned that the terms be relatively pure markers of a

factor. We therefore specified that the terms not have a

secondary loading of I.25I or greater in either analysis.

This reduced the pool of candidate descriptors to 12 for

Positive Affect and 25 for Negative Affect.

Preliminary reliability and validity analyses indicated that 10

terms were sufficient for the higher order Positive Affect

scale. We therefore dropped two terms that had relatively

higher secondary loadings on the Negative Affect factor,

yielding the final set of 10 descriptors (shown in Table 2).

The 25 Negative Affect candidate terms included all 3 terms

from seven of Tellegen's content categories, plus 2 from

each of two others. Because we wanted to tap a broad range

of content, we constructed a preliminary 14-item scale that

included 2 terms from each of the seven complete triads.

We found, however, that the terms from the Contempt and

Revulsion content categories did not significantly enhance

the reliability and validity of the scale. Moreover, these

PANAS-X Manual

terms were less familiar to our subjects (primarily

undergraduates) and were occasionally left unanswered.

Therefore, the final 10-item version consisted of 2 terms

each from the other five triads (displayed in Table 2).

B. Normative and Internal Consistency Data

Between-subjects data. Most of our basic psychometric data

were gathered from undergraduates enrolled in various

psychology courses at Southern Methodist University

(SMU), a private southwestern university. However, we

(and others) have also collected data on various student, adult

and psychiatric patient samples, and we present these results

as well.

We have obtained PANAS-X ratings using eight different

temporal instructions. Subjects have rated how they felt:

(a) "right now (that is, at the present moment)" (Moment

instructions); (b) "today" (Today); (c) "during the past few

days" (Past Few Days); (d) "during the past week" (Past

Week); (e) "during the past few weeks (Past Few Weeks);

(f) "during the past month" (Past Month), (g) "during the

past year (Past Year); and (h) "in general, that is, on the

average" (General).

_________________________________________________________________________________________________

Table 3 Means and Standard Deviations for the General Positive Affect and Negative Affect Scales as a Function of

Rated Time Frame and Subject Population

_________________________________________________________________________________________________

Positive Affect Negative Affect

______________ ______________

Time Frame/Sample N M SD MSD

_________________________________________________________________________________________________

Moment

SMU undergraduates 2,213 29.0 8.0 15.8 5.9

Australian undergraduatesa 279 25.3 7.2 15.2 5.0

VA substance abusers 158 32.0 8.5 23.7 10.1

Psychiatric inpatients 56 32.6 7.8 21.1 9.0

Today

SMU undergraduates 1,664 30.1 8.2 17.6 7.0

Past Few Days

SMU undergraduates 1,577 33.7 7.3 18.5 6.9

Iowa undergraduates 502 31.5 7.4 21.7 7.4

Past Week

SMU undergraduates 1,521 32.4 7.3 20.4 7.0

Dallas-area adults 328 31.1 7.5 18.0 7.1

Australian adult mena 114 32.6 5.9 16.3 4.7

Australian adult womena 115 30.7 7.1 15.8 4.9

Past Few Weeks

SMU undergraduates 2,076 32.6 7.1 20.2 7.2

SMU employees 164 33.1 6.8 17.9 6.4

Past Month

SMU undergraduates 1,006 34.5 7.2 20.2 7.3

Past Year

SMU undergraduates 964 35.9 6.4 22.8 6.6

General

SMU undergraduates 3,622 35.7 6.2 19.5 6.0

SMU employees 202 35.1 7.4 18.9 6.4

Detroit-area adultsb 815 36.0 6.0 18.2 6.3

Australian adult mena 114 33.5 5.9 14.2 4.1

Australian adult womena 115 33.9 5.1 15.5 5.3

Psychiatric inpatients 117 32.4 8.1 25.5 10.0

Mixed clinical sample 107 30.2 6.6 26.3 9.0

Note. These data include those reported in Watson et al. (1988), Table 1.

aUnpublished data reported by Ross Wilkinson, The Australian National University, April, 1993.

bThese data are reported in Quinn (1989).

_________________________________________________________________________________________________

PANAS-X Manual

Table 3 presents basic descriptive statistics for the general

Positive Affect and Negative Affect scales for these different

time instructions in various samples. We have not found

any large or consistent gender differences on these scales (the

issue of gender differences will be discussed in more detail

subsequently), so our own data are collapsed across gender.

Nevertheless, we recommend testing for gender differences

in any new (especially nonstudent) sample.

Inspecting Table 3, one sees that subjects report more

Positive Affect than Negative Affect, regardless of the rated

time frame or the population studied. Moreover, mean

scores on both scales tend to increase as the rated time frame

lengthens. This pattern is not surprising: As the rated

time period increases, the probability that a subject will

have experienced a significant amount of a given affect also

increases. Finally, it is interesting to compare the mean

scores of the college student, adult, and psychiatric patient

samples. Table 3 indicates that the mean Positive Affect

scores of college students and adults are generally quite

similar. The findings for Negative Affect are more variable,

but the bulk of the data suggest that adults report slightly

lower levels of Negative Affect than college students.

Overall, these data suggest that normative data collected on

college student samples can be applied to community-

dwelling adult samples with some confidence, particularly

with regard to Positive Affect.

______________________________________________________________________________

Table 4 Internal Consistency Reliabilities (Coefficient Alpha) and Intercorrelations of the General

Positive Affect and Negative Affect Scales

______________________________________________________________________________

Coefficient a

______________________

Positive Negative Scale

Time Frame/Sample N Affect Affect Intercorrelation

______________________________________________________________________________

Moment

SMU undergraduates 2,213 .88 .85 -.06

VA substance abusers 158 .87 .91 -.21

Psychiatric inpatients 56 .83 .88 -.23

Today

SMU undergraduates 1,664 .89 .87 -.05

Past Few Days

SMU undergraduates 1,577 .88 .87 -.17

Iowa undergraduates 502 .88 .87 -.31

Past Week

SMU undergraduates 1,521 .88 .85 -.14

Dallas-area adults 328 .90 .90 -.38

Australian adultsa 229 .86 .79 -.07

Past Few Weeks

SMU undergraduates 2,076 .87 .87 -.13

SMU employees 164 .86 .87 -.09

Past Month

SMU undergraduates 1,006 .89 .89 -.15

Past Year

SMU undergraduates 964 .87 .84 -.23

General

SMU undergraduates 3,622 .87 .85 -.13

SMU employees 201 .90 .87 -.31

Detroit-area adultsb 815 .84 .88 -.28

Australian adultsa 229 .85 .87 .01

Psychiatric inpatients 117 .87 .93 -.33

Mixed clinical sample 107 .83 .89 -.32

Note. These data include those reported in Watson et al. (1988), Table 2.

aUnpublished data reported by Ross Wilkinson, The Australian National University, April, 1993.

bThese data are reported in Quinn (1989).

______________________________________________________________________________

PANAS-X Manual

The psychiatric patients, however, clearly represent a

distinct population. Most notably, consistent with

previous research (e.g., Watson & Clark, 1984; Watson,

Clark, & Carey, 1988), psychiatric patients report

significantly higher levels of Negative Affect than do

nonpatients. The data for Positive Affect are less clear, but

one can anticipate that many patients will obtain unusually

low scores on this scale.

Internal consistency reliabilities (Cronbach's coefficient

alpha) for the two higher order scales, and also their

intercorrelations, are presented in Table 4. The alpha

reliabilities for both scales are high, generally ranging from

.83 to .90 for Positive Affect, and from .85 to .90 for

Negative Affect. It is noteworthy that the reliabilities of

the scales essentially are unaffected by the time instructions

that are used or by the type of subject population (student,

adult, or patient) that is assessed.

Table 4 also shows that the correlation between the Positive

Affect and Negative Affect scales is generally low, typically

ranging from -.05 to -.35. These discriminant values

indicate quasi-independence, an attractive feature for many

research and data-analytic purposes, and they are somewhat

lower than those of many other brief measures of Negative

Affect and Positive Affect (see Watson, 1988b).

Interestingly, the scale intercorrelation is not systematically

influenced by the rated time frame. Thus, the higher order

scales maintain their quasi-independence regardless of

whether state (shorter term) or trait (longer term) affect is

assessed.

Aggregated within-subject data. The data we have presented

thus far have been derived from single, between-subjects

assessments of large subject samples. Some investigators,

however, may wish to use the higher order PANAS-X scales

in designs that necessitate repeated within-subject

assessments. Accordingly, we present basic descriptive

statistics from three groups of subjects who completed the

higher order PANAS-X scales (using Today instructions) on a

daily basis over a period of several weeks. All subjects

completed a minimum of 30 daily mood assessments; all of

the assessments were completed in the evening, so that the

ratings would provide a reasonable estimate of the subjects'

moods over the course of the day (for more details regarding

this type of design, see Kennedy-Moore, Greenberg,

Newman, & Stone, 1992; Watson, 1988a; Watson, Clark,

McIntyre, & Hamaker, 1992).

The three groups consisted of: (1) 425 SMU undergraduates,

who completed an average of 43.3 mood assessments; (2) 90

adult men (M age = 42.7 years), who completed an average

of 77.9 mood ratings; and (3) 27 adults (age range = 21 to

76 years) diagnosed with chronic fatigue syndrome (CFS).

The CFS patients were drawn from the Minnesota Regional

Chronic Fatigue Syndrome Research Program at the

Hennepin County Medical Center (see Marshall et al., 1994,

for more details); they completed an average of 53.8 mood

assessments.

Mean Positive Affect and Negative Affect scores initially

were calculated for each subject. Overall sample means and

standard deviations were then computed for each scale; these

values are shown in Table 5. Looking first at the college

student data, it is noteworthy that the means for both

Positive Affect and Negative Affect are comparable--but

slightly lower--than the corresponding single-assessment

between-subjects values (using Today instructions) that are

shown in Table 3. Interestingly, the adult men reported

somewhat lower levels of both affects than did the college

students. The most striking data, however, are those of the

CFS patients. Their mean Positive Affect score was

substantially lower than that of the adult men, and more

than a full standard deviation below that of the college

students. In marked contrast, their Negative Affect levels

were unremarkable, falling in between those of the college

students and adults. Thus, these findings indicate clearly

that CFS is associated with a marked reduction in positive

emotional experiences (for further discussion of these data,

see Marshall et al., 1994).

______________________________________________________________________________

Table 5 Descriptive Statistics for the General Positive Affect and Negative Affect Scales Derived from Mean Daily

Mood Scores

______________________________________________________________________________

Positive Affect Negative Affect

No. of Total No. of _______________ _______________

Sample Subjects Observations M SD M SD

______________________________________________________________________________

SMU undergraduates 425 18,420 28.3 5.9 16.4 4.1

Adult mena90 7,013 25.0 7.4 12.7 3.7

CFS patients 27 1,453 20.1 6.1 15.3 5.6

Note. CFS = Chronic Fatigue Syndrome. See text for more details.

aThese data are reported in Kennedy-Moore et al. (1992), Table I.

______________________________________________________________________________

PANAS-X Manual

____________________________________________________________________________________________________

Table 6 Correlations Between the General Positive Affect and Negative Affect Scales and Regression-Based Scores on the

First Two Varimax Factors in Six Samples Assessed with Tellegen's Set of 60 Mood Descriptors

____________________________________________________________________________________________________

Positive Affect Negative Affect

Scale Correlations Scale Correlations

____________________ ____________________

Rated Time Frame NFactor 1 Factor 2 Factor 1 Factor 2

____________________________________________________________________________________________________

Moment 660 -.02 .95 .91 -.15

Today 657 -.02 .95 .93 -.11

Past Few Days 1,002 -.15 .92 .93 -.10

Past Few Weeks 586 -.10 .92 .92 -.18

Past Year 649 -.17 .89 .93 -.09

General 663 -.08 .94 .93 -.12

Note. This table is adapted from Watson et al. (1988, Table 4). See text for details.

____________________________________________________________________________________________________

C. Construct Validity

Factorial validity. An important step in validating these

general scales was to demonstrate that they captured the

underlying higher order dimensions adequately. We have

examined this issue in three series of analyses. First, we

subjected ratings on Tellegen's 60 mood descriptors (see

Zevon & Tellegen, 1982) to a principal factor analysis

(squared multiple correlations as the initial communality

estimates) in six large data sets, each of which was based on

a different rated time frame. The Ns were 660 (Moment),

657 (Today), 1,002 (Past Few Days), 586 (Past Few

Weeks), 649 (Year), and 663 (General).

Two dominant factors emerged in each solution that jointly

accounted for roughly two-thirds of the common variance,

ranging from 62.8% in the Moment solution to 68.7% in

the General ratings. The first two factors in each solution

were then rotated using varimax. Each of the six solutions

generated two sets of factor scoring weights that were used

to compute regression-based estimates of the underlying

Negative Affect and Positive Affect factors. Within each

data set, we then correlated these estimated factor scores

with the PANAS-X Negative Affect and Positive Affect

scales.

The results, shown in Table 6, demonstrate the expected

convergent/discriminant pattern: Both scales are very highly

correlated with their corresponding regression-based factor

scores in each solution, with convergent correlations

ranging from .89 to .95. In contrast, the discriminant

correlations are quite low, ranging from only -.02 to -.18.

Furthermore, Watson et al. (1988) present other data

indicating that the PANAS-X scales offer a better

convergent/discriminant correlational pattern with the

underlying factors than do other commonly used scale pairs.

The second series of analyses involved 10 additional data

sets. These analyses were identical to those already

discussed, except that the subjects were assessed on the 60

PANAS-X terms, rather than Tellegen's set of mood

descriptors. Eight of the data sets were collected from SMU

undergraduates; they differed only on the rated time frame.

The Ns were 1,027 (Moment), 1,007 (Today), 289 (Past

Few Days), 1,278 (Week), 678 (Past Few Weeks), 1,006

(Past Month), 315 (Past Year), and 1,657 (General). The

ninth data set consisted of 502 University of Iowa

undergraduates who rated themselves using Past Few Days

instructions. The final sample was composed of 328 Dallas-

area adults who rated their mood using Past Week

instructions.

The results of these analyses are presented in Table 7 and it

can be seen that they replicate closely those of the first

series. Again, the convergent correlations were quite high,

ranging from .90 to .95 for Positive Affect, and from .92 to

.95 for Negative Affect. As before, the discriminant

correlations generally were quite low, ranging from -.02 to -

.28 for Positive Affect, and from .00 to -.16 for Negative

Affect. In other words, these data again demonstrate that the

general Negative Affect and Positive Affect scales of the

PANAS-X are excellent measures of the underlying higher

order factors.

The data presented in Tables 6 and 7 are based entirely on

between-subjects data. In contrast, the final series of

analyses examined whether the general PANAS-X scales are

also excellent measures of the two higher order factors that

emerge in within-subject data. To investigate this issue we

conducted factor analyses in two very large data sets. The

first consisted of 226 SMU undergraduates who rated their

current, momentary mood on the 60 PANAS-X terms

repeatedly over a 1-2 month period (M = 45.0 assessments

per subject). The subjects were instructed to complete one

mood assessment each day; the times for these ratings varied

from day to day according to a pre-arranged, randomized

PANAS-X Manual

____________________________________________________________________________________________________

Table 7 Correlations Between the General Positive Affect and Negative Affect Scales and Regression-Based Scores on the

First Two Varimax Factors in Ten Samples Assessed with the 60 PANAS-X Mood Descriptors

____________________________________________________________________________________________________

Positive Affect Negative Affect

Scale Correlations Scale Correlations

____________________ ____________________

Rated Time Frame NFactor 1 Factor 2 Factor 1 Factor 2

____________________________________________________________________________________________________

SMU undergraduates

Moment 1,027 -.02 .94 .94 -.05

Today 1,007 -.05 .95 .94 .00

Past Few Days 289 -.01 .93 .95 -.01

Past Week 1,278 -.09 .94 .94 -.03

Past Few Weeks 678 -.09 .92 .93 -.08

Past Month 1,006 -.12 .94 .95 -.05

Past Year 315 -.17 .90 .92 -.05

General 1,657 -.10 .93 .93 -.01

Iowa undergraduates

Past Few Days 502 -.19 .93 .94 -.16

Dallas-area adults

Past Week 328 -.28 .92 .93 -.15

____________________________________________________________________________________________________

schedule. The second sample consisted of 248 SMU

undergraduates who rated their mood on the full PANAS-X on

a daily basis (using Today instructions) over a 1-2 month

period (M = 45.7 assessments per subject); as in previous

analyses of daily affect, all ratings were made in the evening

so that they would provide a reasonable estimate of the

subjects' moods over the course of the day. All subjects in

both data sets completed a minimum of 35 mood

assessments.

The mood ratings in each sample were standardized on a

within-subject basis; that is, each subject's responses were

converted to standard scores (M = 0, SD = 1). This procedure

eliminates between-subjects variability, so that subsequent

analyses reflect only within-subject variation. The data in

each sample were then subjected to a separate principal factor

analysis (squared multiple correlations in the diagonal); note

that this is equivalent to factoring each subject's data

separately and then averaging the resulting solutions in each

sample. The first two factors in each solution were then

rotated using varimax; as before, regression-based estimates

of these factors were computed. Finally, within each data

set, we correlated these factor scores with the PANAS-X

Negative Affect and Positive Affect scales. The results

(shown in Table 8) again strongly demonstrate the factorial

validity of these scales. Replicating the between-subjects

results, the convergent correlations for the Positive Affect

scale were .93 and .90 in the Moment and Today data,

respectively; the corresponding values for the Negative Affect

scale were .89 and .89, respectively. Moreover, the

discriminant coefficients were low, ranging from -.16 to -.23.

Thus, we see clear evidence that the general PANAS-X scales

are excellent measures of the major dimensions underlying

intraindividual mood experience.

Convergence with peer ratings. The data we have presented

thus far consist entirely of subjects' self-reports. In order to

establish construct validity, however, it is important to

consider other types of evidence as well. Accordingly, we

have conducted two studies in which self-ratings on the two

higher order scales (as well as the specific affect scales, to

be discussed later) were correlated with corresponding

judgments made by well-acquainted peers. All affect ratings

in both studies were based on trait (i.e., General) time

instructions. The first study (discussed in more detail by

Watson & Clark, 1991) involved dormitory residents at

SMU. To participate in the study, subjects were required to

that all group members know each other reasonably well.

Each subject rated all five group members, thereby

generating one set of self-ratings and four sets of peer-

ratings. The results reported here were computed from the

data of 89 subjects, each of whom was rated by at least three

well-acquainted peers; the peer judgments for each subject

were averaged to yield a single overall peer rating score on

each scale for that subject.

The second study was an examination of currently dating

couples at various universities in the Dallas-Fort Worth

metropolitan area. To be included in the study, the couple

had to have been dating each other for at least one month (M

time of dating = 21.8 months). All subjects rated both

themselves and their partners on the full PANAS-X.

Before presenting these results, we should note that one can

expect only moderate self-peer convergence on these scales.

Several studies have found that more externally visible

behavioral dispositions show better self-peer convergence

than do more internal, subjective traits (Albright, Kenny, &

PANAS-X Manual

____________________________________________________________________________________________________

Table 8 Correlations Between the General Positive Affect and Negative Affect Scales and Regression-Based Scores on the

First Two Varimax Factors Emerging in Within-Subject Data

____________________________________________________________________________________________________

Positive Affect Negative Affect

Scale Correlations Scale Correlations

Time No. of Total No. of __________________ __________________

Instructions Subjects Observations Factor 1 Factor 2 Factor 1 Factor 2

____________________________________________________________________________________________________

Moment 226 10,169 .93 -.20 -.16 .89

Today 248 11,322 .90 -.23 -.18 .89

____________________________________________________________________________________________________

Malloy, 1988; Funder & Colvin, 1988; Funder & Dobroth,

1987; Kenrick & Stringfield, 1980; Norman & Goldberg,

1966; Watson, 1989; Watson & Clark, 1991). Because

affective experience is strongly subjective, one cannot expect

the convergent correlations to be very high. Furthermore,

earlier research has also found that self-peer convergence

increases as more peer judges are used (McCrae & Costa,

1987; Watson, 1989; Watson & Clark, 1991). Accordingly,

one would predict that convergence would be better in the

dormitory study (in which each target was rated by 3-4 well-

acquainted peers) than in the dating study (in which each

target was rated by only a single peer).

The results of both studies are presented in Table 9. These

data clearly support the convergent and discriminant validity

of the general Positive Affect and Negative Affect scales.

The convergent correlations all are statistically significant;

those for Positive Affect are especially good, with values of

.48 and .35 in the dormitory and dating studies, respectively.

The convergent correlations for Negative Affect (.36 and .21,

respectively) are lower, but still respectable given the

limitations discussed earlier. Finally, it is noteworthy that

the discriminant coefficients are invariably low and non-

significant.

In summary, the two general PANAS-X scales provide

reliable, valid, and largely independent measures of the

higher order Positive Affect and Negative Affect dimensions,

regardless of the subject population or time frame used.

For more information regarding the reliability and validity of

these scales, see Watson (1988b) and Watson et al. (1988).

III. The Specific Affect Scales

A. Construction of the Fear, Sadness, Guilt, Hostility,

Shyness, Fatigue, and Surprise Scales

In constructing the lower order scales, we continued to be

concerned with developing reliable and valid measures that

could be used with different time instructions. Therefore,

the initial scale construction was based on the six large data

sets (with Ns ranging from 586 to 1,002) that were used in

the first series of analyses on factorial validity (see Table 6

and the accompanying discussion). As noted previously,

each of these data sets included Tellegen's list of 60 mood

terms (see Zevon & Tellegen, 1982), differing only in the

time frames used to generate the mood ratings (Moment,

Today, Past Few Days, Past Few Weeks, Past Year,

General).

Each of these data sets was factored separately using

principal factor analysis (squared multiple correlations in the

diagonal) and varimax rotation. To determine the final

number of factors in each data set, a range of solutions--

starting at two factors--was examined, until a solution was

reached that contained an uninterpretable factor (i.e., one

with fewer than three marker terms). In these analyses, we

defined a marker as a variable that loaded I.40I or higher on

the factor and had its highest loading on the factor. These

criteria yielded the following numbers of factors: Moment

(7), Today (8), Past Few Days (8), Past Few Weeks (8),

Past Year (10), and General (6). In each case, this

represents the maximum number of interpretable factors that

could be identified in the data.

________________________________________________

Table 9 Convergence Between Self- and Peer-ratings on the

Two Higher Order PANAS-X Scales

________________________________________________

Self Rated

________________________

Peer Rated PA NA

________________________________________________

Dormitory Study (N = 89)

Positive Affect .48* -.05

Negative Affect -.13 .36*

________________________________________________

Dating Study (N = 137)

Positive Affect .35* -.09

Negative Affect -.03 .21*

*p < .05, two-tailed.

________________________________________________

PANAS-X Manual

____________________________________________________________________________________________________

Table 10 Varimax-Rotated Factor Loadings of the 60 Tellegen Mood Terms in the Past Few Weeks Solution (N=586)

____________________________________________________________________________________________________

Factor

__________________________________________________________

Descriptor 1 2 3 4 5 6 7 8

____________________________________________________________________________________________________

enthusiastic .74

happy .70

joyful .70

interested .69

determined .66

delighted .66

confident .65

excited .65

friendly .63

warmhearted .62

alert .61

proud .59

strong .59

active .58

sociable .57

inspired .56

content .56

at ease .55 -.40

attentive .49

healthy .46

calm .45 -.39

** hostile .69

** scornful .68

disdainful .63

** loathing .62

** angry .61

contemptuous .59

** disgusted .56 .35

** irritable .55

revulsion .53

tormented .39 .36

** scared .76

** afraid .71

** frightened .70

** shaky .57

** nervous .56

** jittery .52

distressed .35 .42 .33 .33

____________________________________________________________________________________________________

(

table continues

)

PANAS-X Manual

Table 10 (cont.)

____________________________________________________________________________________________________

Factor

_______________________________________________________________

Descriptor 1 2 3 4 5 6 7 8

____________________________________________________________________________________________________

** alone .75

** lonely .71

** sad .62

** blue .31 .61

** downhearted .55

rejected .54

upset .39 .32 .49

** dissatisfied with self .74

** angry at self .73

** disgusted with self .68

** guilty .54

** blameworthy .51

** ashamed .51

** sleepy .76

** tired .74

** sluggish .57

** astonished .62

** amazed .62

** surprised .60

** bashful .67

** shy .66

** sheepish .44

Note. Loadings below I.30I are omitted.

**Item was included as a marker in the preliminary version of the scale assessing this factor.

____________________________________________________________________________________________________

Overall, these multifactorial solutions were remarkably

clear and clean. Most terms defined one and only one

factor, and there were few significant secondary loadings.

Moreover, the factors that emerged were both easily

interpretable and broadly consistent with existing theoretical

models of emotion (e.g., Ekman, 1971; Izard, 1972). A

representative factor loading matrix (the Past Few Weeks

solution) is presented in Table 10. Eight factors emerged

clearly in at least four of the six solutions. Seven of these

dimensions--Fear, Sadness, Guilt, Hostility, Shyness,

Fatigue, and Surprise-- represented specific emotional states

and therefore were used as the basis for scale development.

The terms that most strongly and consistently defined these

factors across all of the solutions were selected as the

component scale items, and are indicated in Table 10. The

eighth replicable factor was the higher order Positive Affect

dimension; we describe our further efforts to identify

specific positive emotional factors in the next section.

Initial internal consistency reliabilities (Cronbach's

coefficient alpha) were computed for all seven scales in the

six scale-development data sets. Consistent with their

factor-analytic derivation, the scales were all reasonably

homogeneous. Moreover, the reliability of the scales did

not vary systematically with the rated time frame. Not

surprisingly, however, the shorter (3-item) scales were

somewhat less reliable than those with five or six items

(Fear, Hostility, Guilt, and Sadness). We therefore

decided that they could be improved by adding a fourth term.

Using data from the samples to be described in the next

section, we tested an additional marker for Shyness (timid)

and Fatigue (drowsy), and found that they significantly

improved the psychometric properties of their respective

scales. These terms were therefore added to the Shyness and

Fatigue scales, bringing them to their final 4-item form.

Although the psychometric properties of Surprise also

might have been improved through the addition of a fourth

term, no further revisions were made on this scale because

we were unable to find a commonly used term that did not

greatly affect its valence or correlate strongly with other

lower order scales. We will present psychometric and

normative information on these finalized scales after we have

described the development of the remaining specific affect

scales.

PANAS-X Manual

B. Construction of the Joviality, Self-Assurance,

Attentiveness, and Serenity Scales

The one unexpected aspect of these multifactorial solutions

was the failure of Positive Affect to split into more specific

positive emotional states. Separate positive affect factors

emerged only in the Past Year solution. In these data,

Positive Affect split into Joy/Sociability and

Interest/Energy. To test whether these specific positive

emotional factors should be measured separately, we

constructed scales using the marker variables from the two

factors in the Past Year solution. Across the six data sets,

these two scales correlated from .66 to .77 with one another

(median r = .73). Because of this strong level of

intercorrelation, we did not pursue these two scales further.

More generally, positive mood descriptors appear to be even

more highly interrelated than are negative terms, and we

therefore believe that positive emotional states may prove

ultimately to be less differentiable in self-report.

Nevertheless, we continued to seek an alternative method

for developing scales to assess specific positive affects. To

achieve a desirable level of differentiation, we needed a

broader sample of positive mood descriptors. We therefore

created an expanded pool of 36 positive mood terms. This

list included 20 of the previous set of 21 terms (healthy was

dropped because preliminary analyses showed that it

consistently split across several factors), plus additional

terms assessing venturesomeness, energy, cheerfulness,

sociability, mental alertness, and serenity.

This expanded set of positive mood descriptors was subjected

to principal components analyses (with varimax rotation) in

three new student samples. First, 607 subjects rated

themselves using Past Week instructions; 347 of these

subjects were then retested (using the same instructions) 2

months later. Finally, 327 individuals rated themselves

using General instructions. Each of these solutions

indicated the presence of a large general factor, which

accounted for 30.3% to 37.5% of the total variance.

Nevertheless, four interpretable factors could be identified in

each solution. Terms that did not load clearly or

consistently on one of these factors were eliminated

gradually from subsequent analyses, leading to a reduced set

of 21 clear marker terms. Analyses on this reduced set of

descriptors yielded a clean simple-structure rotation in each

solution. A representative solution (the initial Past Week

sample) is shown in Table 11.

Four specific positive affect scales--Joviality, Self-

Assurance, Attentiveness, and Serenity--were constructed

from the clearest and most consistent markers of each of

these factors. The descriptors comprising these PANAS-X

scales are presented in Table 2.

C. Normative and Internal Consistency Data

Basic between-subjects data. Table 12 reports basic

descriptive statistics and internal consistency reliabilities

(Cronbach's coefficient a) for the lower order PANAS-X scales

________________________________________________

Table 11 Varimax-Rotated Factor Loadings of the

Positive Mood Terms Defining the Joviality,

Self-Assurance, Attentiveness, and Serenity

Factors (Past Week Instructions, N = 607)

________________________________________________

Factor

_______________________________

Descriptor 1 2 3 4

________________________________________________

happy .79

cheerful .78

joyful .78

excited .77

enthusiastic .73

lively .70 .30

energetic .64 .31

delighted .63

bold .71

fearless .69

strong .68

proud .64

confident .60 .34

daring .42 .57

concentrating .79

attentive .77

determined .66

alert .61

calm .78

relaxed .76

at ease .37 .62

Note. Loadings below I.30I are omitted.

________________________________________________

in 11 data sets reflecting eight different time instructions.

Looking first at the descriptive statistics, it should be noted

that these data show a pattern similar to that observed with

the higher order scales (see Table 3); that is, mean scale

scores show a general tendency to increase as the rated time

frame lengthens.

Table 12 also clearly demonstrates that the alpha reliabilities

of the longer (i.e., 5-8 items) PANAS-X scales are high.

Joviality is both the longest and the most reliable of the

lower order scales, with a median internal consistency

estimate of .93 (range = .88 to .94) across the 11 samples.

Furthermore, Fear (median a = .87), Sadness (median a =

.87), Guilt (median a = .88), Hostility (median a = .85),

Fatigue (median a = .88), Self-Assurance (median a = .83)

and Shyness (median a = .83) also consistently show good

reliabilities. In contrast, three of the shorter (i.e., 3-4 item)

scales--Attentiveness, Serenity, and Surprise--yielded

slightly lower reliability estimates; across the 11 samples,

PANAS-X Manual

____________________________________________________________________________________________________

Table 12 Descriptive Statistics and Internal Consistency Reliabilities for the 11 Lower Order PANAS-X Scales

____________________________________________________________________________________________________

Scale M SD Coefficient a

______________________________________________________

Moment (1,027 SMU undergraduates)

Fear 9.9 4.5 .87

Sadness 9.4 4.4 .86

Guilt 8.7 4.0 .86

Hostility 9.3 3.9 .82

Shyness 6.7 2.9 .80

Fatigue 11.1 4.2 .88

Joviality 21.7 7.5 .93

Self-Assurance 16.5 5.1 .83

Attentiveness 12.5 3.1 .72

Serenity 9.8 2.8 .74

Surprise 4.8 2.4 .80

Today (1,007 SMU undergraduates)

Fear 11.0 4.9 .87

Sadness 9.8 4.7 .87

Guilt 10.1 5.2 .89

Hostility 10.7 4.9 .86

Shyness 7.5 3.4 .83

Fatigue 11.5 4.4 .89

Joviality 23.9 7.8 .94

Self-Assurance 16.0 5.2 .84

Attentiveness 13.1 3.3 .79

Serenity 9.4 2.7 .76

Surprise 5.9 2.7 .75

Past Few Days (289 SMU undergraduates)

Fear 11.2 5.2 .90

Sadness 9.9 4.8 .89

Guilt 10.4 5.2 .90

Hostility 11.7 5.5 .89

Shyness 7.0 3.0 .85

Fatigue 11.5 4.1 .89

Joviality 26.5 7.0 .93

Self-Assurance 17.3 5.1 .84

Attentiveness 13.5 3.1 .79

Serenity 9.1 2.6 .79

Surprise 6.7 2.9 .79

Past Few Days (502 Iowa undergraduates)

Fear 12.9 5.0 .86

Sadness 11.5 4.6 .86

Guilt 11.8 5.3 .87

Hostility 12.4 4.7 .83

Shyness 8.2 3.2 .80

Fatigue 12.6 3.8 .86

Joviality 25.7 6.9 .93

Self-Assurance 17.2 5.0 .83

Attentiveness 13.0 3.1 .75

Serenity 9.2 2.8 .79

Surprise 6.8 2.7 .78

________________________________________________

Scale M SD Coefficient a

__________________________________________________

Past Week (1,278 SMU undergraduates)

Fear 12.2 5.2 .88

Sadness 11.2 4.9 .89

Guilt 10.8 4.9 .87

Hostility 11.6 4.6 .85

Shyness 8.1 3.4 .83

Fatigue 11.7 3.8 .87

Joviality 26.5 6.9 .93

Self-Assurance 17.1 4.8 .83

Attentiveness 13.1 3.1 .79

Serenity 8.9 2.5 .74

Surprise 6.6 2.7 .80

Past Week (328 Dallas-area adults)

Fear 9.6 4.1 .88

Sadness 9.5 4.4 .88

Guilt 10.1 4.9 .91

Hostility 11.1 4.6 .88

Shyness 6.3 2.7 .83

Fatigue 9.0 3.8 .89

Joviality 25.0 6.8 .93

Self-Assurance 16.5 4.6 .81

Attentiveness 12.7 3.0 .78

Serenity 8.7 2.4 .75

Surprise 5.4 2.3 .75

Past Few Weeks (678 SMU undergraduates)

Fear 12.3 4.9 .86

Sadness 11.7 4.8 .87

Guilt 12.0 5.2 .86

Hostility 12.9 5.0 .85

Shyness 7.7 3.1 .81

Fatigue 12.7 3.9 .88

Joviality 26.8 6.6 .93

Self-Assurance 17.7 4.7 .81

Attentiveness 13.5 2.9 .75

Serenity 8.9 2.6 .79

Surprise 6.8 2.8 .80

Past Month (1,006 SMU undergraduates)

Fear 11.3 4.6 .86

Sadness 10.6 4.6 .88

Guilt 11.0 5.2 .89

Hostility 12.3 4.8 .86

Shyness 7.7 3.2 .82

Fatigue 11.3 3.7 .87

Joviality 27.7 6.4 .92

Self-Assurance 18.5 4.8 .83

Attentiveness 13.8 3.0 .78

Serenity 9.4 2.6 .79

Surprise 6.9 2.7 .77

________________________________________________

PANAS-X Manual

________________________________________________

Table 12 (cont.)

________________________________________________

Scale M SD Coefficient a

________________________________________________

Past Year (315 SMU undergraduates)

Fear 13.7 4.6 .84

Sadness 12.9 4.7 .86

Guilt 13.1 5.5 .88

Hostility 14.3 4.7 .83

Shyness 8.7 3.3 .84

Fatigue 12.2 3.6 .84

Joviality 28.0 5.8 .91

Self-Assurance 19.0 4.4 .80

Attentiveness 14.0 2.8 .78

Serenity 9.2 2.5 .74

Surprise 7.5 2.3 .74

General (1,657 SMU undergraduates)

Fear 11.3 3.8 .83

Sadness 10.1 3.7 .83

Guilt 10.8 4.3 .85

Hostility 11.5 4.0 .83

Shyness 8.3 3.3 .83

Fatigue 10.3 3.4 .86

Joviality 28.1 6.0 .93

Self-Assurance 19.1 4.3 .81

Attentiveness 14.2 2.6 .76

Serenity 9.8 2.3 .73

Surprise 6.7 2.3 .76

General (107 mixed inpatients/outpatients)

Fear 15.1 9.0 .92

Sadness 14.8 5.3 .88

Guilt 17.7 6.6 .90

Hostility 14.4 4.8 .79

Shyness 9.9 4.3 .86

Fatigue 10.5 4.2 .89

Joviality 21.3 6.1 .88

Self-Assurance 15.8 4.9 .80

Attentiveness 13.1 3.0 .70

Serenity 7.9 2.8 .83

Surprise 6.4 2.5 .72

________________________________________________

these scales had median internal consistency values of .78,

.76, and .77, respectively. Nevertheless, these median

values reflect mean inter-item correlations of .45 or

greater, indicating that the scales are appropriately

homogeneous; thus, their reliability estimates simply

reflect the fact that they have relatively few items. As

mentioned earlier with regard to Surprise, the data suggest

that these scales could be improved through the inclusion

of additional marker terms. Unfortunately, the English

language contains few suitable terms in these content

domains and we have not been able to identify additional

markers for these scales.

Analyses of gender. An issue of general interest concerns

the possibility of systematic gender differences in affective

experience. We examined this issue in the 10 large data sets

(with Ns ranging from 289 to 1,657) that were used in the

second series of analyses on factorial validity (see Table 7

and the accompanying discussion). As was discussed

earlier, subjects in these samples were assessed on the full

set of 60 PANAS-X terms. Accordingly, we were able to

conduct t-tests comparing men and women on all 13 PANAS-

X scales in each sample. Note that all of these data sets

contained at least 100 subjects of each gender (the gender

breakdowns of these samples are reported in Table 13), and

most of them contained many more than that. Given the

large size of these samples, these t-tests will identify

significant effects even when the absolute difference between

groups is quite small.

Generally speaking, these analyses revealed few consistent

gender-related differences in affective experience. Five

scales--general Negative Affect, general Positive Affect,

Attentiveness, Surprise and Sadness--showed virtually no

significant gender-related effects. General Positive Affect

and Attentiveness both yielded significant but inconsistent

group differences in two samples: In each case men scored

higher in one sample and women scored higher in the other.

Sadness (women scored higher in the Past Week/Adult

sample) and Surprise (men scored higher in the Past Month

sample) each produced one significant group difference.

Finally, the general Negative Affect scale showed

significant effects whatsoever.

Five additional scales displayed modest gender-related

differences. Specifically, women scored significantly higher

on Joviality in four samples (Today, Past Week/SMU, Past

Month, General), on Fatigue in three samples (Moment,

Past Month, Past Week/Adult), and on Fear in two samples

(Past Few Days/Iowa, Past Year). Conversely, in four

samples men scored significantly higher on Guilt (Today,

Past Few Days/SMU, Past Week/SMU, General) and

Shyness (Past Few Days/SMU, Past Few Weeks, Past

Month, General).

The three remaining scales showed more consistent gender

effects. Specifically, men reported significantly higher

levels of Self-Assurance in all 10 samples; in addition, they

scored significantly higher on Serenity in nine samples (the

sole exception being the Today data set) and on Hostility in

seven samples (Moment, Today, Past Few Days/SMU, Past

Week/SMU, Past Few Weeks, Past Month, General).

Accordingly, Table 13 reports gender-specific normative data

on these three scales in all 10 data sets. Note that even

though these scales yielded consistent gender differences, the

absolute size of the difference generally is quite small. That

is, across the 10 data sets the mean scores for men are

approximately two points higher on Self-Assurance, one

point higher on Hostility, and a half-point higher on

Serenity. Thus, it appears that the affective experiences of

women and men generally are quite similar, so that the

overall normative statistics presented in Tables 3 and 12 can

be used with some confidence.

PANAS-X Manual

________________________________________________

Table 13 Gender-Specific Normative Statistics for the

Self-Assurance, Hostility, and Serenity Scales

________________________________________________

Men Women

____________ ____________

Time Frame/Sample M SD M SD

________________________________________________

Self-Assurance

Momentab 17.8 5.0 15.6 5.0

Todayab 16.9 5.3 15.3 5.0

Past Few Daysab 18.8 4.7 16.4 5.2

Past Few Daysac 17.7 4.6 16.6 5.3

Past Weekab 18.0 4.8 16.5 4.7

Past Weekad 17.6 4.4 15.5 4.7

Past Few Weeksab 19.1 4.4 16.8 4.6

Past Monthab 19.6 4.8 17.7 4.6

Past Yearab 20.3 4.2 18.0 4.2

Generalab 19.8 4.1 18.6 4.4

Hostility

Momentab 10.0 4.1 8.8 3.6

Todayab 11.4 5.1 10.2 4.7

Past Few Daysab 12.7 5.2 11.0 5.4

Past Few Daysc12.6 4.5 12.2 4.9

Past Weekab 12.4 4.8 11.0 4.4

Past Weekd10.9 4.4 11.3 4.8

Past Few Weeksab 13.7 4.7 12.4 5.2

Past Monthab 12.7 4.8 11.9 4.6

Past Yearab 15.0 5.3 14.0 4.3

Generalab 12.4 4.3 10.8 3.6

Serenity

Momentab 10.0 2.7 9.6 2.8

Todayb9.4 2.7 9.3 2.7

Past Few Daysab 9.6 2.5 8.9 2.6

Past Few Daysac 9.5 2.7 8.7 2.9

Past Weekab 9.1 2.5 8.7 2.5

Past Weekad 9.1 2.4 8.5 2.5

Past Few Weeksab 9.3 2.6 8.7 2.6

Past Monthab 9.8 2.6 9.2 2.6

Past Yearab 9.6 2.7 9.0 2.4

Generalab 10.0 2.3 9.6 2.2

Note. Ns by gender: Moment (437 M, 585 W), Today (420

M, 583 W), Past Few Days/SMU (102 M, 182 W), Past

Few Days/Iowa (281 M, 221 W), Past Week/SMU (502 M,

769 W), Past Week/Adult (142 M, 186 W), Past Few

Weeks (259 M, 409 W), Past Month (391 M, 602 W), Past

Year (128 M, 183 W), General (660 M, 989 W).

aMeans across gender differed (p < .05, two-tailed). bSMU

undergraduates. cIowa undergraduates. dDallas-area adults.

________________________________________________

Within-subject data. As with the higher order scales, some

investigators may wish to use the specific affect scales in

designs that necessitate repeated within-subject assessments.

Accordingly, Table 14 presents basic descriptive statistics

from a sample of 262 SMU undergraduates who completed

the full PANAS-X (using Today instructions) on a daily basis

over a period of several weeks. Paralleling the data reported

in Table 5, all subjects completed a minimum of 30 daily

mood ratings (M = 45.0 observations per subject); all of the

assessments were completed in the evening, so that the

ratings would provide a reasonable estimate of the subjects'

moods over the course of the day. As with the Table 5 data,

mean scale scores initially were calculated for each subject;

overall sample means and standard deviations were then

computed for each scale. It is noteworthy that--paralleling

the pattern observed with the higher order scales--the means

for all 11 scales are lower than the corresponding single-

assessment between-subjects values (using Today

instructions) that are shown in Table 3. These results

strongly suggest that subjects tend to give slightly lower

mood ratings with repeated measurement.

D. Construct Validity

Comparison analyses with the POMS scales. To

demonstrate the convergent and discriminant validity of a

subset of the PANAS-X scales, we created an affect

questionnaire that included the descriptors for Fear,

Hostility, Sadness, Fatigue, general Positive Affect, and

the six scales from the Profile of Mood States (POMS;

McNair, Lorr, & Droppleman, 1971). This questionnaire

was administered to 563 SMU undergraduates using Past

Few Weeks instructions.

These five PANAS-X scales measure the same basic affects as

corresponding scales from the POMS. Specifically, Fear

can be identified with POMS Tension-Anxiety, Hostility

with POMS Anger-Hostility, Sadness with POMS

Depression-Dejection, Fatigue with POMS Fatigue, and

general Positive Affect with POMS Vigor (which, despite

its name, includes a broad sample of positive mood terms).

Intercorrelations among these conceptually overlapping

scales are presented in Table 15. The table shows that each

of the PANAS-X scales is strongly related to its POMS

counterpart, with convergent correlations ranging from .85

to .91. These very high coefficients partly reflect item

overlap, as the corresponding scales have one to three items

in common. Nevertheless, the important point is that the

PANAS-X scales assess affective states that are broadly

similar to those measured in existing multi-affect

inventories such as the POMS.

Table 15 also demonstrates that the PANAS-X scales offer an

important advantage over their POMS counterparts: They

tend to be less highly correlated with one another, and thus

show better discriminant validity. The mean correlation

among the PANAS-X Fear, Hostility, Sadness and Fatigue

scales was .45, which was significantly lower than the

mean correlation (.60) among the corresponding POMS

PANAS-X Manual

________________________________________________

Table 14 Descriptive Statistics for the 11 Lower Order

Scales Derived from Mean Daily Mood Scores

________________________________________________

Scale MSD

________________________________________________

Fear 9.7 2.7

Sadness 8.2 2.4

Guilt 9.0 2.8

Hostility 9.4 2.3

Shyness 5.5 1.5

Fatigue 8.6 2.2

Joviality 22.7 4.8

Self-Assurance 14.7 3.5

Attentiveness 11.3 2.3

Serenity 8.3 1.8

Surprise 5.5 1.8

Note. N = 262. Total number of observations = 11,783.

See text for details.

________________________________________________

scales (p < .01, two-tailed; Fisher's r to z transformation

was used in the computation of these and subsequent mean

correlations). Follow-up comparisons indicated that five of

the six individual correlations were also significantly lower

(p < .01, two-tailed) in the PANAS-X scales; specifically,

only the Fear-Hostility correlation did not differ significantly

between the two instruments. Thus,

the PANAS-X scales generally provide a less redundant, more

differentiated assessment of affect.

Convergence with peer ratings. To document the construct

validity of the lower order PANAS-X scales further, we again

will consider evidence from two studies in which self-ratings

on these scales were correlated with corresponding judgments

made by well-acquainted peers. First, in the dormitory study

that was discussed previously (see Table 9 and the

accompanying discussion), self- and aggregated peer-ratings

(averaged across the responses from three or four well-

acquainted peers) were available on seven of the lower order

scales: Fear, Hostility, Guilt, Sadness, Shyness, Fatigue,

and Surprise (note, however, that preliminary 3-item

versions of Shyness and Fatigue were used). The

heteromethod correlations from this sample are presented in

Table 16. These data produced strong self-peer convergence,

and generally support the convergent and discriminant

validity of the PANAS-X scales. The single exception is

Surprise, which produced a nonsignificant level of self-peer

agreement (r = .14) in these ratings. The six remaining

scales, however, yielded convergent correlations ranging

from .27 to .52, with a mean value of .38. In contrast,

the discriminant coefficients were generally low and

nonsignificant. Using the criterion that the convergent

correlation should be higher than any of the other values in

its row or column of the heteromethod block (Campbell &

Fiske, 1959), five of the scales (Fear, Hostility, Sadness,

Shyness, and Fatigue) showed acceptable discriminant

validity. (For a more detailed discussion of these data, see

Watson & Clark, 1991).

____________________________________________________________________________________________________

Table 15 Correlations among the PANAS-X Scales and Corresponding Scales from the Profile of Mood States (POMS;

McNair, Lorr & Droppleman, 1971)

____________________________________________________________________________________________________

PANAS-X Scales POMS Scales

_________________________________ ____________________________

Instrument/Scale 12345 6789

____________________________________________________________________________________________________

PANAS

cales

1. Fear ---

2. Hostility .58 ---

3. Sadness .61 .49 ---

4. Fatiguea.40 .31 .27 ---

5. Positive Affect .02 -.06 -.25 -.07 ---

POMS

Scales

6. Tension-Anxiety .85 .62 .57 .48 .02 --

7. Anger-Hostility .59 .91 .51 .35 .00 .63 --

8. Depression-Dejection .74 .66 .85 .34 -.24 .69 .66 --

9. Fatigue .53 .46 .40 .89 -.07 .61 .48 .47 --

10. Vigor -.03 -.09 -.28 -.08 .86 -.05 -.03 -.25 -.07

Note. N= 563. Mood ratings based on Past Few Weeks instructions. Convergent correlations are shown in boldface. All

correlations greater than I.10I are significant at p < .01, two-tailed.

aData are based on a preliminary, 3-item version of this scale.

____________________________________________________________________________________________________

PANAS-X Manual

____________________________________________________________________________________________________

Table 16 Self- versus Peer-Ratings on Seven Lower Order PANAS-X Scales (Dormitory Study)

____________________________________________________________________________________________________

Self Rated

____________________________________________________________________________

Peer Rated Fear Hostility Guilt Sadness Shyness Fatigue Surprise

____________________________________________________________________________________________________

Fear

.40

*-.07 .16 .25* .25* .00 .10

Hostility .08

.31

*.10 .22* -.08 -.04 -.12

Guilt

.34

*.04

.34

.39

*.20 .09 .16

Sadness

.31

*-.03

.30

.52

*.27* -.01 .13

Shynessa.23* -.18 .09 .03

.42

*.01 -.02

Fatiguea-.02 .12 -.10 -.02 .08 .27* -.04

Surprise .03 -.19 -.02 .02 .02 -.05 .14

Note. N = 89. Convergent correlations are shown in boldface. Correlations of .30 or greater are underlined.

aData are based on preliminary, 3-item versions of these scales. *p < .05, two-tailed.

____________________________________________________________________________________________________

Second, self- and peer-ratings were available for all 11 lower

order scales from the dating study described earlier. The

heteromethod correlations from this sample are shown in

Table 17. Similar to the results presented in Table 16, these

data yielded strong self-peer convergence and broadly support

the convergent and discriminant validity of the lower order

PANAS-X scales. Eight of the 11 scales (Joviality, Self-

Assurance, Attentiveness, Serenity, Sadness, Hostility,

Shyness and Fatigue) had convergent correlations of .30 or

greater; in fact, the mean coefficient across these scales was

.36. Moreover, using the same criterion described in

connection with Table 16, all eight of these scales displayed

acceptable discriminant validity. Thus, the data from this

study clearly support the construct validity of these eight

scales. A ninth scale (Fear) also had a significant

convergent correlation (.21), but failed to demonstrate

acceptable discriminant validity. Finally, the two remaining

lower order scales (Surprise and Guilt) did not yield

significant evidence of either convergent or discriminant

validity. Taken together with the results of the dormitory

study, these data seriously challenge the validity of trait

ratings on Surprise.

____________________________________________________________________________________________________

Table 17 Self- versus Peer-Ratings on the 11 Lower Order PANAS-X Scales (Dating Study)

____________________________________________________________________________________________________

Self Rated

_________________________________________________________________________________

Peer Rated Jov Assur Atten Seren Fear Sad Guilt Host Shy Fat Surp

____________________________________________________________________________________________________

Joviality

.36

*.28* .15 .13 -.02 -.26* .00 -.16 -.29* -.03 .20*

Self-Assurance .17*

.44

*.25* .09 -.06 -.17* .02 -.11 -.27* -.03 .09

Attentiveness .17* .24*

.34

*.02 -.02 -.13 -.05 -.10 -.25* -.10 .04

Serenity .09 .17* .16

.37

*-.04 -.14 -.08 -.15 .10 .04 .07

Fear .00 -.17* -.06 -.25* .21* .15 .10 .20* .15 .14 .06

Sadness -.02 -.19* .07 -.23* .13

.30

*.13 .20* .19* -.03 -.03

Guilt -.07 -.09 .00 -.20* .04 .14 .15 .19* .10 .07 -.07

Hostility -.11 -.02 .02 -.22* -.05 .11 .04

.32

*.00 .01 -.09

Shyness -.06 -.20* -.09 -.04 .13 .03 .05 .06

.37

*.09 .08

Fatigue .01 -.01 .08 -.06 .05 .02 -.01 .10 .27*

.37

*-.02

Surprise .16 .11 .16 -.04 .01 -.14 -.08 -.10 -.02 .02 .14

Note. N = 137. Convergent rs in boldface; rs of

.30 underlined. Jov = Joviality; Assur = Self-Assurance; Atten =

Attentiveness; Seren = Serenity; Sad = Sadness; Host = Hostility; Shy = Shyness; Fat = Fatigue; Surp = Surprise.

*p < .05, two-tailed.

____________________________________________________________________________________________________

PANAS-X Manual

IV. General Issues Concerning the PANAS-X Scales

A. Testing the Hierarchical Arrangement of the PANAS-X

Scales

As was discussed previously, the PANAS-X was designed

explicitly to reflect the hierarchical structure of self-rated

affect. That is, self-rated affect is characterized by two

broad higher order dimensions (Negative Affect and Positive

Affect), each of which is composed of several correlated,

yet ultimately distinguishable specific affect states. The

PANAS-X includes scales assessing both of these structural

levels--that is, the two higher order dimensions and 11

specific lower order states.

How accurately does the PANAS-X capture this hierarchical

structure? To answer this question, we subjected the 11

lower order PANAS-X scales to second-order principal

components analyses in two large samples of SMU

undergraduates, who rated themselves using Moment (N =

1,027) and General (N = 1,657) instructions. Two large

second-order factors emerged in each solution: These two

dimensions jointly accounted for 52.8% and 54.9% of the

total variance in the Moment and General data, respectively.

Two components were therefore extracted in each solution

and were rotated using varimax; the rotated loadings are

presented in Table 18. In each solution, the first factor

clearly can be identified as general Negative Affect. Fear,

Sadness, Guilt, Hostility and Shyness are all very strong

markers of the high end of this second-order factor, with

loadings ranging from .59 to .83 across the two solutions.

Fatigue and Surprise also have moderate positive loadings

(ranging from .33 to .53) on this dimension. Finally,

Serenity serves to define the low end of the factor, with

loadings of -.50 and -.26 in the Moment and General

solutions, respectively.

The second higher order dimension clearly corresponds to the

general Positive Affect dimension that consistently emerges

in self-report studies. It is most strongly and clearly defined

on its high end by Joviality, Self-Assurance, and

Attentiveness, which have loadings ranging from .70 to .86

across the two solutions. Surprise (loadings of .51 and .62

in the Moment and General solutions, respectively) and

Serenity (.26 and .46, respectively) also are reasonably good

markers of high Positive Affect. Finally, Fatigue is the

best marker of low Positive Affect, with loadings of -.39

and -.25 in the Moment and General solutions, respectively.

How closely do these higher order dimensions correspond to

the general PANAS-X scales? To examine this issue, we

computed regression-based factor scores for the second-order

dimensions in each solution. Within each data set, we

then correlated these computed factor scores with the general

Negative Affect and Positive Affect scales. These results

are displayed in Table 19, and they again confirm the

factorial validity of the higher order scales. Specifically,

scores on Factor 1 correlated .92 and .91 with general

Negative Affect in the Moment and General solutions,

respectively; conversely, scores on Factor 2 correlated .94

and .91 with general Positive Affect in the Moment and

General solutions, respectively. Note also that the

discriminant correlations were invariably low, ranging only

from -.01 to -.11.

________________________________________________

Table 18 Varimax-Rotated Loadings of the 11 Lower

Order PANAS-X Scales with Two Different

Time Instructions

__________________________________________________

Factor 1 Factor 2

_________________ ________________

Scale Moment General Moment General

__________________________________________________

Fear .79 .83 .05 .07

Sadness .77 .79 -.27 -.21

Hostility .78 .77 -.07 -.01

Guilt .69 .78 -.09 -.15

Shyness .59 .62 .04 -.09

Fatigue .33 .53 -.39 -.25

Joviality -.14 -.13 .86 .85

Self-Assurance -.04 -.11 .80 .78

Attentiveness -.06 -.13 .76 .70

Surprise .49 .42 .51 .62

Serenity -.50 -.26 .26 .46

Note. Ns = 1,027 (Moment) and 1,657 (General).

Loadings of I.30I or greater are shown in boldface.

__________________________________________________

These data clearly demonstrate the hierarchical arrangement

of the PANAS-X scales. Furthermore, on the basis of these

and other analyses, we have grouped the 11 lower order

scales into three broad subcategories (see Table 2). First,

Fear, Sadness, Guilt, and Hostility scales are classified as

Basic Negative Emotion Scales. As the data in Table 18

illustrate, these scales are consistently and substantially

intercorrelated, and therefore, are strong and clear markers

of the higher order Negative Affect dimension (see also

Watson & Clark, 1989, 1992a). Second, Joviality, Self-

Assurance and Attentiveness are classified as Basic Positive

Emotion Scales. As can be seen in Table 18, these scales

are highly correlated with one another, and so are strong and

consistent markers of the second-order Positive Affect

dimension.

Finally, Shyness, Fatigue, Surprise and Serenity are

grouped as Other Affective States because they do not

strongly or consistently define either of the second-order

factors. Shyness tends to load moderately to strongly on

Negative Affect (see Table 18), but its loading generally is

somewhat lower than those of Fear, Sadness, Guilt, and

Hostility. That is, Shyness appears to be less strongly

saturated with general Negative Affect variance than are these

other scales (see also Watson & Clark, 1989). In contrast,

Surprise typically has moderate positive loadings on both

higher order factors. Fatigue and Serenity also tend to load

PANAS-X Manual

significantly on both general factors: Fatigue is a marker of

high Negative, low Positive Affect, whereas Serenity is a

marker of low Negative, high Positive Affect. However,

these loadings vary considerably across different samples and

time frames.

________________________________________________

Table 19 Correlations between the General Positive

Affect and Negative Affect Scales and

Regression-Based Scores on the First Two

Varimax Factors Defined by the 11 Lower

Order PANAS-X Scales

________________________________________________

Correlations with

____________________________

Scale Factor 1 Factor 2

________________________________________________

Moment Instructions (N = 1,027)

Positive Affect -.03 .94

Negative Affect .92 -.05

General Instructions (N=1,657)

Positive Affect -.11 .91

Negative Affect .91 -.01

________________________________________________

B. Using the PANAS-X Scales as State Measures

We have presented various types of evidence that establish

the convergent and discriminant validity of the PANAS-X

scales. In these final sections we examine more specific

validation issues, namely, using the PANAS-X scales as

measures of state and trait affect, respectively.

Relations with other measures of mood and

symptomatology. Many researchers will be interested in

using the PANAS-X scales as measures of state affect--that is,

to assess relatively short-term fluctuations in mood. What

evidence suggests that the PANAS-X can be used validly in

this way? First, the PANAS-X scales are strongly correlated

with other existing measures of shorter term affect. We

have already seen, for example, that five of the PANAS-X

scales are very highly correlated (convergent correlations

ranged from .85 to .91) with corresponding measures from

the POMS, which typically uses short-term instructions (in

this case, Past Few Weeks). Moreover, the PANAS-X

scales showed better discriminant validity--that is, they were

less highly intercorrelated than were their POMS counterparts

(see Table 15).

Watson and Clark (1992a, Study 1) report additional

convergent and discriminant validity data for the Fear,

Sadness, and Hostility scales in a sample of 195 SMU

undergraduates (see their Tables 1 and 2). For instance,

scores on the PANAS-X Sadness scale (assessed using Past

Few Weeks time instructions) correlated .59 with the Beck

Depression Inventory (BDI; Beck, Ward, Mendelson,

Mock, & Erbaugh, 1961), .69 with the Depression scale

from the Hopkins Symptom Checklist (HSCL; Derogatis,

Lipman, Rickels, Uhlenhuth, & Covi, 1974). and .75

with the Center for Epidemiological Studies Depression

Scale (CES-D; Radloff, 1977). Similarly, the PANAS-X Fear

scale correlated .74 with the HSCL Anxiety scale and .56

with the STAI State Anxiety Scale (Spielberger, Gorsuch,

& Lushene, 1970). Finally, the PANAS-X Hostility scale

correlated .55 with the Hostility scale from the SCL-90

(Derogatis, 1977; Derogatis & Cleary, 1977) and .45 with

the state form of the State-Trait Anger Scale (STAS;

Spielberger, Jacobs, Russell, & Crane, 1983). Furthermore,

it is important to note that all three PANAS-X scales showed

evidence of significant discriminant validity (see Watson &

Clark, 1992a, Table 2).

Similarly, Watson et al. (1988) present correlations

between the two higher order PANAS-X scales and other

commonly used measures of state affect and psychiatric

symptomatology (see Watson et al., 1988, Table 7). For

example, in a sample of 880 SMU undergraduates, the BDI

correlated .56 and -.35 with general Negative Affect and

Positive Affect, respectively (the PANAS-X ratings were

based on Past Few Days instructions). Similarly, in

another sample of 208 SMU students, the BDI correlated .58

and -.36 with Negative Affect and Positive Affect,

respectively (Past Few Weeks instructions). In this latter

sample (N = 203), the STAI State Anxiety Scale also

correlated .51 and -.35 with Negative Affect and Positive

Affect, respectively. Thus, both of the general PANAS-X

scales are significantly related to state measures of

depression and anxiety. Finally, in a sample of 398 SMU

undergraduates, general Negative Affect (using Past Few

Weeks instructions) correlated .74 with the total score on the

HSCL, a widely used measure of clinical symptomatology.

Studies of intraindividual mood fluctuation. A second line

of evidence indicates that, when used with short-term time

instructions (i.e., Moment or Today), the PANAS-X scales

are sensitive to changing internal or external circumstances.

We have used the general PANAS-X scales in four within-

subjects investigations that illustrate their utility in

studying qualitatively distinctive intraindividual mood

fluctuations. In the first study, 80 SMU students

completed a mood questionnaire consisting of the two higher

order scales each evening for 5-7 weeks, using Today

instructions (Watson, 1988a). At each assessment the

subjects also estimated their daily social activity (number of

hours spent with friends that day), rated the level of stress

they had experienced, and noted whether or not they had

exercised during the day. A total of 3,554 measurements

were collected (M = 44.4 per subject). As hypothesized (see

Clark & Watson, 1988, 1990b; Watson, 1988a; Watson

& Tellegen, 1985), within-subject variations in perceived

stress were strongly correlated with fluctuations in Negative

Affect, but not in Positive Affect. Also, as expected,

social activity and physical exercise were more highly related

to Positive Affect than to Negative Affect.

The second study was primarily concerned with diurnal

PANAS-X Manual

variation in mood (Clark, Watson, & Leeka, 1989). The

subjects were 196 SMU students who completed the two

general PANAS-X scales approximately seven times a day for

1 week. Subjects also rated their current stress, and noted

whether or not they had engaged in a number of activities

(including social interaction and exercise) within the past

hour. A total of 8,700 assessments were collected (M =

44.4 per subject). As hypothesized, Positive Affect

showed significant diurnal variation: It rose sharply from

early morning until noon, remained relatively constant until

9 p.m., and then fell rapidly. Negative Affect, however,

did not exhibit a systematic diurnal pattern in these data.

Furthermore, replicating the results of the earlier study,

perceived stress was again strongly correlated with within-

subject fluctuations in Negative Affect but not Positive

Affect. And, as before, social interaction and exercise

were more strongly related to Positive Affect than to

Negative Affect (Clark & Watson, 1990b).

In the third study (McIntyre, Watson, & Cunningham,

1990), 18 students completed a mood questionnaire

(including the two general scales) at the beginning of a 1-

week period in order to establish baseline levels on the two

higher order dimensions. Three other questionnaires were

completed within the week by each subject: Once after

social interaction, once after physical exercise, and once

prior to a stressful examination. Consistent with the

results of the other studies, within-subjects analyses of

variance revealed that Positive Affect was increased

significantly by social interaction and exercise, but was not

affected by test stress; conversely, Negative Affect was

increased significantly by the stressful examination, but

was not influenced by social activity or exercise.

In the fourth study (Watson et al., 1992, Study 2), 127

SMU undergraduates completed the two higher order PANAS-

X scales each evening for 5-7 weeks using Today

instructions (M = 42.7 assessments per subject); a subset of

these subjects (N = 96) also completed the three lower order

positive emotion scales (i.e., Joviality, Self-Assurance, and

Attentiveness). At each assessment, all subjects also rated

the amount of time they spent in each of 21

social/interpersonal activities during that day. Ratings on

these activities were summed to create three subscales

(Social Entertainment, Active Participation, Social

Responsibilities) as well as a measure of Overall Social

Activity. Replicating and extending previous research in

this area, within-subject correlational analyses indicated that

all four positive affect scales were significantly related to

three of the social activity measures (Social Entertainment,

Active Participation, Overall Social Activity); in contrast,

general Negative Affect was completely unrelated to social

activity.

Finally, data from a sample of 308 undergraduates who were

retested after 2 months on the PANAS-X using Past Week

instructions are shown in the first column of Table 20.

These correlations range from .23 (for Surprise) to .49 (for

Shyness), indicating a moderate level of stability. These

results suggest that subjects have a characteristic range of

affect within which short-term fluctuations occur. In

summary, the PANAS-X scales are strongly correlated with

commonly used measures of state affect and current

psychiatric symptomatology, and are sensitive to changing

endogenous and exogenous conditions. These data indicate

that the PANAS-X scales can be used validly to assess short-

term, state affect.

________________________________________________

Table 20 Test-Retest Reliabilities of the PANAS-X

Scales (2-Month Retest Interval)

________________________________________________

Past Week General Instructions

Instructions _____________________

Scale (N = 308) (N = 502) (N = 399)

________________________________________________

Higher Order Scales

Positive Affect .43 .70 .64

Negative Affect .41 .71 .59

Lower Order Scales

Fear .35 .62 .57

Hostility .39 .65 .58

Guilt .36 .68 .65

Sadness .35 .62 .60

Shyness .49 .70a.64

Fatigue .30 .57a.53

Surprise .23 .56 .52

Joviality .43 --b.64

Self-Assurance .47 --b.68

Attentiveness .42 --b.55

Serenity .32 --b.51

Note. All correlations significant at p < .01, two-tailed.

aData are based on preliminary, 3-item versions of these

scales. bNot assessed in this sample.

________________________________________________

C. Using the PANAS-X Scales as Measures of Trait Affect

Test-retest stability data. What evidence supports the

validity of the PANAS-X scales as trait measures, that is, as

measures of long-term, individual differences in affectivity?

In this regard, we have already presented data indicating that-

-with the exception of Surprise--trait (i.e., General) versions

of the PANAS-X scales are significantly correlated with

corresponding judgments made by well-acquainted peers.

Moreover, most of these scales also displayed acceptable

discriminant validity. Thus, the PANAS-X scales generally

appear to possess adequate external validity.

Test-retest reliability also is an important consideration in

establishing the construct validity of trait measures. Table

20 presents 2-month stability coefficients--based on General

instructions--for nine of the PANAS-X scales in one sample

of undergraduates (N = 502) and all of the scales in another

(N = 399). Scores on all of these scales were quite stable,

PANAS-X Manual

with retest coefficients ranging from .51 (for Serenity) to

.71 (for general Negative Affect in the larger sample). As

expected, the stability coefficients based on General

instructions are consistently higher than those based on Past

Week instructions, which further validates the use of

different time instructions with the PANAS-X scales.

We have also completed one longer-term analysis of

temporal stability on the two higher order scales. A sample

of 239 SMU undergraduates initially completed the general

Negative Affect and Positive Affect scales (using General

instructions) from January, 1985 to September, 1988; these

subjects were reassessed on these scales (again using General

instructions) during the spring and summer of 1993.

Overall, the retest interval ranged from 56 to 99 months (M

retest interval = 72.4 months). It also should be noted that

all subjects graduated from the university during this

intervening time period. Thus, all of the respondents

experienced a major life change over the course of the study.

Stability correlations from this study are presented in Table

21. The most important aspect of these data is that both of

the higher order scales demonstrated significant, moderate

stability over this extended time interval, with retest

correlations of .43 and .39 for Negative Affect and Positive

Affect, respectively. Note also that the discriminant

correlations were substantially lower, again demonstrating

the discriminant validity of these scales (for more details

regarding this study, see Watson & Walker, 1996).

________________________________________________

Table 21 Long-term Retest Stabilities of the Higher

Order PANAS-X scales (Mean Retest Interval =

72.4 Months)

________________________________________________

Time 1 Score

________________________________

Time 2 Score Positive Affect Negative Affect

________________________________________________

Positive Affect .42 -.24

Negative Affect -.18 .43

Note. N =237. All scales were assessed using General

Time Instructions. All correlations significant at p < .01,

two-tailed.

________________________________________________

Convergence with aggregated state measures. Another

important approach to validating the trait versions of the

PANAS-X scales is to demonstrate that they converge with

aggregated measures of shorter term affect. We have

conducted two studies of this type. First, 80 undergraduates

enrolled in an undergraduate personality course at the

University of Texas at Dallas completed the higher order

Positive Affect and Negative Affect scales (using General,

trait instructions) during the first week of class. Subjects

then were reassessed on these scales (using Past Week

instructions) once a week over the next 13 weeks. All

subjects completed a minimum of 7 weekly mood ratings

(M = 12.2 assessments per subject). Mean weekly Positive

Affect and Negative Affect scores then were created by

averaging each subject's responses over the entire weekly

rating period.

Correlations between these average weekly scores and the

trait versions of the scales are reported in the upper half of

Table 22. Again, both of the higher order scales exhibited

an impressive convergent/discriminant pattern. That is, the

convergent correlations for both scales are high (r s = .66

and .48 for Positive Affect and Negative Affect,

respectively), whereas the discriminant correlations are low

and non-significant.

In the second study, 410 SMU undergraduates completed a

trait form of the two higher order PANAS-X scales, and also

rated themselves on these scales (using Today instructions)

on a daily basis over a period of several weeks. As in

previous analyses of this type, all subjects completed a

minimum of 30 daily mood assessments (M = 43.3

assessments per subject); all of the ratings were completed

in the evening, so that they would provide a reasonable

estimate of the subjects' moods over the course of the day.

Mean daily Positive Affect and Negative Affect scores then

were created by averaging each subject's responses over the

entire daily rating period.

Correlations between the trait scales and average daily mood

scores are displayed in the lower half of Table 22, and it can

be seen that they replicate closely the results of the weekly

study. Again, the convergent correlations for both scales are

high (rs = .64 and .53 for Positive Affect and Negative

Affect, respectively), whereas the discriminant correlations

are low and non-significant.

_______________________________________________

Table 22 Correlations Between General Trait Ratings

on the Higher Order PANAS-X Scales and

Aggregated State Ratings

_______________________________________________

General Ratings

Aggregated ______________________________

State Ratings Positive Affect Negative Affect

_______________________________________________

Mean Weekly Ratings (N = 80)

Positive Affect .66* -.10

Negative Affect -.02 .48*

Mean Daily Ratings (N = 410)

Positive Affect .64* -.03

Negative Affect -.06 .53*

*p < .05, two-tailed.

_______________________________________________

A subset of these subjects (N = 211) completed the full, 60-

item PANAS-X, so that comparable data also were available

PANAS-X Manual

for the 11 lower order scales. It is noteworthy that six

scales had convergent correlations of .50 or greater: Joviality

(.59), Guilt (.59), Attentiveness (.58), Sadness (.54), Self-

Assurance (.53), and Shyness (.51). Four additional scales

had convergent coefficients in the .40 to .50 range: Serenity

(.47), Fear (.46), Hostility (.46), and Fatigue (.41). It is

noteworthy that Surprise produced the lowest convergent

correlation (.36), again challenging the validity of the trait

form of this scale.

_______________________________________________

Table 23 Varimax-Rotated Factor Loadings of

Personality, Emotionality, and Higher Order

PANAS-X Scales (General Instructions) in a

Sample of SMU Undergraduates

_______________________________________________

Factor

_____________________

Scale 1 2 3

_______________________________________________

GTS Negative Temperament .93

EPQ Neuroticism .90

Goldberg Neuroticism .84

PANAS-X Negative Affect .83

GTS Positive Temperament .87

Goldberg Extraversion .86

EPQ Extraversion .83

PANAS-X Positive Affect .79

GTS Disinhibition .87

EPQ Psychoticism .74

Goldberg Conscientiousness -.83

Note. N = 231. Loadings below I.30I are omitted.

GTS = General Temperament Survey; EPQ = Eysenck

Personality Questionnaire; PANAS-X = Positive and

Negative Affect Schedule (Expanded Form). Adapted from

Watson & Clark, 1997, Table 4).

_______________________________________________

Relations with measures of personality and emotionality.

We have also collected extensive data relating trait versions

of the higher order scales to various measures of personality

and emotionality. These data consistently yield a clear

convergent/discriminant pattern: Trait Negative Affect is

substantially correlated with measures of Neuroticism or

Negative Emotionality, but is generally unrelated to

measures of Extraversion or Positive Emotionality.

Conversely, trait Positive Affect is strongly related to

Extraversion and Positive Emotionality, but not to

Neuroticism or Negative Emotionality (Watson & Clark,

1984, 1992b, 1997). Table 23 presents illustrative data

from a sample of 231 SMU undergraduates. These subjects

were assessed on 11 measures: [a] trait versions of the two

higher order PANAS-X scales; [b] the Neuroticism,

Extraversion, and Psychoticism scales from the Eysenck

Personality Questionnaire (EPQ; Eysenck & Eysenck,

1975); [c] the Neuroticism, Extraversion, and

Conscientiousness scales developed by Goldberg (1983);

and [d] the Negative Temperament, Positive Temperament

and Disinhibition scales from the General Temperament

Survey (GTS; Clark & Watson, 1990a). It should be

noted that the GTS Negative and Positive Temperament

scales are true-false measures designed to assess trait

Negative and Positive Affectivity, respectively.

Scores on these 11 personality and affectivity measures were

subjected to a principal components analysis. An

inspection of the unrotated eigenvalues revealed three strong

factors with eigenvalues greater than 1.00; together, these

factors accounted for 74.5% of the total variance. These

three factors were rotated using varimax and the resulting

loadings are shown in Table 23. The first dimension

clearly can be identified as Neuroticism/Negative

Emotionality, and it is noteworthy that the PANAS-X

Negative Affect scale loads as strongly on this factor (.83) as

do the other measures. Similarly, the second dimension

can be interpreted as Extraversion/Positive Emotionality;

the PANAS-X Positive Affect scale loads .79 on this factor.

Finally, the third factor can be identified as Disinhibition

versus Conscientiousness.

Similarly, we have also related trait scores on the lower

order scales to measures of personality and emotionality.

These analyses have yielded results that are broadly similar

to those observed at the higher order level. That is, the

specific negative affects are moderately to strongly correlated

with measures of Neuroticism/Negative Emotionality,

whereas the individual positive affects are significantly

related to Extraversion/Positive Emotionality. However,

some interesting additional findings have been noted as well.

Table 24 presents illustrative data from a sample of 325

SMU undergraduates. These subjects were assessed on 20

personality and affectivity scales: [a] trait scores on the basic

negative emotion (Fear, Sadness, Guilt, Hostility) and

positive emotion (Joviality, Self-Assurance, Attentiveness)

scales from the PANAS-X; [b] the three GTS scales; [c] the

Neuroticism, Extraversion, Conscientiousness,

Agreeableness, and Openness scales from the NEO Five-

Factor Inventory (NEO-FFI; Costa & McCrae, 1992); and

[d] factor scales of the same "Big Five" personality traits

constructed from an expanded, 80-item version of

Goldberg's (1983) scales (see Watson & Clark, 1992b, for a

discussion of the development of these factor scales).

An inspection of the unrotated eigenvalues indicated the

presence of five strong factors, which jointly accounted for

74.9% of the total variance. The varimax-rotated loadings

from this solution are shown in Table 24. Corroborating

the results obtained with the general Negative Affect scale,

all four negative emotions loaded strongly on the

Neuroticism/Negative Emotionality factor. Fear and Guilt

were essentially pure markers of this factor, whereas

Sadness also had a modest secondary loading on the

Extraversion dimension. Hostility, however, split

between the Neuroticism and Agreeableness factors, and

actually had its highest loading (-.64) on the latter.

PANAS-X Manual

____________________________________________________________________________________________________

Table 24 Varimax-Rotated Factor Loadings of Personality, Emotionality and Lower-Order PANAS-X Scales (General

Instructions) in Sample of SMU Undergraduates

____________________________________________________________________________________________________

Factor

_____________________________________________________

Scale 1 2 3 4 5

____________________________________________________________________________________________________

NEO-FFI Neuroticism .82 -.21 -.16 .04 -.03

GTS Negative Temperament .80 -.07 .01 -.17 .00

Goldberg Neuroticism .80 -.15 -.13 .00 .02

PANAS-X Fear .79 -.04 .05 -.04 -.10

PANAS-X Guilt .76 -.08 -.17 -.13 -.03

PANAS-X Sadness .75 -.30 .00 -.09 .11

NEO-FFI Extraversion -.07 .88 -.06 .22 .07

Goldberg Extraversion -.20 .87 -.10 .05 .15

PANAS-X Joviality -.13 .84 .13 .27 .00

GTS Positive Temperament -.19 .78 .28 .02 .10

PANAS-X Self-Assurance -.35 .59 .25 -.36 .16

Goldberg Conscientiousness -.06 .06 .89 .04 .02

NEO-FFI Conscientiousness -.11 .10 .87 .02 -.11

PANAS-X Attentiveness -.14 .25 .80 .00 .07

GTS Disinhibition -.03 .32 -.73 -.38 .09

NEO-FFI Agreeableness -.12 .13 .07 .89 .10

Goldberg Agreeableness -.05 .25 .13 .84 .16

PANAS-X Hostility .52 -.02 -.07 -.63 -.06

NEO-FFI Openness .07 .00 -.03 .16 .88

Goldberg Openness -.10 .30 -.04 .08 .82

Note. N= 325. Loadings above I.40I are shown in boldface. NEO-FFI = NEO Five-Factor Inventory. Adapted from

Watson & Clark (1992b, Table 5).

____________________________________________________________________________________________________

Looking at the positive affects, Table 24 demonstrates

that Joviality and Self-Assurance both loaded strongly and

primarily on Extraversion/Positive Emotionality. In

addition, Self-Assurance also had moderate negative

loadings on both Neuroticism and Agreeableness.

Finally, Attentiveness was primarily a marker of

Conscientiousness, but also had a modest loading on

Extraversion (.25). Thus, the data reveal evidence of

relations between personality and affectivity that are both

general (i.e., between the negative affects and

Neuroticism, and between the positive affects and

Extraversion) and specific (e.g., between Hostility and

Agreeableness, and between Attentiveness and

Conscientiousness).

Summary of trait validity evidence. To summarize the

extensive data presented here and elsewhere: trait scores

on the PANAS-X scales (a) are stable over time, (b) show

significant convergent and discriminant validity when

correlated with peer-judgments, (c) are highly correlated

with corresponding measures of aggregated state affect,

and (d) are strongly and systematically related to measures

of personality and emotionality. These data clearly

demonstrate that --with the possible exception of Surprise-

-the PANAS-X scales can be used validly to assess long-

term individual differences in affect.

V. References

Albright, L., Kenny, D. A., & Malloy, T. E. (1988).

Consensus in personality judgments at zero

acquaintance. Journal of Personality and Social

Psychology, 55, 387-395.

Almagor, M., & Ben-Porath, Y. (1989). The two-

factor model of self-reported mood: A cross-cultural

replication. Journal of Personality Assessment, 53,

10-21.

PANAS-X Manual

Beck, A. T., Ward, C. H., Mendelson, M., Mock, J.,

& Erbaugh, J. (1961). An inventory for measuring

depression. Archives of General Psychiatry, 4, 561-

571.

Clark, L. A., & Watson, D. (1988). Mood and the

mundane: Relations between daily life events and self-

reported mood. Journal of Personality and Social

Psychology, 54, 296-308.

Clark, L. A., & Watson, D. (1990a). The General

Temperament Survey. Unpublished manuscript,

Southern Methodist University, Dallas, TX.

Clark, L. A., & Watson, D. (1990b). Predicting

momentary mood from ordinary life events and

personality. Unpublished manuscript, Southern

Methodist University, Dallas, TX.

Clark, L. A., Watson, D., & Leeka, J. (1989).

Diurnal variation in the positive affects. Motivation

and Emotion, 13, 205-234.

Costa, P. T., Jr., & McCrae, R. R. (1989). The

NEO-PI/NEO-FFI manual supplement. Odessa, FL:

Psychological Assessment Resources.

Derogatis, L. R. (1977). The SCL-90 Manual I: Scoring,

administration, and procedures for the SCL-90.

Baltimore: Johns Hopkins U School of Medicine.

Derogatis, L. R., & Cleary, P. A. (1977).

Confirmation of the dimensional structure of the SCL-

90: A study in construct validation. Journal of

Clinical Psychology, 33, 981-989.

Derogatis, L. R., Lipman, R S., Rickels, K.,

Uhlenhuth, E. H., & Covi, L. (1974). The

Hopkins Symptom Checklist (HSCL): A self-report

symptom inventory. Behavioral Science, 19, 1-15.

Ekman, P. (1971). Universals and cultural differences in

facial expressions of emotions. In J. K. Cole (Ed.),

Nebraska Symposium on Motivation (Vol. 19, pp.

207-283). Lincoln: University of Nebraska Press.

Eysenck, H. J., & Eysenck, S. B G. (1975). Manual

of the Eysenck Personality Questionnaire. San

Diego, CA: Educational & Industrial Testing Service.

Funder, D. C., & Colvin, C. R. (1988). Friends and

strangers: Acquaintanceship, agreement, and the

accuracy of personality judgment. Journal of

Personality and Social Psychology, 55, 149-158.

Funder, D. C., & Dobroth, K. M. (1987). Differences

between traits: Properties associated with interjudge

agreement. Journal of Personality and Social

Psychology, 52, 409-418.

Goldberg, L. R. (1983, June). The magical number

five, plus or minus two: Some conjectures on the

dimensionality of personality descriptions. Paper

presented at a research seminar, Gerontology Research

Center, Baltimore, MD.

Izard, C. E. (1972). Patterns of emotions: A new

analysis of anxiety and depression. New York:

Academic Press.

Kennedy-Moore, E., Greenberg, M., Newman, M., &

Stone, A. A. The relationship between daily events

and mood: The mood measure may matter. Motivation

and Emotion, 16, 143-155.

Kenrick, D. T., & Stringfield, D. O. (1980).

Personality traits and the eye of the beholder: Crossing

some traditional philosophical boundaries in the search

for consistency in all of the people. Psychological

Review, 87, 88-104.

Marshall, P. S., Watson, D., Steinberg, P., Cornblatt,

B., Peterson, P. K., Callies, A., & Schenck, C. H.

(1994). An assessment of cognitive function and

mood in chronic fatigue syndrome. Manuscript

submitted for publication.

Mayer, J. D., & Gaschke, Y. N. (1988). The

experience and meta-experience of mood. Journal of

Personality and Social Psychology, 55, 102-111.

McCrae, R. R., & Costa, P. T., Jr. (1987). Validation

of a five-factor model of personality across instruments

and observers. Journal of Personality and Social

Psychology, 52, 81-90.

McIntyre, C. W., Watson, D., & Cunningham, A. C.

(1990). The effects of social interaction, exercise,

and test stress on positive and negative affect.

Bulletin of the Psychonomic Society, 28, 141-143.

McNair, D. M., Lorr, M., & Droppleman, L. F.

(1971). Manual: Profile of Mood States. San

Diego: Educational and Industrial Testing Service.

Meyer, G. J., & Shack, J. R. (1989). Structural

convergence of mood and personality: Evidence for new

and old directions. Journal of Personality and Social

Psychology, 57, 691-706.

Norman, W. T., & Goldberg, L. R. (1966). Raters, ratees,

and randomness in personality structure. Journal of

Personality & Social Psychology, 4, 681-691.

Nowlis, V. (1965). Research with the Mood Adjective

Check List. In S. S. Tomkins & C. E. Izard (Eds.),

Affect, cognition, and personality (pp. 352-389).

PANAS-X Manual

New York: Springer Publishing.

Quinn, B. A. (1989). Religiousness and psychological

well-being: An empirical investigation. Unpublished

doctoral dissertation, Wayne State Univ., Detroit, MI.

Radloff, L. S. (1977). The CES-D Scale: A new self-

report depression scale for research in the general

population. Applied Psychological Measurement, 1,

385-401.

Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E.

(1970). Manual for the State-Trait Anxiety Inventory.

Palo Alto, CA: Consulting Psychologists Press.

Spielberger, C. D., Jacobs, G. A., Russell, S., & Crane,

R. S. (1983). Assessment of anger: The State-Trait

Anger Scale. In J. N. Butcher & C. D. Spielberger

(Eds.), Advances in personality assessment (Vol. 2,

pp. 52-76). Hillsdale, NJ: Erlbaum.

Tellegen, A. (1985). Structures of mood and

personality and their relevance to assessing anxiety,

with an emphasis on self-report. In A. H. Tuma and

J. D. Maser (Eds.), Anxiety and the anxiety disorders

(pp. 681-706). Hillsdale, NJ: Erlbaum.

Watson, D. (1988a). Intraindividual and interindividual

analyses of Positive and Negative Affect: Their relation

to health complaints, perceived stress, and daily

activities. Journal of Personality and Social

Psychology, 54, 1020-1030.

Watson, D. (1988b). The vicissitudes of mood

measurement: Effects of varying descriptors, time

frames, and response formats on measures of Positive

and Negative Affect. Journal of Personality and

Social Psychology, 55, 128-141.

Watson, D. (1989). Strangers' ratings of the five robust

personality factors: Evidence of a surprising

convergence with self-report. Journal of Personality

and Social Psychology, 57, 120-128.

Watson, D., & Clark, L. A. (1984). Negative affectivity:

The disposition to experience aversive emotional states.

Psychological Bulletin, 96, 465-490.

Watson, D., & Clark, L. A. (1989, August). General

affective dimensions versus discrete emotional factors:

A hierarchical model. In L. A. Clark & D. Watson

(Chairs), The emotional bases of personality.

Symposium presented at the 97th Annual Convention

of the Amer. Psychological Assoc., New Orleans, LA.

Watson, D., & Clark, L. A. (1991). Self- versus peer-

ratings of specific emotional traits: Evidence of

convergent and discriminant validity. Journal of

Personality and Social Psychology, 60, 927-940.

Watson, D., & Clark, L. A. (1992a). Affects separable

and inseparable: On the hierarchical arrangement of the

negative affects. Journal of Personality and Social

Psychology, 62, 489-505.

Watson, D., & Clark, L. A. (1992b). On traits and

temperament: General and specific factors of emotional

experience and their relation to the five-factor model.

Journal of Personality, 60, 441-476.

Watson, D., & Clark, L. A. (1997). Extraversion & its

positive emotional core. In R. Hogan, J. Johnson, &

S. Briggs (Eds.), Handbook of personality psychology

(pp. 767-793). San Diego, CA: Academic Press.

Watson, D., Clark, L. A., & Carey, G. (1988).

Positive and negative affectivity and their relation to

anxiety and depressive disorders. Journal of Abnormal

Psychology, 97, 346-353.

Watson, D., Clark, L. A., McIntyre, C. W., &

Hamaker, S. (1992). Affect, personality, and social

activity. Journal of Personality and Social

Psychology, 63, 1011-1025.

Watson, D., Clark, L. A., & Tellegen, A. (1984).

Cross-cultural convergence in the structure of mood:

A Japanese replication and a comparison with U. S.

findings. Journal of Personality and Social

Psychology, 47, 127-144.

Watson, D., Clark, L. A., & Tellegen, A. (1988).

Development and validation of brief measures of

Positive and Negative Affect: The PANAS Scales.

Journal of Personality and Social Psychology, 54,

1063-1070.

Watson, D., & Tellegen, A. (1985). Toward a

consensual structure of mood. Psychological

Bulletin, 98, 219-235.

Watson, D., & Walker, L. M. (1996). The long-term

stability and predictive validity of trait measures of

affect. Journal of Personality and Social Psychology,

70, 567-577.

Zevon, M. A., & Tellegen, A. (1982). The structure

of mood change: An idiographic/nomothetic analysis.

Journal of Personality and Social Psychology, 43,

111-122.

Zuckerman, M., & Lubin, B. (1965). Manual for the

Multiple Affect Adjective Check List. San Diego,

CA: Educational and Industrial Testing Service.

The PANAS X: Manual For Positive And Negative Affect Schedule Expanded Form X

Navigation menu

Versions of this User Manual:

Views

Navigation