Aging and performance on an everyday-based visual search task

Acta Psychologica 140 (2012) 208–217

Contents lists available at SciVerse ScienceDirect

Acta Psychologica journal homepage: www.elsevier.com/ locate/actpsy

Aging and performance on an everyday-based visual search task Lauren M. Potter a,⁎, Madeleine A. Grealy b, Mark A. Elliott b, Pilar Andrés c a b c

Department of Psychology, School of Life Sciences, John Muir Building, Heriot–Watt University, Edinburgh EH14 4AS, Scotland, UK School of Psychological Sciences and Health, University of Strathclyde, Graham Hills Building, 40 George Street, Glasgow G1 1QE, Scotland, UK Department of Psychology, University of the Balearic Islands, Ctra. Valldemossa, km 7.5, 07122 Palma de Mallorca, Spain

a r t i c l e

i n f o

Article history: Received 4 August 2011 Received in revised form 10 April 2012 Accepted 1 May 2012 Available online 2 June 2012 PsycINFO codes: 2323 Visual perception 2340 Attention 2860 Gerontology Keywords: Aging Middle age Visual search Everyday-based task

a b s t r a c t Research on aging and visual search often requires older people to search computer screens for target letters or numbers. The aim of this experiment was to investigate age-related differences using an everyday-based visual search task in a large participant sample (n = 261) aged 20–88 years. Our results show that: (1) old–old adults have more difficulty with triple conjunction searches with one highly distinctive feature compared to young– old and younger adults; (2) age-related declines in conjunction searches emerge in middle age then progress throughout older age; (3) age-related declines are evident in feature searches on target absent trials, as older people seem to exhaustively and serially search the whole display to determine a target's absence. Together, these findings suggest that declines emerge in middle age then progress throughout older age in feature integration, guided search, perceptual grouping and/or spreading suppression processes. Discussed are implications for enhancing everyday functioning throughout adulthood. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Older adults often report difficulties when searching for items within cluttered visual scenes (e.g., Kline et al., 1992; Kosnik, Winslow, Kline, Rasinski, & Sekuler, 1988). Research has been conducted on visual search and aging in everyday domains such as driving (e.g., Bédard et al., 2006), reading medication labels (e.g., Markowitz, Kent, Schuchard, & Fletcher, 2008), navigating web pages (e.g., Grahame, Laberge, & Scialfa, 2004), and face recognition (e.g., Hahn, Carlson, Singer, & Gronlund, 2006). Much of the research on visual search and aging though is laboratorybased and involves visual search displays on computer screens, using stimuli such as target letters within an array of other letters (BurtonDanner, Owsley, & Jackson, 2001; Foster, Behrmann, & Stuss, 1995). As older people often perform more poorly in laboratory compared to more real world contexts (Park & Gutchess, 2000), the first objective of this study was to investigate visual search abilities with older age in an everyday-based task involving a more naturalistic array of items than symbols on a computer screen, namely, searching shelves of pasta-filled jars for particular target items. Previous research also tends to compare younger and older adults' performance while neglecting the middle-aged, making it difficult to ascertain how performance changes across the adult lifespan. However,

⁎ Corresponding author. Tel.: + 44 131 451 8439; fax: + 44 131 451 3735. E-mail address: [email protected] (L.M. Potter). 0001-6918/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.actpsy.2012.05.001

age-related differences emerging before older age may be precursors to future declines and hence may provide important information not just for theories of cognitive aging, but also for the design and implementation of early interventions. A second objective of this study was therefore to contribute to filling this gap in the literature. For this purpose we investigated the performance of a large sample of participants across adulthood (20s to 80s). This research was run as a single experiment, but its presentation and results are divided into two sections for the sake of clarity. The first analysis extends the laboratory-based literature to examine target search times on everyday feature, double conjunction, and triple conjunction searches across the adult lifespan. The second analysis also investigates performance across a range of ages, but extends previous work by examining age differences in search time on target absent trials in a feature search as well as conjunction search. The aim was to investigate the extent to which adults of different ages used the self-terminating strategy of stopping searching when they realised that the target was absent. The rationale and hypotheses for each analysis are detailed below. 1.1. Visual search in feature and conjunction searches throughout adulthood In laboratory-based visual search tasks, participants scan lists of letters, pictures, or words in search of a particular target item. In a typical feature search the target differs from distractor items in terms of a single feature, and search times tend not to differ much between younger and older adults, or change as increasing numbers of distractors are added

L.M. Potter et al. / Acta Psychologica 140 (2012) 208–217

to the display (Burton-Danner et al., 2001; Foster et al., 1995; Oken, Kishiyama, & Kaye, 1994). When age differences have been found, they tend to be small and partially accounted for by age-related slowing and other processes such as distractibility and spatial resolution (e.g., Davis, Fujawa, & Shikano, 2002). However, the more distinct targets are from distractors in feature search, the less likely age differences are to be found (e.g., Whiting, Madden, Pierce, & Allen, 2005). This shows that younger and older adults can often detect a target similarly fast, and suggests that they do not have to search items in the display one by one. According to feature integration theory (Treisman, 1993; Treisman & Gelade, 1980), this is because elementary perceptual features such as colour and shape are extracted rapidly and in parallel over broad spatial areas (Burton-Danner et al., 2001; Plude & Doussard-Roosevelt, 1989; Scialfa, Esau, & Joffe, 1998), resulting in ‘perceptual pop-out’ (Treisman & Gelade, 1980) of the target. Thus, successful detection of targets in feature searches seems to occur through bottom-up processes (Donner et al., 2002; Grossberg, Mingolla, & Ross, 1994). In a typical conjunction search though, the target differs from distractors in terms of a conjunction of two or more features, such as searching for a red X within an array of green Xs and red Os. Older compared to younger adults often demonstrate significantly slower search times as additional distractors are added to the display (Burton-Danner et al., 2001; Donner et al., 2002; Wolfe & Cave, 1999). Feature integration theory (Treisman, 1993; Treisman & Gelade, 1980) proposes that this is because the features registered during the extraction stage are not sufficient for perceptual pop-out. Instead, features have to be matched between target and distractors in a predominantly serial manner, that is through feature integration processes. According to Wolfe's guided search model (Wolfe, Cave & Frenzel, 1989; Wolfe, 1994), these serial processes may only need to function across a limited region of the visual field at any one time (depending on the features being searched for), as information from parallel processors can be used to guide the deployment of spatially limited resources. For example, in the conjunction search, only half of the items (the red items) are potential targets, because a parallel colour processor can guide search to the red items to look for the red X, thus eliminating the green items. Both theories agree that the detection of targets requires more effortful or top-down processing in conjunction compared to feature searches, which is more difficult for older than younger adults. Some visual searches, however, require levels of processing not previously acknowledged by feature integration theory. For example on triple conjunction searches requiring the comparison of three features between target and distractors, search times are sometimes no different to or faster than double conjunction searches requiring the comparison of two features (Humphrey & Kramer, 1997; Quinlan & Humphreys, 1987; Wolfe et al., 1989). The guided search model explains these findings by considering the activation of certain features. That is, one parallel processor may activate all locations for one feature, while second and third parallel processors activate all locations for the second and third features of the target item within the triple conjunction search. The more distinctive one or more of the features are, the faster these processes will operate. This process reduces the set of potential target items to a subset, and a spatially limited process then searches for the target item. This could lead to search times for triple conjunction searches approximating those of double conjunction searches, as serial search would not take much more time for three compared to two features (particularly if one or more of the features in the triple search were particularly distinct). In the current study we explored visual search performance in a large sample and across different ages (20s to 80s) instead of focusing on just the performance of one younger group versus one older group. This approach builds on the few previous studies which have investigated visual search across the lifespan. For example, Hommel, Li, and Li (2004) examined trends in visual search from 6 to 89 years old to investigate whether the processes which improve during child development are those which decline with older age. They used a computer task in

209

which participants searched for a target (a filled white circle) among an array of distractors (unfilled white circles in a feature search, and filled white squares as well as unfilled white circles in a conjunction search). They found that age decrements were more pronounced in early and later life, more so for conjunction than feature searches, yet the developmental trend was asymmetrical; while children's performance was particularly affected by the mere presence of distractors, performance was particularly impaired in older age on target absent trials and with increasing numbers of distractors. Hommel et al. argued that while children have difficulty with distractibility, older people's performance is influenced more by a more cautious search style (i.e., exhaustive search under target absent conditions) which may be the result of compensating for neurocognitive decline. The present study extends Hommel et al.'s approach by focusing on performance throughout adulthood and older age, using a more everyday context, and comparing performance between different forms of conjunction search. Specifically, the current study aimed to investigate search times not just between feature and conjunction searches, but also between double and triple conjunction searches in which the triple conjunction search contained one feature more distinctive than the others. To do this, we created one feature search, one double conjunction search and one triple conjunction search. The feature search was designed to have a high level of distinctiveness between target and distractors to ensure that feature extraction processes would be sufficient to detect the target. We designed the double conjunction search so that the target item would differ in one gross feature from half of the distractor items and would share another gross feature with the rest. This was to ensure that two features would have to be compared between target and distractors and that serial search (feature integration or aspects of guided search) would be required to detect the target. Lastly, we designed a triple conjunction search so that the target would differ from distractor items in terms of three gross features. One of these features, though, was made more distinctive than the others to test whether this made the triple conjunction search the same difficulty level as the double conjunction search. It was predicted that age would have an increasing effect on visual search speed as task demand increased. Search times would not significantly differ between age groups for the feature search, but might become significantly slower on the double and triple conjunction search before older age, and would significantly slow throughout older age itself. For all ages it was predicted that search times would be significantly faster on the feature search compared to the double and triple conjunction searches, and would show no significant difference between the double and triple conjunction searches. 1.2. Exhaustive search versus stopping at target in feature and conjunction searches Previous laboratory-based search tasks show that in double conjunction searches, younger and older adults use the self-terminating strategy of stopping searching the rest of the display once the target is detected. This is evidenced by an approximate 2:1 ratio of search times on target absent to target present trials, which indicates a serial search of the whole display on absent trials (Madden, Pierce, & Allen, 1996; Plude & Doussard-Roosevelt, 1989; Van Zandt & Townsend, 1993). Serial searches are not always required however on target absent trials. For example, in feature searches younger adults detect a target almost as quickly as detecting its absence and produce absent-topresent ratios of approximately 1:1 (Burton-Danner et al., 2001; Wolfe et al., 1989). This suggests that younger adults stop searching the display as soon as the target is rapidly extracted (through parallel processing) on present trials, and as soon as target absence is rapidly detected on absent trials. It is not clear, however, whether older adults can quickly detect the absence of a target like younger people during feature searches. In Hommel et al.'s (2004) lifespan study, older people in their 60s to 80s (but not in younger age groups) showed significantly longer search times on absent trials in laboratory-based feature

210

L.M. Potter et al. / Acta Psychologica 140 (2012) 208–217

searches, which was attributed to increased cautiousness in older age. Few other studies have examined absent trials on feature searches with older age, possibly because it is accepted that feature searches do not typically require the serial search processes that are known to decline with older age. Thus the focus within the aging and visual search literature has been on age-related differences in conjunction searches instead. Such an approach has failed to acknowledge that the processes involved in detecting the absence of a target on feature searches may change with older age and be more complex than simple feature extraction or parallel search processes. In particular, it is not clear in feature searches whether the pattern of stopping the search quickly (upon realising that the target is absent) is maintained throughout adulthood and older age, or whether exhaustive strategies become used at some ages to search the whole display before deciding that the target is absent. Our aim was to investigate this throughout each decade of adulthood by measuring search times to target absent and present trials on a feature search as well as a double conjunction search. Based on previous literature, we predicted for the double conjunction search that the effect of age on search time would be stronger in target absent trials than in target present trials. Regardless of age, participants would stop searching the rest of the display once the target was detected on present trials, but on the target absent trials search times would be longer and older adults would show more exhaustive searching than younger adults. This would be evidenced by significantly longer search times for absent compared to present trials, and a ratio of absent to present search times that would approximate 2:1 in younger age groups and then increase with age. For the feature search, we predicted that each of the seven age groups would rapidly locate the target on present trials, and that younger age groups would rapidly detect the absence of the target on absent trials. This would be evidenced in younger adulthood by no significant difference between absent and present search times and an approximate 1:1 ratio of absent to present search times. Based on the rationale that processes beyond parallel search may be involved in detecting target absence in feature searches, we also predicted that the detection of target absence would slow down with older age in the feature search. This would be evidenced in older adulthood by a significant difference between absent and present search times, and a ratio of absent to present search times in excess of 1:1.

were not taking any medications likely to affect performance, and had normal or corrected vision. Ethical approval was granted locally, and all participants gave written informed consent. Table 1 summarises participant characteristics. One-way between subjects ANOVAs confirmed that there were no significant differences between age groups in MMSE scores (F (6, 254)=2.03, p=n.s.) or full time educational levels from secondary school onwards (F (6, 254)=1.74, p=n.s.). 2.2. Design We used a mixed design in which participants in each decade from the 20s to 80s completed 30 trials searching for a target item on each of 5 different visual search displays. Three search displays were used to assess feature, double and triple conjunction search (target present trials only). Two search displays were used to assess exhaustive searching versus stopping in feature and conjunction searches (target absent versus present trials). One hundred and fifty trials were completed in total. The variables and search displays used in each analysis, and the stimuli in each display are detailed in Table 2. Each of the 5 displays comprised one target item and 15 distractor items arranged 4 × 4 on a set of shelves. Thus, the target could be positioned in one of 16 possible locations on each trial. To specify the location in which the experimenter would position the target item on each trial for each of the displays, we designed two ‘target location’ charts. The first chart was for the first analysis to determine target locations for trials on the feature, double conjunction and triple conjunction searches. The second chart was for the second analysis in which the target item was either present or absent, and this determined target locations on present trials. On absent trials the display comprised solely of distractor items. Each chart was divided into two sections: part (a) listed target locations for practice trials and part (b) for the test trials. Target locations were randomised for each chart. We then adjusted parts (a) and (b) to ensure that each of the 16 locations were roughly equally represented within each. On the target absent trials the distractor items were randomly distributed. We also ensured that the type of target item differed between the feature and conjunction displays to minimise target priming effects. 2.3. Apparatus

2. Method 2.1. Participants We recruited 261 male and female adults aged 20–88 years (mean number of participants per decade of age=37) by contacting various organisations, groups and clubs in Southern and Central Scotland. Participants were recruited over a broad section of socio-economical strata and educational and occupational backgrounds. Volunteers were screened cognitively using the Mini-Mental State Examination (MMSE, score≥27/30; Folstein, Folstein, & McHugh, 1975), and in terms of physical health using a self-report medical history questionnaire and a thorough examination by a physiotherapist and biomechanist. Participants

The test apparatus consisted of a shelving unit (approximately 1 m × 1 m) with 4 shelves, search time equipment, and glass jars containing a variety of pasta shapes and colours. Although in real-world scenarios pasta tends to be sold in branded packets, transparent glass jars were used to avoid contaminating effects of familiarity with different brands on search times. The shelving unit was positioned so that when seated, participants' eye heights were at the centre of the display. The search time apparatus consisted of an electronic sensor and push down button, both wired to a millisecond timer. A spring loaded roller blind was secured to the top of the unit. When the experimenter released the roller blind to reveal the display, the millisecond timer started, and was stopped when the participant lifted their hand from

Table 1 Participant characteristics. Age-group

n Female

n Male

N total

Age (years)a

MMSE/30a

20–29 30–39 40–49 50–59 60–69 70–79 80–89 Overall

17 16 20 25 34 24 17 153

19 14 09 13 19 18 16 108

36 30 29 38 53 42 33 261

25.42 34.12 46.48 55.25 65.95 74.13 83.28 54.95

29.94 29.93 29.83 29.87 29.85 29.69 29.67 29.82

Educationa (years) Secondary school+

a

Values show means (and standard deviations).

(2.38) (2.39) (2.30) (2.87) (3.01) (2.78) (2.43) (21.09)

(0.23) (0.25) (0.47) (0.41) (0.50) (0.60) (0.54) (0.46)

6.75 6.77 6.28 6.26 5.28 5.79 5.03 5.96

(1.57) (2.31) (3.34) (3.35) (3.16) (3.82) (3.58) (3.16)

L.M. Potter et al. / Acta Psychologica 140 (2012) 208–217

211

Table 2 Details of target and distractor items in each visual search display. Display

Target

Comparison between feature and conjunction searches Feature search (colour) Yellow straws Double conjunction search (colour and Multi-coloured straws shape) Full jar of multi-coloured Triple conjunction search (colour, shape, quantity) straws Comparison of visual search strategies: stopping versus exhaustive searches Feature search (colour) with absent trials Yellow straws included Double conjunction search (colour and shape) Multi-coloured straws with absent trials included

Distractors

Level of processing

15 jars multi-coloured straws 8 jars multi-coloured twists;7 jars yellow straws

Feature extraction Serial search

5 full jars multi-coloured twists; 5 full jars yellow straws; 5 half jars multi-coloured straws

Perceptual grouping and suppression of distinctive items; serial search

15 jars multi-coloured straws when target present; 16 when target was absent 8 jars multi-coloured twists;7 jars yellow straws when target present, 8 when target was absent

Feature extraction on present trials and grouping on absent trials Serial search

the push down button to point to the target item. For the double and triple conjunction searches, which had more than one type of distractor, the different types were displayed in counterbalanced positions. Displays with target absent trials consisted of 16 distractor items. 2.4. Procedure Participants sat facing the shelving unit and placed their dominant hand on the push down button. They were informed that their task over a number of trials would be to find a target jar of pasta within the display, and to ensure that they would not forget the target item an example was placed on the side of the desk. On each trial participants were told that the experimenter would position the target jar in a display that was behind a blind. When the experimenter released the blind to reveal the display, their task was to look for the target jar as quickly as they could but without rushing (to minimise the risk of pointing to the wrong jar), and to lift their hand off the button and point to the target once they had found it. They were told that the target jar might be absent on some trials, and were instructed that if this happened they were to lift their hand and say “it's not there” as soon as they decided it was absent. Participants completed 30 test trials in each display (i.e., the trials were blocked for each display). Early checks confirmed that search times did not speed up or slow down throughout trials within each display (minimising learning effects across trials). The order in which the five displays (feature, double conjunction, triple conjunction, feature with absent trials, and double conjunction with absent trials) were completed was randomised for each participant. Early checks confirmed that search times between displays changed in accordance with display complexity rather than order of display presented. Practice trials were administered to participants in the first display allocated to them to familiarise them with the apparatus. The first test trial began once search times on the practice trials remained within 2000 ms of the previous trial for 5 trials in a row. This occurred on average after 20–30 trials for the younger and older adults. For each of the 30 trials in each of the 5 displays, search times were recorded. Mean search times were calculated and outliers (search times outside two standard deviations of the mean for that decade of age) and errors (pointing to the wrong jar) were removed. No errors were made by participants in their 20s–50s, although a small proportion of older adults (2 in their 60s and 70s, and 4 in their 80s) made one or two errors in the first display allocated to them only. To ensure that the declines found on our everyday-based test of visual search were due to visual search speed, rather than the speed with which participants were able to lift their hand from the response button, a measure of movement speed was included as a covariate. Movement speed was taken from an independent reaching task using the same sample. Participants were presented with a cup across a number of trials with marked contact points on the rim. The presentation of the cup triggered a millisecond timer which was stopped when they lifted their dominant hand from a response button to reach forward and grasp the cup (for

details see Potter & Grealy, 2006). The use of this movement task as a covariate has been previously published (see Potter & Grealy, 2008). 3. Results 3.1. Visual search in feature and conjunction searches throughout adulthood To test the hypothesis that search times would increase with age, and more so for conjunction than feature search, we first performed a series of multiple linear regression analyses. Search times on the feature trials (regression 1), the double conjunction trials (regression 2) and the triple conjunction trials (regression 3) were the dependent variables in the analyses. Movement speed was included as an independent variable in each analysis, along with age, to partial out the effect of motor ability on search times (meaning that any effects of age would be attributable to age-related differences in visual search speeds and not movement speed). A scatterplot of the relationship between age and search times on the feature, double conjunction and triple conjunction trials is presented in Fig. 1. As Table 3 shows, the regression models accounted for a significant proportion of the variance in feature, double conjunction and triple conjunction search times. In each model, age was the sole

Fig. 1. Relationship between age and search time for feature, double conjunction, and triple conjunction searches.

212

L.M. Potter et al. / Acta Psychologica 140 (2012) 208–217

Table 3 Linear regressions predicting feature, double-conjunction and triple-conjunction search times from age and movement speed. R2

SE

Β

6.91 0.03

1.42 0.17

0.32* 0.01

Regression predicting double-conjunction search times Age 0.18 29.03* 21.28 Movement speed 0.12

3.22 0.38

0.42* 0.02

Regression predicting triple-conjunction search times Age 0.40 84.71* 36.19 Movement speed 0.00

3.14 0.37

0.63* 0.00

Variables

F

Regression predicting feature search times Age 0.11 15.54* Movement speed

B

*p b .001. Betas for age differ significantly between all conditions: feature vs. double conjunction, t(518) = 4.75, p b .001; feature vs. triple conjunction, t(518) = 9.80, p b .001; double conjunction vs. triple conjunction, t(518) = 3.32, p b .001.

significant predictor of search times. As hypothesised, search times on all tasks increased with age (as indicated by the positive beta weights shown in Table 3). We also tested the differences between the beta weights for age across the three conditions using the procedure recommended by Edwards (1984). This showed that age had a stronger effect on double conjunction search times than on feature search times (t(518)=4.75, pb .001), and a stronger effect on triple conjunction search times than on double conjunction search times (t(518)=3.32, pb .001). Thus, in line with predictions, age had an increasing effect on visual search speed as task demand increased.

Mean Search Times (with Standard Deviations) Age Feature Double Triple conjunction conjunction 20’s 1268 (276) 2931 (590) 2629 (419) 30’s 1649 (438) 3466 (955) 3176 (862) 40’s 1928 (463) 3672 (981) 3802 (1128) 50’s 1727 (399) 3765 (994) 3483 (885) 60’s 1609 (329) 3491 (716) 3795 (774) 70’s 1795 (377) 3914 (795) 4317 (857) 80’s 1798 (316) 4573 (1119) 5022 (1098) Means Adjusted for Movement Speed (with Standard Deviations) Age

Feature

20’s 30’s 40’s 50’s 60’s 70’s 80’s

1286 (378) 1661 (372) 1940 (372) 1741 (376) 1605 (371) 1779 (376) 1768 (396)

Double conjunction 2952 (900) 3480 (887) 3686 (883) 3781 (894) 3487 (881) 3895 (901) 4539 (936)

Triple conjunction 2632 (894) 3178 (882) 3805 (878) 3485 (888) 3795 (874) 4314 (894) 5018 (931)

Fig. 2. Mean search times for feature and conjunction searches across age groups. Table shows means and standard deviations before and after adjusting for movement speed.

To ensure that the decrements found on our everyday-based test of visual search were due to age-related differences in specific visual search processes and not just baseline speed, we also re-ran the multiple linear regressions, this time using feature search times (in addition to age and movement speed) as a predictor of double and triple conjunction search times. Consistent with the regressions reported above, age was a significant independent predictor of both double conjunction search times (β=.22, pb .01) and triple conjunction search times (β=.45, pb .01). Feature search times also independently predicted double conjunction search times (β=.63, pb .01) and triple conjunction search times (β=.56, pb .01). Movement speed was not an independent predictor of either double conjunction search times (β=.01, n.s.) or triple conjunction search times (β=−.01, n.s.). As the age effects did not change when feature search times were included in the regressions predicting conjunction search times, these findings show that the age differences in conjunction searches were not due to baseline speed alone. To examine possible interaction effects between age and search task, the mean search times for the feature search, and double and triple conjunction searches were then examined using ANOVA. Fig. 2 illustrates mean search times for each of the three displays. A mixed ANOVA with a between-subjects factor of age group (20s through to 80s) and a within-subjects factor of display (feature, double conjunction, triple conjunction) revealed a significant main effect of display (F (2, 508) = 1902.07, p b .0005, ηp2 = .88). A post hoc Tukey's HSD test found that search times were significantly faster for the feature search compared to the double and triple conjunction searches (both p b .05), but did not differ between the double and triple conjunction searches. A significant main effect of age group was also shown (F (6, 254) = 18.04, p b .0005, ηp2 = .30), and a post hoc Tukey's HSD test showed that search times were significantly faster for participants in their 20s compared to all older age groups except those in their 30s (all pb .05), and significantly slower for older adults in their 70s and 80s compared to younger age groups (all pb .05). No other comparisons were significant. Of particular interest was a significant interaction between display and age group (F (12, 254) = 21.17, p b .0005, ηp2 = .33). A post hoc Tukey's HSD test revealed that, in line with predictions, search times were significantly faster for the feature search compared to the double and triple conjunction searches for all age groups (all p b .05). As predicted, there were no significant differences in search times between the double and triple conjunction searches for participants in their 20s– 60s. However, older adults in their 70s and 80s were significantly slower on the triple compared to the standard conjunction. This was contrary to our predictions. The results also showed that there were no significant differences in search times between age groups on the feature search, except for participants in their 20s who produced significantly faster search times than all other age groups except the 30s (all p b .05). The prediction that search times might become significantly slower before older age on the conjunction searches, and slow further throughout older age was also confirmed by the results. For the double conjunction search, search times were significantly slower for participants in their 50s–80s compared to those in their 20s and 30s (all p b .05). For the triple conjunction search, search times were significantly slower for most age groups compared to those in their 20s (all p b .05), for middle aged participants in their 40s to 60s compared to most younger age groups (all p b .05), and for older adults in their 70s and 80s compared to all younger age groups (all p b .05). Finally, a mixed between-subjects ANCOVA revealed that the main effects of display, age group, and the interaction between display and age group remained significant after adjusting for movement speed (respectively: F (2, 506) = 136.20, p b .0005, ηp2 = .35; F (6, 253) = 14.09, p b .0005, ηp2 = .25; F (12, 506) = 17.29, p b .0005, ηp2 = .29). Furthermore, the pattern of significant post hoc comparisons reported did not change after adjusting for movement speed. Fig. 2 includes the mean search times after adjusting for movement speed.

L.M. Potter et al. / Acta Psychologica 140 (2012) 208–217

Fig. 3. Relationship between age and search time for feature and conjunction searches on target present and target absent trials.

3.2. Exhaustive search versus stopping at target in feature and conjunction searches A series of multiple linear regressions was performed to test the effects of age on both feature and conjunction search times in the target present and target absent trials, separately. As in the earlier reported regressions, both age and movement speed were entered together into the analyses as independent variables, meaning that any significant effects of age on search times would be attributable to agerelated differences in visual search speed and not motor ability. A scatterplot showing the relationship between age and search time for the feature and conjunction searches on the target present and target absent trials is presented in Fig. 3. Table 4 shows that a significant proportion of the variation in search times was accounted for in all regression models. In all models, age was the only significant independent predictor of search times. Search times increased with age, as Table 4 Linear regressions predicting target present and target absent feature and conjunction search times from age and movement speed. R2

Β

Regression predicting target present feature search times Age 0.17 26.73* 7.37 Movement speed − 0.09

1.09 0.13

0.43* − 0.04

Regression predicting target absent feature search times Age 0.69 289.95* 37.20 Movement speed 0.12

1.77 0.21

0.82* 0.02

Regression predicting target present conjunction search times Age 0.42 93.20* 27.74 Movement speed 0.11

2.33 0.28

0.64* 0.02

Regression predicting target absent conjunction search times Age 0.56 163.37* 72.13 Movement speed 0.70

4.69 0.56

0.72* 0.06

F

B

expected. In line with predictions, Edwards's (1984) procedure for testing differences between beta weights showed that the effect of age on search times was stronger in target absent trials than in target present trials. This was the case for the feature (t(518) = 15.19, p b .001) as well as conjunction search (t(518) = 9.50, p b .001). The mean search times for the target absent and present trials on the feature and double conjunction search were then analysed using ANOVA. Fig. 4 illustrates the mean search times, and details the ratio of absent to present search times for each age group and display. A mixed ANOVA was conducted on search times with a betweensubjects factor of age group (20s through to 80s) and within-subjects factors of display (feature, double conjunction) and target presence (present, absent). This revealed a significant main effect of display (F (1, 254)= 4543.44, pb .0005, ηp2 =.95), and a post hoc Tukey's HSD test showed that search times were significantly faster for the feature compared to the double conjunction search (p b .05). A significant main effect of target presence was found (F (1, 254) = 4489.43, p b .0005, ηp2 = .95), and a post hoc Tukey's HSD test showed that search times were significantly slower for absent compared to present trials (p b .05). As expected there was also a significant main effect of age (F (6, 254) = 117.88, p b .0005, ηp2 = .74). A post hoc Tukey's HSD test showed that search times were significantly faster for participants in their 20s and 30s compared to all older age groups (all p b .05), and significantly slower for older adults in their 80s compared to those below 70 (all p b .05). Of particular interest was a significant interaction between display, target presence, and age group (F (6, 254)=8.14, pb .0005, ηp2 =.16). A post hoc Tukey's HSD test was used to test the hypothesis that search times would be significantly slower for absent compared to present trials on the conjunction search for each age group. This hypothesis was supported (all pb .05) and the ratio of absent to present search times was around 2:1 for both older and younger adults. It was also predicted that search times for the feature search would not differ significantly between absent and present trials for adults in their 20s–50s, and this was found to be the case (all pb .05), with mean absent-to-present ratios not greater than 1.25:1. This pattern, however, differed for older age groups; search times were significantly slower for absent compared to present trials for older adults in their 60s–80s (all pb .05), who showed absent-topresent ratios from 1.45:1 to just above 2:1. A mixed between-subjects ANCOVA with movement speed as the covariate revealed that the main effects of display, target presence, and age, and the interaction between them remained significant (respectively: F (1, 253) = 277.00, p b .0005, ηp2 = .52; F (1, 253) = 283.42, p b .0005, ηp2 = .53; F (6, 253) = 91.41, p b .0005, ηp2 = .68; F (6, 253) = 6.96, p b .0005, ηp2 = .14) after controlling for movement speed. In addition, the pattern of significant post hoc comparisons did not change after adjusting for movement speed. 4. Discussion 4.1. Visual search in feature and conjunction searches throughout adulthood

SE

Variables

213

*p b .001. Betas for age differ significantly between target present and target absent conditions: feature search times (target present vs. absent), t(518) = 15.19, p b .001; conjunction search times (target present vs. absent), t(518) = 9.50, p b .001.

Search times were examined between feature, double, and triple conjunction searches. Search times were significantly faster for the feature search compared to the double and triple conjunction searches for all age groups, which mirrors previous findings from laboratorybased aging studies, including Hommel et al.'s (2004) lifespan study. This suggests that feature extraction processes are faster than serial search processes for younger and older adults (Burton-Danner et al., 2001; Foster et al., 1995; Scialfa et al., 1998) on everyday-based as well as laboratory-based tasks. Previous lifespan studies tended not to compare performance on double and triple conjunction searches throughout adulthood and older adulthood. Our results showed that for participants in their 20s–60s, there were no significant differences in search times between double and triple conjunction searches as predicted. These findings can be

214

L.M. Potter et al. / Acta Psychologica 140 (2012) 208–217

Mean Search Times (with Standard Deviations) and target absent to target present search time ratios Age

Feature: target present

Feature: target absent

20’s

1308 (207)

1461 (225)

Absent: Conjunction: Present target present Ratio 1.12:1 2707 (523)

4707 (822)

Absent: Present Ratio 1.74:1

30’s

1589 (353)

1701 (241)

1.07:1

3267 (484)

6011 (813)

1.84:1

40’s

1823 (322)

2007 (211)

1.10:1

3456 (796)

6514 (678)

1.89:1

50’s

1780 (282)

2230 (262)

1.25:1

3914 (860)

7515 (1211)

1.92:1

60’s

1741 (282)

2519 (483)

1.45:1

3618 (539)

7089 (127)

1.96:1

70’s

1789 (347)

3380 (738)

1.89:1

4272 (569)

8553 (1667)

2.01:1

80’s

1823 (108)

3714 (446)

2.03:1

4687 (205)

10022 (1165)

2.14:1

Conjunction: target absent

Fig. 4. Mean search times for target present versus target absent trials on feature and conjunction searches across age groups. Table shows means, standard deviations, and absent to present search time ratios.

accounted for by Wolfe's guided search model (Wolfe et al., 1989; Wolfe, 1994); one parallel processor may have activated all locations for the target feature of full jars in the triple conjunction search, while second and third parallel processors activated all locations for the second and third target features of multicoloured pasta and straw shapes. This activation process could have reduced the set of potential target items to a subset and allowed a spatially limited process to search this subset and find the target item in a search time that approximated that of the double conjunction search. An alternative explanation focuses on processes of suppression rather than activation. Duncan and Humphreys (1989) proposed a model to explain performance on simple feature searches, in which a decrease in the activation strength of one item becomes distributed to other items by a spreading suppression mechanism, and these perceptual grouping processes allow homogenous distractor items to be eliminated from the search en masse. Applying this model to performance in triple conjunction searches, participants in their 20s to 60s may have quickly grouped together items with the third and most distinctive feature in the triple conjunction search (half-full jars) and eliminated them through spreading suppression processes. This would leave only two gross features to be integrated, thus rendering the same level of effortful processing required as that on the double conjunction search. Contrary to predictions however, participants in their 70s–80s showed significantly slower search times for the triple conjunction search compared to the double conjunction search. This novel finding is the first to acknowledge that the ‘old–old’ may have more difficulty than ‘young–old’ participants in eliminating a third distractor feature from visual search, even though the distractor feature is more distinctive than the other two. Based on Wolfe's model (Wolfe et al., 1989; Wolfe, 1994), it could be that ‘old–old’ adults have difficulty with the simultaneous activation of more than two parallel processors, and/or serially searching among three features within albeit a spatially limited subset of potential target items. In contrast, Duncan and Humphreys' (1989) model would suggest that ‘old–old’ adults have particular difficulty with perceptually grouping and eliminating distinctive features en masse. Indeed, findings from previous studies suggest that older

adults have particular difficulty in perceptually grouping homogenous items. Older adults only benefit from distractor homogeneity when display size is small (Plude & Hoyer, 1985), and the benefit of distractors which move coherently over those which move incoherently is significantly larger for younger rather than for older adults (Folk & Lincourt, 1996; Madden et al., 1996). Further explanations include that the oldest adults may be susceptible to failures in memory binding when encoding multiple features (e.g., Kessels, Hobbel, & Postma, 2007) like those within our triple conjunction search. This is consistent with previous research showing that older people show age-related declines in the ability to inhibit distractor items from working memory (Andrés, Van der Linden, & Parmentier, 2004; Hasher & Zacks, 1988), and that older adults are more susceptible than younger adults to distraction during triple conjunction search, even though they use a similar feature selection strategy (Dennis, Scialfa, & Ho, 2004). The extent to which older adults have difficulty with the simultaneous activation of more than two parallel processors, or the perceptual grouping or inhibition of homogenous distractor items from working memory, requires further investigation. As expected, age had a stronger effect on search times in the triple and double conjunction searches compared to the feature search. While participants in their 20s showed faster search times than those in their 40s and over in the feature search, there were no other significant age differences in search times for the feature search, which is in line with previous findings (Burton-Danner et al., 2001; Foster et al., 1995; Scialfa et al., 1998). For the double and triple conjunction searches, older adults in their 70s and 80s produced significantly slower search times than younger age groups, supporting previous findings showing slower serial search processes for older compared to younger adults (Burton-Danner et al., 2001; Foster et al., 1995; Oken et al., 1994). These age-related differences in serial search could be explained by a general reduction in speed of processing (Salthouse, 1996), or declines in the ability to focus or shift visuospatial attention onto targets (McDowd & Birren, 1990), or to inhibit distracting information (Andrés, Parmentier, & Escera, 2006; Andrés et al., 2004; Hasher & Zacks, 1988).

L.M. Potter et al. / Acta Psychologica 140 (2012) 208–217

If the difficulty lay with processing speed or shifting/focusing attention, however, then one would expect to see these difficulties in searches involving feature extraction as well as feature integration, which was not the case. Furthermore, when feature search time was included as a measure of baseline speed in the regressions predicting conjunction search times, it was a significant predictor of both double and triple conjunction search times, but the effect of age as a significant predictor remained unchanged. These findings suggest that the age effects on conjunction searches cannot solely be explained by a global decline across cognitive subsystems due to a slowing of the functional efficiency of the central nervous system (e.g., Rabbitt, 1996; Salthouse, 1996). Instead, age effects on the conjunction searches may also be explained by local declines in the specific cognitive subsystems underlying visual search processes, and in their underlying neural networks (e.g., Andrés & Van der Linden, 2000; Kramer, Hahn, & Gopher, 1999). In relation to inhibitory abilities, previous research shows older adults have slower search times as the number of distractors is increased in conjunction but not in feature searches (e.g., Burton-Danner et al., 2001; Foster et al., 1995; Humphrey & Kramer, 1997), suggesting that older adults' difficulty with conjunction searches may be due to problems with inhibiting distractor items that share some of the same features as the target item. Thus, the slower search times of older compared to younger adults is consistent with Hasher and Zacks' (1988) theory of inhibitory declines with older age in that inefficient inhibitory processes permit “the initial entrance to working memory of information that is off the goal path… and the prolonged maintenance of such information in working memory” (p. 213). A second novel finding was the significant declines in visual search from as early as the 50s on the double conjunction search, and even earlier (in the 40s) for the triple conjunction search, with further declines evident in the 70s for the triple conjunction search. This suggests that difficulties with the serial search processes integral to conjunction searches show early decline and are not exclusive to older adults. These early changes in visual search abilities also mirror results from largescale longitudinal studies showing early decline across a range of cognitive tests including memory, reasoning and spatial abilities (for reviews see Salthouse, 2009; Singh-Manoux et al., 2012). These results have important implications for informing the design of everyday items. For example, household items should incorporate distinctive features to direct the user's attention to the relevant part of the object, while minimising conjunctions of features (such as multiple symbols, colours, and shapes) which are characteristic of visual displays on mobile telephones and computer screens. The design of shop displays should also be informed by visual search abilities, as older people often complain that they cannot see what is in front of them (e.g., Kline et al., 1992; Kosnik et al., 1988). Searching shop shelves could be made easier by stacking the same product (say a particular type of coffee) in a distinct ‘block’, rather than in a single location within a larger array of distracting items (such as other types of coffee). This would speed search time as well as minimise the chance of selecting a similar but incorrect item by mistake. Our findings suggest that older adults may have difficulty perceiving a display of homogenous items as a perceptual whole, particularly with large display sizes (Plude & Hoyer, 1985), thus shop displays could benefit from being small and perceptually distinct from one another by using a vertical barrier between displays of e.g., different types of coffee, and by clearly labelling each separate display. 4.2. Exhaustive search versus stopping at target in feature and conjunction searches Search times on target present versus target absent trials were compared between feature and conjunction searches. For the conjunction search, search times were significantly slower for target absent compared to target present trials for all age groups as predicted, and both younger and older adults produced target absent to target present

215

search time ratios of around 2:1. Together, these findings are consistent with previous laboratory-based tasks (Madden et al., 1996; Plude & Doussard-Roosevelt, 1989; Van Zandt & Townsend, 1993), and suggest that conjunction searches were serially searched on absent trials, but that on present trials a self-terminating strategy was used to stop searching as soon as the target was detected. As expected, the effect of age on search time was stronger in target absent trials than target present trials for the conjunction search. Target absent to target present search time ratios increased from the 20s (1.74:1) to 80s (2.14 to 1), which is consistent with the findings from Hommel et al.'s (2004) lifespan study suggesting that exhaustive search in target absent trials reflects a more cautious and possibly compensatory search style with older age. For the feature search, search times did not differ significantly between absent and present trials as predicted for participants in their 20s–50s, who showed an absent to present ratio of around 1:1. This is in line with previous laboratory-based findings, showing that the process of extracting a target was equally as fast as detecting its absence. According to Chun and Wolfe (1996), the guided search model explains how the absence of a target can be rapidly detected. The probability that each item is a target (expressed as an activation) is calculated by computing in parallel differences between items and their similarity to the target item, then these activations are checked in decreasing order until the target is located, or until an activation threshold is reached (which becomes more conservative when targets are overlooked, and less strict following successful trials), or through guessing (the likelihood of which increases with trial duration). Duncan and Humphreys' (1989) model also explains the rapid detection of target absence; the search is stopped as soon as the target is detected through feature extraction processes, and as soon as its absence is detected through spreading suppression processes, that is, through perceptual grouping of distractor items as a homogenous set without the target item. Few studies though have explored whether older adults show the same pattern of rapid detection of target absence in feature searches. Our findings revealed that older people in their 60s–80s showed significantly longer search times on target absent than on target present trials in a feature search, with absent-to-present ratios of around 2:1. These findings from an everyday-based task mirror those from Hommel et al.'s (2004) computer-based study, and suggest that in older age the absence of a target even in feature search is not immediately apparent. Furthermore, this could not be accounted for by movement speed. It is not clear, however, why older adults had difficulty detecting the absence of a target in a feature search and why they would have to serially search the display to determine that the target was not there. Hommel et al. (2004) suggested that cognitive and attentional difficulties may lead to more cautious performance among the oldest adults. They argued that cortical noise increases with older age, therefore calling signals which indicate the presence of a target may be too weak for a response. Thus, when targets are absent in a feature search, older people may check at least those calling signals that are strong enough to represent possible targets, which increases search times. These calling signals seem to be akin to the target activations in Wolfe's (1989, 1994) guided search model which could also explain our findings. There may be age-related difficulties for the oldest adults in the parallel computation processes (of differences between items and their similarity to the target item) involved in calculating the probability that each item is a target (expressed as an activation). Or, there may be difficulties in the serial processes involved in checking these activations in decreasing order, either to detect the target or to reach the threshold at which target absence is decided. According to Duncan and Humphreys' (1989) spreading suppression model the oldest adults may have been slower in detecting the absence (compared to presence) of a distinctive target in a feature search because of difficulty in perceptually grouping the identical distractors as a homogenous set without the target item, and then inhibiting them from entering

216

L.M. Potter et al. / Acta Psychologica 140 (2012) 208–217

working memory. Indeed, whereas feature extraction seems to be an early bottom-up process (Donner et al., 2002; Grossberg et al., 1994), the serial process of checking activations in the guided search model (Wolfe, 1989, 1994), and the perceptual grouping processes characteristic of spreading suppression (Duncan & Humphreys, 1989) are controlled more by top-down processes and may underlie age-related declines in performance, particularly for the oldest groups of older adults. Problems in either set of processes from guided search or spreading suppression models could have resulted in more exhaustive serial search of the whole display. A further explanation is based on evidence that older adults set more stringent criteria in simple twochoice reaction time tasks (Ratcliff, 2001) and produce significantly longer response times compared to younger adults. Thus, in absent trials older participants may have set a higher criterion for saying that the target was not there, resulting in longer search times. Our findings emphasise that while there are no age-related declines in the feature extraction processes characteristic of feature searches, there are difficulties with other processes that may be involved in feature searches, specifically detecting the absence of a target. The specific processes underlying these declines, however, require further investigation, especially as these could be the types of processes which contribute to certain difficulties in everyday contexts with older age. For example, when searching supermarket displays, there is no guarantee that the target item we are searching for will be present, thus in older age it could take twice as long to realise that the target is absent than to find it when it is present. Even these seemingly small increases in search times are not a minor issue for older people in the real world, as older people report that each inconvenience experienced in day to day life adds up to discourage them from going out (e.g., Rantakokko et al., 2009; Sakari-Rantala, Heikkinen, & Ruoppila, 1995) and participating in tasks such as shopping in the first place. In addition, the problems faced by older people in everyday visual search tasks can be greater than just requiring more time to search. For example, if an older person is unsure about a target's absence then they are not likely to be satisfied by guessing or basing their judgement on previous ‘trials’ as outlined in the guided search model; instead they are more likely to check the display over again until they are sure, which can be a further time-consuming and frustrating process. These problems reinforce the arguments made in the previous section for simplifying complex visual search displays in everyday settings such as supermarket shelves. 4.3. Conclusion This study, using a range of ages and an everyday-based task, shows that: old–old adults have more difficulty than young–old and younger participants on triple conjunction searches with one distinctive feature; age-related declines on conjunction searches emerge as early as middle age and then progress throughout older age; and age-related declines are evident in feature searches when target absent trials are examined, in that older people seem to search a display serially and exhaustively to determine that a target is absent. There were several limitations to this study, however. First, while our participants had normal or corrected vision, we did not control for declines in perceptual processes such as visual acuity. Therefore the extent to which declines in acuity contribute to slowed search times, particularly for the oldest groups and in the most complex conditions would benefit from further investigation. Second, it could be argued that older people may be less able to switch search strategies between displays, and may persist in using the strategy they adopted for the first display presented to them, even though more effective strategies could be used in subsequent displays. However, as the order of displays was randomised for each participant, and search times slowed as the displays became more complex for both younger and older people, it seems unlikely that the display presented first and subsequent order of presentation influenced our results. Nevertheless, further studies should control for this possibility.

Future research is required to determine the processes responsible for the age effects found in our study, and examinations of more everyday tasks are also required to establish the generality of our findings. The addition of eye-tracking techniques to these paradigms would allow for more precise measurements of search strategies, further our knowledge of how visual search abilities change throughout adulthood, and inform the application of early interventions and improvements in the design of everyday items and environments which could facilitate functional ability throughout the whole adult lifespan. Acknowledgements This study was funded by the EPSRC as part of a large-scale multidisciplinary project entitled ‘The integration of biomechanical and psychological parameters of functional performance of older adults into a new computer aided design package for inclusive design’. We thank the many younger and older adults who kindly volunteered as participants. Pilar Andrés is supported by a grant from the Spanish Ministry of Science and Innovation (PSI2010-21609-C02-02). References Andrés, P., Parmentier, F., & Escera, C. (2006). The effect of aging on involuntary capture of attention by irrelevant sounds: A test of the frontal hypothesis of aging. Neuropsychologia, 44, 2564–2568. Andrés, P., & Van der Linden, M. (2000). Age-related differences in supervisory attentional system functions. Journal of Gerontology: Psychological Sciences, 55, 373–380. Andrés, P., Van der Linden, M., & Parmentier, F. (2004). Directed forgetting in working memory: Age-related differences. Memory, 12, 248–256. Bédard, M., Leonard, E., McAuliffe, J., Weaver, B., Gibbons, C., & Dubois, S. (2006). Experimental Aging Research, 32, 119–135. Burton-Danner, K., Owsley, C., & Jackson, G. R. (2001). Aging and feature search: The effect of search area. Experimental Aging Research, 27, 1–18. Chun, M. M., & Wolfe, J. M. (1996). Just say no: How are visual searches terminated when there is no target present? Cognitive Psychology, 30, 39–78. Davis, E. T., Fujawa, G., & Shikano, T. (2002). Perceptual processing and search efficiency of young and older adults in a simple feature search task: A staircase approach. Journal of Gerontology, 57B, 324–337. Dennis, W., Scialfa, C. T., & Ho, G. (2004). Age differences in feature selection in triple conjunction search. Journals of Gerontology Series B — Psychological Sciences and Social Sciences, 59, 191–198. Donner, T. H., Kettermann, A., Diesch, E., Ostendorf, A. V., Villringer, A., & Brandt, S. A. (2002). Feature and conjunction searches of equal difficulty engage only partially overlapping frontoparietal networks. NeuroImage, 15, 16–25. Duncan, J., & Humphreys, G. W. (1989). Visual search and stimulus similarity. Psychological Review, 96, 433–458. Edwards, A. L. (1984). An introduction to linear regression and correlation. New York: W. H. Freeman and Company. Folk, C. L., & Lincourt, A. E. (1996). The effects of age on guided conjunction search. Experimental Aging Research, 22, 99–118. Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Mini-mental state. A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189–198. Foster, J. K., Behrmann, M., & Stuss, D. T. (1995). Aging and visual search: Generalised cognitive slowing or selective deficit in attention? Aging and Cognition, 2, 279–299. Grahame, M., Laberge, J., & Scialfa, C. T. (2004). Age differences in search of web pages: The effects of link size, link number and clutter. Human Factors: The Journal of the Human Factors and Ergonomics Society, 46, 385–398. Grossberg, S., Mingolla, E., & Ross, W. D. (1994). A neural theory of attentive visual search: Interactions of boundary, surface, spatial, and object representations. Psychological Review, 101, 470–489. Hahn, S., Carlson, C., Singer, S., & Gronlund, S. D. (2006). Aging and visual search: Automatic and controlled attentional bias to threat faces. Acta Psychologica, 123, 312–336. Hasher, L., & Zacks, R. T. (1988). Working memory, comprehension, and aging: A review and a new view. In G. G. Bower (Ed.), The psychology of learning and motivation, Volume 22. (pp. 193–225)San Diego, CA: Academic Press. Hommel, B., Li, K. Z., & Li, S. C. (2004). Visual search across the lifespan. Developmental Psychology, 40, 545–558. Humphrey, D. G., & Kramer, A. F. (1997). Age differences in visual search for feature, conjunction, and triple-conjunction targets. Psychology and Aging, 12, 704–717. Kessels, R. P. C., Hobbel, D., & Postma, A. (2007). Aging, context memory and binding: A comparison of “what, where and when” in young and older adults. International Journal of Neuroscience, 117, 795–810. Kline, D. W., Kline, T. J. B., Fozard, J. L., Kosnik, W., Scheiber, F., & Sekuler, R. (1992). Vision, aging, and driving: The problems of older drivers. Journals of Gerontology: Psychological Sciences, 47, 27–34.

L.M. Potter et al. / Acta Psychologica 140 (2012) 208–217 Kosnik, W., Winslow, L., Kline, D., Rasinski, K., & Sekuler, R. (1988). Visual changes in daily life throughout adulthood. Journals of Gerontology: Psychological Sciences, 43, 63–70. Kramer, A. F., Hahn, S., & Gopher, D. (1999). Task coordination and aging: Explorations of executive control processes in the task switching paradigm. Acta Psychologica, 101, 339–378. Madden, D. J., Pierce, T. W., & Allen, P. A. (1996). Adult age differences in the use of distractor homogeneity during visual search. Psychology and Aging, 11, 454–474. Markowitz, S. N., Kent, C. K., Schuchard, R. A., & Fletcher, D. C. (2008). Ability to read medication labels improved by participation in a low vision rehabilitation program. Journal of Visual Impairment & Blindness, 102, 774–777. McDowd, J., & Birren, J. E. (1990). Aging and attentional processes. In J. E. Birren, & K. W. Schaie (Eds.), Handbook of the psychology of aging (pp. 222–233). New York: Academic Press. Oken, B. S., Kishiyama, S. S., & Kaye, J. A. (1994). Age-related differences in visual search task performance: Relative stability of parallel but not serial search. Journal of Geriatric Psychiatry and Neurology, 7, 163–168. Park, D. C., & Gutchess, A. H. (2000). Cognitive aging and everyday life. In D. Park, & N. Schwartz (Eds.), Cognitive aging: A primer (pp. 217–232). Hove: Psychology Press. Plude, D., & Doussard-Roosevelt, J. (1989). Aging, selective attention, and feature integration. Psychology and Aging, 4, 98–105. Plude, D., & Hoyer, W. (1985). Attention and performance: Identifying and localising age deficits. In N. Charness (Ed.), Aging and human performance (pp. 47–99). London: Wiley. Potter, L. M., & Grealy, M. A. (2006). Aging and inhibitory errors on a motor shift of set task. Experimental Brain Research, 171, 56–66. Potter, L. M., & Grealy, M. A. (2008). Aging and inhibition of a prepotent motor response during an ongoing action. Aging, Neuropsychology and Cognition, 15, 232–255. Quinlan, P., & Humphreys, G. (1987). Visual search for targets defined by combinations of colour, shape and size: An examination of the task constraints on feature and conjunction searches. Perception & Psychophysics, 41, 455–472. Rabbitt, P. (1996). Do individual differences in speed reflect global or local differences in mental abilities? Intelligence, 22, 69–88. Rantakokko, M., Manty, M., Iwarsson, S., Tormakangas, T., Leinonen, R., Heikkinen, E., et al. (2009). Fear of moving outdoors and development of outdoor walking difficulty in older people. Journal of the American Geriatrics Society, 57, 634–640.

217

Ratcliff, R. (2001). The effects of cognitive aging on reaction time in a signal detection task. Psychology and Aging, 16, 323–341. Sakari-Rantala, R., Heikkinen, E., & Ruoppila, I. (1995). Difficulties in mobility among elderly people and their association with socioeconomic factors, dwelling environment and use of services. Aging Clinical and Experimental Research, 7, 433–440. Salthouse, T. A. (1996). The processing speed theory of adult age differences in cognition. Psychological Review, 103, 403–428. Salthouse, T. A. (2009). When does age-related cognitive decline begin? Neurobiology of Aging, 30, 507–514. Scialfa, C. T., Esau, S. P., & Joffe, K. M. (1998). Age, target-distractor similarity, and visual search. Experimental Aging Research, 24, 337–358. Singh-Manoux, A., Kivimaki, M., Glymour, M., Elbaz, A., Berr, C., Ebmeier, K. P., et al. (2012). Timing of onset of cognitive decline: Results from Whitehall II prospective cohort study. British Medical Journal, http://dx.doi.org/10.1136/bmj.d7622. Treisman, A. (1993). The perception of features and objects. In A. Baddeley, & L. Weiskrantz (Eds.), Attention: Selection, awareness, and control (pp. 5–35). New York: Oxford University Press. Treisman, A., & Gelade, G. (1980). A feature integration theory of attention. Cognitive Psychology, 12, 97–136. Van Zandt, T., & Townsend, J. T. (1993). Self-terminating versus exhaustive processes in rapid visual memory search: An evaluative review. Perception & Psychophysics, 53, 563–580. Whiting, W. L., Madden, D. J., Pierce, T. W., & Allen, P. A. (2005). Searching from the top down: Ageing and attentional guidance during singleton detection. Quarterly Journal of Experimental Psychology. A, Human Experimental Psychology, 58, 72–97. Wolfe, J. M. (1994). Guided search 2.0 — A revised model of visual search. Psychonomic Bulletin & Review, 1, 202–238. Wolfe, J. M., & Cave, K. (1999). The psychophysical evidence for a binding problem in human vision. Neuron, 24, 11–17. Wolfe, J. M., Cave, K. R., & Frenzel, S. L. (1989). Guided search: An alternative to the feature integration model for visual search. Journal of Experimental Psychology. Human Perception and Performance, 15, 419–433.