Corrigendum

This article refers to:

The Emotional Stroop as an Emotion Regulation Task

Article title: The Emotional Stroop as an Emotion Regulation Task

Author: Cathleen Kappes and Christina Bermeitinger

Journal: Experimental Aging Research

Bibliometrics: Volume 42, Issue 2, Pages 161–194 (2016)

DOI: 10.1080/0361073X.2016.1132890

Due to reanalyses prompted by the careful reading of a colleague, the authors realized a mistake in the assignment of words presented to one group of participants (four words in the positive word list of both lists for young adults) as well as a few trials with a technical error in their experiment that was published as Study 2 in Kappes and Bermeitinger (2016). In detail, first, for young participants, eight words of the positive list (four in word list A and four in word list B; between-subject) were erroneously assigned to the wrong arousal category (that is, four words that are actually medium arousing—i.e., Fahrt, Leistung, Zirkus, Ansporn—were assigned to the high arousing words and vice versa for four other words, i.e., Anreiz, Genie, Weltruhm, Abitur; see updated Appendix A and Appendix B). Second, we noticed that due to technical reasons in a few trials, for half of the older participants, no target was shown (M = 1.2% of trials, ranging from 0.9% to 1.6% between participants). We now rearranged the words in the positive lists for young participants and excluded the trials without any target for old participants. Of course, thus, the statistical results slightly changed. In the following, we present the corrected results, tables, and figures. Note that the general pattern and the main results (including statistical significance) were the same as presented in our original article. So, our interpretation of the data remains unaltered. According to the new results, the following two sentences of the original discussion of Experiment 2 had to be deleted: “Younger adults’ mean RTs were faster in the positive word blocks than in the negative word blocks for words presented in the background, while[whereas] there was no difference for the older age group and the other word positions. As this effect did not interact with arousal level (and therefore the effect also includes differences in response to neutral words), it is difficult to interpret this finding and [it] is in need of replication (for example[,] in a study with more power).” The authors very much regret any inconveniences the errors and corrections may have caused and/or cause.

RESULTS OF STUDY 2

Emotional Stroop Effects (ESE)

Based on the same outlier criteria as used in Experiment 1, 8.7% of all trials were discarded (2.3% of all trials were error responses).

A 3 (arousal) × 2 (valence) × 3 (word position) × 2 (level) × 2 (age group) multivariate analysis of variance (MANOVA) was conducted (means and standard deviations are displayed in Table 3). Moreover, order (block order: negative/positive vs. positive/negative) as a between-subject factor was included to control for order effects. As expected, we obtained a main effect of age group, F(1, 76) = 48.78, p < .001, η_p² = .39. Again, older adults were significantly slower than young adults (M_young = 918 ms, SD_young = 130 ms vs. M_old = 1,190 ms, SD_old = 208 ms).

Table 1. ESEs separately for young and old adults in Study 1 and Study 2

	Study 1		Study 2
Age group	Med/Neg	High/Neg	Med/Neg	High/Neg	Med/Pos	High/Pos
Young	−5 (26)	−5 (31)	−1 (39)	3 (51)	7 (45)	7 (40)
Old	20 (31)	10 (29)	0 (40)	−4 (42)	1 (48)	−3 (51)

Note. ESEs in ms; SD in parenthesis. ESE (= emotional Stroop effect) for: Med/Neg = medium arousing negative words; High/Neg = high arousing negative words; Med/Pos = medium arousing positive words; High/Pos = high arousing positive words. Neg/Pos = negative word blocks preceded positive word blocks; Pos/Neg = positive word blocks preceded negative word blocks.

Table 3. Means and standard deviations of RT (in ms) separately for valence and arousal of words and word position for young and older adults in Study 2

			Young adults				Older adults
Valence	Arousal		BG	IR	REL	Total	BG	IR	REL	Total
Negative	Low	M	916	923	940	926	1164	1179	1217	1187
	(neutral)	(SD)	(136)	(142)	(147)	(136)	(203)	(220)	(230)	(213)
	Medium	M	914	918	942	925	1162	1191	1208	1187
		(SD)	(137)	(131)	(133)	(128)	(213)	(222)	(230)	(217)
	High	M	927	927	933	929	1171	1186	1193	1184
		(SD)	(128)	(141)	(150)	(135)	(198)	(219)	(220)	(209)
Positive	Low	M	883	915	918	905	1175	1192	1219	1195
	(neutral)	(SD)	(142)	(146)	(147)	(138)	(224)	(222)	(233)	(221)
	Medium	M	894	916	925	912	1180	1194	1216	1196
		(SD)	(134)	(154)	(146)	(139)	(208)	(224)	(236)	(217)
	High	M	897	911	928	912	1185	1192	1200	1193
		(SD)	(123)	(151)	(153)	(138)	(201)	(221)	(219)	(209)

Note. BG = background; IR = irrelevant object; REL = relevant object.

There was no evidence for any ESEs—the main effect of arousal and the interaction between arousal and valence were nonsignificant (F(2, 75) = 0.10, p = .90 and F(2, 75) = 0.21, p = .81). Moreover (and in contrast to Study 1), we neither obtained a significant interaction effect between age and arousal, F(2, 75) = 0.59, p = .55, nor a three-way interaction between age, arousal, and valence, F(2, 75) = 0.12, p = .89. The Stroop effects (none of them were significant) are displayed in Table 1.¹ Furthermore, we obtained an interaction effect between valence, age group, and order, F(1, 76) = 10.07, p = .002, η_p² = .12. Older adults were faster on average in the second valence block irrespective of which one they started with. In contrast, young adults were slower in both negative valence blocks compared with the positive valence block.

As in Study 1, we found a main effect of word position, F(2, 75) = 26.76, p < .001, η_p² = .42, which indicated longer RTs with words in the relevant object (M = 1069 ms) than with words in the irrelevant object (1053 ms) or in the background (1038 ms). This main effect was qualified by an interaction with level and order of word blocks, F(2, 75) = 4.66, p = .012, η_p² = .11. Differences between levels were larger for words in the background and irrelevant object in the negative/positive order than in the positive/negative order, but the levels did not differ depending on order when the word appeared in the relevant object.

We conducted the same MANOVA for error rates. There was a significant main effect of age group, F(1, 76) = 22.31, p < .001, η_p² = .23, indicating that older adults overall made fewer mistakes than young adults (M_young = 3.1% vs. M_old = 1.5%). Moreover, besides significant effects of valence, level, Level × Valence, Level × Age × Order, Position × Order × Arousal, and Arousal × Valence × Order, we obtained a four-way interaction between age, valence, arousal, and level, F(2, 75) = 4.55, p = .014, η_p² = .11. As in Study 1, the error rate is very low, which is why we refrain from an interpretation of this four-way interaction.

Relationship Between ESEs, Positive and Negative Affect, and FGA

To test group differences, we conducted three univariate analyses of variance (ANOVA) with age (young vs. old) and order of blocks (negative/positive vs. positive/negative) as between-subject factors and negative affect, positive affect, and flexible goal adjustment (FGA), respectively, as dependent variables. Only age was significant for negative affect and FGA. As in Study 1, older adults reported significantly less negative affect in the last 12 months than young adults, F(1, 76) = 7.10, p < .01 (see Table 2 in Kappes & Bermeitinger, 2016). Moreover, older adults reported significantly higher flexible goal adjustment values than young adults, F(1, 76) = 8.50, p < .01.

As in Study 1, we were interested in the relationship between ESE, positive and negative affect, and FGA. In a first step, we tested correlations between the four ESEs, positive and negative affect, and FGA. As we also wanted to test whether the results of Study 1 with negative (compared with neutral) stimuli could be replicated, we also conducted analyses separately for age groups and order of valence blocks (see Table 4).

Table 4. Summary of intercorrelations of NA, PA, and FGA with the different ESEs overall and separately for age and order groups in Study 2

		ESEhigh/neg		ESEmed/neg		ESEhigh/pos		ESEmed/pos
Measure	Block order	Young	Old	Young	Old	Young	Old	Young	Old
NA	Neg/Pos	.33	.02	.19	−.24	−.17	−.33	−.12	−.54*
	Pos/Neg	−.37	.25	−.12	−.01	.27	.03	.28	−.05
	Overall	.11		−.03		.01		−.06
PA	Neg/Pos	−.48*	−.06	−.11	.21	.11	.14	−.14	.31
	Pos/Neg	−.02	−.27	−.12	−.05	−.31	−.25	.16	−.11
	Overall	−.20^†		−.03		−.12		.04
FGA	Neg/Pos	−.24	−.19	−.01	.40^†	−.10	.22	.20	.41^†
	Pos/Neg	−.17	−.46*	.00	−.06	−.28	−.49*	−.12	−.13
	Overall	−.26*		.08		−.18		.08

Note. ESE (= emotional Stroop effect) for: Med/Neg = medium arousing negative words; High/Neg = high arousing negative words; Med/Pos = medium arousing positive words; High/Pos = high arousing positive words. NA = negative affect; PA = positive affect; FGA = flexible goal adjustment. Neg/Pos = negative word blocks preceded positive word blocks; Pos/Neg = positive word blocks preceded negative word blocks.

^† p < .10; *p < .05.

Negative Affect

Overall, there was no significant correlation between ESEs and negative affect obtained. The correlation effect between the ESE_med/neg and negative affect for the old age group in the negative/positive block order was in the same direction as in Study 1 but was not significant (r(17) = −.24, p = .33).

Positive Affect

In contrast to Study 1, we obtained a negative correlation between the ESE_high/neg and positive affect in the young age group with negative/positive order of blocks (r(19) = −.48, p = .029), meaning that a larger ESE with these words was related to less positive affect in the last 12 months.

Flexible Goal Adjustment

Regarding FGA, we only found a significant negative relationship between the ESE for high arousing negative words and FGA, r(78) = −.26, p = .019, that is, the more pronounced flexible goal adjustment was, the smaller was the ESE for negative high arousing words. Replicating the findings of Study 1, the old age group in the block order group confronted with negative word blocks first followed by positive word blocks exhibited a marginally significant positive correlation between ESE_med/neg and FGA (r(17) = .41, p = .081). However, this effect was not significantly different from the old age group in the opposite block order group (r(18) = −.13; Z = 1.61, p = .11). For correlations between ESE_high/pos and FGA, they showed the same pattern. Older adults in the negative/positive order group had a positive correlation (r(17) = .22, p = .37), whereas the other old age group had a negative correlation (r(18) = −.49, p = .029), and the correlations differed significantly from each other, Z = 2.18, p = .03. Thus, higher values in flexible goal adjustment were related to a smaller ESE for high arousing positive words in the older age group, when they saw positive word blocks first.

Relationship Between ESEs and Change in Momentary Affect

In a last step, we were interested in the relationship between the ESEs and (change in) momentary affect in response to the emotional Stroop task. First, we analyzed differences between baseline valence level and valence after the first two and second two blocks. We conducted a 2 (age group: young vs. old) × 2 (block order: negative/positive vs. positive/negative) × 3 (time: baseline valence vs. valence after first two blocks vs. after second two blocks; within-subject) repeated-measures ANOVA with valence as dependent variable. We obtained a significant interaction between time and age group, F(2, 152) = 7.45, p < .01, η_p² = .09 (see Table 5 in Kappes & Bermeitinger, 2016). Although valence decreased in the young age group between baseline and the Stroop task (M_baseline = 5.7 vs. M_{1st blocks} = 5.2 vs. M_{2nd blocks} = 5.2), it increased in the older age group (M_baseline = 5.4 vs. M_{1st blocks} = 5.7 vs. M_{2nd blocks} = 5.7). Moreover, there was a significant interaction effect between age and order of blocks, F(1, 76) = 4.23, p < .05, η_p² = .05. Young adults in the negative/positive words block order experienced the least positive affect overall (young: M_neg/pos = 4.8 vs. M_pos/neg = 5.6; old: M_neg/pos = 5.9 vs. M_pos/neg = 5.4).

To test the relationship between changes in momentary affect and the different ESEs, we calculated the difference between valence levels after the first two blocks and baseline and the difference between valence levels after the second two blocks and valence after the first two blocks. In order to examine changes in momentary affect after confrontation with either negative or positive words, we restructured the data set to obtain one variable signifying change after negative word blocks and one representing change after positive word blocks.² We conducted a regression analysis and regressed the difference variable after negative word blocks on the ESE_med/neg (z-standardized), age (effect-coded: young = −1, old = 1), and their interaction. The age groups differed significantly in the change of momentary affect after completing negative word blocks (B = 0.29, β = 0.23, p = .043), with older adults experiencing less decrease in momentary affect than young adults (M_young = −0.66 vs. M_old = −0.08). The influence of ESE_med/neg differed for the age groups, as indicated by a marginally significant interaction effect (B = 0.25, β = 0.20, p = .077). As displayed in Figure 3a, in the older age group, the ESE_med/neg was positively related to change in momentary affect (B = 0.27), which means the larger the ESE, the more positive is the change in momentary affect (i.e., the slower participants responded to negative medium arousing words in comparison with neutral words, the more increased their momentary affect). In contrast, younger adults showed the reversed pattern (B = −0.24). We repeated this analysis with ESE_high/neg; only the age group effect was significant (see Table 6).

Table 6. Predictors of change in momentary affect after negative word blocks in Study 2

	Change in momentary affect
	ESEmed/neg			ESEhigh/neg
Variable	B	β	95% CI	B	β	95% CI
Constant	−0.37*		[−0.65, −0.09]	−0.37*		[−0.65, −0.08]
ESE	0.15	0.01	[−0.27, 0.30]	0.14	0.11	[−0.16, 0.43]
Age	0.29*	0.23	[0.01, 0.57]	0.30*	0.23	[0.02 0.59]
ESE × Age	0.25^†	0.20	[−0.03, 0.54]	0.006	0.01	[−0.29, 0.30]
R^2corr	.05			.02

Note. CI = confidence interval. ESE z-standardized. Age effect coded (age: young = −1, old = 1). ESE (= emotional Stroop effect) for: Med/Neg = medium arousing negative words; High/Neg = high arousing negative words.

^† p < .10; *p < .05.

Table 7. Predictors of change in momentary affect after positive word blocks in Study 2

	Change in momentary affect
	ESEmed/pos			ESEhigh/pos
Variable	B	β	95% CI	B	β	95% CI
Constant	0.07		[−0.32, 0.46]	0.02		[−0.35, 0.38]
ESE	0.13	0.08	[−0.26, 0.53]	0.17	0.09	[−0.21, 0.55]
Age	0.39^†	0.22	[−0.01, 0.78]	0.39*	0.22	[0.02, 0.76]
ESE × Age	−0.20	−0.11	[−0.59, 0.19]	−0.65**	−0.37	[−1.03, −0.27]
R^2corr	.03			.14

Note. CI = confidence interval. ESE z-standardized. Age effect coded (age: young = −1, old = 1). ESE (= emotional Stroop effect) for: Med/Pos = medium arousing positive words; High/Pos = high arousing positive words.

^† p < .10; *p < .05; **p < .01.

Figure 3. Change in momentary affect in Study 2 after (a) negative word blocks as a function of ESE for negative medium arousing words separately for the age groups, and (b) after positive word blocks as a function of ESE for positive high arousing words separately for the age groups.

We conducted the same analyses for change in momentary affect after completing positive word blocks. As can be seen in Table 7, only age group was marginally significant (B = 0.39, β = 0.22, p = .053). Young adults momentary affect decreased, whereas that of older adults increased (M_young = −0.29 vs. M_old = 0.44). More importantly, the ESE_high/pos showed differential effects depending on age (Figure 3b). In the young age group, the ESE was positively related to change in momentary affect (B = 0.82), that is, the more delayed response times to high arousing positive words were in comparison with neutral words, the more increased their momentary affect. In contrast, momentary affect decreased in the older age group as a function of the ESE_high/pos (B = −0.49).

Notes

¹ As in Study 1, a main effect of object level was qualified by an interaction with age, F(1, 76) = 10.36, p = .002, η² = .12. Participants responded faster when the relevant object was in the front than in the rear. This difference was larger for older adults.

² Prior analyses included order of blocks (effect-coded: negative/positive = −1, positive/negative = 1) and its interaction terms. Order had no significant influence.

Appendix A. Stimuli used in Studies 1 and 2

Table

Age group		Age group
Young	Old	Young	Old
High/Neg		High/Pos
ALARM	ALARM	ABITUR	ANSPORN
ANGST	ANGST	ANREIZ	APPLAUS
ANKLAGE	ANKLAGE	APPLAUS	EROTIK
ARSCH	FLUCHT	EROTIK	EUPHORIE
AUFSTAND	FURCHT	EUPHORIE	FAHRT
FLUCHT	GEFAHR	FINALE	FINALE
FURCHT	GEGNER	FLIRT	FLIRT
GEFAHR	GEWALT	FREIHEIT	FREIHEIT
GEWALT	HETZER	GEBURT	GEBURT
HETZER	IRRSINN	GENIE	GEWINN
IRRSINN	KRIEG	GEWINN	JAZZ
KRIEG	LÜGE	JAZZ	JUBEL
LEICHNAM	OHRFEIGE	JUBEL	LEISTUNG
LEID	OPFER	LIEBE	LIEBE
LÜGE	RADAU	LIEBELEI	LIEBELEI
MAHNUNG	RAUB	LUST	LUST
OHRFEIGE	SARG	PARTY	PARTY
OPFER	SCHAUDER	POTENZ	POTENZ
RAUB	SCHIMPF	REKORD	REKORD
SARG	SCHREI	RETTER	RETTER
SCHAUDER	SCHULD	ROMANZE	ROMANZE
SCHLAMPE	SKLAVE	SIEG	SIEG
SCHREI	SORGE	SIEGER	SIEGER
SEKTE	SPITZEL	SPAß	SPAß
SKLAVE	STRAFE	THEATER	THEATER
SPITZEL	STREIT	TRIUMPH	TRIUMPH
STREIT	TRENNUNG	VEREHRER	VEREHRER
TOTE	TROTZ	VORSPIEL	VORSPIEL
TRENNUNG	UNFALL	WAGEMUT	WAGEMUT
UNFALL	UNFRIEDE	WELTRUHM	WUNDER
UNTREUE	UNTREUE	WUNDER	ZIRKUS
WAHNSINN	WAHNSINN	ZUKUNFT	ZUKUNFT
Med/Neg		Med/Pos
ABSCHIED	ABSCHIED	ANSPORN	ABITUR
AMPEL	ARSCH	BALANCE	ANREIZ
ASCHE	AUFPREIS	BEGABUNG	BALANCE
AUFPREIS	BAHRE	BEWEGUNG	BEGABUNG
BAHRE	BEHÖRDE	DENKER	BEWEGUNG
BEHÖRDE	BEILEID	FAHRT	DENKER
BEILEID	FRIEDHOF	FILM	FILM
BEULE	FRIST	GEDICHT	GEDICHT
DUMMHEIT	GEJAMMER	INSTINKT	GENIE
FRIEDHOF	HABGIER	KAMERAD	IMPULS
FRIST	IDIOT	KOLLEGE	INSTINKT
GEJAMMER	KÄ€FIG	LEISTUNG	KAMERAD
GEREDE	KLINIK	LUXUS	KOLLEGE
IDIOT	KUMMER	MADEIRA	LUXUS
KÄ€FIG	MINE	MANEGE	MADEIRA
MINE	REINFALL	PRIVILEG	MANEGE
NACHTEIL	RUDEL	RUBIN	NEUBAU
REINFALL	SCHLAMPE	RÜCKKEHR	PRIVILEG
SÄGE	SEKTE	SCHILLER	RUBIN
SCHIMPF	SNOB	SILBER	RÜCKKEHR
SNOB	TOTE	SPIEL	SCHILLER
SORGEN	TRITT	SPRACHE	SILBER
TRITT	UMWEG	TALENT	SPIEL
UMWEG	UNDANK	TOLERANZ	SPRACHE
UNDANK	URNE	UNSCHULD	TALENT
UNFRIEDE	VERBOT	URLAUBER	UNSCHULD
URNE	VORWAND	VOGEL	URLAUBER
VORWAND	WAISE	VORLIEBE	VORLIEBE
WAISE	WARTEREI	VORZUG	VORZUG
WARTEREI	WITWER	WEIN	WEIN
WITWER	ZOLL	WERT	WELTRUHM
ZWEIFEL	ZWEIFEL	ZIRKUS	ZAUBER
Neutral (negative word list)		Neutral (positive word list)
ABSTAND	ABDRUCK	ABLAUF
ABTEI	ABLAUF	ABLEGER	AUSSAGE
ANZAHL	ABSTAND	ABSENDER	BALKEN
ARCHIV	ABTEIL	ABTEIL	BALTIKUM
BATTERIE	ACHSE	ACHSE	BEREICH
BEREICH	ADRESSAT	ADRESSAT	DECKEL
DAUMEN	ANZAHL	AUSSAGE	GÜRTEL
DECKEL	BATTERIE	BALKEN	HEBEL
DIELE	DAUMEN	BALTIKUM	HOCKER
FLASCHE	DIELE	FALTE	KARTEI
GEGEND	FLASCHE	GIEBEL	KARTON
HOCKER	GIEBEL	HAUBE	KRAWATTE
KARTEI	HAUBE	HOBEL	KREIDE
KARTON	HOBEL	KREIDE	LEITER
LOCHER	LATEIN	LENKER	METER
PAPPE	LENKER	LINEAL	PAPPE
PENDEL	LINEAL	METER	QUADER
QUADER	MECHANIK	ORDNER	REIHE
REIHE	MIETE	PATER	SCHRANK
SATZ	OBJEKT	PERSON	SCHRAUBE
SCHRANK	ORDNER	PUMPE	SILBE
SCHRAUBE	PATER	ROLLE	SITZUNG
SIEB	PENDEL	SITZUNG	SPULE
SILBE	PUMPE	SKAT	STEMPEL
STEMPEL	RASIERER	SORTE	TABLETT
STUHL	SATZ	SPULE	TAFEL
TABLETT	SIEB	STUFE	TASTE
TAFEL	SORTE	TASTE	THEMA
TISCH	TISCH	ZETTEL	WEBER
WEBER	WESTE	ZIEGEL	ZIEGEL
WESTE	ZETTEL	ZUSTAND	ZWECK
ZWIEBACK	ZUSTAND	ZYLINDER	ZYLINDER

Note. Med/Neg = medium arousing negative words; High/Neg = high arousing negative words; Med/Pos = medium arousing positive words; High/Pos = high arousing positive words.

Appendix B. Mean arousal ratings, valence rating, and word frequency by arousal and valence condition for age groups separately

Table

	Young age group			Old age group
Parameter	M (SD)	Min	Max	M (SD)	Min	Max
Arousal
High/Neg	4.15^a (0.21)	3.83	4.57	4.10^a (0.27)	3.70	4.8
Med/Neg	3.05^b (0.17)	2.68	3.28	3.01^b (0.25)	2.50	3.4
Low (Negative word blocks)	1.86^c (0.08)	1.71	2.00	1.87^c (0.08)	1.70	2.00
Low (Positive word blocks)	2.11^d (0.06)	2.00	2.22	1.82^c (0.40)	1.40	3.40
Med/Pos	2.71^e (0.32)	2.30	3.50	2.82^b (0.35)	2.10	3.40
High/Pos	3.77^f (0.24)	3.30	4.44	3.83^d (0.34)	3.40	4.80
Valence
High/Neg	−1.97^a (0.47)	−2.90	−1.10	−2.42^a (0.33)	−3.00	−1.63
Med/Neg	−1.44^b (0.43)	−2.20	−0.06	−1.67^b (0.61)	−2.71	−0.60
Low (Negative word blocks)	0.13^c (0.25)	−0.50	0.50	0.11^c (0.18)	−0.38	0.37
Low (Positive word blocks)	0.09^c (0.25)	−0.45	0.50	0.02^c (0.47)	−2.33	0.44
Med/Pos	1.36^b (0.46)	0.60	2.30	1.56^b (0.36)	1.00	2.47
High/Pos	1.72^a (0.56)	0.70	2.90	1.90^b.d (0.48)	0.84	2.75
Ftot1MIL
High/Neg	40.50^a (66.32)	0.50	315.33	50.04^a (70.63)	0.50	315.33
Med/Neg	13.54^a (22.51)	0.50	86.33	13.39^a (19.43)	0.50	72.50
Low (Negative word blocks)	30.19^a (47.37)	0	167.17	24.36^a (37.89)	0	163.17
Low (Positive word blocks)	31.29^a (55.52)	0.83	209.50	36.69^a (58.63)	0	209.50
Med/Pos	49.44^a (65.41)	1.33	202.83	36.78^a (53.90)	1.33	202.83
High/Pos	38.38^a (58.21)	1.17	226.50	45.77^a (64.20)	1.17	226.50
Syllables
High/Neg	1.78 (0.61)	1	3	1.78 (0.66)	1	3
Med/Neg	2.03 (0.54)	1	3	1.94 (0.50)	1	3
Low (Negative word blocks)	1.88 (0.42)	1	3	2.06 (0.50)	1	3
Low (Positive word blocks)	2.19 (0.47)	1	3	2.06 (0.44)	1	3
Med/Pos	2.19 (0.69)	1	3	2.25 (0.62)	1	3
High/Pos	2.13 (0.66)	1	3	2.06 (0.67)	1	3

Note. Med/Neg = medium arousing negative words; High/Neg = high arousing negative words; Med/Pos = medium arousing positive words; High/Pos = high arousing positive words. Ftot1MIL = total frequency of appearance per million words.

Norm ratings for the young age group (Vö et al., 2009). Ratings for the older age group based on a prestudy (see text for further details). Arousal was rated on a 5-point scale, and valence was rated on a 7-point scale (see Vö et al., 2009).

Comparison between word types separately for arousal, valence, and frequency and age groups. Bonferroni-adjusted comparisons. Superscripts identify differences. For valence, absolute values were compared. Bold = age group comparisons t test with p < .05 for each row. Analysis of neutral words in positive blocks in the older age group is based on only 31 words, as one word was included falsely.