Exporing the Effect of Visual Complexity on Recall

Exporing the Effect of Visual Complexity on Recall

Hayleigh Proctor

08/09/2021

This study was conducted to explore the effect of visual complexity on an individuals` recall of product brands and their attributes in either simple or complex adverts . Within the field of visual complexity, there has been contradiction as to whether complexity helps or hinders recall, this study aims to resolve this question. A survey was conducted to measure their free and cued recall for adverts that varied in their visual complexity. The complex advertisements were defined as having three objects included whilst the simple advertisements had only one object included. This was decided to align with the industry standard for defining visual complexity as set by Attneave (1954), Snodgrass & Vanderwart (1980) and Chikhman et al., (2012). A percentage scoring system was used to compare overall memory performance. The data showed that those in the simple condition performed better compared to those in the complex condition. However, this was not the case for every individual. The results found the effects of complexity to be marginally significant (p < 0.09); however, the study had limited power, and a replication with a larger population could provide a more complete picture of the influence of the independent variable. Whilst this study does not provide a definitive conclusion towards the effect of visual complexity, it does explore and provide an insight into the effects of complexity on recall of product attributes in advertisements.

#visualcomplexity #recall #free-recall #cued-recall #advertisements #simple #complex

PARTICIPANTS
The larger the number of participants in a study, the better-protected results will be from extraneous variables. For this reason, the participants were collected through random snowball sampling (Emerson, 2015). Each condition had 22 participants, a minimum age of 16 being the only participation condition. The participants were randomly allocated to each one of the four experimental conditions, providing 88 total participants (N= 88). There were no gender requirements for participation (Females (N = 47), Males (N = 31), Other (N = 4)).
The majority of participants were born in the U.K. (N = 46) or Poland (N = 35). The majority are currently residing in England (N = 57) or Poland (N = 21), but responses were still collected from further afield, such as France and the U.S.A. (N = 10). The majority of participants fell into the two youngest age categories, 16 to 18-year-olds (N = 22) and 22 to 27-year-olds (N = 37).
General demographic information provided insight into the advertisement exposure in participants' generic routines. The majority of participants were native English speakers (N = 49). The majority of participants use streaming services (N = 76), of which just under half of the respondents said their service had adverts (N = 38). Participants also use ad blockers (N = 49). Just over a quarter of participants use cable T.V. (N = 27). When asked whether they pay for premium applications, the majority said ‘never’ (N = 60), occasionally (N = 16), sometimes (N = 9), usually (N = 2), whilst only one participant always pays for premium applications (N = 1).
MATERIALS
Firstly, two product categories were chosen, bottled water and soap bars, four brands were then selected per category (see table 1). There were 16 advertisements in total, eight for the simple and complex conditions, respectively. (APPENDIX A) The editing software Gimp was used to design the advertisements to enable the selected products to be presented in the controlled advert setting. This 'controlled setting' ensured that the backgrounds were consistent across the adverts, e.g., they all used the same blue background. Additionally, no text or fonts were added, and the objects included had the same position as their counterparts. There were two experimental groups wherein participants were presented the advertisements. Within those two groups participants would view one of the product categories e.g., the water products. To account for confounding variables advertisements were counterbalanced, randomizing their order of appearance. Participants only saw one product category (e.g., soap or water) and one variation of the advert e.g., if they saw the simple A1 Aveeno advert, they were not be presented with the complex B1 Aveeno advert. If participants saw the complex B5 Buxton advert, they were not presented with the simple B1 Buxton version. If participants saw the soap adverts, they did not see the water and vice versa.
The web-based software Qualtrics was used to create the surveys (APPENDIX B) and a generalized report of the results. After extracting the data, SPSS was used to dummy code and manipulate the data to measure the effect of visual complexity on recall.
DESIGN
This experiment used a between-group design wherein participants were allocated either the simple or complex condition to examine which level of complexity had the larger effect (Turkeltaub et al., 2011). The type of complexity, simple or complex, is the independent variable of the experiment. The dependent variable is the effect this has on participants' recall (Atinc et al., 2011). In this project, simple advertisements are defined by having only one object included in the background, whereas complex advertisements are defined by having three objects.
Participants were first asked questions pertaining to free recall of product attributes before then being presented with the cued recall questions. This was to allow a distinction between non prompted (free) and prompted (cued) responses, enabling me to mark each survey and allocate a combined percentage recall score to each participant.
To control for confounding variables, the surveys were counterbalanced. Participants were shown the adverts randomly within each experimental group so that I could isolate the sequence effects that participants are exposed to. However, I could not control for extraneous variables such as the time of day participants completed the survey, their emotional state, or their level of intelligence. Additionally, situational factors such as the location they were in, e.g., whether the room they were in was too loud, too hot, too cold, could not be accounted for.
To prevent participants from rehearsing the material, distraction tasks were provided before requesting question responses (APPENDIX C). These were designed to be cognitively engaging by requiring participants to read sections of text and 'fill in' the missing words and select the 'odd word out' in a listing task. When completing these tasks, participants would not necessarily be aware that they were not an essential part of the study and thus, in processing their responses, would have to pause. For example, 'which word does not belong with the others?' had the response options of ‘Dog’, ‘Cat’, ‘Donkey’, and ‘Dragon’. There are actually two responses that could be deemed correct; however, participants are told to select one. The correct responses were ‘Cat’ as it is the only word beginning with the letter 'C' and ‘Dragon’ as it is the only creature with wings. Participants could not advance to the next section if there were any responses left blank.
All of the advertisements had the same consistent blue background, no fonts were used, and all objects had the same positioning between the simple and complex conditions. For example, A2 and B2 Dove both had the blue ribbon object included in the same position. All simple advertisements had one object; all complex advertisements had three objects to allow a comparison of the effect of complexity on consumers' explicit recall.
PROCEDURE
Participants were found and randomly allocated to one of the experimental groups. They were first presented with the participant information sheet (APPENDIX D) in which general information about the experiment was explained without revealing that it was the level of complexity being measured. Participants were also required to complete the consent form. (APPENDIX E) Thus ensuring the participant is aware that their data will be collected anonymously and that they have the right to withdraw at any time should they please.
Participants then viewed four advertisements for 30 seconds per advert. They were not able to advance to the next image until the timer ended . The counterbalancing of questionnaires meant that the adverts were viewed in random orders. The distraction task then engaged participants for a few minutes as they could not advance until the distraction tasks were complete.
Participants were then asked the free recall questions in which they are expected to list the brands they can remember and list the product attributes for said brands. The soap category had 26 points available for free recall, and the water category had 15 points available. This is due to more attributes generally being included on the packaging of the soap comparatively to a generic product like water. Ergo, a more comprehensive list of features was able to be asked.
Once the participant had submitted the free recall section, they moved onto the cued recall questions. This section provided prompts in the questions, for example, ‘name the products, if any, that were moisturizing?’ participants may not have been able to recall this attribute freely. Therefore, these questions had to be presented separately so as not to influence each other. Furthermore, the free recall had to be asked first for the same reason of not influencing responses. If participants had filled the cued responses first, this would invalidate any free recall questions which may have followed. The soap and water categories respectively had 16 points available for the cued recall questions.
Once the survey was completed, participants were shown the debrief sheet (APPENDIX F) in which the aim of the study was fully explained, and they were provided with details should they have any questions about their role and wish to discuss it further.

Lancaster University

Data/SPSS.sav

Proctor2021

Lydia Brooks

Open

Field of visual complexity

English

Data

LA1 4YW

Sally Linkenauger

MSC

Cognitive, Perception; Marketing

88

ANOVA; T-Test